Psychological Bulletin

1970, Vol. 73, No. 4, 258-302

ALTRUISM—AN EXAMINATION OF THE CONCEPT AND A

REVIEW OF THE LITERATURE1
DENNIS L. KREBS
Harvard University

Literature relating to altruism is reviewed. It is suggested that the study of

altruism is important at three levels: as it relates to the main goal of socializa-
tion, to a core attribute of personality, and to theories concerned with human
nature. The concept of altruism is examined. Independent variables associated
with altruism are organized in a 2 X 4 framework on the basis of the source
of experimental measurement and level of generality. Characteristics of the
benefactor and characteristics of the recipient are categorized as state variables,
trait variables, social roles and demographic attributes, and characteristics in-
fluenced by social norms. Research at each level is critically reviewed. Posi-
tive and negative affective states, and states induced by the observation of
models are found to influence the altruism of benefactors; and dependency
and interpersonal attractiveness are found to influence the altruism-eliciting
capacity of recipients. Research relevant to personality traits is criticized.
Effects are found for sex, age, ordinal position, social class, and nationality in
relation to benefactors, and for friendship status, ingroup affiliation, and social
class in relation to recipients. The normative level of analysis is criticized.

In 1956 Louis J. Budd reviewed the history tenance of the social system. The study of
of a "forgotten aspect of social thought" in altruism is important at at least three dis-
America. He concluded that although concern tinct (yet interrelated) levels.
with altruism had "threatened to become a At the first level, it supplies information
fad" in the 1890's, it had quickly died away. about a set of behavior that constitutes the
Of late, though, there has been a renewed in- central goal of early socialization. As pointed
terest in altruism—less, however, as a social out by Anna Freud (1963):
panacea or religious ideal, and more as an
We know that the child acts throughout the period
aspect of behavior worthy of scientific in- of development above described [from birth to age
vestigation. five] as if there was nothing more important than
It is not difficult to see why altruism has the gratifying of his own pleasures and fulfilling of
captured the interest of social scientists. In his powerful instincts, whereas education proceeds as
if the prevention of these objects was its most im-
view of its practical and theoretical impor- portant task [p. 101].
tance, it may well be wondered why the study
of altruism has only recently gained popular- At the second level, altruism is important as
ity. As the anthropologist Bohannan (1963) a personality attribute. People who are con-
pointed out, the most basic tenet of all major sidered altruistic are reacted to differently
religions of the world is that "unselfishness is from those who are considered selfish. Be-
the primary virtue and that selfishness lies at cause the way people act is not always a
the root of the world's ills [p. 336]." But veridical representation of the way they feel,
altruism is of interest for reasons other than the attribution of altruism poses a particularly
its status as a virtue or its role in the main- difficult problem for self-theory and the study
1
The author appreciatively acknowledges the help of person-perception.
and inspiration supplied by Judith Anne Krebs and Finally, the study of altruism raises im-
Robert Rosenthal, both of whom worked overtime portant questions about the ability of several
in his behalf. The critical comments of Terence D.
Creighton at the University of British Columbia are influential theories to account for the ap-
also acknowledged. parently altruistic aspects of general human
Requests for reprints should be addressed to Dennis behavior. Reinforcement theory, psychoana-
L. Krebs, Department of Social Relations, William
James Hall, Harvard University, Cambridge, Massa- lytic theory, and the theory of evolution seem
chusetts 02138. to suggest that human behavior is essentially
258
 ALTRUISM 259

egoistic. Yet, behavior that seems quite al- sonality attribute have usually sought a defi-
truistic is apparent in everyday life. nition in the attributional processes of average
The bulk of research on altruism has not people. It could, of course, be argued that the
attempted to establish the existence of al- layman's implicit definition of altruism is oi
truism. The problems associated with the little concern to the social scientist. It must
identification of altruism, however, have not be realized, though, that whatever errors the
escaped researchers. An examination of the layman may make in his attribution of al-
contribution of the study of altruism to the truism in specific situations, the category of
understanding of social behavior, personality, altruism is what he thinks it is—it has no
and human nature must orient itself around existence outside of the consensual agreements
the ways investigators have dealt with the of people (cf. Bruner, 19S8); and it is the
specification of the phenomenon. layman's definition that determines his reac-
tions.
Altruism as a Social Behavior Although new research may supply a more
Most behavioral research has skirted the meaningful and precise definition of altruism,
problem associated with the specification of it is interesting to note that social scientists
altruism by employing operational definitions. have generally employed everyday definitions.
It has examined antecedents of behavior that Leeds (1963), for example, defined altruism
seemed altruistic, and assumed that motiva- on the basis of the same three criteria that
tions were congruent. Although the antecedents Heider (1958) identified as determinants of
of operations with undefined motivational the attributions of laymen. According to Leeds
bases can be elucidated, it is important to (1963), an altruistic act (a) is an end in it-
make sure that the conclusions that are drawn self; it is not directed at gain, (b) is emitted
about them do not relate back to the motiva- voluntarily, and (c) does good. Heider's
tion implied by the category of behavior in (1958) treatment of the determinants of
question. Motivational specification is par- gratitude (where gratitude is the result of
ticularly important in relation to a moral be- the attribution of altruism) mirrors Leeds'
havior such as altruism. As Piaget (1932), (1963) definitional outline:
Kohlberg (1964), and others have demon- We do not feel grateful to a person who helps us
strated, it is more the intention behind an act [Leeds' point el fortuitously [al or because he was
forced to do so [b], or because he was obliged to do
than its consequences that determines its so [6]. Gratitude is determined by the will, the in-
moral value. tention of the benefactor [p. 265].
Fortunately, due to the constraints of ex-
Several recent studies have supported
perimental situations, the range of ulterior
Heider's (1958) propositions and Leeds'
motives in laboratory studies is usually limited
(1963) assumptions. Tesser, Gatewood, and
and the motivation behind the behavior in
Driver (1968), for example, found that under-
question quite apparent. It must be realized,
graduates thought they would feel more grati-
however, that although terms such as pro-
tude when the benefit they received was in-
social behavior (Bandura & Walters, 1963),
tentional, when it was valuable, and when it
helping behavior (Berkowitz, 1967), volun-
cost the benefactor a great deal. Goranson
teering (Rosenbaum, 1956), gift-giving (Blake
and Berkowitz (1966), Frisch and Greenberg
& Rosenbaum, 19SS), sharing (Staub &
(1968), Lerner and Lichtman (1968), and
Sherk, in press), and aiding (Midlarsky, Hornstein, Fisch, and Holmes (1968) found
1968b) all imply self-sacrificial other-directed
that help was most liable to be reciprocated
behavior, they do not establish it. (and, by implication, considered altruistic)
Altruism as a Personality Attribute when it was perceived as voluntary and inten-
tional. Brehm and Cole (1966), Kiesler
While researchers concerned with the ante- (1966), and Schopler and Thompson (1968)
cedents of altruistic behavior have generally demonstrated that inappropriate favors (fa-
chosen to define altruism operationally, in- vors whose intentions were in doubt) elicited
vestigators concerned with altruism as a per- less help than appropriate favors. When the
260 DENNIS L. KREBS

legitimacy of apparent altruism is questioned, relation to internal standards. Inasmuch as

reciprocity is less likely to prevail. altruistic behavior is moral behavior, the at-
Although research on attribution of al- tribution of altruism should pass through the
truism has elucidated some common principles three qualitatively different stages of moral
of judgment, it has not supplied any informa- development, and mature judgments of al-
tion about the validity of common attribu- truism should be based on internal standards
tions. It has classically been the problems as- of reciprocity and justice.
sociated with the validity of judgments of Reinforcement theorists have not generally
altruism, and the existence of altruism itself, concerned themselves with the attribution of
that have concerned thinkers—it has seemed altruism. They have, rather, attempted to
less important to know why judgments are demonstrate that although specific rewards do
made than whether they are correct. not always follow altruistic responses, altruis-
tic behavior can still be a function of rein-
Altruism as a Part oj Human Nature—Its forcement.
Challenge to Grand Theories
In a study with Paskel (reported by Aron-
In relation to the theory of reinforcement, freed, 1968), Aronfreed found that children
psychoanalysis, and the theory of evolution, who were exposed to expressions of joy (con-
the existence of behavior that seems altruistic ditioned stimulus) at the same time as they
has created problems. Critics have asked, for were hugged (unconditioned stimulus) and
example, how behavior that sacrifices rewards experienced positive affect (unconditioned re-
can be consistent with the principle of rein- sponse) sacrificed candy in order to elicit ex-
forcement. Intrinsic to the challenge that the pressions of joy (which in turn evoked rein-
existence of altruism presents to antithetical forcing positive affect). Once learned, it was
theoretical positions is the challenge that such expected that internal representation of ex-
positions present to the existence of altruism pressive signs in the recipient would be enough
itself. to reinforce altruistic responses. Aronfreed's
Altruism in the theory oj reinforcement. position implies that altruism involves a
The problem of the place of reinforcement in Cyrano de Bergerac type of sacrifice.
altruistic behavior really involves two often Goldiamond (1968) identified three ways
confused questions. The first asks whether by which behavior that seems to receive no
self-sacrificial behavior can continue in the immediate gains can be perpetuated. In the
absence of positive reinforcement. The second first place, even though particular responses
asks whether the average person includes be- are not rewarded, the net gains of the be-
havior that benefits another yet gains rewards havioral strategy may ultimately be maximal.
in his category of altruism. The first question Second, the reinforcer, as in the case of
cannot be answered in the form in which it masochism, may be subtle. And finally, the
stands—it requires proof of the null hypothe- behavior may endure in the absence of rein-
sis. The second question is a matter of at- forcement because the original program of re-
tribution. inforcement rendered it resistant to extinction.
Although no research has been reported Altruism in psychoanalytic theory. Although
which is directly relevant to the attribution of the challenge that altruism has presented to
altruism in situations of varying rewards, reinforcement theory is more serious than its
some research on moral judgment seems rele- challenge to psychoanalysis, Freud's conten-
vant. Kohlberg (in press) outlined three great tion that children are basically selfish, id-
stages in the development of moral judgment. driven animals has had repercussions. As early
At the least mature stage (Stage 1) judg- as 1929, Wodehouse objected to
ments are made on the basis of the hedonic
consequences of an act (whether it elicits re- a certain doctrine which, ostensibly connected with
ward or punishment from external sources). [Freudl, does seem to be bringing real danger into
educational psychology. It comes often in a broad
At the second stage, morality is seen as a and vague form which needs a broad and vague
function of the approval and disapproval of name, and I will call it the doctrine of Natural Self-
others. And at the third stage it is judged in ishness [p. 39].
 ALTRUISM 261

Partially as a result of psychoanalytic the- piricists have not fared so well. Empiricists
ory's inability to explain altruistic behavior have argued that the demonstration of a
and other adaptive undefensive behavior phylogenetic increase in altruism would im-
(Maddi, 1968), new developments tended to plicitly support the likelihood of innate al-
stress adaptations of the ego (e.g., Erikson, truism in man. With this in mind, researchers
1950; Murray, 1938; Sullivan, 1953; White, attempted to demonstrate the existence of
1963). Deemphasis on id-driven behavior altruism in infrahumans.
served to soften the theory's insistence on Starting low on the phylogenetic scale, Rice
natural egoism. and Gainer (1962) found that albino rats
Altruism in the theory of evolution—the- were more likely to press a bar when it low-
oretical arguments and animal research. While ered a struggling rat than when it lowered
the reaction to the egoistic assumptions of a styrofoam block. Rice (1965) also found
psychoanalysis was largely theoretical, the that rats (but not guinea pigs) would press
reaction to the apparent egoism of the theory a bar to remove a companion from a tank of
of evolution came from both theoretical and water. The attribution of altruism in these
empirical sources. Theoreticians attempted to cases, though, can be questioned. It is pos-
show that the theory of evolution did not sible that the rat's reaction was a function of
propose, or even imply, the existence of in- the presence of another rat or the noxiousness
nate egoism; and empiricists attempted to of the screams of the distressed rats. The
demonstrate that infrahumans are capable of latter suggestion received support from a
altruism. study by Lavery and Foley (1963), who
Holmes (1945) argued against the popular found that rats which were exposed to white
egoistic interpretation of Darwinian theory noise made more noise-terminating bar-presses
and insisted that altruistic instincts were as than rats exposed to the recorded squeals of
basic as egoistic instincts. Derived from "in- other rats. The fact that more bar-presses
stincts subsidiary to the basic function of were elicited in the white-noise condition sug-
reproduction [p. Ill]," he saw the function gests that the white noise was more noxious
of altruism as the preservation of the species, than recorded squeals, and that the "altru-
even at the cost of individual lives. ism" of the Rice and Gainer (1962) and Rice
Campbell (1965) argued against the preva- (1965) studies was simply a reaction to a
lence of "skin-surface hedonism" and, on the noxious stimulus.
basis of research demonstrating ingroup More doubt was placed on the existence of
solidarity and sacrifice in the face of outgroup altruism in the albino rat by another study by
threat, suggested that altruistic motives are Rice (1964). Rats pressed a bar less when it
not only a function of sociocultural evolution served to terminate a companion's shock than
(are acquired), but also have a basis in bio- when it did not. Rats which were exposed to
logical evolution (are innate). Altruism, he a shocked companion crouched in fear at the
suggested, is likely to have genetic determi- other side of their cages. A further finding
nants because of the survival value of in- showed that rats failed to press a bar to
group identification and outgroup hostility: terminate the recorded squeals of other rats,
"The tremendous survival value of being so- as reported by Lavery and Foley (1963). All
cial makes innate social motives as likely on considered then, no real support for the ex-
a priori grounds as self-centered ones [p. istence of altruism in rats has been supplied.
301]." Both Holmes (1945) and Campbell More support for the existence of altruism
(1965) suggested that egoism, whose aim is in infrahumans, though, has appeared in rela-
the survival of the individual, and altruism, tion to higher forms of animals. Summarizing
whose aim is the survival of the group, are research on primates, Hebb and Thompson
in conflict, and that the conflict offers the (1954) suggested that "there is definite evi-
optimal evolutionary compromise. dence in other animals of a phylogenetic de-
While the arguments of theoreticians seem velopment of something we call altruism, de-
cogent (and, in fact, are supported by a close fined as intrinsically motivated concern for
reading of Darwin) the arguments of em- others [p. 744]." Unfortunately, though, evi-
262 DENNIS L. KREBS

dence for altruism came from mainly anec- principles which underlie the attribution of
dotal sources. In one of the only studies on altruism. They have not, though, attempted
altruism in primates, Nissen and Crawford to determine whether the altruism that was
(1936) found that two of six preadolescent attributed actually met the criteria of at-
chimpanzees sometimes gave food, and tokens tribution. Finally, research has been con-
which could be exchanged for food, to chimps ducted and theoretical positions amended in
in adjoining cages. The chimpanzees tended, order to accommodate altruistic behavior.
though, to give only their least preferred Positions which seemed to pose challenges to
food, and to give only when food was solicited. the existence of altruism have come to ac-
Although solicited giving (and prostitution) cept the fact that altruistic behavior occurs
in wild chimps is common (Yerkes & Yerkes, (within the definitional constraints of the
193S), it is possible that the small amount of theory in question), and have set out to ex-
unsolicited giving was an experimenter effect plain it.
—a result of modeling. Nissen and Crawford All in all, though, the definitional problem,
(1936) reported, in fact, that experimenters which involves the status of altruism as a
were able to facilitate unsolicited food passing. dependent variable, has attracted very little
It is very difficult to draw any conclusions attention from recent researchers. They have
from comparative research. Infrahuman ani- generally been content to assume that be-
mals do perform acts which benefit other ani- havior that seems altruistic is altruistic, and
mals, but the reasons behind these acts are to concern themselves with its determinants.
almost as ambiguous as those behind the It is to determinants and correlates of appar-
behavior of humans. Although the existence ent altruism, then, that a review of the litera-
of genetic or biochemical determinants of ture must turn.
altruism is a fascinating possibility, it must
remain only that until further evidence ap- AN ORGANIZATIONAL FRAMEWORK FOR
pears. VARIABLES ASSOCIATED WITH ALTRUISM
In summary, the study of altruism is im- In the past few years researchers on altru-
portant at three conceptually distinct levels, ism have studied so many variables that an
relating to behavior, personality, and human integrated perspective is already difficult. It
nature. These levels are interrelated—judg- seems possible, however, to attain some
ments about human nature are made on the integrative clarity by ordering the variables
basis of representative personalities, and per- along two dimensions. To begin with, the
sonalities are determined by samples of be- prototypical altruistic situation involves some-
havior. The study of altruism is particularly one who gives (a benefactor), and someone
important to the understanding and imple- who receives (a recipient). In some cases,
mentation of socialization practices. It is also characteristics of the benefactor affect altru-
significant to person-perception and self- ism, and in other cases it is characteristics of
theory as a core personality attribute. And the recipient. Independent variables, then,
finally, it is important because it has stimu- can be divided into those which relate to
lated examination and change in several im- characteristics of the benefactor, and those
portant theoretical positions. which relate to characteristics of the recipient.
The question of the definition of altruism, It is, of course, true that all variables have an
and the related question of whether altruism ultimate effect on the benefactor, but the
exists, have posed difficult problems. Be- effect is often achieved by varying character-
havioral researchers have generally avoided istics of recipients. The first dimension of
the definitional issue, which involves establish- classification, then, separates variables which
ing the intention behind apparently self- relate to the characteristics of benefactors
sacrificial other-oriented acts, by employing that cause or correlate with altruism from the
operational definitions. Investigators con- altruism-eliciting characteristics of recipients.
cerned with altruism as a personality char- Independent variables can also be classified
acteristic have generally dealt with the defini- according to their level of generality. Research
tional issue by attempting to determine the on altruism has manipulated independent vari-
 ALTRUISM 263

TABLE 1
A CLASSIFICATION OP INDEPENDENT VARIABLES EMPLOYED IN RESEARCH ON ALTRUISM

Source of experimental variation

Categories of
independent
variables
Characteristics of the benefactor Characteristics of the recipient

