Journal of Counseling Psychology

Vol. 2, No. A, 1955

Relationships Among Acceptance of Self,

Acceptance of Others, and MMPI Scores
Emanuel M. Berger
University of Minnesota
The general purpose of this exploratory pointed in me." "1 seem to have a real
study is to increase the diagnostic value of inner strength in handling things. I'm on
the Minnesota Multiphasic Personality In- a pretty solid foundation and it makes me
ventory (6) in terms of Expressed Accept- pretty sure of myself."
ance of Self (SA) and Expressed Accept- Examples of AO items are: "There's no
ance of Others (AO); and to indicate how sense in compromising. When people have
various scores or patterns on the MMPI values I don't like, I just don't care to
reflect these other variables for a college have much to do with them." "I enjoy
population. doing little favors for people even if I don't
The definition of SA is in terms of rely- know them well."
ing on one's own standards and values; More complete definitions of SA and AO
having faith in one's capacity to cope with and data on the reliability and validity of
life; assuming responsibility for one's own these scales are given in (1).
behavior; accepting criticism or praise ob- In the interpretation of any results hav-
jectively; not denying or distorting feel- ing to do with SA and AO, it is important
ings, motives, abilities, and limitations, but to bear in mind that what is obtained is
accepting them; considering oneself a per- what individuals express about their ac-
son of worth, equal to others; not expect- ceptance of self and others, and that a
ing rejection by others without reason; not basic and realistic self-acceptance or other-
regarding oneself as different from others, acceptance is not necessarily implied. In-
"queer"; and not being shy or self-con- flated self-pictures that may represent un-
scious. realistic overcompensation for basic inse-
The definition of AO is in terms of not curity can also be obtained.
rejecting or passing judgment against oth- However, such a phenomenological ac-
ers whose standards or behavior are con- count may be helpful in understanding per-
tradictory to one's own; not attempting to sonality dynamics, or at least in suggesting
dominate others; not assuming responsi- fruitful hypotheses. In some cases the ob-
bility for others; not denying the worth of tained results are contrary to what might
others or their equality as persons with be expected by counselors who adhere too
oneself; showing a desire to serve others; strictly to some modification of the inter-
taking an active interest in others; being pretations in the manual for the MMPI (6),
careful not to infringe on the rights of which are based on psychiatric populations,
others in attempting to advance one's own or to interpretations based on a single scale
welfare. without regard to pattern. The lack of in-
There are thirty-six items on the SA formation on which to base interpretation
scale, twenty-eight on AO. Respondents for a relatively normal population is prob-
are asked to indicate on a five-point scale, ably a factor here. In other instances,
the extent to which each statement is true where there may be no particular expec-
for them. tations, the data can be informative.
Examples of SA items are: Tin afraid The subjects were unselected testees at
for people that I like to find out what the Student Counseling Bureau of the Uni-
I'm really like for fear they'd be disap- versity of Minnesota for whom the MMPI

279
 280 Emanuel M. Berger

Table 1
Product-Moment Correlations Among SA, AO, and the MMPI
(N for men = 109. N for women = 76)

