Soc Indic Res (2018) 136:663–692

https://doi.org/10.1007/s11205-017-1580-x

Measuring Attitudes Toward Distributive Justice:

The Basic Social Justice Orientations Scale

Sebastian Hülle1 • Stefan Liebig1 • Meike Janina May1

Accepted: 13 February 2017 / Published online: 23 March 2017

 Springer Science+Business Media Dordrecht 2017

Abstract Previous research on social inequalities relied primarily on objective indicators.

According to recent studies, however, subjective indicators that reflect a person’s per-
ceptions and evaluations of inequalities are also relevant. Such evaluations depend on an
individual’s normative orientation, so respective attitudes toward distributive justice need
to be accounted for appropriately. This article introduces a short scale for measuring such
order-related justice attitudes. The introduced Basic Social Justice Orientations (BSJO)
scale comprises current insights into the empirical justice research and measures indi-
viduals’ attitudes toward the following four basic distributive principles: equality, need,
equity, and entitlement. The BSJO scale has four dimensions that measure support for
these four justice principles on the basis of eight items. We assess the quality of the scale
using data from three general population surveys conducted in Germany: the first wave of
the panel ‘‘Legitimation of Inequality Over the Life Span’’ (LINOS-1), the Innovation
Sample of the Socio-Economic Panel (SOEP-IS 2012), and the German General Social
Survey (ALLBUS 2014). The scale was found to be a valid instrument that can be used to
measure order-related justice attitudes toward distributive justice. The BSJO scale is a
short and therefore time-efficient instrument that can be implemented in general population
surveys.

Keywords Social justice
Justice attitudes
Distributive justice
Justice ideologies

Attitude measurement
Survey research

Electronic supplementary material The online version of this article (doi:10.1007/s11205-017-1580-x)

contains supplementary material, which is available to authorized users.

& Sebastian Hülle

[email protected]
1
Faculty of Sociology, University of Bielefeld, Bielefeld, Germany

123
664 S. Hülle et al.

1 Introduction

Inequalities in wealth and income have increased over the past few decades in many
countries (Nolan et al. 2014; Piketty 2014; OECD 2015; see also Atkinson and Bour-
guignon 2015). This increase can have certain implications for both the individual and
society. Previous studies that focused on the consequences of inequalities on well-being
relied primarily on objective indicators of inequality and produced research results that are
inconclusive, with positive, negative, or insignificant effects (e.g., Verme 2011; Clark and
d’Ambrosio 2015). We argue that the reason for these seemingly contradictory results lies
in different perceptions and evaluations of objective inequalities (see Kluegel and Smith
1986; Nielsen 2017). Social inequalities are the result of distributions of goods and burdens
in a society. However, people can evaluate the same objective social inequality differently
depending on their individual preferences for specific distributive principles (Hegtvedt and
Isom 2014). Hence, when studying support for socio-economic equality policies and the
consequences of inequality, there are good reasons to not exclusively rely on objective
indicators of inequality but to additionally consider subjective indicators (see Roller 1995).
However, because such evaluations depend on an individual’s normative orientation
(Hadler 2005), the respective justice attitudes must also be accounted for appropriately.
Following Liebig and Sauer (2016), we call such attitudes order-related justice atti-
tudes. These normative attitudes relate to the rules or norms people think should guide the
allocation and (re)distribution of goods and burdens within social groups and society.1
Normative ideas as to how benefits and burdens should be distributed in a society have
been an integral part of large-scale population surveys since the early 1970s (e.g., General
Social Survey [GSS]) and 1980s (e.g., International Social Survey Programme [ISSP]).
Since these surveys are designed to observe changes in attitudes over time, they often use
measurement concepts that were developed in the 1970s and 1980s; but these measurement
concepts do not capture the developments in empirical justice research over the last two
decades. Furthermore, they often lack a thorough theoretical and conceptual framework.
For example, common measurements for order-related justice attitudes only distinguish
between preferences for equality or inequality as basic normative orientations. This is
insufficient to understand how an individual evaluates the allocation of benefits and bur-
dens in society and the resulting social inequalities.
In this article, we introduce the Basic Social Justice Orientations (BSJO) scale, which
measures an individual’s order-related justice attitudes in a more comprehensive and
differentiated way than common instruments in survey research so far. The BSJO scale
reflects the current state of theory building in social justice research and allows for the most
recent empirical insights into the structure of order-related justice attitudes (Deutsch 1975;
Konow 2003; Gollwitzer and van Prooijen 2016; Liebig and Sauer 2016). It measures an
individual’s preference for the four basic justice principles: equality, need, equity, and
entitlement. It is both sufficiently differentiated and, with only eight items, is a short scale,
so it can be applied easily and efficiently in large-scale population surveys.