Situational state Positive affective states Dependency

variables Negative affective states Interpersonal attractiveness
States induced by the observation of models As an independent
Increased salience of social norms and behavior alter- variable
natives As a mediating variable
Information about appropriateness
Information about consequences
Internalization of altruistic dispositions
Experimenter effects
Trait variables Variables that relate to rating-scale measures of altruism Locus of dependency
Variables that relate to pencil-and-paper measures of altruism Other trait characteristics
Variables that relate to behavioral measures of altruism
Social roles and Sex Friendship status
demographic vari- Sex differences in children Ingroup affiliation
ables Sex differences in adults Social class
Age
Ordinal position
Social class and group affiliation
Nationality
Social norms Norm of social responsibility Norm of reciprocity
Norm of giving Reciprocity
Generalized reciprocity

ables at four more or less distinct levels of tionnaire-derived measures of personality

generality. The first level involves temporary traits with an index of altruism.
psychological states, such as those that ac- At the third level—that which involves
company experiences of success, failure, de- social roles and demographic variables such as
pendency, interpersonal attraction, and the social class, age, and sex—the level of gen-
observation of models. Most of the research on erality is even greater. Social roles and demo-
altruism relates to state variables, probably graphic variables differ from trait variables
because they are the easiest to manipulate in because they are more general, permanent,
laboratory experiments. Independent variables and basically characteristic. Social roles, of
of the state type are largely situational. They course, often relate to personality traits and
have an immediate, temporary, and relatively psychological states. Women and children, for
limited effect, and they usually say little example, are expected to react differently from
about the nature of the people whom they men and adults.
affect. The final level, which deals with social
The second level of generality involves norms, is the most general of the four. Norms
personality traits. Although states and traits such as the norm of social responsibility and
may well interrelate, trait variables such as the norm of reciprocity, if not universal
cyclothymia, need for approval, and conserva- (Gouldner, 1960), affect most people in most
tism refer to more general and lasting at- cultures. In fact, it could be argued that
tributes of people. In some cases, traits seem internalized social norms are so general that
to correspond to the characteristic states of they supply no information about the variance
people. Studies which examine trait variables in incidences of altruism. Their effect may
are usually less manipulatively experimental only be of interest as it relates to temporary
than studies which examine state variables. states, personality traits, and general social
They generally correlate rating-scale or ques- roles. Several researchers, however, have at-
264 DENNIS L. KREBS

tributed variations in altruism to the effect of mental differences in reactions to success and
social norms. The problems presented by the failure. Although fifth-grade children tended
normative approach will be examined more to leave more candy for a hypothetical other
closely when related research is reviewed. after they succeeded (versus failed or did
A classificatory framework that involves average) on a bowling game task, fourth-grade
the interaction between two sets of criteria children left more after they failed. The au-
has been outlined. Table 1 presents the result- thor suggested that a "norm of deserving"
ing eight categories and an outline of the motivated the fourth graders, but that the
variables within the categories that have been fifth graders were motivated by "norms or
examined. standards or values directly related to shar-
ing." No reason was given, though, why the
ALTRUISM AS A FUNCTION OF TEMPORARY two norms should differentially affect the
STATES OF THE BENEFACTOR two particular age groups in question. The
The preponderance of research on altruism additional finding that children who saw
has manipulated situational variables which themselves as having internal control over
induce states in benefactors that mediate their environment shared more after success
altruistic responses. The state may be a simple than those who felt externally controlled sug-
affective state, or a cognitive state which gests that perceived competence is related to
relates to particular response dispositions. sharing.
Research which has manipulated situa- A relationship between competence and
tional variables and their corresponding psy- altruism was found by Midlarsky (1968a).
chological states can be divided into three Subjects who were told that they adapted well
categories. The first two relate to affective to electric shock (high competence) took
states, and the third relates to cognitive states more shocks for another than those who were
induced by the observation of models. Positive told they adapted poorly. Unfortunately, the
states have been created by supplying experi- shocks were not of equal intensity across con-
ences which involve success and the percep- ditions. A later study by Kazdin and Bryan,2
tion of competence. Negative states have been though, which controlled for the cost of help-
created by supplying experiences which in- ing, found essentially the same thing. Subjects
volve failure, unintentional harm to another, who were told they were highly competent on
and acts of transgression. Finally, states have tasks which were both relevant and irrelevant
been induced by the presentation of altruistic to the dependent variable offered to donate
models. Table 2 contains an outline of re- more blood than those who were told they
search which has manipulated positive and were incompetent. The notion that a tem-
negative affective states. porary state mediated the altruism was sup-
ported by the fact that very few volunteers
Positive States of the Benefactor followed through with their commitment to
give blood. Once they had a chance to recover
Four studies have tested the effect of ex- from the positive experience, it would appear,
periences of success and competence on their altruistic inclinations decreased. (It is,
altruism. Berkowitz and Conner (1966) tested of course, possible that some subjects gen-
the hypothesis that success increases the eralized their perceived competence to getting
salience of the social responsibility norm, rid of the solicitor, with no intention of ever
which leads to altruism toward dependent giving blood.)
others. They found that success on a simple Some naturalistic reports relating to re-
task resulted in greater effort on behalf of a actions to disasters (e.g., Form & Nosow,
highly dependent peer than did failure or no 1958; Torrance & Ziller, 19S7) suggest that
experience at all. Success did not result in observers who perceive themselves as com-
more helping for others of low dependency. petent in emergency situations help more than
The Berkowitz and Conner (1966) study 2
Kazdin, A. E., & Bryan, J. H. Competence and
used undergraduates. A later study by Staub volunteering. Unpublished manuscript, Northwestern
(1968) suggested that there may be develop- University, 1968.
 ALTRUISM 265
TABLE 2
STUDIES WHICH EXAMINED THE EFFECT OF AFFECTIVE STATES OF THE BENEFACTOR ON ALTRUISM

Subjects
Author and date Main independent variables Main dependent variables
Sex Agea N

Positive states of the benefactor

Berkowitz & Conner (1966) M CS 108 Success, failure, or no experi- Number of envelopes made
ence on jigsaw task ; high, for recipient
intermediate, low de-
pendency of recipient.
Staub (1968) M/F» 9, 10 196 Success, failure, moderate Weight of candy left for
performance on bowling recipient.
game task, locus of con-
trol.
Midlarsky (1968a) M CS 80 Ability to adapt well (high Number of shock contingent
competence) or poorly problems completed for
(low competence) to recipient.
shocks; high vs. low de-
pendency of recipient ;
visibility.
Kazdin & Bryan M/F b CS 96 Success vs. average perform- Frequency of volunteering
ance on task ; task rele- to give blood.
vant or irrelevant to de-
pendent variable.
i
Negative states of the benefactor

Darlington & Macker (1966) F CS 39 Failure which harms (vs. Frequency of volunteering
does not harm) another. to give blood (after three
requests).
Rawlings (1968) V CS 40 Failure which harms another Duration of reciprocal
vs. observation of harm to shocks.
another
Krebs & Baer \ V CS 40 Success or failure on intelli- Amount of help volunteered
i gence test ; harm or help for charitable cause.
to another.
Lerner & Matthews (1967) V CS 66 Fate of recipient dependent Frequency of choices to
vs. independent of draw of comfort partner.
benefactor.
Lerner F CS 61 Fate of recipient dependent Frequency of choices to
on draw of benefactor vs. take partner's place ; fre-
draw of experimenter. quency of choices to
comfort partner
Freedman et al. (1967)
Experiment 1 M HS 16 Telling a lie vs. not telling Frequency of volunteering
a lie. for pleasant or unpleasant
experiment.
Experiment 2 F CS 67 Responsibility for upsetting Frequency of volunteering
index cards. for victim's vs. non-
victim's experiment.
Experiment 3 F CS 74 Responsibility for upsetting Frequency of volunteering
index cards. to help victim in person
vs. not in person.
Berscheid & Walter (1967) F adult 240 Opportunity to compensate Frequency of compensating
harm done adequately. victim and nonvictim.
Epstein & Hornstein (1969) M CS 60 Punishment (vs. no punish- Frequency of selfish choices
ment) for selfish choice ; (choices which earned 10 j!
liked, disliked, neutral but shocked another).
recipient.
Wallace & Sadalla (1966) M CS 55 Public, private, or no trans- Frequency of volunteering
b
gression. for stress experiment.
•Silverman (1967) M/F 11 199 Private high cheating, pri- Amount of free-play time
vate low cheating, no volunteered for experi-
cheating. ment.

Note.—Abbreviations are: HS = high school student; CS =» college student.

a
b
Age of first-grade children estimated at 6, etc.
Sex constituted a variable.
266 DENNIS L. KREBS

those who do not. Competence in disasters, the effect that experiences of success, failure,
though, is different from most experimentally and helping and harming another had on
manipulated competence. In experiments, the charitable behavior. Altruism was greatest
experience of success seems to increase self- after failure which harmed another, and least
esteem, which leads to increased positive af- after success which benefited another. There
fect and altruism. In disaster situations, was no difference between the straight success
however, it would seem that it is the implicit and failure conditions. The findings were
role requirements associated with competence, interpreted as support for a self-concept
especially when the competence is the result equilibrium model which suggests that people
of special training, that mediate helping be- whose self-images are unrepresentatively low
havior. are likely to seize an opportunity to be-
In summary, although none of the relevant haviorally reassert a more favorable self-
experiments supplied unequivocal evidence, definition. Unrepresentatively high self-im-
they all found indications that altruistic re- ages, on the other hand, are not maximized.
sponses on behalf of dependent others are In the Krebs and Baer (see Footnote 3)
more probable after success than after failure, study, the success and failure of the potential
or after a neutral experience. benefactor had a corresponding effect on his
partner, and the altruism was directed to a
Negative States of the Benefactor
charitable cause. Studies by Lerner and his
Of the studies that compared the effects of associates demonstrated that partner-oriented
success and failure on altruism, only one altruism occurs in situations where success
(Staub, 1968) found a positive relationship for self results in failure for another. Subjects
between failure and altruism, for fourth-grade who drew a slip of paper that assigned them
children, and that relationship reversed itself to a control condition and their partner to a
in the fifth grade. Other studies, though, have shock condition (fates interdependent) were
found that failure which has a particular more prone to comfort the other and volunteer
consequence—harm to another—leads to al- to take his place than subjects who deter-
truistic responses. Darlington and Macker mined only their own fate (fates independent;
(1966), for example, found that failure to Lerner and Matthews, 1967) or subjects whose
complete a pencil-and-paper task correctly fates were determined by the experimenter
resulted in more agreement to give blood (Lerner 4 ).
when the failure hurt a helpful other than Other studies that did not involve suc-
when it did not. The findings were interpreted cess and failure have supplied further sup-
as evidence for displacement of guilt-produced port for the notion of reparative altruism.
altruism. Because it was only after the third Two studies investigated the effect of com-
of three appeals for blood that any difference pliant and unintentional harm-doing on altru-
was found, and because 13 subjects were dis- ism. Carlsmith and Gross (1969) found that
carded, the results of the study must be subjects in a Milgram (1963) type of situation
viewed with caution. Moreover, a later study who delivered shocks to another were mor&
(Rawlings, 1968) found that the observation likely to volunteer to support a humani-
of a person receiving harm is enough in itself tarian project than those who did not shock
to induce altruism. Although subjects whose another. Freedman, Wallington, and Bless
errors on a task caused their partners to re- (1967, Experiment 2) found that subjects
ceive shocks delivered reciprocal shocks of who knocked over a pile of index cards were
short duration to a third person (and, there- more willing to volunteer for an experiment
fore longer duration to themselves), reciprocal to help another than those who did not,
shocks of short duration were also given when providing that the solicitor was not the owner
they only observed their partners getting
shocked. another on altruism. Paper submitted for publica-
A study by Krebs and Baer 8 compared tion.
4
Lerner, M. S. The effect of a negative outcome-
3
Krebs, D. L., & Baer, R. The effect of perceived on cognitions of responsibility and attraction. Un-
competence and unintentional help and harm to published manuscript, University of Kentucky, 1968..
 ALTRUISM 267

of the index cards. In a similar situation is relieved by punishment, then it can be pre-
(Experiment 3), subjects who harmed another dicted that punished responses are less likely
were more likely to volunteer to help him if to extinguish than responses that are not
they did not expect to meet him than if they punished. This prediction was partially sup-
did. Although harm-doing elicited altruism, it ported by Epstein and Hornstein (1969).
was not oriented directly toward its victim. They found that selfish behavior toward a
The Darlington and Macker (1966), Krebs disliked other (pressing a lever which maxi-
and Baer (see Footnote 3), Carlsmith and mized chances of gaining ten cents, while
Gross (1969), and Freedman et al. (1967) delivering an electric shock to another) in-
studies suggest that reparative altruistic creased, and altruistic behavior decreased,
behavior relieves an unpleasant negative when selfish behavior was punished by a
state associated with lowered self-esteem third person. Due to the fact that selfish be-
by supplying a situation in which a wrong havior toward a liked other decreased after
can be righted and self-esteem elevated. The punishment, the generality of the findings is
Rawlings (1968) and Lerner and Matthews limited.
(1967) studies, though, lend themselves to a Two final studies suggest that private trans-
slightly different interpretation. It may be gressions are not as likely to lead to altruistic
that harming another creates a feeling of reparation as public transgressions. Wallace
guilt, which results in expiative self-punitive and Sadalla (1966) found that subjects who
responses. In cases where altruism was di- broke an expensive machine were more likely
rected toward the expiation of guilt it would to volunteer for a painful experiment than
be expected that (a) private transgressions those who did not, but only if their trans-
would lead to expiative responses, and (b) gression was discovered. Silverman (1967)
self-punitive expiation would be preferred to failed to find a higher incidence of volunteer-
nonpunitive reparation. These predictions do ing from children who cheated on a task but
not follow from a reparative self-esteem model did not consider themselves caught than from
because the function of altruism is to reassert those who did not cheat.
a new self-definition, with no necessity for In summary, many studies have supported
pain. the notion that public transgression, whether
Although it is possible that transgression intentional or unintentional, whether immoral
elicits reparative responses in some situations or only situationally unfortunate, leads to
by some people, and expiative responses in reparative altruism. Reparative altruism would
other situations by other people, it seems seem to alleviate a negative state associated
that the altruistic responses in most of the with lowered self-esteem. When amends can-
relevant studies were of the reparative nature. not be made to the victim, reparative re-
Freedman et al. (1967, Experiment 1), for sponses are generalized to others; in fact,
example, failed to find, as expected, that sub- in some situations reparative responses are
jects who lied about their knowledge of an made only if they can be directed toward a
experiment chose the more unpleasant of two third party.
other experiments. And Berscheid and Walster
(1967) found that harm-doers were most States Induced by the Observation oj Al-
likely to compensate their victims when they truistic Models
could make exact reparation. Female mem- The observation of models, according to
bers of church auxiliaries who caused their Bandura and Walters (1963), affects behavior
partners to lose needed books of green stamps in two distinct ways: by inducing the acquisi-
subsequently awarded bonuses to them if the tion of long-term behavioral dispositions, and
bonuses were neither insufficient nor excessive by inducing the performance of imitative be-
relative to the original loss. The behavior of havior. The acquisition of response disposi-
the church ladies, in the experiment at least, tions, which, in relation to moral behavior, is
hardly seemed self-punitively expiative. referred to as internalization, forms the basis
There is one study, though, that lends of behavioral analogues in which models cor-
itself to an expiation interpretation. If guilt respond to parents, and modeling effects are
268 DENNIS L. KREBS

thought to lead to long-range changes in attention to particular courses of action, and

personality. Performance, on the other hand, increase the salience of social norms. Second,
refers to situation-specific behavior that oc- they supply information about what is ap-
curs as the result of the induction of tem- propriate in various situations by setting an
porary states. Although most of the research example, by helping to create a normative
on modeling relates to the performance of standard, and by helping to supply a defini-
altruistic behavior, much of it is interpreted tion of the situation. Third, models supply
as evidence for the internalization of altruistic information about the consequences of courses
dispositions. It is, of course, difficult to sort of action. Although the three aspects of the
out acquisition and performance effects in modeling situation are conceptually distinct,
one-shot laboratory experiments, but to draw they are not, in most cases, unrelated. A
conclusions about the development of person- rough hierarchical relationship, in fact, seems
ality traits on the basis of situational con- to exist between them, with the third sub-
formity is clearly unwarranted. Before con- suming the second, and the second subsuming
clusions can be drawn about the acquisition the first. Although models who make be-
of behavioral dispositions, two criteria should havioral alternatives salient need not supply
be met: (a) The behavior in question should information about what is appropriate, nor
be general to situations other than that in need they supply information about be-
which it was elicited, and (b) it should be havioral consequences, models who supply in-
relatively enduring. Because the preponder- formation about behavioral consequences
ance of studies on modeling have failed to usually supply information about what is ap-
meet the two criteria of internalization, they propriate, and they usually make a course
seem best interpreted in relation to para- of action salient. Studies which do not un-
meters of performance, as a function of equivocally supply information about ap-
temporary states. This is not to say that propriateness seem most properly reviewed
they should be considered irrelevant to social- in the first category, and studies in which the
ization, but only to suggest that they have perception of consequences is in doubt are
not established that modeling produces long- reviewed in the second. Table 3 contains an
term personality changes; or even that it outline of studies on modeling which fall in
would, on a prolonged basis. the three categories.
Although the modeling approach has en-
couraged research efforts, it has generally Modeling as a Function of the Increased
failed to supply explanations for modeling Salience of Behavioral Alternatives and Social
effects. The label modeling has been used Norms
to explain modeling effects, but modeling, like A series of three naturalistic experiments
imitation, merely describes a sequence of by Bryan and Test (1967) demonstrated that
congruent actions—it does not explain it. observation of helpful and charitable models
Before any real explanatory power is achieved, elicits congruent altruistic behavior. In the
the reasons why models induce imitative be- first experiment, motorists were more likely to
havior must be elucidated; and these reasons stop to help a lady fix a flat tire if they drove
may well vary across situations. past a model helping a confederate fix a flat
A survey of the studies on modeling sug- than if they did not. Two other experiments
gests that the performance of altruistic be- found that shoppers contributed more to a
havior may be based on one of several con- Salvation Army kettle in the 20 seconds or
ceptually distinct aspects of modeling. An half minute after a model gave than in the
attempt will be made to sort out the aspects 20 seconds or half minute before he gave.
of modeling situations that elicit modeling Although the most parsimonious interpretation
effects by organizing the relevant studies in for these findings would seem to relate to the
terms of what seems to be their most appro- increased salinence of the altruistic behavioral
priate explanatory base. alternative, Bryan and Test (1967) offered
At the most elementary level, models make two other interpretations. The first suggested
behavioral alternatives salient: They draw that the effects were due to the information
 ALTRUISM 269

that was supplied about the pleasant con- money was given in the altruistic model con-
sequences of helping, and the second sug- dition, least was given in the selfish model
gested that they were due to a comparison condition, and the amount given in the con-
between observer and model, which resulted trol condition fell in between. Although the
in shame. The first interpretation does not responses of the models may have supplied
seem tenable in the flat-tire situation, where some hints about the appropriateness of vari-
the consequence of helping was hard work ous responses, there was nothing about the
(unless of course those who stopped had a models that legitimized their responses. Their
pick-up in mind). The second interpreta- main function seems to have been to increase
tion runs into some difficulty in all three ex- the salience of behavioral alternatives.
periments due to the fact that most of the The three studies reviewed above used
shoppers and motorists did not act altruist- adult subjects. When children are exposed to
ically. Moreover, the sight of the Salvation adult models, however, the behavior of adults
Army solicitor in itself would seem to be is likely to serve as a standard of appropriate-
enough to elicit shame in those who were ness. The performances of authoritative or
prone to react with shame. The more con- prestigious adults not only indicate what can
servative interpretation, therefore, seems be done, but also what should be done.
justified.
In another study, Test and Bryan (in Modeling as a Function of Information about
press) compared the effect of three modeling the Appropriateness of Behavioral Alternatives
situations on helping behavior. They found
that potential benefactors helped a stooge Appropriateness as a function of prestige.
more after they observed a generous model, Several studies have exposed children to
and after they were helped themselves than altruistic adult models. Rosenhan and White
after they were exposed to an indifferent (1967) found that 63% of a sample of ele-
model or to no model at all. There were no mentary school children donated gift certifi-
differences between the observation of help cates to charity during a game that they
and receipt of help conditions, nor between played with a charitable model. Forty-eight
the indifferent-model and no-model condi- percent also gave when they subsequently
tions. Although Test and Bryan (in press) played alone. Although the authors inter-
stated that their failure to find a difference preted the second finding as evidence for in-
between the indifferent-model and no-model ternalization, the temporal contiguity between
conditions was not in accord with the results tasks and the similarity of situations sug-
of studies by Rosenbaum and Blake (19SS) gests that the children simply behaved in
and Rosenbaum (1956), this may not be the what seemed to be an appropriate way.
case. The studies are not strictly comparable A study by White (1967), which satisfied
because the Test and Bryan study employed the endurance (but not generality) criterion,
a model who failed to help, whereas the latter suggested possible limitations on the modeling
studies employed models who actively refused of altruistic responses. Fourth- and fifth-grade
to help. Although the Test and Bryan in- children were exposed to one of four con-
different-model condition did involve con- ditions: enforced rehearsal (ER), in which
straints against helping, they were not as they practiced charitable behavior; observa-
strong as those in the Rosenbaum studies. tion plus voluntary rehearsal (OVR), in
In an effort to obtain an effect for both which they practiced charity with a generous
altruistic and selfish models, Wheeler and model; observation of a generous model alone
Wagner (1968) exposed enlisted men and (0); and a no-model control condition (C).
navy recruits to a situation in which they After performing the experimental task, half
overheard a solicitor request a twenty-five- the children were left alone to win some
dollar donation for one of two causes: to help gift certificates and to donate, and half were
fly the family of a dying serviceman to his sent back to class without playing the game.
bedside, or to build up the monetary reserve All subjects returned for a second testing after
of a local serviceman's fund. As expected, most several days. Both groups of children reacted
270 DENNIS L. KREBS

TABLE 3
STUDIES WHICH EXAMINED THE EFFECT op MODELS ON ALTRUISM

Subjects Models

Author and date Pres- Main Independent Main dependent

ent variables variables
Sex Age N Sex Age vs.
ab-
sent"

Increased salience of behavioral alternatives and social norms

Bryan & Test (1967)

Experiment 1 M/F adult 4,000 M adult Ab Model helping lady fix Frequency with which
a flat tire vs. not. motorists stopped to
help lady fix a tire.
Experiment 2 M/F adult 720 M adult Ab Model donating to Sal- Number of observa-
vation Army vs. not. tion periods in which
donation occurred.
Test & Bryan F CS 80 F CS Ab Helping model, non- Number of problems
(in press) helping model, or no done for recipient.
model; dependency;
reciprocity.
Wheeler & Wagner M adult 144 M adult Ab Generous model, selfish Amount donated.
(1968) model, or no model;
cost of helping;
strength of appeal
for help.