Hs D Hy Pd Mf Pa Pt Sc Ma Sc
Men .58"—.08 —45* . 2 5 * ' — 03 .12 — . 3 0 " — . 5 2 " — .40** — .11 —.63"
SA
Women .57** — . 2 5 " — 5 4 " .00 —.26* —.03 —.29* — . 5 5 " — . 4 9 " — .12 —.70"
Men .56* .22* —.11 . 3 8 " —.03 .09 —.01 —.19 14 — . 2 8 " —.46*
AO
Women .24* —.15 —.11 .01 —.16 —.27' —.11 — . 3 2 " —.29* —32. —.20*
* P < .05
•• P < .01
was assigned largely as part of some rou- significantly with SA for girls, but not for
tine procedure or as an adjunct to coun- men. His finding that the Mf scale was
seling, rather than in connection with a more important in the security of boys is
presented personal problem. There were not paralleled in the present results, but
109 males and 76 females, most of whom Mf was found to be significantly corre-
were students and within the 18-22 age lated with AO for girls, the correlation be-
range. ing a negative one.
Half of the subjects were asked to re- In the present findings Pd is signifi-
spond to the SA and AO Scales before cantly related to SA for girls but not men,
taking the MMPI and half after taking it. and Pa is significantly related to both SA
Their mean scores on the MMPI are and AO for men only.
essentially similar to those for 608 men The comparative results for Hy will be
and 508 women freshmen at the Univer- discussed in a separate section that follows.
sity,1 and in no respect more deviant in
the direction of maladjustment. K and SA
Pearson product-moment correlations The fairly high and positive correlations
were computed among SA, AO, and scales between K and SA for both men (r = .58)
of the MMPI using the Fisher r into z and women (r = .57) suggest that K may
transformation (8) in the process of de- function as more than a measure of test-
termining the significance with which ob- taking attitude, and is reflecting to an ap-
tained correlations departed from zero. preciative extent an expressed acceptance
The correlations shown in Table 1 are of self.
similar, with some exceptions, to those This is consistent with inferences made
found by Gough (4), using the Social Se- by Sweetland and Quay (9) who found a
curity-Insecurity Test (SI) of Maslow, correlation of .72 between K and the num-
Hirsch, and Honigman (7), and getting its ber of symbolized dreams and that the
•correlations with the scales of the MMPI. latter were produced by well-adjusted peo-
Since the scale Gough used attempts ple. They also found correlations of .68
to measure "insecurity," his positive cor- and .74 between K and the same measure
relations correspond in general meaning to of social security referred to previously.
negative correlations reported here for SA. On the basis of this and other data they
The similarities in the correlations in- conclude that K measures in the same area
clude K, D, Pt, and Sc, although the cor- as social security and something opposite
relations of SA with these variables tend to Pt and Sc. In this study, similarly, SA
to be higher. Gough also found that the shows a fairly high positive correlation
Hs scale was more important to the se- with K and fairly high negative correla-
curity of the girls and Hs does correlate tions with Pt and Sc.
It is possible to think of SA as repre-
Unpublished data from Donald P. Hoyt. senting defensiveness, too, in the sense of
 Self-Others-MMPI Scores 281
denial of self-rejecting attitudes Most of meaning that, in their judgment, the item
the items on the SA scale are posed in could be answered cither way by such
such a way that the higher the degree of a person.
denial, the higher the scores. There is The range of raw scores on K was 14-21,
some evidence that extremely high scores the mean 17.9 If '">" responses are count-
are analogous to extremely high K in im- ed as K responses, the range becomes 16-
plying overly self-accepting (and other-ac- 25, with a mean of 20 0 Tim would rep-
cepting) attitudes of an improbable sort, resent maximal K responses th.it could con-
such that the validity of the scores is ques- ceivably be made by "a normal, realistically
tionable This in the fact that the frequen- self-confident" person, as judged by this
cy of elevated L scores (T > 60) is sig- group of counselors.
nificantly greater among menTvhose scores In terms of T scores, the lange of scores
on SA are equal to or greater than one on K would be 53-66, with a mean of 61.
sigma above the mean as compared with When the "?" responses are counted, the
scores in any other range for SA How- range becomes 57-74, the mean, 64.
ever, the point here is that this is a two- Interpreting these results conservatively
sided coin and that what has been con- and relying more on the definite K re-
sidered normal defensiveness (normal K) sponses rather than including the "?" re-
may also be considered normal self-ac- sponses, it is suggested that K scores in
ceptance or social security. Likewise the the range of 55-65 could be made by nor-
tendency to get poor scores (low K) may mal, realistically self-confident people as
be considered as self-rejection or social in- opposed to some interpretation in terms of
security within the limits suggested by the "defensiveness." The converse, that peo-
correlations with K. ple who make such K scores are self-con-
It seems likely that K is functioning most fident, does not always follow, because the