1
In line with the distinction first proposed by Wegener (1992) and later by Liebig and Sauer (2016), we can
distinguish between order-related, procedure-related, and outcome-related justice attitudes. Order-related
justice attitudes are distinct from procedure-related and outcome-related justice attitudes: Procedure-related
justice attitudes are preferences concerning decision-making procedures that are used to ensure just allo-
cation and distribution (e.g., lotteries, majority decisions). By contrast, outcome-related justice attitudes
reflect assessments of allocation outcomes or distributions (e.g., the amount of money that would be
considered a just income). In this article, we introduce a measure of order-related justice attitudes.

123
Measuring Attitudes Toward Distributive Justice: The Basic… 665

The article is organized as follows: We first provide details of how order-related justice
attitudes were previously measured in survey research and the problems associated with
these methods (Sect. 2), followed by a discussion of the theoretical foundations of the
BSJO scale and how it overcomes the problems of earlier measurement approaches
(Sect. 3). We give an overview of the three population surveys in which the BSJO scale
was implemented—the first wave of the panel ‘‘Legitimation of Inequality Over the Life
Span’’ (LINOS-1), the Innovation Sample of the Socio-Economic Panel (SOEP-IS 2012),
and the German General Social Survey (ALLBUS 2014)—and discuss the operational-
ization of the BSJO scale (Sect. 4). We provide evidence for the validity of the BSJO scale
in Sect. 5. The analyses follow the recommendations for quality standards of survey
instruments by the German Data Forum (Rat für Sozial- und Wirtschaftsdaten [RatSWD])
(see Rammstedt et al. 2015). Finally, we offer suggestions for future use of the BSJO scale
and discuss its research potential.

2 Measurement of Order-Related Justice Attitudes in Population Surveys

Perceptions and evaluations of social inequalities as well as normative ideas of how

benefits and burdens should be distributed in a society have been among the key com-
ponents of nearly all major national and international survey programs in social science
research since the 1970s and 1980s (e.g., GSS, ALLBUS, and ISSP). In the beginning,
researchers’ primary purpose was to examine the distribution of preferences with respect to
equality and inequality, whereas in the 1990s, a more differentiated approach began to
develop that was inspired by the advances being made in psychological and social-psy-
chological justice research (see Liebig and Sauer 2016). Wegener and Liebig (1995)
proposed an instrument for measuring justice attitudes that was more differentiated than
those that had been previously used. The instrument was theoretically built on the grid-
group theory of cultural anthropologist Mary Douglas (Douglas 1982). In accordance with
the theory, Wegener and Liebig (1995) distinguished four ‘‘ways of thinking about jus-
tice,’’ or ‘‘justice ideologies’’: ascriptivism, egalitarianism, individualism, and fatalism,
each of which relates to two structural dimensions of social life: membership in social
groups (group) and restrictions on individual freedom of choice imposed by social regu-
lations (grid). The authors detailed these four justice ideologies as follows: Ascriptivists,
first, will have a vested interest in keeping the responsibility for prevailing social
inequalities impersonal. They maintain that the factual distribution of privileges and
goods, although in fact ascribed, is natural and self-evident and therefore good. We
call this type of justice ideology ascriptivism. Enclavists, on the other hand, appeal to
authority. They hold the state ultimately responsible for the unjust distribution of
wealth. Hence, the redistribution of resources that aim [sic] to fulfill the egalitarian
wish for maximum equality should, in their view, also be the responsibility of the
state. This is therefore the justice ideology of egalitarianism. Individualists, in
contrast, hold that only achievers are rewarded with success and that a system of free
competition is fair and functional; they support individualism as justice ideology.
This is in direct contrast to the point of view of the isolates who blame the ‘‘system’’
for his or her unfortunate situation. Feeling at the mercy of a ‘‘system’’ that denies
them justice they tend to accept their situation fatalistically (Wegener and Liebig
2000, pp. 184–185, emphasis in original).