Information about appropriateness: prestige of models

Rosenhan & White M/Fb 9-10 130 M adult P/Ab Prior nurturance of Frequency of donation
(1967) model; generous in presence and ab-
model vs. no model. sence of model.
White (1967) M/F" 10,11 210 M CS Ab Generosity of models; Number of gift cer-
rehearsal of altruistic tificates donated to
behavior; durability charity.
of altruism.
M. Harris (1968) M/Fb 9-10 168 F CS Ab Object of generous Number of chips
model's donations ; shared with model
model reinforced vs. and charity; fre-
not; selfish model, quency of sharing.
and no model.
Bryan & Walbek
(1969)
Experiment 1 M/F" 8-10 91 same adult Ab Model who preached Frequency of dona-
sex charity, greed, or tions, amount do-
neutral, and prac- nated.
ticed charity or
greed.
Experiment 2 M/F" 8-9 168 same peer Ab
sex

Experiment 3 M/Fb 8-9 132 same peer Ab

sex
Grusec & Skubiski M/F 8,10 80 same adult Ab Prior nurturance of Number of marbles
(in press) sex model ; model who shared.
preached charity or
practiced charity.
Note.—Models for Blake et al. (1955) were symbolic. Abbreviations are: CS = college student.
• during measure of altruism,
b sex constituted a variable.
 ALTRUISM 271

Table 3— (Continued)

Subjects Models

Author and date Pres- Main independent Main dependent

ent variables variables
Sex Age N Sex Age vs.
ab-
sent"

Information about appropriateness: group standards

Blake et al. (1955) M/F CS 50 Mean amount of prior Amount donated

donations; variation toward gift for re-
among prior dona- tiring secretary.
tions.
Schachter & Hall M/F cs 444 M/F peers P/Ab Conditions of volun- Frequency of volun-
(1952) teering: high re- teering for an ex-
straint vs. low re- periment; frequency
straint. of showing up.
Blake et al. (1956) M/F CS 362 M/F peers P/Ab Public vs. private com- Frequency of volun-
mitment; attractive- teering for an ex-
ness of alternative. periment.
Rosenbaum & Blake M/F cs 45 M/F" CS P Agreement vs. refusal Frequency of volun-
(1955) of models to volun- teering for an ex-
teer. periment.
Rosenbaum (1956) M cs 135 M CS P Agreement vs. refusal Frequency of volun-
of models; intensity teering for an ex-
of request. periment.

Information about appropriateness: groups standard and diffusion of responsibility and blame

Darley & Latane M/Fb CS 72 M/F CS Ab Size of group, status of Frequency and speed
(1968) other; personality of reporting epilep-
measures. tic fit.
Latane" & Darley M CS 58 M CS P Size of group ; response Frequency and speed
(1968) of others. of reporting smoke.
Latane" & Rodin M cs 120 M CS P Waiting alone vs. with Frequency and speed
(1969) another; status of of helping an in-
other (friend, jured woman.
stranger, stooge).
Korte (1969) M cs 60 M CS Ab Focused vs. diffuse re- Frequency of helping
sponsibility; other's an asthmatic ex-
definition of situa- perimenter.
tion ; personality
measures.
Kaufmann (1968) M cs 186 M CS P Status of subject ; ex- Frequency of stopping
pectation of future a Milgram-type
role ; culpability of experiment.
learner; legitimacy of
authority.
Staub & Feagans M/F 5-12" 153 M/F peer P Presence vs. absence of Frequency of trying to
(1969) another. help an injured other.

Information about consequences

Hornstein et al. M/F adult 105 M adult Ab Foreign vs. compatriot Number of returned
(1968) model; positive, nega- wallets.
tive, neutral effect
on model of acting
generously.
Midlarsky & Bryan M 6-10" 160 F CS P/Ab Reaction of model to Number of self-sacri-
(1967) self-sacrifice; affec- ficial responses;
tionate responses or amount donated.
not ; expressive cues
during test, or not.
272 DENNIS L. KREBS

similarly in the delayed session. The effect of by the model than information about ap-
time, however, served to reduce the number propriateness. As suggested earlier, actions
of altruistic responses, especially in the ER which indicate appropriateness also increase
condition. A noteworthy finding was that the salience of behavioral alternatives. Show-
there was little reduction in giving for the ing children what to do, however subtly,
girls in the OVR condition. It suggests that clearly supplies more information than telling
for girls observation of models may have only them what to do.
temporary effects on performance, whereas It has been argued that although children
observation plus rehearsal may result in the see most adults as authorities, adults gen-
acquisition of behavioral dispositions. Al- erally do not. In some situations, though,
though other research has not supported this adults do supply standards of appropriateness,
suggestion (e.g., Rosenhan & White, 1967), especially when they are in groups. Studies
it serves to emphasize the importance of which explored the effect of groups of models
differentiating situational performance effects on altruism have used two types of models:
from internalized behavioral dispositions. those who are actively altruistic, and those
A study by M. Harris (1968) demon- who are passively selfish. A series of early
strated that modeling effects sometimes have studies of the first type demonstrates the ef-
little generality. She found that children who fect of group standards on gift-giving and
were exposed to a model who gave tokens to volunteering.
them returned the favor, but children who Appropriateness as a junction of group
were exposed to a model who donated to standards—the modeling of action. Blake
charity also donated to charity. The children ct al. (19SS) found that the amount which
imitated the specific behavior of the models, graduate students donated toward a gift for
probably because it seemed appropriate, but a retiring secretary depended on the amount
failed to internalize a general altruistic dis- they thought others had donated. Although no
position. model was physically present, a list containing
A study by Bryan and Walbek (1969) "prior donations" supplied information about
tested the relative strengths of behavioral and the behavior of the reference group.
verbal indications of appropriate behavior. Schachter and Hall (1952) examined the
Groups of elementary school children were effect of group influence on volunteering be-
exposed to one of six situations. They ob- havior. Students who observed half of a class
served models who practiced charity and appear to volunteer (low restraint) were more
preached either charity, greed, or gave a likely to sign up for an experiment than those
neutral statement; or they observed models in high restraint conditions. Subjects in low
who practiced greed and preached either restraint conditions, though, were not more
charity, greed, or gave a neutral statement. likely than subjects in high restraint condi-
A series of three experiments indicated that tions to fulfill their commitment. Another
although verbalizations affected the attractive- study (Blake, Berkowitz, Bellamy, & Mou-
ness of models, it was what models did that ton, 19S6) found that students who were
affected the responses of the children. No asked to raise their hands in class (without
effect was found for hypocrisy (discrepancy half the class appearing to volunteer) were
between word and deed). less likely to volunteer than students who
Grusec and Skubiski (in press) supported were asked to sign their names in private.
the findings of Bryan and Walbek (1969). The failure of other class members to volun-
They found that third- and fifth-grade children teer seemed to establish a group standard of
were more prone to donate marbles after noncompliance.
observing a model donate than after hearing Field studies have supported the findings
him verbalize the appropriate behavior. Only of the classroom studies. Rosenbaum and
girls who had had a prior interaction with Blake (19SS) found that students were more
a nurturant model donated as much in the liable to volunteer for an experiment if they
preach charity as the practice charity con- observed another student volunteer than if
dition. It seems likely that more was supplied they did not. Interpretation of the finding,
 ALTRUISM 273

though, is difficult due to the fact that it interpreted these results as support for the
was only the subjects who observed another hypothesis that people in groups fail to re-
person volunteer (and return to his studies) spond because responsibility and blame are
who gained information concerning the diffused among group members. The diffusion
(short) duration of the experiment. A later of responsibility and blame explanation,
study by Rosenbaum (1956) controlled for though, was cast in doubt by the further
perceived duration of commitment and at- finding that females in three-person groups
tained comparable results. Unfortunately the did not respond any slower when the other
latter study contained a further confound: bystander was a male or a premed student—
Subjects, who were exposed to a model who people who would be expected to take re-
volunteered, anticipated participating in the sponsibility in emergencies.
experiment with the model, whereas subjects Even though there seems to be more in
in the other condition expected to take part effect than diffusion of responsibility and
in the experiment alone. The added informa- blame, studies by Allen (1968) and Korte
tion may well have influenced the volunteer- (1969) suggest that diffusion does influence
ing behavior. helping behavior in some situations. In a
It should be apparent that the behavior situation similar to that of Darley and Latane
reviewed above is interpretable in a Sherif (1968), Korte (1969) found a greater inci-
(1947) and Asch (1956) type of conformity dence of helping for subjects who believed
paradigm. People behave altruistically in or- that the other members of their three-person
der to conform to group standards. group were strapped down to a table than
In the same way that students model the for subjects who did not. In the Allen (1968)
noncompliance of their fellow students, by- study, commuters in New York subways were
standers may model the failure of other by- given a chance to correct a stooge who gave
standers to respond to pleas for help. In both incorrect directions. The communters were
cases the reaction (or lack of reaction) of the least likely to correct a stooge when the re-
group serves as a standard of behavior. A quest for directions was directed to the stooge.
series of studies by Barley and Latane dem- They were most likely to correct the stooge
onstrated that the presence of others inhibits when they were asked for directions; and
altruistic responses in emergency situations. moderately likely to correct him when the
Appropriateness as a function of group request for directions was directed at a
standards—the modeling of inaction. Puzzled group of bystanders.
by the shockingly irresponsible behavior of In a second series of experiments (Latane
the 38 witnesses who remained behind their & Darley, 1968), bystanders were placed in
apartment doors while Kitty Genovese was emergency situations that contained a per-
being murdered (A. Rosenthal, 1964), Darley sonal threat to their safety. Male college stu-
and Latane ran a number of experiments on dents who found themselves in a room that
bystander intervention in emergencies. In the began to fill with smoke were more likely to
first study, Darley and Latane (1968) plased report the emergency when they were alone
college undergraduates alone in a room with in the room than when they were in the
an intercom and told them that it was pos- presence of two other subjects or two passive
sible for only one person in their group (of stooges. These findings do not fit well with
either two, three, or six) to be on the air the diffusion of responsibility and blame in-
at a time. After a short conversation, the sub- terpretation, due to the fact that it was the
jects heard a confederate appear to have an safety of the subjects themselves that was in
epileptic fit. It was the subjects who thought question. It would seem that part of the
that they were the only other ones present inhibiting effect of passive bystanders occurs
who responded most consistently and rapidly; because they supply, by their lack of concern,
and subjects who thought they were in three- an innocuous definition of the situation. Their
person groups responded more consistently passivity may, as a study by Latane (1967)
and rapidly than those who thought they suggests, even help to validate defensively
were in six-person groups. The investigators distorted perceptions of the situation. Sub-
274 DENNIS L. KREBS

jects whose tasks were interrupted by the diffusing responsibility and blame when the
tape-recorded sounds of a "bully" assaulting welfare of the subject himself was in question.
a young child accounted for their failure to Although most of the bystander-interven-
help by saying that they did not believe that tion studies found that bystanders were prone
the fight was real. The obvious inference— to help when they were alone, one study
that the tape-recorded sounds of the fight (Latane, 1967) found that virtually all un-
were unconvincing—was cast in doubt by the dergraduates failed to intervene in behalf of
further finding that subjects who were freed an assaulted child, even when they were alone.
from responsibility believed the fight was real. Similar results were obtained in a study by
Detracting responsibility, though, may have Kaufmann (1968). Only 11% of subjects who
changed the situation in other ways. The were asked to observe a "teacher" deliver
group which was freed from responsibility had what seemed to be increasingly severe and
no reason to doubt the reality of the fight— dangerous shocks to a "learner" (cf. Mil-
what purpose could it serve? A faked emer- gram, 1963) responded to the learner's pleas
gency would make more sense to subjects for help. No differences were found among
who were on the spot. conditions in which the bystander was given
A further study (Latane & Rodin, 1969) a position of high status, was led to believe
extended the generality of prior findings. he would later be the "teacher," or had rea-
Male undergraduates who heard a female son to doubt the legitimacy of the teacher's
"market research representative" appear to authority. People fail to help in some situa-
fall from a chair and hurt her ankle were tions, it would appear, on the slightest excuse.
more likely to help her when they were alone A developmental study by Staub and Fea-
than when they were in the presence of a gans (1969) found that children also failed
stranger or an unconcerned stooge. Subjects to help in emergency situations. Unlike their
who were paired with a stranger were more adult counterparts, though, they were less
likely to help the "representative" than sub- prone to help when they were alone than
jects who were paired with a passive stooge. when they were with another child. Nursery
Both Latane and Rodin (1969) and school, first- and second-grade children tended
Barley (1967) found that subjects who to help another child who appeared to fall
were paired with friends were more likely to and hurt himself more when they were in
help than those who were paired with strang- pairs than when they were alone. At the
ers. However, subjects who were paired with fourth and sixth grade the trend tended to
friends were not as likely to help as subjects reverse itself and the children behaved more
who were exposed to the emergency alone. In like their adult counterparts. It is possible
view of the fact that the passivity of subjects that the presence of a partner reduced the
to emergency situations is probably more ap- fear associated with doing something wrong
parent than real (they are upset but they in the younger children, but served as a
do not show it), the greater incidence of source of potential negative evaluation for the
helping with friends might be due to the older children.
fact that they are able to anticipate each Whatever the precise reasons, it seems
other's true reactions. Acting together, they likely that the behavior of bystanders is pre-
also have consensual support for their defini- dicted on their ignorance concerning the con-
tion of the situation. sequences of helping. If they were assured
Surveying all the studies in the bystander that no harm would come to them, their
intervention series, it would appear that both apparently heartless inaction would seem less
a diffusion of responsibility and blame, and
likely. And if they were assured that they
a group influence explanation are needed (cf.
Latane & Rodin, 1969). In the seizure situa- would be rewarded, helping behavior would
tion, where subjects were isolated from seem virtually certain. It is surprising, there-
one another, no information about group fore, that none of the studies which manipu-
reaction was supplied. Yet in the smoke-filled lated perception of consequences has achieved
room situation, little good could come from a clear effect.
 ALTRUISM 275

Modeling as a Function oj Perception oj their effect on the performance of altruistic

Consequences behavior, some investigators (e.g., M. Harris,
1968; Midlarsky & Bryan, 1967) analyzed
Models who supply information about the them in relation to the internalization of rela-
consequences of altruistic behavior almost al- tively permanent behavioral dispositions.
ways supply other information. Information
None of the studies, however, met the two
about the consequences of particular choices criteria of internalization—situational gen-
of action usually supplies an indication of its erality and longitudinal stability.
appropriateness. The observation of the act
also, of course, increases the salience of re- Several studies, though, have set out to
lated behavioral alternatives. examine the antecedents of modeling effects.
Instead of explaining imitative behavior with
Of the experiments which tested for the
influence of perception of consequences, one what is essentially a descriptive rubric, they
used adult subjects and two used children. have attempted to find out what it is about
In the study on adults, Hornstein et al., observing a model that results in the acquisi-
(1968) presented passersby with a wallet that tion of congruent behavioral dispositions.
had apparently been found by a model and Three general approaches have been taken.
then relost. Passersby who were led to believe The first examined modeling effects in terms
they were similar to the model were more of identification, the second examined them
likely to return the wallet when an attached in terms of secondary reinforcement, and the
note indicated that the model felt good or third examined them in terms of empathy.
neutral about returning the wallet than when
the note indicated he felt bad. Modeling and the Internalization of
In the first of the studies on children, Mid- Altruistic Dispositions
larsky and Bryan (1967) failed to find any Although several investigators have viewed
difference between the number of self-sacri- modeling in terms of identification (see Ban-
ficial and charitable responses of elementary dura & Walters, 1963), there has been only
school girls who observed a model who one study which examined the relationship
emitted expressive signs of joy after making between identification and altruism. Ruther-
self-sacrificial choices, and models who did ford and Mussen (1968) found that generous
not. M. Harris (1968), using an age group nursery school boys saw their fathers as more
similar to that of Midlarsky and Bryan warm and nurturant than less generous boys.
(1967), failed to find a difference between These findings were interpreted as support
the altruistic behavior of children who were for the hypothesis that generosity results from
exposed to models who were praised for their identification with warm and nurturant same-
altruism, and those who were not. Because sexed parents. Although the findings indicated
the praising agent was absent when the chil- that the generous boys saw their fathers as
dren performed, however, they had reason not warm and nurturant, no information was sup-
to expect similar consequences. plied about the generosity of the fathers, nor
None of the studies that examined the was the assumption that the boys were identi-
effect of perception of consequences found a fied with their fathers substantiated. More-
significant difference between positive and over, the hypothesis that identification-
neutral consequence conditions. The Horn- caused the boy's generosity was supported by
stein et al. (1968) study, though, found that only correlational data.
the perception of negative consequences had Mowrer (1960) suggested that rewarding
an inhibiting effect on altruism. It is possible models are imitated because stimuli that are
that the lack of similarity between observer associated with the performance of imitative
and model in the studies which used children responses, especially proprioceptive feedback,
made the consequences of the model's actions are secondarily reinforcing. Hartup and
irrelevant. Coates (1967) tested Mowrer's suggestion by
Although situations in which consequences exposing nursery children who had histories
can be observed were analyzed according to of frequent and infrequent peer reinforcement
276 DENNIS L. KREBS