as a validity indicator at its upper extremes content of the items that make up any par-
when it involves improbably "good" atti- ticular K score can vary a great dual.
tudes which imply that one never feels
like swearing or smashing things. Where Hy and SA
one indicates a generous attitude toward According to Cottle (2), a high average
others as persons and denies feeling use- score on Hy indicates a lack of self-conti-
less or worrying frequently or being shy, dence. His interpretation is apparently
there seems little reason to deduce an eva- based on the findings of Cough (5), with
sive or deceptive attitude toward the test which the present findings are in general
rather than to recognize this as denial of agreement, although this is a striking ex-
distrusting others and an indication of ception. Gough reported a low positive
something like self-confidence. It is gen- correlation between Hy and insecurity,
eral knowledge that low K is associated whereas Table 1 indicates a low positive
with the profiles indicative of severe emo- correlation between SA and Hy which is
tional disturbance, and this seems logical significantly different from zero at the .01
in terms of "true" responses to the items level. Furthermore, in examining the data,
just described as suggesting self-confidence it was observed that an Hy elevation
when denied. (T ^ 60) in otherwise normal profiles
Some evidence was obtained regarding was~accompanied by high SA. A scatter-
one aspect of the above discussion of K. gram for Hy and SA for the 109 men re-
Eleven counselors on the staff of the Stu- vealed that most of the very deviant plot's
dent Counseling Bureau were asked to were for the pattern Hy high, SA low, and
respond to the K items as they believed variables such as D, Pt, Sc, and Sie high.
"a normal, realistically self-confident per- When Hy is low for men SA tends to be
son" would. low, but when Hy is high, SA may be
They were instructed to respond using either high or low depending on other
"true," "false," and "?" with the latter scores in the profile.
282 Emanuel M. Berger
Table 2
Means and SD's for Total Groups on SA and AO, and Comparisons of Means on SA and AO
Accompaming Certain Patterns of the MMPI, with Means for Total Croup of Men
SA AO
N M SD M SD
Total (Men) 10') 151.0 19.7 109.3 10.4
I otal (V\ omen; 76 146 6 19 4 1150 109
t
Patterns SA AO
Hs > 60, \ s a 65 10 156 6 2.1 112.8 9.4 2 0* 11
Pa 60, Vsa 65 11 152 3 17 1 1134 11 1 12
60, Vsa 65 12 107.9 165 90.8 15.3 8.1" 3.9*
Si ^ 65. Other Vsa < 65. Pa < 60 6 165.2 3.0 1148 5.2 6.0" 2 2*
Pt, Sc ^ 65, Other Vsa < 65, Pa < 60 9 131.7 14.3 103 2 13 5 3.6" 1.3
' P < .05
•• p < .01
Note Vsa includes clinical scales of the MMPI significantly correlated with SA at .01 level for both men and
women D, Pt, Sc, Sle. Vsa ,> 65 means that one or more of these MMPI scales equals or exceeds a T
score of 65. —
Certain Patterns of the MMPI considered (—.40), for the pattern in Table
Table 2 shows how the group means on 2 the direction is reversed, with high SA
SA and AO compare with SA and AO ac- and AO accompanying the high Sc. This
companying certain patterns on the MMPI. probably is contrary to what might be ex-
No particular expectations for the relation- pected a priori, and also contradicts the
ship of Hs to SA appear in the literature usual interpretation of "withdrawal" since
to the writer's knowledge. The data for the accompanying Sie scores are more than
the condition of Hs shown above indicate a sigma below the mean for the whole
that higher-than-average SA and average group of men. This holds to the extent
AO may accompany elevated Hs. that Sie is valid as an indicator of sociabil-
Cottle (2) says that there is little evi- ity and there is evidence for this in (3)
dence concerning the meaning of the Pa and (4).
scale and suggests an interpretation of The results for Pt and Sc elevations for
above-average Pa as indicating the indi- the described conditions probably are in
\ ldual who "feels he is being picked on" accord with what would be expected. How -
and "more concerned with self than others ever, the extremely low scores on SA and
and less able to become interested in AO that accompany a Pa elevation when
jthers." The results for the above pattern Vsa are elevated may not be.
for Pa suggest average SA and some tend- Table 3
ency for higher than average AO. This is
Frequencies with Which Certain Cut-off Scores
in line with a stud)' by H G Gough2 in on the MMPI Are Associated With
which Air Force officers who had Pa ele- Lo\ver-Than-A\trage SA and AO
vated around T = 60, but otherwise nor-
Number below
mal MMPI's, were described as "self-con- mean X-
fident" and "tolerant of others" by their
N SA AO SA AO
fellows.
Men 18 15 14 8.0" 5.6'
This evidence suggests a revision of one K < 50
traditional interpretation for Pa for the Women 13 13 8 13.0" 0.7
stated conditions, although the psycholo- 17 14
Men 13 7,2" 4.8*
gical meaning of such an elevation proba- H> < 50
bly would have to be approached by ana- Women 30 18 14 1.2 0.1
lyzing responses to Pa items.
Men 23 23 18 23 0 " 7.3"
Although Sc shows a fair negative cor- Sie ;> 60
relation with SA when patterns are not Women 16 16 10 16 0 " 1.0
.05
2
Personal communication. .01
 Self-Others-MMPI Scores 283