123
666 S. Hülle et al.

The measurement of these four justice ideologies was not based on a specifically developed
scale; instead, it relied on eight items from the ISSP module on social inequality (1987).
The items of the justice ideologies have been validated in the course of several survey
waves of the International Social Justice Project (ISJP) (Stark et al. 2000). Originally, the
items were embedded in different item batteries with different introductory questions (see
Table 13 in Appendix).
The justice ideologies derived from Douglas’ cultural theory represent a major step
toward the measurement of order-related justice attitudes for three reasons. First, the
measurement is based on a clear theoretical conceptualization that not only defines the
scope of order-related justice attitudes but also formulates hypotheses regarding the (socio-
structural) causes of varying justice attitudes (Liebig and Schlothfeldt 2002). Second, the
justice ideologies take into consideration the fact that there are other principles of distri-
bution besides preferences towards equality and inequality that are regarded as legitimate
or just. Third, the scale has certain validity, as it correlates and predicts other justice-
related attitudes and behaviors.2
However, the concept of justice ideologies and its measurement has five major
shortcomings:
1. The items used to measure the preference for a distributive principle (egalitarianism:
equality; individualism: equity; ascriptivism: entitlement) are not one-dimensional: the
dimension of a preferred principle is connected with a dimension that is related to the
institution that is responsible for the distribution in question. It is generally assumed
that people who lean toward egalitarianism prefer a redistributing state, whereas
individualistically inclined people prefer distribution by the market. No such
equivalent exists for ascriptivism in the respective items. Due to this confounding
of distributive principle and distributing institution an item might measure a low value
of support for a justice principle, even if this principle is strongly supported by a
respondent. The reason for this may be that the respondent disapproves of the
distributing institution.
2. The four justice ideologies cannot be used to measure preferences related to the need
principle as a separate principle. The different applications of the concept show that
the equality principle and the need principle are confounded in the ideology of
egalitarianism. As a result, there are two rule preferences that are regarded as distinct
in justice and welfare state research and that cannot be measured independently of
each other with the justice ideology scale.
3. From the perspective of social psychological research, fatalism is a dimension that
cannot be derived logically and therefore does not belong to the set of basic world
views that are the subject of justice research (Deutsch 1975, 1985; Miller 1999;
Konow 2003).
4. The individual items listed in Stark et al. (2000) do not clearly and unambiguously
relate to societal distribution problems. In addition, most of these items do not use a
framing that refers to justice; only the two items for the ideology of fatalism
incorporate the words ‘‘just’’ or ‘‘justice’’. Some of the items do not include any
explicit stimulus that would indicate that what is concerned are questions related to the
distribution of benefits and burdens within society rather than within partnerships,
families, groups, or organizations. Thus, it remains unclear whether the respondents

2
See the extensive body of literature provided by the German workgroup of the International Social Justice
Project at https://www.sowi.hu-berlin.de/de/lehrbereiche/empisoz/forschung/archiv/isjp.

123
Measuring Attitudes Toward Distributive Justice: The Basic… 667

are actually aware that they are being asked to state their preferences concerning
distributive justice in society (see Liebig 2001).
5. The individual items of the scale are not measured together as parts of a single item
battery but instead are measured as parts of several different item batteries. Each of
these item batteries involved a different introductory text (see Table 13 in the
Appendix and Stark et al. 2000) in the surveys of the International Social Justice
Project (ISJP) in 1991, 1996, 2000, and 2006. What makes this issue even more
difficult is that these different item batteries were not included in successive order in
the questionnaire but were placed in different parts. One consequence of this is that the
use of the scale—resp. the related items—is less time-efficient, as different
introductory texts have to be used for each item battery. Another methodological
problem is that the placement of the different items at different points throughout the
questionnaire may lead to uncontrolled order or context effects.
Given the above limitations, an instrument obviously needs to be developed that (1)
measures order-related justice attitudes on the basis of a more differentiated set of
dimensions; (2) is more compatible with previous normative and empirical justice research
by capturing the theoretical developments of recent years (Jasso 2015; Liebig and Sauer
2016); (3) is clearly related to societal distribution problems and is linked to distributive
principles of welfare states; (4) is more theoretically driven; (5) uses a clear stimulus
related to justice and fairness; and (6) ensures time-efficient application when used in
large-scale population surveys. The BSJO scale was developed to meet these requirements.