to rewarding and nonrewarding peer models. expressive cues were exposed to the most
Both history of reinforcement and rewarding- warmth, and because they were given the
ness of model were determined by time-sam- most indication of what was desired of them,
pled observations of a nursery school class. interpretation of the results is difficult.
The altruism of rewarding peer models was A later study by Midlarsky and Bryan
imitated more by children with histories of (1967) controlled for pretask warmth and
frequent peer reinforcement, and the altruism added a condition in which the model failed
of nonrewarding peer models was imitated to emit expressive cues in the test situation.
more by children with histories of infrequent It also took a second measure of altruism in
peer reinforcement. No differences in attrac- a different situation. Although the investiga-
tiveness of model or social acceptance of sub- tors did not interpret the results as support
jects were found among groups. The results for the position of Aronfreed and Paskel, their
suggested that Mowrer's (1960) theory ap- results were very similar. Both studies found
plies only to subjects with histories of fre- that exposure to the EH condition resulted in
quent reinforcement. However, on the basis more self-sacrificial responses than exposure
of the positive correlation between giving and to E or H alone. Moreover, as Aronfreed and
receiving reinforcement from peers (Charles- Paskel would have predicted, more self-sacri-
worth & Hartup, 1967) the results can also ficial responses were made when the model
be interpreted as support for the notion that emitted expressive cues during testing than
subjects modeled the responses of those who when she did not. The findings, though, that
were similar to them. The latter interpretation self-sacrificial responses occurred in condi-
is consistent with the Hornstein et al. (1968) tions other than the HE and EH conditions,
findings. However, several other studies, and that most of the girls who sacrificed
which did not control for history of reinforce- candy also donated to charity in the absence
ment (e.g., Grusec & Skubiski, in press; Ros- of the model, suggest that more was involved
enhan & White, 1967), failed to support the than the empathic transmission of positive
notion that rewarding models are imitated affect.
more than nonrewarding models. There has been no concrete support for
The results of a study by Aronfreed and positions that view modeling as a function
Paskel (1968) on first- to third-grade girls of identification, secondary reinforcement, or
can be interpreted in accord with a secondary empathy. Studies which appear to support
reinforcement position. Aronfreed and Paskel each position have alternative interpretations.
(1968) found that children who were exposed Moreover, although all of the studies were
to a female model, who emitted both ex- concerned with the internalization of altruistic
pressive signs of joy and hugs (EH) after dispositions, none established the situational
she made self-sacrificial responses, evidenced generality and longitudinal stability of the
more imitation than children who were ex- effects. Some researchers would argue that the
posed to expressive cues (E) or hugs (H) two criteria are beyond the range of labora-
alone. If the model who emitted expressive tory studies. What is needed, perhaps, is
cues and affection responses is seen as the supplementary evidence from naturalistic
most rewarding of the three models (and the studies. If, for example, correlates of altruism
fact that she was the only one who dis- could be found in the behavior of parents or
played pretask nurturance makes this likely), friends, relevant experimental analogues would
then the modeling effect can be interpreted as receive support. Such correlates have been
a function of the secondary reinforcement reported in three studies. Rosenhan (1967)
effects of imitation. found that a group of active civil rights
Aronfreed and Paskel (1968) interpreted workers had a close relationship with at least
their results differently—as support for the one altruistic parent; Tomkins (1965) re-
notion that the self-sacrificial responses of the ported that prominent abolitionists were in-
girls were reinforced by empathically experi- fluenced by altruistic friends; and Rettig
enced positive affect. Because the girls who (1956) found that altruism in parents was
were exposed to both affective responses and positively correlated with scores made by
 ALTRUISM 277

college students on an altruism scale. Some —dependency and interpersonal attractive-

evidence was found in the Rettig (1956) ness—have been studied in any detail.
study for the notion that the effects of direct A close consideration of the effect of inter-
reinforcement of altruistic behavior by par- personal attractiveness on altruism suggests
ents dissipates over time, whereas modeling that many altruistic responses result not from
effects are more enduring. characteristics of benefactors or recipients
Although more precise studies are needed, alone, but rather from an interaction between
the results of studies which have found a the characteristics of benefactors and recipi-
relationship between attributes of parents ents. Recipients, for example, may be attrac-
and altruism in children are encouraging, es- tive to different benefactors for different rea-
pecially in view of the failure of most studies sons, and benefactors may differ in attracta-
to find a relationship between moral behavior bility. In the modeling studies, almost all re-
in parents and personality correlates in chil- cipients were dependent, and their dependency
dren (cf. Kohlberg, in press). probably interacted with the modeling effects
to elicit altruism. Even though characteristics
Experimenter Effects of both recipients and benefactors influence
An appropriate way of ending the review altruism, they can be separated for classifi-
of modeling and altruism is to draw attention cation due to the fact that most experiments
to the relationship between modeling and ex- vary one set of characteristics and hold the
perimenter effects. R. Rosenthal (1966) and other constant.
others have demonstrated that experimenters
unintentionally effect the responses of their Dependency of the Recipient
subjects, even when they do not perform The essential attribute of a recipient of
the tasks in question. Modeling can be con- altruism is his perceived need for aid. In
sidered a type of experimenter effect—it almost all of the reported studies on altruism,
causes subjects to act in ways congruent with altruistic behavior was elicited by a dependent
the behavior (and, usually, expectations) of other. Most studies held dependency con-
models. The similarity between modeling and stant, or allowed it to vary randomly. Some
experimenter effects has three implications. studies, however, gave specific attention to
The first emphasizes the suggestion that the effect of dependency.
modeling effects are temporary and situation-
A set of ten studies (see Table 4) by
specific. The second emphasizes the subtlety Berkowitz and his colleagues found a con-
of modeling effects—a decade of research has sistent effect for dependency. All studies used
been unable to isolate the means by which the same basic experimental design. In the
expectations are communicated in experi- standard situation, subjects were recruited for
ments. And finally, the necessity for methodo- an experiment on supervisory ability. A
logical caution in modeling studies becomes "worker" was required to construct paper
obvious. Most studies attribute modeling ef- boxes or envelopes for a "supervisor." In the
fects to more than experimenter bias. In spite high dependency condition, the worker was
of this, few studies have controlled for ex- told that the supervisor's chance of winning a.
pectations of the model. prize depended on the worker's productivity.
In the low dependency condition, the worker
ALTRUISM AS A FUNCTION OF TEMPORARY
was told that it was the quality of the super-
STATES OF THE RECIPIENT
visor's instructions that would determine his
When it is the temporary states of recipi- reward. The measure of altruism was either
ents of altruistic responses that are studied, the number of boxes constructed in the ex-
the question that becomes relevant is what perimental session, or the difference between
kinds of states, and what situations associated the number constructed in the experimental
with these states, elicit altruistic responses. session and the number in a practice ses-
The range of recipient state variables that sion. Berkowitz and his colleagues (see Table
has been studied is small. Only two variables 4) consistently found that more boxes were
278 DENNIS L. KREBS

TABLE 4
STUDIES WHICH DEMONSTRATED THE EFFECT OF DEPENDENCY OP THE RECIPIENT ON ALTRUISM

Subjects
Author and date Relationship between dependency of the recipient and altruism
Sex Age N

Berkowitz & Daniels (1963)

Experiment 1 M cs 80 More produced for high vs. low dependent supervisor; no
main effect for awareness; no main effect for amount of
reward.
Experiment 2 M cs 32 More produced for high vs. low dependent supervisor; less.
produced in Low Awareness X Low Dependency condi-
tion.
Daniels & Berkowitz (1963) M cs 80 More produced for high vs. low dependent supervisor; less in
Low Awareness X Low Dependency; more in High Liking
X High Dependency.
Berkowitz & Daniels (1964) F cs 80 More produced for high vs. low dependent supervisor; no
main effect for awareness.
Berkowitz et al. (1964) M/F cs 160 More produced for high vs. low dependent supervisor; no
main effect for awareness; no effect for sex of subject or
sex of experimenter.
Berkowitz & Conner (1966) M cs 108 More produced for high vs. low (but not medium) dependent
supervisor; more produced in High Dependency X Success
than High Dependency X Failure.
Berkowitz (1966) — — — Study later elaborated by Berkowitz & Friedman (1967) and
Berkowitz (1968) ; see below.
Goranson & Berkowitz F cs 89 (High dependency condition only) More produced after vol-
(1966) untary vs. compulsory help for same (vs. different) person;
least produced after refused help.
Berkowitz & Friedman M 13-16 345 (All in high dependency condition) ; sons of bureaucrats pro-
(1967) duced more than working-class boys; sons of entrepreneurs
produced less after receiving low (vs. high) help from an-
other.
Berkowitz (1967) M/F CS — (1): More produced for externally-caused high dependency.
(2) : More produced for highly dependent others. (4) : Sex-
differences in high and low dependency condition.
Berkowitz (1968) M 13-16 192 (All in high dependency condition) ; working-class boys (vs.
bureaucratic boys) who received high (vs. low) prior help
produced more for the same person.
Schopler & Bateson (196S) M/F CS 196 Males gave more to lowly dependent (vs. highly dependent)
recipients, and females gave more to highly dependent
recipients, when cost of helping was low.
Wheeler & Wagner (1968) M adult 144 More tended to be given in order to help fly relatives to a
dying sailor than to help build up a serviceman's fund."
Midlarsky (1968a) M CS 80 More shock-contingent problems were solved for high (vs.
low) dependent recipients.

Note.—Abbreviations are: CS < • college student.

"P < .10.

built by workers in the high dependency con- It might be argued that the altruism shown
dition than workers in the low dependency by the workers was a function of expectations
condition. One study (Berkowitz & Conner, of approval or fear of disapproval. To test
1966) created three levels of dependency— these possibilities, Berkowitz and Daniels
the supervisor was dependent on the worker (1963) and Daniels and Berkowitz (1963)
for 20%, 50%, or 80% of the points that added an "awareness" condition. Workers in
could earn him a cash prize. Workers whose the high awareness condition were told that
supervisors were 80% dependent made more their supervisors would be informed of their
paper envelopes than those whose supervisors productivity during the experiment, and those
were 20% dependent. in the low awareness condition were told
 ALTRUISM 279

that their work would not be examined until press) failed to find an effect for dependency.
after a month. In no case was a main effect A posttest questionnaire, however, revealed
found for awareness. However, a combina- that the dependency manipulation had failed.
tion of (low) dependency and (low) aware- In a situation similar to that of Test and
ness was associated with low production. Bryan, Midlarsky (1968a) found that more
Moreover, production also failed to increase help was given to a partner with broken eye-
when the experimenter's awareness was ma- glasses than to a. less dependent recipient,
nipulated (Berkowitz, Klanderman, & Harris, even though helping involved the receipt of
1964). Berkowitz interpreted the lack of dif- electric shocks.
ference between the productivity of the high It should be pointed out that the depen-
and low awareness groups in the high de- dency manipulations in the Schopler and
pendency condition as support for the as- Bateson (1965), Wheeler and Wagner (1968),
sumption that his workers were not motivated Test and Bryan (in press), and Midlarsky
toward the attainment of approval or avoid- (1968a) studies differed from those in the
ance of punishment. It is also possible that Berkowitz studies. In the Berkowitz studies
the low awareness manipulation was ineffec- the supervisor was specifically dependent on
tive. Subjects may have found it difficult to the worker—if the worker did not work hard,
believe that their output would remain un- the supervisor did not get rewarded. In the
examined for a month. Schopler and Bateson (1965, Experiment 1)
The notion that dependency elicits helping and Wheeler and Wagner (1968) studies, on
behavior in the Berkowitz type of situation the other hand, the dependency of the other
has not received full support from other was general—each subject could have told
studies (outlined in Table 4). Schopler and himself that if he did not help, someone else
Bateson (1965, Experiment 2 and 3) and would. The Midlarsky (1968a) and Test and
Schopler (1967) found that females yielded Bryan (in press) studies fall in between. Sub-
more money to a partner when he was in a jects were not asked to help the dependent
state of high (versus low) dependency, but other, but they were the only ones who could
only if the cost of yielding was low. Males help him.
in the low cost of yielding condition, on the Before turning to studies which manipu-
other hand, yielded more money when their lated interpersonal attractiveness of the re-
partner was in a state of low dependency. cipient, it should be mentioned that depen-
In a different situation, Schopler and Bateson dency need not be viewed as a unitary vari-
(1965, Experiment 1) found that although able. Studies by Schopler and Matthews
females were more inclined to volunteer to (1965) and Horowitz (1968) demonstrated
help a student finish his thesis when he was that internally caused dependency (depen-
"desperate," males were more inclined to dency caused by the subject) tended to elicit
help when he had a year to work on it (p less altruism than externally caused depen-
< .10). The measure of altruism in the Scho- dency. Locus of dependency, though, seems
pler studies seems more powerful than that best examined as a trait variable.
in the Berkowitz studies because of the mate-
rial sacrifice involved. The more the bene- Interpersonal Attractiveness of the Recipient
factor yielded, the less he kept for himself. Common sense would predict that more is
Other studies which involved a variety of given to liked others than disliked others.
situations, though, have supplied support for Because the prediction seems so obvious, per-
the notion that dependency in the recipient haps, only three studies on altruism (Daniels
elicits altruism. Wheeler and Wagner (1968), & Berkowitz, 1963; Epstein & Hornstein,
for example, found that Navy men were more 1969; Staub & Sherk, in press) have ma-
likely to donate money when they were ex- nipulated interpersonal attractiveness as a
posed to a personal appeal which involved a main independent variable. The apparent
highly dependent family than one which in- paucity of research, however, may be mis-
volved a condition of low dependency (p leading. It is possible that interpersonal at-
< .10). A study by Test and Bryan (in tractiveness has exerted an unrecognized in-
280 DENNIS L. KREBS

TABLE 5
STUDIES WHICH MEASURED THE EFFECT OF INTERPERSONAL ATTRACTIVENESS OF THE
RECIPIENT ON ALTRUISM

Subjects
Author and date Relationship between attractiveness and altruism
Sex Age N

Daniels & Berkowitz (1963) M cs 80 Greatest production for highly dependent well-liked super-
visors.
Staub & Sherk (in press) M/F 9 94 Crayon shared longer with preferred partners.
Epstein & Hornstein (1969) M CS 60 Liked recipients helped most when selfishness punished
(vs. not).
Berkowitz & Friedman (1967) M 13-16 345 Partners who gave high (vs. low) prior help rated as more
likable and received more help, from entrepreneurial but
- not bureaucratic boys.
Schopler & Matthews (1965) M CS 48 Externally dependent subjects rated in more attractive
terms and helped more.
Brehra & Cole (1966) M CS 60 Subjects who did appropriate favor (vs. no favor) rated
as more friendly (but not as generally more attractive)
and received more help.
Kiesler (1966) M HS 120 Partners who did not share after a cooperative (vs. com-
petitive) game rated as unattractive.
Schopler & Thompson (1968) F CS 38 "Salesman" who gave flower in appropriate (vs. inappro-
priate) circumstances rated as more generous (but not
as more attractive).
Walster & Prestholdt (1966) F CS 88 Person who was rated too harshly in low commitment (vs.
high commitment) condition was subsequently rated
as more attractive; he also tended to elicit more help.
Lerner & Lichtman (1968) F CS 140 Partners who performed illicitly (vs. legitimately) gracious
act were rated as less attractive and helped less. At-
tractiveness and helping did not go together in other
conditions.
Lerner & Matthews (1967) F CS 66 When subjects draw placed him in a control and his part-
ner in a shock condition (vs. fates independent) partner
was rated as less attractive, but was comforted more.
Lerner F CS 61 When subjects (vs. experimenter) caused their partners
to serve in shock condition, partners were rated as less
attractive, yet subjects were more prone to take their
place.
Berkowitz & Daniels (1963) M CS 32 High dependent (vs. low dependent) supervisors were
helped more, but liked less.

fluence in other studies. Studies that manipu- them than when they thought they would not.
late characteristics of the recipient usually Staub and Sherk (in press) found that fourth-
affect the attractiveness of the recipient. grade children shared a crayon longer with
Table 5 contains an outline of studies which liked than with disliked partners.
investigated the effect of interpersonal at- In the third study, Epstein and Hornstein
tractiveness on altruism. (1969) found a more complex relation-
Interpersonal attractiveness as an indepen- ship between liking and altruism. Sub-
dent variable. Only three studies have focused jects who liked their partners made fewer
on the altruism-eliciting effect of attractive- selfish responses than subjects who disliked
ness of the recipient. Daniels and Berkowitz their partners when they were punished for
(1963) told "workers" that a questionnaire their selfishness by a third person. When they
revealed they would either like or dislike were not punished, however, they made fewer
their "supervisors." As expected, the workers selfish responses for a disliked partner. Al-
made more boxes for highly dependent super- though the results are difficult to interpret,
visors when they thought they would like it is possible that punishment from a liked
 ALTRUISM 281

other served to remind the subjects of the about attractiveness. Walster and Prestholdt
harm their acts did. Punishment from a dis- (1966) found that subjects who increased
liked other, on the other hand, may have their attractiveness ratings of targets in order
antagonized them into more selfish behavior. to compensate for an unfair rating or justify
Interpersonal attractiveness as a mediating a high rating tended to be more likely to
variable. It seems likely that most studies volunteer to help the target person. Unfor-
which manipulated characteristics of the re- tunately, though, the imminence of summer
cipient incidentally varied his interpersonal vacation and final exams truncated their sam-
attractiveness. Recipients are usually more or ple to the point that statistical analysis of
less attractive depending on their association the relationship between attractiveness and
with moral transgression or need for psycho- volunteering was not feasible.
logical help (Bryan & Davenport 8 ; Nun- Although there are suggestive indications
nally, 1961), their race and nationality that attractiveness mediates altruism, the re-
(Bryan & Test, 1967; Feldman, 1968), the lationship is surprisingly weak. Several
legitimacy of their need (Frisch & Green- studies, in fact, have found a negative rela-
berg, 1968; Horowitz, 1968; Schopler & tionship between helping and attractiveness.
Matthews, 1965), and the amount of prior Lerner and Matthews (1967) and Lerner (see
help attributed to them (Pruitt, 1968). Sev- Footnote 4), for example, found that subjects
eral of the Berkowitz studies (e.g., Berkowitz who perceived themselves (versus their part-
& Friedman, 1967) found incidental relation- ner or an experimenter) as responsible for the
ships between the amount of help given to suffering of another tended to devalue the
supervisors and their rated attractiveness. other in order to preserve their belief in a
Other studies, though (e.g., Berkowitz & just world. In spite of the devaluation,
Daniels ,1963), found a negative relationship though, they were more willing than those
between helping and attractiveness. who did not devalue their partners to take
Research on the effect of inappropriate his place in a shock condition. It is possible
favors on altruism demonstrates that although that in cases where altruism is reparative or
recipients who are helped tend to be seen as part of a role requirement, it is not given as
more attractive than those who are not, the much in behalf of the recipient as in spite of
reverse is sometimes true. Kiesler (1966) him.
found that partners who did appropriate
favors were rated as more attractive than PERSONALITY TRAITS OF THE BENEFACTOR
those who did inappropriate favors, but no —CORRELATIONAL STUDIES
measure of altruism was taken. Brehm and
Cole (1966), Lerner and Lichtman (1968), Research which has dealt with personality
and Schopler and Thompson (1968) found traits of benefactors differs from most of the
that inappropriate favors elicited less al- research reviewed thus far because it is con-
truism than appropriate favors. They also cerned with natural correlations rather than
found that recipients who had done appropri- experimentally induced relationships. Trait-
ate favors tended to be rated more positively. oriented correlational studies (see Table 6)
But in every case, the recipients failed to be have used three different criteria of altruism.
rated as better liked. Lerner and Lichtman Some have defined altruism according to the
(1968), however, found a positive relationship ratings of others. Some have used scores on
between selfishness and unattractiveness—re- pencil-and-paper tests; and some have used
cipients who seemed "illicitly gracious" were behavioral measures. Personality variables in
rated as unattractive and were helped little. each of the three categories have been drawn
A final study, which set out to examine a from several different sources. Trait-oriented
dissonance effect, supplied some information
correlational studies attempt to find out what
8
Bryan, J. H., & Davenport, M. Donations to the personality traits and syndromes are typical
needy: correlates of financial contributions to the
destitute. (Research Bulletin No. 68-1) Princeton, of altruists, and, in general, what kind of
N. J.: Educational Testing Service, 1968. people altruists are.
282 DENNIS L. KREBS

TABLE 6
STUDIES WHICH EXAMINED THE RELATIONSHIP BETWEEN PERSONALITY TRAITS OF THE
BENEFACTOR AND ALTRUISM
Subjects
Author and date Source of personality Positive correlations Negative correlations
traits with altruism with altruism
Sex Age N

Studies which used rating-scale measures of altruism

Turner (1948) M 9-16 116 Ratings of parents Self-adjustment, Antisocial tendencies.

and social workers. grasp of social
standards, social
skills, good com-
munity relations,
adjustment, emo-
tional stability,
ethical goodness.
Cattell & Horowitz F CS 60 Dormmates' ratings, Cyclothymia Paranoic-schizoid.
(1952) 16 PFQ, "objec-
tive tests."
Friedrichs (1960) M CS 280 Dormmates' ratings, Attractiveness as a Economic involve-
self-ratings, ques- friend, political ment.
tionnaires. conservatism, au-
thoritarianism,
theism, sociability,
ingroup involve-
ment.
MacDonald F CS 19 Self-ratings; ques- Need for nurturance, Economic values,
tionnaires. need for autonomy, political values.
social values, reli-
gious values.