Single Scales of the MMPI patterns of the MMPI to SA and AO and

Table 3 shows how K, Hy, and Sie alone the effectiveness of K, Hy, and Sie as
are associated with lower-than-average SA single predictors of lower-than-average SA.
and AO. Sie is the most effective predic- Received May 20, 1955.
tor of low SA for both men and women.
This becomes understandable when the References
content of the items for each of these 1. Berger, E. M. The relation between expressed
MMPI scales is considered. The Sie scale acceptance of self and expressed acceptance
is much more heavily loaded than the of others. J. abnorm. soc. Psychol., 1952, 47,
778-782.
other two scales with items that have 2. Cottle, W. C. The MMPI: A review. Kansas
to do with confidence in social situations Studies m Education, 1953. Vol. 3, No. 2.
and one's ability and other items related 3. Drake, L. E., & Thiele, W. B. Further valida-
in content to self-acceptance as defined tion of the social 1. E. scale for the MMPI.
J. educ. Res., 1948, 41, 551-556.
here. The K and Hy scales are more com- 4. Gough, H. G. A research note on the MMPI
plex in content and, while there is an over- social I. E. scale. ]. educ. Res., 1949, 43, 138-
lap of from eight to ten items of each of 141.
the three scales on each other, the K and 5. Gough, H. G. A note on the security-insecurity
Hy scales have fewer items related to tost. /. soc. Psychol, 1948, 28, 257-261.
6. Hathaway, S. R., & McKinley, J. C. Manual
self-acceptance. for the Minnesota Multtphasic Personality In-
It is not clear why Hy is not effective ventory. New York: Psychological Corp., 1943.
in predicting low SA and AO for women, 7. Maslow, A. H., Hirsch, E., Stein, Marcella, &
or why none of these three scales predicts Honigman, Irma. A clinically derived test for
measuring psychological security-insecurity. J.
low AO for women under the stated con- gen. Psychol., 1945, 33, 21-41.
ditions. A comparative analysis of men 8. McNemar, Q. Psychological statistics. New
and women on the content of the items York: Wiley, 1949.
on the three scales that accompany low 9. Sweetlaod, A., & Quay, H. A note on the K
scale of the Minnesota Multiphasic Personality
SA and AO probably would be fruitful Inventory. ]. consult. Psychol, 1953, 17, 314-
for this purpose. 316.