3 Conceptualization of the BSJO Scale

3.1 Theoretical Background

Starting in the mid-1970s, psychologists began to understand that individuals use a small,
limited set of basic rules when allocating and evaluating the distribution of benefits and
burdens. In contrast to equity theory (Adams 1963), which had been the predominant
theory until then and which is based on the assumption that justice is assessed solely on the
basis of the principle of proportionality, the new ‘‘multi-principle approach’’ emphasized
the important role of the principles of equality and need (Deutsch 1975). This set of three
basic principles of distributive justice—equity, equality, and need—was later extended to
include a fourth principle, namely, the principle of entitlement, according to which the
allocation and distribution of benefits and burdens are considered to be just if the benefits
and burdens in question are allocated and distributed on the basis of ascribed or acquired
status characteristics (Miller 1976, 1999). For a long time, these four basic principles of
distributive justice were only used to study justice attitudes and justice behavior within
social aggregations below the societal level (e.g., dyads, groups, organizations) (Gollwitzer
and van Prooijen 2016). However, to a varying extent and in different combinations, these
four justice principles can be identified as guiding norms for allocating and distributing
goods and burdens within societal institutions of welfare states (e.g., Clasen and van
Oorschot 2002; Taylor-Gooby 2005; Ullrich 2008).
Accounting for these considerations, the Basic Social Justice Orientations (BSJO) scale
introduced here applies these four principles to examine normative attitudes to the reso-
lution of distribution problems on the societal level. Therefore, the BSJO scale comprises
the following four dimensions or justice principles:

123
668 S. Hülle et al.

• Equality According to the equality principle, the allocation of benefits and burdens is
just if everyone is allocated the same share.
• Equity According to the equity principle, the distribution of benefits and burdens is just
if the benefits and burdens in question are allocated according to individuals’ current
individual contributions and efforts.
• Need According to the need principle, benefits are allocated according to people’s
individual needs.
• Entitlement According to the entitlement principle, benefits and burdens should be
allocated on the basis of specific entitlements that are themselves based on ascriptive
characteristics (e.g., social origin, sex) or on status characteristics that have been
acquired in the past (e.g., occupational status). The main difference from the equity
principle is that benefits are not allocated according to individuals’ current contribu-
tions or efforts.

3.2 Development of the Instrument

The BSJO scale aims at measuring individuals’ distributive justice attitudes regarding the
four basic distributive principles of equality, need, equity, and entitlement. These four
dimensions encompass distributive principles that are identified by theoretical and
empirical research as universal (Deutsch 1975; Miller 1976; Fiske 1992). Therefore, the
developed instrument has the potential to be used for various populations in different
cultures and countries and, thus, in international surveys for country comparisons. Fur-
thermore, the BSJO scale is not limited to specific resources that are distributed; instead, it
captures all sorts of benefits and burdens allocated in a society. The scale is also not limited
to a specific target population: The wording of the items was carefully designed with the
intention to ensure its application in general population surveys. It is easy to understand
and unambiguous for everyone, regardless of individual socio-structural characteristics.
Finally, the BSJO scale was developed as a short eight-item scale to facilitate time effi-
ciency in large-scale population surveys.
The BSJO scale originally consisted of 12 items, with three items allocated to each
dimension. The items were developed based on experiences from item testings of earlier
(unpublished) studies (see, e.g., Jäckle 2002). For example, item B is identical to an item
that was measured by the ISJP in 2006 (Legewie et al. 2007), and item F is based on an
item used in the same survey (Gerlitz et al. 2007, p. 23). All other items were specifically
developed for the BSJO scale. Each item of the BSJO scale was subject to several pretests:
Following the LINOS-1 pretest, Item G was reworded slightly for clarity.3 Following the
SOEP-IS 2012 pretest, no modifications were made to the tested survey tool (see Sect. 4.2).
The pretest for the main survey of ALLBUS 2014 involved the testing of eight of the
original twelve items. Furthermore, cognitive pretesting was applied for the ALLBUS
2014. There were no comprehension problems with the items. However, following the
pretesting, item B was reworded (ALLBUS committee meeting on November 19, 2013;
also see Sect. 4.2).
Using the LINOS-1 dataset, the final BSJO short scale with only eight items was
constructed on the basis of the original 12 items. Table 1 provides an overview of the four
dimensions of the BSJO scale and the items used for each dimension (see Table 1; Sauer