Studies which used pencil-and-paper test measures of altruism

Friedrichs (1960) M CS 280 Self-ratings, ques- Church attendance, Ethnocentrism, neu-

tionnaires. theism. roticism.
Ribal (1963) M/F CS 194 Edwards Personal Need for endurance Need for achieve-
Preference Scale (males) ; needs for ment and
affiliation and dominance.
interception (fe-
males).
Saywer (1966) M CS 122 Questionnaire, occu- YMCA orientation —
pational aspiration.

Studies which used behavioral measures of altruism

Rutherford & Mussen M 4-6 31 Doll-play ratings; Dependency, kind- Hostility, competi-
(1968) racing game; teach- ness". tiveness, gregari-
er's ratings. ousness, quarrel-
someness," aggres-
siveness".
Staub & Sherk M/F 9 94 Need for approval (Need for approval),
(in press) questionnaire. activity.
Gore & Rotter (1963) M/F CS 116 Locus of control Internal locus of
questionnaire. control.
Midlarsky (1968a) M CS 80 Locus of control Internal locus of
questionnaire. control.
Staub (1968) M/F 9-10 196 Locus of control Internal locus of Internal locus of
questionnaire. control (following control (following
success). failure).
*t < .10.
 ALTRUISM 283

Rating-Scale Measures of Altruism Pencil-and-Paper Measures of Altruism

Four studies have defined altruism accord- Three studies defined altruism according to
ing to the ratings of acquaintances. Turner scores on pencil-and-paper tests. Friedrichs
(1948), using boys from the Cambridge- (1960) found low positive correlations be-
Somerville Youth Study, found that altruism tween a "Likert-type questionnaire" measure
ratings of parents, teachers, and social work- of altruism and measures of theism and church
ers correlated most highly with ethical good- attendance; and low negative correlations for
ness, emotional stability, and social adjust- ethnocentrism and neurotic symptomology.
ments. Cattell and Horowitz (1952) found Ribal (1963) found that altruistic females
that the factor cyclothymia (social extraver- had high needs for affiliation and intraception,
sion) was most highly correlated with al- and low needs for achievement and domi-
truism in college females. Friedrichs (1960) nance, as defined by the Edwards Personal
found that fraternity members who were Preference Schedule (Edwards, 1954). Al-
rated as most "attractive as a friend" were truistic males scored high on need for en-
considered most altruistic. Factors which re- durance. Finally, Sawyer (1966) found that
lated to altruism were conservatism, socio- YMCA students were more altruistic than
economic security, and social acceptability. social science students or business school stu-
And finally, a group of Radcliffe girls who dents, and that altruism did not relate to
were rated as altruistic by their dormmates authoritarianism.
(McDonald 6 ) scored high on social and re- The results of all three of the studies that
ligious values (and low on economic and used pencil-and-paper measures of altruism
political values) as measured by the Allport- can be questioned. Friedrichs' (1960) Likert-
Vernon Study of Values (Allport, Vernon, & type questionnaire failed to correlate highly
Lindzey, 1960). with dormmates' and self-ratings of altruism.
There are problems with each study which In the Ribal (1963) study, subjects who
used a rating-scale measure of altruism. The scored in the upper quartile of nurturance and
content of Turner's (1948) Altruism Scale the lower quartile of succorance on the Ed-
indicated that it was actually more a measure wards Personal Preference schedule were
of ethical goodness than altruism. A direct called altruistic. McDonald (see Footnote 6),
rating of ethical goodness was more highly though, found that while the girls who were
correlated with the scale than a direct measure judged altruistic by their dormmates scored
of altruism. In the Cattell and Horowitz high on need for nurturance, they did not
(1952) study, different dormmates rated dif- score low on need for succorance. Ribal's
ferent girls, which allowed for an unequal (1963) measure of altruism, then, did not
distribution of response sets. In the Fried- correspond to McDonald's rating-scale mea-
richs (1960) study, the author himself pointed sure. (It is interesting to note that in accord
out that "most of the surface correlations with Ribal's findings, McDonald's females
[were] so small that they must be viewed as scored low on needs for achievement and
no more than suggestive [p. 507]." Because dominance. They failed, however, to score
of her small sample, most of McDonald's (see high on needs for affiliation and intraception.)
Footnote 6) findings were also only suggestive. Finally, the Sawyer (1966) study can be
criticized due to the fact that it defined al-
Moreover, none of the studies controlled for
truism as behavior which maximized gain for
favorability sets and halo effects. The fact
self and other—behavior that seems more
that altruism seemed to correlate most highly adequately defined as cooperative.
with sociability may mean little more than In most cases, pencil-and-paper tests are
that people gratuitiously rate others as al- thinly disguised self-rating measures with no
truistic because they are friendly. check for validity. The general dangers as-
6
MacDonald, S. L. Altruism: A study in means
sociated with pencil-and-paper measures of
and ends. Unpublished honors thesis, Harvard Uni- altruism can, perhaps, be exemplified by men-
versity, April 1966. tioning a controversy associated with the So-
284 DENNIS L. KREBS

cial Responsibility Scale (SRS) (Berkowitz two favorite classmates. Teachers rated the
& Daniels, 1964)—a scale that was meant as generous boys as more generous, less gregari-
a measure of altruism. Stone (196Sa) found ous, and less competitive than the others. A
that the SRS correlated highly with three racing-game measure of competitiveness sup-
measures of social desirability. The finding ported the final rating. Generous boys were
that people who give socially desirable answers also rated as less quarrelsome and aggressive,
receive high altruism scores casts doubt on and more kind (p < .10). The authors in-
the validity of the scale. Although it is en- terpreted their findings as support for the
tirely possible that altruists are the type of notion that generosity is part of a pattern of
people who do what is socially desirable moral characteristics that includes altruism,
(Berkowitz, 1965), the SRS scale does not kindness, and cooperation. The fact that the
establish the relationship (Stone, 196Sb). investigator was present while all behavioral
Scales such as those of D. Harris (1957) for measures were taken (the child in fact played
children and Gough, McClosky, and Meehl the racing game against the investigator)
(1952) for adolescents are subject to the raises the possibility that the altruism of the
same criticism. generous children was due to conformity, or
need for approval.
Behavioral Measures of Altruism A study by Staub and Sherk (in press) sup-
Studies which correlated personality traits plied a behavioral measure of altruism which
with rating-scale and pencil-and-paper mea- was less amenable than that of Rutherford
sures of altruism have generally found a large and Mussen (1968) to alternate interpreta-
number of relationships. Studies which cor- tion. Children in the Staub and Sherk (in
related personality traits with behavioral mea- press) study were not asked to give, and the
sures of altruism, on the other hand, have experimenter was absent when they gave.
generally been less productive. The difficul- Children who scored high on an adapted
ties associated with studies which use be- Crowne-Marlowe need for approval scale
havioral measures arise largely because of the (Crandall, Crandall, & Patkovsky, 1965)
uncontrolled effect of situational variables. As tended to give less than those who scored
was demonstrated in the review of modeling low. Because high need-for-approval children
studies (and pointed out in relation to other also tended to eat fewer candies themselves,
moral behaviors by, for example, Hartshorne though, their "selfishness" was uncertain.
and May, 1928), behavioral indexes of al- Other studies have found a relationship be-
truism are largely situation-specific. It would tween behavioral measures of altruism and
seem that the more unusual the situation, the locus of control (as defined by Rotter, 1966).
less the effect of personality variables. Such, Gore and Rotter (1963) found that students
at least, seemed the case in the study by from a Southern college were more willing to
Darley and Latan£ (1968). Subjects who re- help in a civil rights project if they thought
sponded to emergency pleas for help did not they had internal control over their fates than
differ from those who failed to respond on if they thought their fate was externally con-
measures of Machiavellianism—anomie—au- trolled. Midlarsky (1968a) found a positive
thoritarianism, need for approval, or social correlation between helping (at the cost of
responsibility. In a similar situation, Korte receiving shocks) and internal locus of con-
(1969) failed to find a relationship between trol. Staub (1968) found that fourth-grade
helping in an emergency and deference, au- children who were characterized by high in-
tonomy, or submissiveness. ternal control shared more after an experience
Other studies which have employed less of success, and less after a neutral experience
unusual situations have found relationships or experience of failure than those charac-
between personality traits and behavioral mea- terized by low internal control.
sures of altruism. Rutherford and Mussen Although more faith can generally be put
(1968), for example, differentiated high and in behavioral measures of altruism than rat-
low generous nursery school boys on the basis ing-scale or pencil-and-paper measures, they
of the number of candles they left for their also have problems. Difficulties associated
 ALTRUISM 285

with operational measures of altruism have example, found that subjects who were judged
already been reviewed. The problem, basically, as deserving of undesirable fates failed to
is that the intent behind apparently altruistic elicit as much help as those who were not. In
acts is not easily determined. a different context, Schopler and Matthews
Considered as a whole, no general conclu- (196S) and Berkowitz (1967) found that de-
sions can be drawn about personality traits pendency that was thought to have been
of benefactors. The variations in subject sam- caused by external forces (forces beyond the
ples, independent variables, and measures of control of the recipient) elicited more al-
altruism make interstudy comparison difficult. truism than internally caused dependency. Ex-
Yet, in spite of the fact that trends must be ternally caused dependency was also seen as
considered only suggestive, there is some con- more legitimate than internally caused de-
sistency among findings. On the basis of the pendency. Bryan and Davenport (see Foot-
available evidence, altruistic children seem to note 5 ) , though, failed to find a difference be-
be better adjusted socially than others—they tween contributions to needy cases that were
are less aggressive, quarrelsome, and com- seen as internally caused and those that were
petitive, and they are more emotionally stable. seen as externally caused.
They are not, however, more intelligent than Although a study by Horowitz (1968)
others, nor do they achieve more. They tend offered some additional support for the notion
to be relatively shy and unassertive. that externally caused dependency elicits more
College-age female altruists are socially altruism than internally caused dependency, it
oriented—they are cyclothymic and have so- also found that the reverse is sometimes true.
cial (versus political or economic) values. Undergraduates who were given little choice
They are nurturant people with low needs but to help a graduate student with his PhD
for achievement and dominance. College-age thesis, gave more help when the graduate
male altruists also tend to be socially student's need was perceived as externally
oriented; they are free from neuroticism, and caused. Because the internal locus of de-
tend to think they control their fates. They pendency manipulation described the solicitor
are well liked by others, slightly on the con- as "characteristically dependent and weak,"
servative side, and may tend to be authori- it is likely that his dependency was seen as
tarian. Although they are not more liable to a personality trait. It was also found that
be in the social sciences than business or en- subjects who had a lot of choice about help-
gineering, they are more likely to train for ing volunteered more time to the internally
theYMCA. dependent other. It is possible that people
who are seen as characteristically dependent
ALTRUISM-ELICITING TRAITS OF RECIPIENTS do not threaten the status of others, and,
There has been little research on altruism- therefore, elicit more help. A study by
eliciting traits of recipients. Variables cate- Schopler (1967) found that male undergrad-
gorized as temporary states of recipients often uates (the same subject sample used by
relate to personality traits. Situational de- Horowitz, 1968) failed to help a highly de-
pendency and interpersonal attractiveness of pendent other when he was threatening to
the recipient, for example, may be indicative their status. Female undergraduates, on the
of more stable personality characteristics. other hand, helped the dependent other.

Locus of Dependency Other Characteristics

The characteristics of recipients that have Attributes that affect the interpersonal at-
been found to influence altruism most are tractiveness of recipients may also relate to
those which relate to the legitimacy of their their personality traits. Subjects in the Lerner
need for help. Although situational depen- and Lichtman (1968) study who perceived
dency generally elicits altruism, dependency others as making illicitly gracious responses,
that is perceived as illegitimate often has the for example, probably attributed different
reverse effect. Lerner (see Footnote 4), for personality traits to them than to the "gen-
286 DENNIS L. KREBS

uinely gracious" others. Information about found the reverse. Rosenhan and White (1967)
the needy cases in the Bryan and Davenport found a suggestive tendency for fourth- and
study (see Footnote S), the race of the fifth-grade boys to give more than girls after
solicitors in the Bryan and Test study (1967), prior contact with a female model when the
and the nationality of those who asked for model was absent during the giving period.
help in the Feldman study (1968) also prob- White (1967), however, found that fourth-
ably affected judgments about the personality and fifth-grade girls gave more than boys after
traits of recipients, which, in turn, probably a few days delay, especially when they both
influenced their ability to elicit altruism. observed an altruistic model and rehearsed
altruistic responses. Bryan and Walbek (1969,
SOCIAL ROLES AND DEMOGRAPHIC Experiment 2, 3) found that third- and
ATTRIBUTES OF THE BENEFACTOR fourth-grade boys (but not girls) gave more
after they were exposed to a charitable (versus
Social roles and demographic attributes nondonating) same-sexed peer model. Grusec
such as sex, age, ordinal position, social class,
and Skubiski (in press), however, found that
and nationality have usually been studied as
girls donated more than boys after exposure
incidental correlates of altruism. Although re-
to a charitable model, especially when the
lationships between demographic variables and
model was nurturant, and preached charity.
altruism supply hints concerning the anteced-
Finally, Staub and Sherk (in press) found
ents of altruism, the relationships are usually that fourth-grade boys shared more candy
difficult to interpret. The fact that people of
than fourth-grade girls. The boys, however,
different sex, ordinal position, social class, and also tended to eat more candy in front of
nationality share a large number of traits
their partners. It is apparent that there are
makes specification of precise antecedents dif- no clear trends in the conditions which affect
ficult.
sex differences in altruism in children.
Contradictory findings also occur when rec-
Sex Differences in Altruism
iprocity is the measure of altruism. Floyd
Sex differences in the altruism of children. (1964) found that more first- and third-grade
Of the 17 studies that examined children of girls than boys increased the amount given
both sexes, 11 found no sex differences. Of the to a partner after receiving a lot from him.
studies which found sex differences, only 2 Staub and Sherk (in press), however, failed
(Grusec & Skubiski, in press; White, 1967) to find a difference in reciprocity between
reported main effects that approached signifi- fourth-grade boys and girls, although the
cance ( ^ < . 1 0 ) . The remaining studies re- boys were more liable to reciprocate selfishness
ported only interaction effects. Although the than girls.
common notion that females are more al- Sex differences in the altruism of adults. As
truistic than males was supported more often was the case with studies on children, most
than the reverse, support was by no means studies on adults have failed to find sex dif-
consistent (see Table 7). ferences in altruism (e.g., Berkowitz et al.,
No studies found sex differences in altruism 1964; Blake et al., 19SS; Bryan & Test,
for nursery school children (Fischer, 1963; 1967; Hornstein et al., 1968; Rosenbaum &
Gewirtz, 1948; Hartup & Keller, 1960; Blake, 19SS). A series of studies by Schopler
Murphy, 1937). Of the studies which used and his colleagues, though (Schopler, 1967;
elementary school children of both sexes, Schopler & Bateson, 1965; Schopler & Mat-
those of Handlon and Gross (19S9), L. Har- thews, 1965) found consistent sex differences
ris (1967), M. Harris (1968), Shure (1968), in low-cost dependency-elicited helping. Schop-
Staub (1968), Ugurel-Semin (1952), and ler and Bateson (1965, Experiment 1) found
Wright (1942) failed to find sex differences that although females were most likely to
in altruism. volunteer for an unpleasant experiment when
In the other studies which used elementary the solicitor was highly dependent, males
school children, some found a greater incidence were most likely to volunteer when he was of
of altruism in boys than girls, and some low dependency. Two further experiments
 ALTRUISM 287
(Schopler & Bateson, 1965, Experiments 2, A final study (Schopler, 1967) tested the hy-
3) supported the findings in situations similar pothesis that males were reluctant to help a
to those used by Berkowitz and his colleagues, highly dependent male because he threatened

TABLE 7
STUDIES WHICH FOUND SEX DIFFERENCES IN ALTRUISM

Subjects
Author and date Independent variables Dependent variables
Age N

Studies in which sex differences occurred that favored males

Rosenhan & White (1967) 9, 10 130 Prior contact with model—model Frequency of donating to
absent. orphans.
Bryan & Walbek (1969)
Experiment 2 8,9 168 Observation of charitable vs. Frequency of donating to March
nondonating model." of Dimes.
Experiment 3 8,9 132 Observation of charitable vs. Same as above.
greedy model."
Staub & Sherk (in press) 9 94 Need for approval ; reciprocity." Number of candies shared.
Schopler & Bateson (1965)
Experiment 1 CS 98 (Main effect) dependency. Frequency of volunteering for an
unpleasant experiment.
Experiment 2 CS 88 Low dependency, low cost of Frequency of yielding possibility
helping.*'1" to win money.
Experiment 3 CS 56 Same as above." Same as above.
Berkowitz (1967) CS — Same as above." Amount produced for supervisor.

Studies in which sex differences occurred that favored females

White (1967) 10, 11 210 (Main effect0) consistency of al- Number of gift certificates
truism over time; observation donated to charity.
of charitable model plus re-
hearsal of charity; other con-
ditions."1
Grusec & Skubiski (in press) 9,11 80 (Main effect") ; prior nurturance Number of marbles given to poor
plus model who verbalized children.
charity.
Rosenhan & White (1967) 9, 10 130 Prior contact with model—model Frequency of donating to or-
present. phans.
Floyd (1964)
Experiment 2 6-9 60 Receipt of trinkets from friends Increase in trinket-giving over
and nonfriends. trials.
Schopler & Bateaon (1965)
Experiment 1 CS 98 High dependency vs. low de- Frequency of volunteering for an
pendency.0'11 unpleasant experiment.
Experiment 2 CS 88 Same as above, in low cost con- Frequency of yielding the possi-
ditiond (with significant p bility to win money.
value).
Experiment 3 CS 56 Same as above. Same as above.
Schopler (1967) CS 60 Main effect across note-receipt Same as above.
conditions; note from partner
vs. note from experimenter.
Berkowitz (1967) CS — Evaluative feedback vs. not, in Amount produced for dependent
high dependency condition.11 other.