Summary Comment
The purpose of this study was to add The functional unities of personality are
to the diagnostic value of the MMPI as so poorly mapped at the present time that
used with college students by exploring the validation of any test which yields
its relationships with Expressed Accept- scores on multiple trait names becomes a
ance of Self (SA), and Expressed Accept- life-time task. The Rorschach, with its
ance of Others (AO). The three measures present list of several thousand references,
were administered to 109 men and 76 wo- offers the most striking illustration of the
men, testees at a student counseling bu- current problems in validating personality
reau. Most of the subjects were college test hypotheses whether they are derived
students. SA correlated significantly with from projectively or psychometricalry con-
K, D, Pt, Sc, Sie, and Pa for both men and structed tests. Interest in the meaning of
women, as well as with three other scales MMPI scores, if gauged in terms of the
for which the correlations were significant literature devoted to MMPI studies, is
for men or women but not both. AO cor- rapidly increasing and may soon exceed
related significantly with K. Hs, Hy, Ma, the interest in the Rorschach.
and Sie for men; and with K, Mf, Pt, Sc, For purposes of validation, both of these
and Sie for women. Evidence for con- tests suffer from the difficulties involved
sidering K as measuring something more in securing representative samples of sub-
than "defensiveness," i.e., something akin jects who can be definitely utilized as ex-
to SA, was discussed. Data were pre- ternal criterion groups. The lack of repre-
sented regarding the relation of certain sentativeness produces a large number of
284 Emanuel M. Berger

"statistically significant" correlations which tained correlations are edging us closer and
in cross-validation studies turn out to be closer to a useful grasp of functional
chance-determined. The lack of external unities.
criterion groups forces the research-minded An additional source of confusion is
to correlate unvalidated test with unvali- emerging as the analyses of test relation-
dated test or to resort to factor analysis ships become more refined and more so-
which only too often turns up unnameable phisticated. Men and women do not seem
factors. to be quite so homogeneous where their
Despite these encumbrances, the abun- personality structures are concerned. In
dant research on the MMPI is beginning Berger's study below it appears that the
to yield some valuable hypotheses which same score on the Hy-scale may have dif-
are becoming of increasing importance in ferent meanings when obtained by a male
psychopathology; and, as illustrated by the than when it is obtained by a female. The
Berger study, are making the MMPI more same discrepancy might be true for scores
useful in understanding the relatively on Pt. Somewhat analogous difficulties
normal. The innocent appearing K-scale, have been showing up in the experimental
which was originally developed as a kind analysis of the Anxiety scale by the Iowa
of suppressor variable for estimating test- group.
taking attitudes, now appears to have a Up to this point we have been consider-
variety of functional relationships with se- ing the question of how the Berger study
curity feelings, with self-acceptance, with throws additional light upon MMPI scores.
acceptance of others, and with general ad- One might also ask what understanding of
justment—to mention only a few relation- Berger's scales for Self-Acceptance and Ac-
ships which have emerged in recent studies. ceptance of Others is provided by his
Similarly, the Pt-scale which correlates correlations with the MMPI. Happily,
negatively with Berger's Self-Acceptance most of his correlations are comfortingly
Scale also correlates almost perfectly with negative, so that one may still believe
the Taylor Anxiety Scale. Thus, bit by that the self-accepting person and the per-
bit and with many doubts and confusions, son who accepts others can be non-patho-
these laborious and time-consuming studies logical personalities. The fly in the oint-
are etching a pattern of understandable ment shows up when one perceives that
psychological relationships for which meas- for men there is a significantly positive
uring instruments are also available. It is correlation between acceptance feelings
not surprising to find that self-acceptance and the Hy-scale. Does more than aver-
is probably related to security and to gen- age Hy influence the results on the Berger
eral adjustment. It is, however, comfort- scales? Should Berger devise a K-scale
ing to know that relatively objective in- for his acceptance scales? Such questions
struments have been employed to establish may, when taken in their flood, lead on
the correlation and that these same instru- to madness; at other times they have led
ments will be available for research in to the development of K scales and of
other areas. One may bewail the low cor- other questionnaire techniques which pro-
relations and curse the semantic problems vide for control of social acceptability in
involved in separating self-acceptance from the selection of test responses by subjects.
security and then security from adjustment
(if they can be distinguished!) without los- Victor Raimy
ing sight of the fact that consistently ob- University of Colorado-