• Difference found between conditions for males but not for females j no difference between males and females within each
condition.
b p - .07.
° « < .10.
* Difference found between conditions for females but not for males.
288 DENNIS L. KREBS

their status. It found that although males and Examining altruism in the first three
females helped their partners about the same grades, three studies found evidence for a
amount on an unequal outcomes task, males developmental increase. Midlarsky and Bryan
(but not females) helped less after receiving a (1967) found that first graders who were
note from their "partner" saying, "I think exposed to expressive cues gave less than
you should give me more money." The effect second, third, and fourth graders. (They gave
did not occur when the note came from the more, however, when they received expres-
experimenter. sive cues and hugs.) In a control condition,
Although the influence of sex seemed pro- first and second graders gave less than third
nounced in the Schopler studies, Berkowitz and fourth graders. Staub and Feagens (1969)
et al., (1964) failed to obtain a sex effect. found an increase in helping in an emergency
In an attempt to explain the discrepancy be- situation from kindergarten to the second
tween findings, Berkowitz and his colleagues grade. Ugurel-Semin (1952) found an in-
(1967) ran a series of experiments. Although crease in generosity from age 6 to 8. One
the first three experiments failed to find sex study (Floyd, 1964) failed to find an increase
differences, a final study achieved the desired in sharing from the first to third grades. The
effect. Undergraduates were exposed to a brief bulk of the evidence, however, suggests a de-
autobiographical statement by a same-sex velopmental increase.
peer, then to a neutral or unfavorable evalua- Advancing to the third, fourth, and fifth
tion of it. They were then asked to rate its grades, more support for a developmental
author. The males reacted in a manner simi- advance in altruism is received. Ugurel-Semin
lar to that of the males in the Schopler (1952) found that the number of children
studies: They failed to evidence the customary who demonstrated selfish responses virtually
increase in productivity for the dependent ceased after age 8. There was also a
supervisor. And, in agreement with Schopler, tendency at this time for equalitarian re-
the females did. Berkowitz (1967) accounted sponses to outweigh generous responses. In a
for the results by suggesting that the males study patterned after that of Ugurel-Semin
acted egocentrically as a result of their pre- (1952), Handlon and Gross (1959) found
test personality evaluation. He went on to that the number of pennies or seals given to
suggest that the emphasis on personality test- a partner increased from kindergarten to the
ing in Schopler's university (versus that of fourth grade. A further increase occurred at
Berkowitz) caused the males in the Schopler the fifth grade, and leveled off at the sixth.
studies to be concerned with themselves. Midlarsky and Bryan (1967) found that the
number of self-sacrificial responses made by
Age Differences in Altruism third and fourth graders in a control condi-
Relatively few studies have compared the tion exceeded that made by first and second
incidence of altruism in children of different graders. This pattern prevailed (and was in
ages on the same task (see Table 8). Com- fact more pronounced) for the number of
parison of altruism across studies with differ- M&Ms donated in the absence of a model. M.
ent age groups is difficult for two reasons. In Harris (1968) found that fifth-grade children
the first place, most experimental situations gave more to charity in a control condition
differ in the extent to which they elicit al- than fourth-grade children. Staub (1968)
truism. Second, the dependent measure in found a tendency (p < .10) toward an in-
some studies is the amount given, while in crease in sharing from the fourth to the fifth
others it is the number of altruistic subjects. grade after an experience of success, but a
Table 8, indicates that in spite of the dif- decrease in sharing after an experience of
ferences among situations, some consistency failure. In an emergency situation, however,
can be found. To begin with, although there helping responses decreased in fourth and
are exceptions, few nursery school children sixth grades (Staub & Feagans, 1969).
give, and those who do generally give little. Although the evidence from most studies
As the children get older, incidence and supports a developmental increase in altruism,
amount of altruism tends to increase. studies which have employed generous models
 ALTRUISM 289

TABLE 8
STUDIES WHICH EXAMINED AGE DIFFERENCES IN ALTRUISM FOR CHILDREN

Subjects
Author and date Incidence of Main independent Main dependent
altruism variables variables
Age Sex W
Ugurel-Semin (1952) 4-6 M/F 18 33% Age, sex, economic Number of children
6-7 M/F 23 69% status, family size. who shared 9 nuts
7-8 M/F 44 84% generously or evenly
8-9 M/F 43 77% vs. number who did
9-10 M/F 45 100% not.
10-11 M/F 41 93%
11-12 M/F 28 97%
Handlon & Gross (1958) 4-6 M/F 18 28% Age, sex, ordinal posi- Number of jointly
9-10 M/F 25 42% tion. earned pennies or
10-11 M/F 25 57% seals given to a
11-12 M/F 25 60% partner.
Midlarsky & Bryan (1967) 6-7 F 16 16% (32%) Age; expressive cues Amount donated to
7-8 F 16 29% (21%) and affectionate re- needy children in the
8-9 F 16 43% (47%) sponse of models. observation condi-
9-10 F 16 46% (40%) tion, and number of
self-sacrifical re-
sponses (in paren-
theses) .
Staub & Feagans (1969) 4-6 M/F 16 24% Age; presence or ab- Mean help score (per-
6-7 M/F 16 31% sence of another centage of total pos-
7-8 M/F 16 47% child. sible) when alone.
9-10 M/F 16 50%
11-12 M/F 19 38%
Floyd (1964) 4-5 M/F 32 35% Friendship status of Number of trinkets
6-9* M/F 60 38% recipient; number of given to partner (all
trinkets received conditions com-
from partner bined).
Grusec & Skubiski 8, 10» M/F 80 24% Model's nurturance; Number of marbles
(in press) performance vs. ver- donated to charity.b
balization by model.
Bryan & Walbek (1969)
Experiment 1 8, 9, 10 M/F 91 44% Interaction between Number of subjects who
Experiment 2 8,9 M/F 168 41% model's verbaliza- were generous on at
Experiment 3 8,9 M/F 132 31% tions and perform- least one occasion.1*
ance of generous and
selfish responses.
White (1967) 9, 10« M/F 210 17% Observation of models, Number of gift cer-
rehearsal of altru- tificates donated to
ism, immediate vs. orphans'" immediate
delayed testing. testing.
62% Same as above. Number of subjects who
donated1" immediate
testing.
Rosenhan & White (1967) 9, 10" M/F 65 48% Positive, negative, neu- Number of subjects who
tral model; model donated in model's
present—absent. absence.11
Staub (1968) 9 M/F 196 70% Age; success, failure, Amount of candy left
10 M/F 60% moderate experience. for another in
moderate experience.
M. Harris (1968) 9,10 M/F 168 20% Age, donation to subject Number of chips do-
or charity by chari- nated.11
table model, selfish
model, no model.
61% Same as above. Number of subjects who
gave.b

> No difference between age groups in altruistic responses.

b
Across all conditions.
290 DENNIS L. KREBS

have usually failed to find the age-dependent istic than children from small families, but
increase in altruism. No consistent age-de- that other-than-only children are not m'ore
pendent increases in altruism were found in altruistic than only children. Friedrichs (1960)
the modeling conditions in the Midlarsky and and Handlon and Gross (1959) found no dif-
Bryan (1967), Aronfreed and Paskel (1968) ference between the altruism evidenced by
or M. Harris (1968) studies. Models may in- only and other-than-only children. Ribal
fluence younger children more than older chil- (1963) found that altruistic college students
dren and thereby counter the effect of a de- (those who scored high on need for nurturance
velopmental increase in altruism. and low on need for succorance) tended to
Three final studies which measured verbal come from large families, and to be oldest
indications of altruistic predispositions, found children. Ugurel-Semin (19S2) found that
that altruistic responses increased with age. Istanbul children from small families were
L. Harris (1967) extracted a group of most more selfish than those from large families.
altruistic and least altruistic children from a And Sawyer (1966) found that altruism was
sample of 406 elementary school children on correlated with number of older sisters. Fi-
the basis of their responses to three incom- nally, L. Harris (1967) failed to find an ef-
plete stories. He found that the ratio of chil- fect for ordinal position in elementary school
dren categorized as most and least altruistic children.
was about two to one for third and fourth
graders and five to one for fifth and sixth Social Class and Group Affiliation
graders. Durkin (1961) found that more
fifth-grade children gave altruistic responses Muir and Weinstein (1962) compared up-
to make-believe situations than second grad- per middle- and lower-class norms of social
ers. Detailed questioning, however, suggested obligation. They found that although house-
that their altruism was designed to build wives from high and low socioeconomic strata
credit. And finally, Shure (1968) found that (SES) reacted similarly in most situations,
the judgments of fairness, generosity, and upper middle-class women were more likely
selfishness of 8-year-olds were similar to those than lower-class women to cut off the credit
of both 10-year-olds and adults, but different of dilatory debtors and to feel obliged to
from those of 4-year-olds and 6-year-olds extrafamilial creditors. They viewed favor-
(which were similar). Although a clear rela- doing in an economic idiom: They were harsh
tionship between moral judgment and moral creditors, and they tended to avoid the role
behavior has not been established (Kohlberg, of debtor. Members of low SES, on the other
1964), verbal indications of altruism support hand, tended to give when they were able. In
the developmental trend suggested by studies general, women from high SES seemed most
on altruistic behavior. influenced by norms of reciprocity and ex-
In summary, although the findings are not change, while women from low SES seemed
entirely consistent, there is support for the most influenced by considerations of "mutual
notion that altruism increases with age. Al- aid" and the norm of social responsibility (as
truism in nursery school children was con- defined by Daniels and Berkowitz, 1963).
sistently low. Verbal reports of altruism Berkowitz and Friedman (1967) disagreed
showed clear age differences. And finally, sev- with the implications of the findings of Muir
eral studies found a developmental increase and Weinstein (1962). Categorizing members
in altruism in conditions that did not contain of the middle class as entrepreneurs or bu-
models. reaucrats, Berkowitz and Friedman (1967)
suggested that entrepreneurs are influenced by
the norm of reciprocity (and tend, therefore,
Ordinal Position and Altruism
away from altruism) and that bureaucrats
Although they are not entirely consistent, are influenced by the more altruistic prescrip-
findings relevant to the relationship between tions of the norm of social responsibility.
ordinal position and altruism suggest that Findings for entrepreneurs and bureaucrats
children from large families are more altru- were in accord with predictions: Entrepre-
 ALTRUISM 291

neurial middle-class boys were most likely to International Differences in Altruism

think that "supervisors" would expect them to Only two investigators (Berkowitz, 1966;
work hard after they (the subjects) received Feldman, 1968) tested for international differ-
prior help. They were also most likely to help ences in altruism. Establishing international
less after receiving a small (versus large) differences, of course, is a very large and gen-
amount of prior help. Members of the bu- eral task—one which few investigators have
reaucratic middle class and members of the been willing to tackle. Berkowitz (1966),
working class produced the same amount while recognizing the difficulty of creating
after receiving a little help and a lot of help. comparable international experimental situa-
The findings relating to the working class, tions, found some evidence that low SES boys
though, are not in strict accord with the find- from Oxford, England are more reciprocity
ings of other studies. Two investigations sug- oriented than their Madison, Wisconsin coun-
gested that members of the working class are terparts. No general differences in altruism,
not oriented to social responsibility. Almond however, were found between countries.
and Verba (1963) found that the positive Feldman (1968) found differences among
value of generosity and considerateness in- the reactions of Bostonians, Parisians, and
creased with education; and Kohn (1959) Athenians to requests for aid. Bostonians were
suggested that the working class is more con- the most likely to give directions to, and mail
cerned with immediate consequences of their a stamped letter for, a compatriot. Although
children's behavior than the inculcation of it made little difference to Athenians whether
abstract moral principles. the letter contained a stamp or not (their rate
More in accord with the Almond and of refusal was so high), Parisians and Bos-
Verba (1963) and Kohn (1959) studies are tonians were much less likely to mail a letter
the results of a study by Berkowitz (1966, when it was unstamped.
1968). Working-class boys from Oxford, En- The findings of Feldman (1968) are note-
gland worked hardest for those who had previ- worthy because they contain the only con-
ously helped them, especially when their vincing indication of large scale differences in
helpers came from a different social class. altruism among international cities. The fact
Bureaucratic middle-class boys acted in ac- that no real trend appeared across experiments
cord with the norm of social responsibility. seems to indicate that altruism is largely a
(It is interesting to note that Berkowitz was function of the specific situations, with people
unable to get enough "entrepreneurial" boys from different countries reacting more or less
for the experiment even though he offered to altruistically according to the circumstances
pay them for participating.) surrounding requests for help.
In contrast to the suggestions of Almond Although the results of the Berkowitz
and Verba (1963), however, Ugurel-Semin (1966) and Feldman (1968) studies sug-
(1952) found that poor children from Istanbul gested international differences in altruism, it
were less selfish than children from middle- should be noted that a more conservative (and
class or rich families. The middle-class chil- precise) interpretation would limit them to a
dren were least generous, and the children comparison among cities. It is not known
from rich families shared evenly least often. whether residents of most cities within En-
Sawyer (1966) also found that YMCA stu- gland, France, Greece, and America would
dents of the lower middle-class (versus busi- react to the experimental manipulations in
ness school or social science students) indi- the same way as residents of Oxford, Paris,
cated the greatest willingness to react gen- Athens, Madison, and Boston.
erously to friends, strangers, and antagonists. In summary, the findings relating to social
Business school students indicated they were roles and demographic attributes of benefac-
most prone to help themselves while hinder- tors demonstrated that sex, age, ordinal posi-
ing antagonists; and social science students tion, social class, and nationality sometimes
were most oriented towards helping their affect altruism. Although no clear sex differ-
friends. ences were found for children, adult males
292 DENNIS L, KREBS

were less prone to help highly dependent Friendship as an Elicitor of Altruism

others (versus others of low dependency),
It would seem that a simple relationship be-
especially when their status was threatened.
tween friendship and altruism should exist.
Females, on the other hand, were more prone
As suggested by Daniels and Berkowitz
to help others when they were highly de-
(1963), "We learn that we can usually expect
pendent. A fairly consistent increase in al-
help from our friends if we need their aid but,
truism was found with age. Some evidence
similarly, we also know that we should help
suggested that although only children are not
our friends if they are dependent upon us [p.
less altruistic than other-than-only children,
142]." The relationship between friendship
children from large families are more altru-
and altruism seemed so obvious to Sawyer
istic than those from small families. Evidence
(1966), in fact, that he validated his altruism
was supplied which suggested that members
scale by examining differences among altruism
of the bureaucratic middle class are oriented
directed to friends, strangers, and antagonists.
towards social responsibility, and members of
He found, as expected, that all groups in all
entrepreneurial middle class to reciprocity.
experimental situations exhibited most al-
Working-class boys from Oxford, England
truism toward friends and least toward an-
also evidenced a reciprocity orientation.
tagonists. Other studies, however, have found
In view of the fact that different situations more complex relationships between altruism
tend to elicit different amounts of altruism, and friendship.
and that people who share social roles differ
In one of the earliest experimental studies
along other dimensions, it is not surprising
on altruism, Wright (1942) reported a some-
that general trends are not frequent. The
what counterintuitive finding. Third graders
trends that are found are usually difficult to
were more likely to give the more desirable of
interpret, and further, more precise, research
two toys to a stranger than to a friend. They
is usually demanded. were also more likely to favor a stranger when
they were asked to serve as "umpires" and
SOCIAL ROLES AND DEMOGRAPHIC assign two toys of unequal attractiveness. Al-
ATTRIBUTES OF THE RECIPIENT though the first finding was not replicated in
It is difficult to separate social roles of the a subsequent test, two experiments by Floyd
recipient from interaction variables. Social (1964) also found a tendency for children to
roles such as foreigner and compatriot, in- give less to friends than nonfriends in some
grouper and outgrouper, friend and enemy situations. The Floyd (1964) experiments are
imply a dual classification—the attributes of not strictly comparable to those of Wright
both benefactor and recipient are assigned. (1942), however, because Floyd's nonfriend
To say that the recipient is a friend, in other groups consisted of sociometrically neutral
words, implies that the benefactor is also a and disliked peers, and Wright's nonfriend
friend. Once again the criterion relating to groups consisted of strangers. In the first ex-
the experimental source of variation is evoked. periment by Floyd (1964), friends were com-
The present section reviews studies which pared with ignored others; in the second ex-
manipulated and measured social roles and periment nonfriends were divided into neutral
demographic attributes of recipients. and disliked groups. It was found that nursery
Although it is likely that social roles and school children (Experiment 1) and first- to
demographic attributes of the recipient such third-grade children (Experiment 2) gave
as age, sex, and ordinal position help elicit more trinkets to friends than nonfriends. They
tended, however, to decrease the amount given
altruistic behavior, such attributes have not,
to friends after receiving many trinkets from
as yet, been examined. Most studies have held them (statistically significant in Experiment
them constant, or allowed them to vary ran- 1, not significant in Experiment 2), and to
domly. The four social roles that have been increase the number given after receiving few
studied are friendship status, ingroup af- (not significant in Experiment 1; p < .10 in
filiation, social class, and nationality. Experiment 2). The reverse occurred with
 ALTRUISM 293

neutral and disliked others: An increase oc- ferent conditions. The finding that punishment
curred after receiving many trinkets (signifi- (and loss of trinkets) elicited increments in
cant in Experiment 1, and for disliked others altruism from friends was probably the result
in Experiment 2), and a decrease occurred of the negative reinforcement of one choice
after receiving few trinkets (P < .10 in Ex- in a two-choice situation. Punishment (and a
periment 1 and significant in the neutral few trinkets) indicates a potential loss of
groups in Experiment 2). Floyd (1964) in- friendship. Punishment from an enemy, how-
terpreted his findings as support for a gain- ever, may have created a situation of an-
loss model which was based on findings relat- tagonism in which the subject had the last
ing to level of adaptation. The model predicts word—he, in effect, got even. Reward (or
that people act in ways that increase gains, lack of punishment) in relation to a nonfriend
and minimize losses, of interpersonal reward. was probably interpreted as a gesture of
In the two situations where expectations were friendship (much like a person extending his
met (a lot from a friend and little from a hand) and was, therefore, reciprocated. The
nonfriend) no gain was made in interpersonal reaction to reward from a friend, though, is
rewards, so the amount reciprocated was re- difficult to interpret. It would seem that when
duced. When a little was received from a friends behave generously they are taken for
friend, it was interpreted as punishment for granted and their partners benefit themselves.
prior selfishness, and a lot was returned; when Whether these interpretations are correct or
a lot was received from a nonfriend, it was not, it is interesting to note that the behavior
interpreted as an invitation to friendship (and displayed in these studies seems most easily
interpersonal gain), and a lot was given in re- explained from a reinforcement perspective.
turn. This interpretation is consistent with
the finding that some of the children in Ingroup Affiliation as an Elicit or of Altruism
Wright's (1942) study who gave to strangers
explained their behavior by saying that they Common experience demonstrates that peo-
wanted to gain a friend. ple are more prone to give to people who are
A study by Epstein and Hornstein (1969) close to them than to people who are not. The
on undergraduates is of interest in this sacrifices of one family member for another,
connection. Subjects who liked their part- or of a lover for his mate, for example, are
ners made more self-sacrificial helping choices expected to exceed those in other relation-
when they anticipated punishment (by a third ships. Some studies have examined the effect
party) for selfish responses than when they of ingroup affiliation on altruism. Friedrichs
did not anticipate punishment. Subjects who (1960) found that more altruism was directed
disliked their partners, however, reversed this toward ingroup recipients than outgroup re-
trend—they helped more when they did not cipients, especially when the benefactors were
anticipate punishment and less when they did. authoritarian. Campbell (1965) demonstrated
If the reciprocation of a few trinkets by that ingroup sacrifice was common in the face
friends is equated with punishment by a liked of outgroup threat.
other, and the reciprocation of many trinkets The series of experiments by Feldman
by nonfriends with lack of punishment by a (1968) is relevant to the effect of ingroup
disliked other, the results of the Floyd (1964) affiliation on altruism. Foreigners can be con-
and Epstein and Hornstein (1969) studies sidered outgroup members, and compatriots
are comparable. Examination of the trend ingroup members. Feldman (1968) found that
across trials in the Epstein and Hornstein Parisians and Athenians were more willing to
study indicates that increments and decre- give directions to compatriots than foreigners,
ments in giving were similar to those in the but that Bostonians treated them the same.
Floyd (1964) study. Bostonians, however, were less likely to mail
Epstein and Hornstein (1969) offered an unstamped letter for a foreigner than a
several interpretations for their data, none compatriot. Hornstein et al. (1968), with
of which explained it all. It is possible American subjects, found that fewer wallets
that different principles were in effect in dif- were returned to their owners when the finders
294 DENNIS L. KREBS

were led to believe that a foreigner had found he is to elicit altruism, unless the benefactor
and relost the wallet than when they thought is in a position of prestige. When more was
the previous finder was not foreign (i.e., was given to nonfriends, outgroupers, and mem-
similar to them). In some situations, then, bers of different social classes, it may have
compatriots are helped more, but in others been because the recipients were of high
(especially in Athens) foreigners are given prestige, and, therefore, also attractive.
most help. Whether because of similarity or prestige,
then, the attractiveness of the recipient may
Social Class as an Elicitor of Altruism have elicited altruism.
Although most of the research on class dif-
ferences focused on the social class of bene- NORMS THAT AFFECT BENEFACTORS
factors, some research suggests that recipients Some investigators have suggested that
from the working class tend to elicit less help people act altruistically because of the pre-
than middle-class recipients. Questionnaire scriptions of social norms. It has been sug-
responses in the Berkowitz and Friedman gested that altruism is regulated by two social
(1967) study indicated that middle-class boys
norms—the norm of social responsibility, and
were more attractive than working-class boys the norm of giving. The level of analysis at
to peers of both classes. Entrepreneurial mid-
which normative explanations occur is thought
dle-class boys who received little prior help to be appropriate to sociological theory, but
from a working-class partner worked less hard
its contribution to psychological research is
for a working-class (versus middle-class) questioned.
supervisor. In general, though, most help was
given to partners of the same social class. In The Norm of Social Responsibility
Oxford, England (Berkowitz, 1966, 1968)
working-class boys tended to work hardest for Research has been reviewed which demon-
middle-class supervisors. Middle-class part- strated that people tend to help those who
ners, on the other hand, worked least hard for are dependent on them. Berkowitz and his
working-class partners. The trends, however, colleagues (e.g., Berkowitz & Daniels, 1963)
were not consistent throughout conditions. attributed such helping to the influence of the
In summary, several studies suggest that "norm of social responsibility," which pre-
potential recipients are sometimes more likely scribes that people should help those who
to elicit benefits when they are friends, in- need help. Normative analysis suggests that
groupers, and members of the same social people act altruistically in particular situa-
class and nationality as benefactors. Other tions because it is the proper thing to do.
studies, though, suggest that nonfriends, out- The notion that people act in accord with
groupers, foreigners, and members of higher normative standards of conduct seems quite
social classes elicit more altruism. It is pos- sound as far as it goes. It would seem, though,
sible that the seemingly contradictory find- that it does not go far enough. Although
ings support the same general principle. normative analysts (e.g., Berkowitz 7 ) have
Studies on interpersonal attraction may offer criticized the explanatory circularity of rein-
a clue. Daniels and Berkowitz (1963) and forcement theory, the postulation of social
Epstein and Hornstein (1969) manipulated norms can also end in tautology. A particular
interpersonal attractiveness by varying simi- response, for example, can be predicted on
larity between potential benefactor and the basis of a norm. If it occurs, the norm is
recipient. Other studies (e.g., Heider, 1958; said to have had an effect. If it does not
Newcomb, 19S6; Novak & Lerner, 1968) occur, the situation is said to fall outside the
have found that similarity results in at- range of the norm. In cases where the norm
traction. Friendship and ingroup affiliation is established on the basis of the behavior it
imply a sharing of roles, a similarity, as does is supposed to predict, it is in the same posi-
congruity between social class. The evidence 7
Berkowitz, L. Beyond exchange: ideals and other
suggests that the more similar the potential factors affecting helping and altruism. Unpublished
recipient is to his benefactor, the more likely manuscript, University of Wisconsin, 1968.
 ALTRUISM 295
tion as Skinner's (1953) reinforcer—its ex- NORMS WHICH AFFECT THE ALTRUISM-
istence is established by the effect it produces. ELICITING CAPACITY OF RECIPIENTS
The danger with normative analysis is that The Generation of Debt and the Norm of
norms can be invented post hoc to explain Reciprocity
almost anything. Although Berkowitz has
taken steps toward the specification of elicit- The norm of reciprocity, as postulated by
ing situations—dependency is said to evoke Gouldner (1960), prescribes that people should
the norm of social responsibility—the increase help those who have helped them, and that
in understanding supplied by the postulation people should not injure those who have
of norms can still be questioned. Because it is helped them. Research pertaining to the effect
assumed that the norm is internalized, and of the norm of reciprocity is thought to relate
that it gains its power from its ability to to the altruism-eliciting capacity of recipients
evoke positive and negative affect, normative because in the prototypical experimental situa-
analysis ultimately contains a cognitive-af- tion it is the generosity of the potential re-
fective base. Understanding of the effects of cipient that is varied. The fact that the gen-
norms, it would seem, awaits elucidation of erosity of the recipient is an attribute that
their cognitive-affective representations. At he acquired in the role of benefactor does not
any rate, the methods used to study social alter the fact that the altruism-eliciting vari-
norms should be different from the operations able lies with the recipient. The characteristic
used to establish their effects. And once norms of the recipient of importance here relates to
are identified, focus should shift to the dy- his role as creditor. Implicit in the examina-
namics of their internalization and the con- tion of research relevant to the norm of re-
ditions of their effect. ciprocity is the notion of a less general level
of analysis—one which identifies eliciting
The Norm of Giving characteristics of recipients within the norma-
When Gouldner (1960) outlined the range tive framework of their culture. The state
of what he called the norm of reciprocity, he variable reference of reciprocity, therefore,
mentioned some situations in which it did not should be kept in mind.
apply. The situations contained one essential It might, of course, be wondered whether
characteristic: They all involved dependent the payment of debt constitutes altruism. Ac-
recipients such as the very young, the very cording to the definition employed here (and,
old, and the sick, Leeds (1963) suggested that it is thought, the results of common attribu-
another norm applies in situations which are tion) strict reciprocity falls outside the range
beyond the range of the norm of reciprocity. of altruistic behavior. Behavior which repays
She called this norm the norm of giving, and more than it owes, or repays favors that did
suggested that situations in which the norm not generate expectation of return, on the
of giving was influential involved "role vac- other hand, seems altruistic. A relatively un-
uums" or "social vacuums." Role vacuums discriminating examination of the research on
consist of nonprescribed opportunities for al- reciprocity will be made, though, in the hope
truism. The "whore with a heart" is an ex- of elucidating general determinants of altru-
ample. Social vacuums occur in situations ism, and, if nothing else, clarifying the dis-
where institutionalized means of action are tinction between reciprocity and altruism.
not available. Emergency situations are cases Research on the effects of reciprocity and
in point. In short, the norm of giving applies altruism has examined two distinct relation-
in situations where help is needed, but no ships. The first is straightforward—it involves
institutionalized means of help is available. a recipient to whom something is owed. In
Leeds' (1963) concern, like that of Gould- the second situation, the owed recipient is
ner (I960), was with the sociological function not available for repayment, and payment is
of norms. Little psychological research has generalized to someone else.
examined the norm of giving, even though it As described by Gouldner (1960), the socio-
may have been antecedent to many altruistic logical purpose of the norm of reciprocity is
situations. to prevent exploitation of the weak by the
296 DENNIS L. KREBS

powerful and to insure that people pay their tive—subjects did not give any more to those
debts. The force of the norm of reciprocity who had large (versus small) future resources.
is variable according to (a) the need state of Studies on inappropriate favors (e.g.,
the original recipient (who becomes the bene- Brehm & Cole, 1966; Schopler & Thompson,
factor), (b) the resources of the original 1968) supported Gouldner's point c. Reci-
donor (who becomes the recipient), (c) the procity was inhibited when the motives of
motives imputed to the original donor, and the original benefactors were questioned.
(d) the extent to which the original donor Two studies extended the generality of the
gave of his own free will (p. 171). Gouldner findings on reciprocity to children. Staub and
(1960) recognized the fact that reciprocal be- Sherk (in press) found that fourth-grade
havior is usually expedient—the building of children shared a crayon longer with children
credit increases security. Research has demon- who had previously shared a lot (versus a
strated that all four factors influence re- little) candy with them. M. Harris (1968)
ciprocity. found that fourth- and fifth-grade children
In the only sociological study relevant to gave to models who gave to them.
Gouldner's (1960) discussion, Muir and Wein- Limitations on the occurrence of reciprocal
stein (1962) tested the applicability of Gould- help-giving were revealed in several studies.
ner's (1960) model to the exchange of small Berkowitz and Friedman (1967), for example,
favors. "Intensive interviews" showed that found that although the norm of reciprocity
females from high and low SES thought that seemed to influence the helping behavior of
the magnitude of a debt was related to the sons of middle-class entrepreneurs and work-
original recipient's need or desire (Gouldner's ers, sons of middle-class bureaucratics were
point a) and to the resources of the original influenced by the norm of social responsibil-
donor (Gouldner's b). Females from the high ity. Floyd (1964) found that although chil-
SES thought debts were greater when the dren tended to increase the number of trinkets
original favors were given freely (Gouldner's given to nonfriends after receiving a lot from
d). them, they tended to decrease the number
Other investigators, who used laboratory after receiving a lot from friends. Moreover,
manipulations, supplied further support for although selfish nonfriends elicited a decrease
Gouldner's contentions. Goranson and Berko- in giving, selfish friends elicited an increase.
witz (1966) found that undergraduate females It would seem that the history of reciprocal
did more work for supervisors after they exchanges between friends alters the set of
received voluntary (but not compulsory) help expectations that governs the norm of reci-
from them, even though they did not expect procity. It is also possible, of course, that
to see them again (Gouldner's d). Frisch and friends planned to even up after the experi-
Greenberg (1968) found that undergraduate ment.
males who competed with a partner for two
dollars were more likely to help him accumu- Different Recipients—Generalized Reciprocity
late "proficiency credits" when they thought Reciprocity refers to an obligatory bond
they had received a lot (versus a little) prior between two people. In some cases, though,
help, especially when the help was perceived a third person is involved. When one person
as intentional (Gouldner's d). helps another person, then leaves the field, the
Pruitt (1968) offered further support for recipient of the benefits may be more likely
the effect of reciprocity, and for Gouldner's to help a third person. This behavior does
point b. Male undergraduates gave more to a not, as some investigators have assumed, con-
partner during a mixed-motive game when stitute reciprocity, at least in the sense that
they had previously received a lot (versus a Gouldner (1960) meant it. It is possible that
little) from him. They also gave more when the dynamics of owing are not operative at
they had received 80% of one dollar than all. The norm of reciprocity prescribes that
20% of four dollars. Strictly expedient ex- people should help those who help them—it
pectations (using reciprocity as a tactic to says nothing about third parties. To remain
get future favors) did not seem to be opera- consistent with the language of relevant in-
 ALTRUISM 297
vestigators, the term generalized reciprocity had little, when he was sincere, and when
is employed. It is meant as a descriptive, not he gave voluntarily. Moreover, several studies
an explanatory title. found that when people are unable to con-
Berkowitz and Daniels (1964) found that tact their original benefactors, they are prone
workers who received help from a confederate to give to others.
helped a highly dependent supervisor more
than those who did not. The investigators CONCLUSIONS
interpreted these results as support for the Although large segments of the research
hypothesis that the observation of helping on altruism failed to supply consistent find-
increases the salience of the social responsi- ings, several trends have appeared. New de-
bility norm, which results in further helping. velopments in reinforcement theory, and the
It is also possible that those who were helped rise of concern with the process of attribution
felt better, and the resulting positive affect have begun to supply new ways of looking
mediated altruism. A further study (Goran- at altruism. The concept, however, is still
son & Berkowitz, 1966) found some differ- unclear and no way has been found to meas-
ences among voluntary-, compulsory-, and ure its motivational base.
refused-help conditions. Workers in the vol- Elucidation of antecedents of altruistic be-
untary-help (same person) condition produced havior has not awaited the specification of
more for their supervisors than workers in the phenomenon; operational definitions have
the other two conditions. Workers in the re- supplied a way of bypassing the issue. Re-
fused-help condition indicated that they did search projects have appraised the correlates
not think that people would expect them to and determinants of altruism from several
work hard for their supervisors. A later study different perspectives. Most research examined
(Berkowitz & Friedman, 1967) found social the effect of temporary states of the bene-
class differences in generalized reciprocity. factor. Both positive affective states asso-
Boys from the entrepreneurial middle class ciated with success and competence, and nega-
(but not from the bureaucratic middle class tive affective states associated with harming
or working class) worked hard for their su- another resulted in altruistic behavior. States
pervisors after they received a lot of help—• associated with the observation of altruistic
whether they thought they had received it models also mediated altruistic responses.
from their supervisor or a different supervisor. Although the bulk of studies on modeling
A study by Test and Bryan (in press) com- seem best analyzed in relation to the tem-
pared the effects of modeling, dependency, porary effect of models, some studies sug-
and generalized reciprocity on helping be- gested that the observation of models leads
havior in female undergraduates. Experimen- to the acquisition of behavioral dispositions
tal conditions in which a model helped a third and personality traits.
person, and in which the subject himself Research which examined the effect of tem-
was helped, resulted in more subsequent help- porary states of the recipient suggested that
ing than no-model and negative-model condi- the basic altruism-eliciting attribute of re-
tions. The investigators concluded that gen- cipients is their dependency. The trait as
eralized reciprocity had little effect beyond well as the state of dependency was found to
that of modeling. The fact that situations elicit altruism. In most cases externally
which involve reciprocity also involve models caused dependency elicited more altruism
suggests an overlap in determinants of al- than internally caused dependency, but in
truism. cases where the status of the recipient was
Although there may be limitations in cer- seen as a potential threat, internally caused
tain social classes and with long range dependency elicited the greater amount of
acquaintances, the bulk of the evidence sup- altruism. Besides dependency, interpersonal
ports Gouldner's (1960) suggestion that attractiveness of the recipient was found to
people tend to return favors. They are most relate to altruism. Although few studies pur-
likely to reciprocate when the original bene- posefully manipulated the variable, it seems
factor gave something he needed, when he likely that interpersonal attractiveness inci-
298 DENNIS L. KREBS

dentally influenced the responses of benefac- receipt of favors results in a general disposi-
tors in many studies. Some studies, however, tion toward altruism, in some situations. Al-
indicated that attractiveness and altruism do though research was analyzed in the category
not always go together. of social norms, a state variable reference
Studies which examined trait variables was assumed.
were plagued with difficulties. Because altru- Considered as a whole, research on altru-
ism is a positive trait, it is difficult to sepa- ism has begun to identify general correlates
rate it from other positive characteristics. and antecedents of particular beneficial be-
Although trait studies generally found that haviors. It has helped supply hints concern-
altruists were well adjusted, sociable, and un- ing socialization and the effect of judgments
assuming people, the validity of their meas- of altruism on everyday behavior. It has also
ures of altruism was usually in doubt. Studies cast light on important theoretical issues such
which used behavioral measures of altruism as the viability of the psychoanalytic and
were less likely than studies which used rat- reinforcement perspectives.
ing-scale and pencil-and-paper test measures Philosophers were originally concerned with
to find trait correlates. altruism because it related in an essential way
In relation to social roles and demographic to the nature of man. Although the method-
attributes of benefactors, several trends ap- ology of current researchers is different, and
peared. Although altruism did not consistently their focus more specific, they are, in the
relate to sex in children, adult males were final analysis, also trying to understand hu-
found to act less altruistically toward highly man nature. Many have welcomed the new
(versus lowly) dependent others, especially research on the positive aspects of man, but
if they seemed threatening. Altruistic females it should be realized that the study of bene-
acted more altruistically toward highly de- ficial behaviors does not establish the exist-
pendent others. A fairly consistent increase in ence of altruism. Elucidation of the phenome-
altruism with age was found in children. non of altruism is, in fact, as capable of
There was some indication that people from shaking the foundation of man's self-concep-
large families are more altruistic than people tion as elucidation of such things as aggres-
from small families. Social-class differences sion and anxiety. Unfortunately, the fact
were also found. Members of the work- that man acts altruistically does not mean
ing class and entrepreneurial middle class that he is altruistic. The hopeful thing about
tended to behave in accord with the norm of research, however, is that at the same time
reciprocity. Bureaucratic middle classers, on as it sorts out antecedents, it identifies mecha-
the other hand, were more socially respon- nisms that can lead to change.
sible. Finally, although some international
differences were found, they did not form a REFERENCES
general pattern. When it came to social roles ALLEN, H. Bystander intervention. Unpublished doc-
of recipients, an examination of the effects toral dissertation, New York University, 1968.
of friendship status, ingroup affiliation, and Reported in J. M. Darley. The sharing of respon-
social class suggested that people give to sibility. Paper presented at the meeting of the
American Psychological Association, Washington,
those who are similar to them, to those who D. C., September 1967.
are prestigious, and to those from whom they ALLPOET, G. W., VERNON, P. E., & LINDZEY, G.
stand to gain. Manual for the study of values. (3rd ed.) Cam-
The level of normative analysis was re- bridge: Houghton Mifflin Co., 1960.
ALMOND, G. A., & VERBA, S. The civic culture.
jected as it related to benefactors, and vari- Princeton: Princeton University Press, 1963.
ables associated with social norms were ARONFREED, J. Conduct and conscience. New York:
analyzed according to their state reference. Academic Press, 1968.
In relation to recipients, research was re- ARONFREED, J., & PASKEL, V. Altruism, empathy and
viewed which supported Gouldner's (1960) the conditioning of positive affect. Reported in
J. Aronfreed, Conduct and conscience. New York:
notions relating to the norm of reciprocity. Academic Press, 1968.
An examination of what has been called ASCH, S. E. Studies of independence and conformity:
generalized reciprocity demonstrated that the A minority of one against a unanimous majority.
 ALTRUISM 299

Psychological Monographs, 1956, 70(9, Whole No. BRUNER, J. S. Social psychology and perception. In
416). E. E. Maccoby, T. M, Newcomb, & E. L. Hartley
BANDURA, A., & WALTERS, R. H. Social learning and (Eds.), Readings in social psychology. New York:
personality development. New York: Holt, Rine- Holt and Co., 1958.
hart & Winston, 1963. BUDD, L. J. Altruism arrives in America. American
BERKOWITZ, L. Response to Stone. Journal of Per- Quarterly, 1956, 8, 40-52.
sonality and Social Psychology, 1965, 2, 757-758. CAMPBELL, D. T. Ethnocentric and other altruistic
BERKOWITZ, L. A laboratory investigation of social motives. Nebraska Symposium on Motivation.
class and national differences in helping behavior. 1965, 13, 283-311.
International Journal of Psychology, 1966, 1, 231- CARLSMITH, J. M., & GROSS, A. E. Some effects of
242. guilt on compliance. Journal of Personality and
BERKOWITZ, L. Judgmental standards in situational Social Psychology, 1969, 11, 232-240.
and sex differences in helping behavior. Paper pre- CATTELL, R., & HOROWITZ, J. Objective personality
sented at the meeting of the American Psychologi- tests investigating the structure of altruism in re-
cal Association, Washington, D. C., September lation to source traits. A, H, and L. Journal of
1967. Personality, 1952, 21, 103-117.
BERKOWITZ, L. Responsibility, reciprocity, and so- CHARLESWORTH, R., & HARTUP, W. W. Positive so-
cial distance in help-giving: An experimental in- cial reinforcement in the nursery school peer group.
vestigation of English social class differences. Child Development, 1967, 38, 993-1002.
Journal of Experimental Social Psychology, 1968, CRANDALL, V. D., CRANDALL, V. J., & PATKOVSKY, W.
4, 46-63. A children's social desirability questionnaire. Jour-
BERKOWITZ, L., & CONNOR, W. H. Success, failure nal of Consulting Psychology, 1965, 29, 27-36.
and social responsibility. Journal of Personality DANIELS, L., & BERKOWITZ, L. Liking and response
and Social Psychology, 1966, 4, 664-669. to dependency relationships. Human Relations,
BERKOWITZ, L., & DANIELS, L. Responsibility and 1963, 16, 141-148.
dependency. Journal of Abnormal and Social DARLEY, J. M. The sharing of responsibility. Paper
Psychology, 1963, 66, 429-436. presented at the American Psychological Associa-
BERKOWITZ, L., & DANIELS, L. Affecting the salience tion, Washington, D. C., September 1967.
of the social responsibility norm: Effects of past DARLEY, J. M., & LATANE, B. Bystander intervention
help on the response to dependency relationships. in emergencies: diffusion of responsibility. Journal
Journal of Abnormal and Social Psychology, 1964, of Personality and Social Psychology, 1968, 8, 377-
68, 275-281. 383.
BERKOWITZ, L., & FRIEDMAN, P. Some social class DARLINGTON, R. B., & MACKER, C. E. Displacement
differences in helping behavior. Journal of Per- of guilt-produced altruistic behavior. Journal of
sonality and Social Psychology, 1967, 5, 217-225. Personality and Social Psychology, 1966, 4, 442-
BERKOWITZ, L., KLANDERMAN, S., & HARRIS, R. Ef- 443.
fects of experimenter awareness and sex of sub- DURKIN, D. The specificity of children's moral judg-
ject and experimenter on reactions to dependency ments. Journal of Genetic Psychology, 1961, 98,
relationship. Sociometry, 1964, 27, 327-337. 3-13.
BERSCHEID, E., & WALSTER, E. When does a harm- EDWARDS, A. L. Edwards personal preference sched-
doer compensate a victim? Journal of Personality ule. New York: The Psychological Corporation,
and Social Psychology, 1967, 6, 435-441. 1954.
BLAKE, R. R., BERKOWITZ, H., BELLAMY, R., & EPSTEIN, Y. M., & HORNSTEIN, H. A. Penalty and
MOUTON, J. S. Volunteering as an avoidance act. interpersonal attraction as factors influencing the
Journal of Abnormal and Social Psychology, 19S6, decision to help another person. Journal of Ex-
53, 1S4-1S6. perimental Social Psychology, 1969, 5, 272-282.
BLAKE, R., ROSENBAUM, M., & DURYEA, R. Gift- ERIKSON, E. H. Childhood and society. New York:
giving as a function of group standards. Human W. W. Norton & Co., 1950.
Relations, 1955, 8, 61-73. FELDMAN, R. E. Response to compatriot and for-
BOHANNAN, P. Social anthropology. New York: Holt, eigner who seek assistance. Journal of Personality
Rinehart & Winston, 1963. and Social Psychology, 1968, 10, 202-214.
BREHM, J. W., & COLE, A. H. Effect of a favor FISCHER, W. F. Sharing in preschool children as a
which reduces freedom. Journal of Personality and function of amount and type of reinforcement.
Social Psychology, 1966, 3, 420-426. Genetic Psychological Monographs, 1963, 68, 215-
BRYAN, J. H., & TEST, M. A. Models and helping: 245.
Naturalistic studies in aiding behavior. Journal FLOYD, J. Effects of amount of reward and friend-
of Personality and Social Psychology, 1967, 6, ship status of the other on the frequency of shar-
400-407. ing in children. Unpublished doctoral dissertation,
BRYAN, J. H., & WALBEK, N. Words and deeds about University of Minnesota, 1964,
sacrifice: Their impact upon children's judgment FORM, W. H., & Nosow, S. Community in disaster.
and behavior. Paper presented at the Biennial New York: Harper, 1958.
meeting of the Society for Research in Child De- FREEDMAN, J. L., WALLTNOTON, S. A., & BLESS, E.
velopment, Santa Monica, March 1969. Compliance without pressure: The effect of guilt.
300 DENNIS L. KREBS

Journal of Personality and Social Psychology, HORNSTELN, H. A., FISCH, E., & HOLMES, M. The
1967, 7, 117-124. influence of a model's feeling about his behavior
FREUD, A. The infantile instinct life. In H. Hernia and his relevance as a comparison other on ob-
& G. M. Karth (Eds.), A handbook of psycho- servers' behavior. Journal of Personality and Social
analysis. New York: World Publishing Co., 1963. Psychology, 1968, 10, 222-226.
FRLEDRICHS, R. W. Alter versus ego: An exploratory HOROWITZ, I. A. Effect of choice and locus of de-
assessment of altruism. American Sociological Re- pendence on helping behavior. Journal of Personal-
view, 1960, 25, 496-S08. ity and Social Psychology, 1968, 8, 373-376.
FEISCH, D. M., & GREENBERO, M. S. Reciprocity and KAOTMANN, H. The unconcerned bystander. Proceed-
intentionality in the giving of help. Proceedings of ings of the 76th Annual Convention of the Ameri-
the 76th Annual Convention of the American can Psychological Association, 1968, 3, 387-388.
Psychological Association 1968, 3, 383-384. KIESLER, S. The effect of perceived role require-
GERWITZ, J. L. Succorance in young children. Un- ments on reactions to favor-doing. Journal of Ex-
published doctoral dissertation, State University perimental Social Psychology, 1966, 2, 198-210.
of Iowa, 1948. KOHLBERG, L. Development of moral character and
GOLDIAMOND, I. Moral behavior: A functional analy- moral ideology. In M. L. Hoffman & L. W.
sis. Psychology Today, 1968, 2, 31-34. Hoffman (Eds.), Review of child development
GORANSON, R., & BERKOWITZ, L. Reciprocity and research. Vol. 1. New York: Russel Sage Founda-
responsibility reactions to prior help. Journal of tion, 1964.
Personality and Social Psychology, 1966, 3, 227- KOHLBERG, L. Stages in the development of moral
232. thought and action. New York: Holt, Rinehart &
GORE, P. M., & ROTTER, J. B. A personality correlate Winston, in press.
of social action. Journal of Personality, 1963, 31, KOHN, M. L. Social class and parental values. Ameri-
S8-64. can Journal of Sociology, 1959, 64, 337-351.
GotroH, H. G., McCLOSKEY, H., & MEEHL, P. E. A KORTE, C. Group effects on help-giving in an
personality scale for social responsibility. Journal emergency. Proceedings of the 77th Annual Con-
of Abnormal and Social Psychology, 1952, 47, 73- vention of the American Psychological Associa-
80. tion, 1969, 4, 383-384
GOULDNER, A. The norm of reciprocity: A prelimi- LATANE, B. The urge to help. Paper presented at the
nary statement. American Sociological Review, meeting of the American Psychological Associa-
1960, 25, 161-178. tion, Washington, D. C., September 1967.
GRUSEC, J. E., & SKUBISKI, S. L. Model nurturance, LATANE, B., & DARLEY, J. Group inhibition of by-
demand characteristics of the modeling experi- stander intervention in emergencies. Journal of
ment, and altruism. Journal of Personality and Personality and Social Psychology, 1968, 10, 215-
Social Psychology, in press. 221.
HANDLONT, B. J., & GROSS, P. The development of LATANE, B., & RODIN, J. A lady in distress: In-
sharing behavior. Journal of Abnormal and Social hibiting effects of friends and strangers on by-
Psychology, 1959, 59, 425-428. stander intervention. Journal of Experimental So-
HARRIS, D. B. A scale for measuring attitudes of cial Psychology, 1969, 5, 189-202.
social responsibility in children. Journal of Ab- LAVERY, J. J., & FOLEY, P. J. Altruism or arousal
normal and Social Psychology, 1957, 55, 322-326. in the rat? Science, 1963, 140, 172-173.
HARRIS, L. A. A study of altruism. Elementary LEEDS, R. Altruism and the norm of giving. Merrill-
School Journal, 1967, 68(3), 135-141. Palmer Quarterly, 1963, 9, 229-240.
HARRIS, M. Some determinants of sharing in chil- LERNER, M. J., & LICHTMAN, R. R. Effects of per-
dren. Unpublished doctoral dissertation, Stanford ceived norms on attitudes and altruistic behavior
University, 1968. toward a dependent other. Journal of Personality
HARTSHORNE, H., & MAY, M. A. Studies in the na- and Social Psychology, 1968, 9, 226-232.
ture of character. Vol. 1. Studies in deceit. New LERNER, M. J., & MATTHEWS, G. Reactions to the
York: Macmillan, 1928. suffering of others under conditions of indirect
HARTOT, W. W., & COATES, B. Imitation of a peer as responsibility. Journal of Personality and Social
a function of reinforcement from the peer group Psychology, 1967, 5, 319-325.
and rewardingness of the model. Child Develop-
ment, 1967, 38, 1003-1016. MADDI, S. Personality theory—a comparative analy-
HARTUP, W. W., & KELLER, E. D. Nurturance in sis. New York: Appleton-Century-Crofts, 1968.
preschool children and its relation to dependency. MIDLARSKY, E. Some antecedents of aiding under
Child Development, 1960, 31, 681-689. stress. Proceedings of the 76th Annual Convention
HEBB, D. O., & THOMPSON, W. R. The social sig- of the American Psychological Association, 1968,
nificance of animal studies. In G. Lindzey (Ed.), 3, 385-386. (a)
Handbook of social psychology. Vol. 1. Theory MIDLARSKY, E. Aiding responses: An analysis and
and method. Cambridge: Addison-Wesley, 1954. review. Merrill-Palmer Quarterly, 1968, 14, 229-
HEIDER, F. The psychology of interpersonal rela- 260. (b)
tions. New York: Wiley, 1958. MIDLARSKY, E., & BRYAN, J. H. Training charity in
HOLMES, S. J. The reproductive beginnings of altru- children. Journal of Personality and Social Psy-
ism. Psychological Review, 1945, 52, 109-112. chology, 1967, 5, 408-415.
 ALTRUISM 301

MILGRAM, S. A behavioral study of obedience. Jour- ROSENTHAL, A. M. Study of the sickness called
nal of Abnormal and Social Psychology, 1963, 67, apathy. New York Times Magazine Section, May
371-378. 3, 1964, 24.
MOWRER, 0. H. Learning theory and the symbolic ROSENTHAL, R. Experimental effects in behavioral
process. New York: Wiley, 1960. research. New York: Appleton-Century-Crofts,
MUIR, D., & WEINSTEIN, E. The social debt: An in- 1966.
vestigation of lower-class and middle-class norms of ROTTER, J. B. Generalized expectancies for internal
social obligation. American Sociological Review, versus external control of reinforcements. Psycho-
1962, 27, 532-539. logical Monographs, 1966, 80(1, Whole No. 609).
MURPHY, L. B. Social behavior and child person- RUTHERFORD, E., & MUSSEN, P. Generosity in nursery
ality: An exploratory study of some roots of school boys. Child Development, 1968, 39, 755-765.
sympathy. New York: Columbia University Press, SAWYER, J. The altruism scale: A measure of co-
1937. operative, individualistic, and competitive interper-
MURRAY, H. A. Explorations in personality. New sonal orientation. American Journal of Sociology,
York: Oxford University Press, 1938. 1966, 71, 407-416.
XEWCOMB, T. M. The prediction of interpersonal at- SCHACHTER, S., & HALL, R. Group-derived restraints
traction. American Psychologist, 1956, 11, 575-586. and audience persuasion. Human Relations, 1952, 5,
NISSEN, H. W., & CRAWFORD, M. P. A preliminary 397-406.
study of food-sharing behavior in young chimpan- SCHOPLER, J. An investigation of sex differences on
zees. Journal of Comparative Psychology, 1936, the influence of dependence. Sociometry, 1967, 30,
22, 383-419. 50-63.
NOVAK, D. W., & LERNER, M. S. Rejection as a con- SCHOPLER, J., & BATESON, N. The power of de-
sequence of perceived similarity. Journal of Per- pendence. Journal of Personality and Social Psy-
sonality and Social Psychology, 1968, 9, 147-152. chology, 1965, 2, 247-254.
NUNNALLY, J. C. Popular conceptions of mental SCHOPLER, J., & MATTHEWS, M. The influence of
health. New York: Holt, Rinehart & Winston, perceived causal locus of partner's dependence on
1961. the use of interpersonal power. Journal of Per-
PIAGET, J. The moral judgment of the child. New sonality and Social Psychology, 1965, 2, 609-612.
York: Harcourt, Brace & World, 1932. SCHOPLER, J., & THOMPSON, V. The role of attribu-
PRUITT, D. G. Reciprocity and credit building in a tion process in mediating amount of reciprocity
laboratory dyad. Journal of Personality and Social for a favor. Journal of Personality and Social Psy-
Psychology, 1968, 8, 143-147. chology, 1968, 10, 243-250.
RAWLINGS, E. I. Witnessing harm to other: A re- SHERIF, M. Group influences upon the formation of
assessment of the role of guilt in altruistic be- norms and attitudes. In T. M. Newcomb, & E. L.
havior. Journal of Personality and Social Psy- Hartley (Eds.), Readings in social psychology.
chology, 1968, 10, 377-380. New York: Holt and Co., 1947.
RETTIG, S. An exploratory study of altruism. Dis- SHURE, M. B. Fairness, generosity and selfishness: The
sertation Abstracts, 1956, 16, 2229-2230. naive psychology of children and young adults.
RIBAL, J. E. Social character and meanings of selfish- Child Development, 1968, 39, 875-886.
ness and altruism. Sociology and Social Research, SILVERMAN, L. W. Incidence of guilt reactions in
1963, 47, 311-321. children. Journal of Personality and Social Psy-
RICE, G. E., JR. Aiding behavior versus fear in the chology, 1967, 7, 338-340.
albino rat. Psychological Record, 1964, 14, 165-170. SKINNER, B. F. Science and human behavior. New
RICE, G. E., JR. Aiding responses in rats: Not in York: Free Press, 1953.
guinea pigs. Paper presented at the meeting of the STAUB, E. The effects of success and failure on chil-
American Psychological Association, Chicago, Sep- dren's sharing behavior. Paper presented at the
tember 1965. 39th annual meeting of the Eastern Psychological
RICE, G. E., JR., & GAINER, P. "Altruism" in the Association, Washington, D. C., April 1968.
albino rat. Journal of Comparative and Physio- STAUB, E., & FEAOANS, L. Effects of age and number
logical Psychology, 1962, 55, 123-125. of witnesses in children's attempts to help another
ROSENBAUM, M. The effect of stimulus and back- child in distress. Paper presented at the 40th annual
ground factors on the volunteering response. Jour- meeting of the Eastern Psychological Association,
nal of Abnormal and Social Psychology, 1956, 53, Philadelphia, Pennsylvania, April 1969.
118-121. STAUB, E., & SHERK, L. Need for approval, chil-
ROSENBAUM, H., & BLAKE, R. Volunteering as a dren's sharing behavior, and reciprocity in snaring.
function of field structure. Journal of Abnormal Child Development, in press.
and Social Psychology, 1955, 50, 193-196. STONE, L. A. Social desirability correlates of social
ROSENHAN, D. The origins of altruistic social au- responsibility. Journal of Personality and Social
tonomy. Princeton, N. J.: Educational Testing Psychology, 1965, 2, 756-757. (a)
Service, 1967. STONE, L. A. Rejoinder to Berkowitz: Social de-
ROSENHAN, D., & WHITE, G. M. Observation and re- sirability or social responsibility? Journal of Per-
hearsal as determinants of prosocial behavior. Jour- sonality and Social Psychology, 1965, 2, 758. (b)
nal of Personality and Social Psychology, 1967, 5, SULLIVAN, H. S. An interpersonal theory of psy-
424-431. chiatry. New York: Horton, 1953,
302 DENNIS L, KREBS

TESSER, A., GATEWOOD, R., & DRIVER, M. Some de- WALSTER, E., & PRESTHOLDT, P. The effect of mis-
terminants of gratitude. Journal of Personality and judging another: Overcompensation or dissonance
Social Psychology, 1968, 9, 233-236. reduction ? Journal of Experimental and Social Psy-
TEST, M. A., & BRYAN, J. H. Dependency, models chology, 1966, 2, 85-97.
and reciprocity. Journal of Personality and Social WHEELER, L., & WAGNER, C. M. The contation of
Psychology, in press. generosity. Paper presented at the 39th annual
TOMKINS, S. S. The psychology of commitment: I. meeting of the Eastern Psychological Association,
The constructive role of violence and suffering for Washington, D. C., April 1968.
the individual and for his society. In S. S. Tomkins WHITE, G. The elicitation and durability of altruistic
& C. E. Izard (Eds.), Affect, cognition, and per- behavior in children. Unpublished doctoral dis-
sonality. New York: Springer, 1965. sertation, Princeton University, 1967.
TORRANCE, E. P., & ZHXER, R. C. Risk and life ex- WHITE, R. W. Ego and reality in psychoanalytic
perience: development of a scale for measuring theory. Psychological Issues, 1963, 3(3, Whole No.
risk-taking tendencies. (Rep. No. 57-23v) United
11).
States Air Force Personnel Training Research Re- WODEHOUSE, H. Natural selfishness, and its position
port. Washington, D. C.: United States Govern-
in the doctrine of Freud. British Journal of Medi-
ment Printing Office, 1957. cal Psychology, 1929, 9, 38-59.
TURNER, W. D. Altruism and its measurement in
children. Journal of Abnormal and Social Psy- WRIGHT, B. Altruism in children and the perceived
chology, 1948, 43, 502-516. conduct of others. Journal of Abnormal and Social
UGUREL-SEMIN, R. Moral behavior and moral judg- Psychology, 1942, 37, 218-233.
ment of children. Journal of Abnormal and Social YERKES, R., & YERKES, A. Social behavior in infra-
Psychology, 1952, 47, 463-474. human primates. In G. Lindzey (Ed.), A hand-
WALLACE, J., & SADALLA, E. Behavioral consequences booh of social psychology. Worchester, Massa-
of transgression: The effects of social recognition. chusetts: Clark University Press, 1935.
Journal of Experimental Research in Personality,
1966, 1, 187-194. (Received April 21, 1969)