3
German pretest wording of item G (LINOS-1): ‘‘Eine Gesellschaft ist gerecht, wenn Einkommensun-
terschiede gering sind.’’

123
Measuring Attitudes Toward Distributive Justice: The Basic… 669

Table 1 BSJO scale for measuring order-related justice attitudes (Sauer et al. 2014)
Justice Item Item text
principle

Equality C It is just if all people have the same living conditions

K It is just if income and wealth are equally distributed among the members of our
society
G A society is just if there are only minor income disparities between people
Need E A society is just if it takes care of those who are poor and needy
J It is just if people taking care of their children or their dependent relatives receive
special support and benefits
A A society is just if all people have sufficient nutrition, shelter, clothing as well as
access to education and medical care
Equity B It is just if hard working people earn more than others
I It is just if every person receives only that which has been acquired through their own
efforts
H A society is just if differences in income and assets reflect performance differences
between people
Entitlement D It is just if members of respectable families have certain advantages in their lives
L It is fair if people on a higher level of society have better living conditions than those
on the lower level
F It is just if people who have achieved good reputation and wealth profit from this
later in life

Items were asked in alphabetical order. Items in italics are items that are not included in the short version of
the BSJO scale with eight items. The introductory text reads: ‘‘There are different ideas about how a society
can be fair and just. What is your personal opinion about this?’’
Source: Question 2 in the LINOS-1 questionnaire (Sauer et al. 2014, p. 142)

et al. 2014, p. 142). The criterion used for item selection was factorial validity (see
Sect. 5.3).
Please note that the instrument was developed in the German language and has been
used so far only in German-speaking contexts (see Table 14 in the Appendix for the
German version of the scale). The provided English translations of the items in this article
are as accurate as possible, as they involved native English speakers, but they follow no
standards for translations of survey instruments (see, e.g., Harkness 2008, pp. 68–74;
Mohler et al. 2016).

4 Implementation of the BSJO Scale in Population Surveys in Germany

4.1 Data

The BSJO scale was implemented in three large-scale surveys in Germany that used
probability sampling. First, the BSJO scale was part of a split sample in 2012/2013 of the
Innovation Sample of the Socio-Economic Panel (SOEP Innovations-Stichprobe [SOEP-
IS]). The SOEP-IS is an independent, household-based longitudinal survey and supple-
ments the main SOEP survey by providing a respondent infrastructure for testing inno-
vative survey modules and field processes. The target population of the SOEP-IS 2012

123
670 S. Hülle et al.

(Richter and Schupp 2012; SOEP 2014; DIW 2015) is members of the resident population
of Germany living in private households who were at least 16 years old. Second, the BSJO
scale was included in the first wave of the long-term panel ‘‘Legitimation of Inequality
Over the Life Span’’ (LINOS-1) that was conducted in winter 2012/2013. The LINOS
panel investigates the conditions under which inequalities are perceived as problems of
justice and how social contexts influence the formation of justice attitudes over the life
span (Liebig et al. 2014; Sauer and Valet 2014; Valet et al. 2014). The target population
consists of members of the resident population of Germany who were subject to social
security contributions (i.e., employees) on December 31, 2011, and who were between 18
and 57 years old. Third, the BSJO scale was implemented in the German General Social
Survey (Allgemeine Bevölkerungsumfrage der Sozialwissenschaften [ALLBUS]). The
ALLBUS is a biennial survey that has collected data on individual attitudes and behaviors
and social structure in Germany since 1980. The target population of the ALLBUS 2014
(GESIS 2015a, b) is composed of members of the resident population of Germany who
were living in private households and who were at least 18 years old. While SOEP-IS 2012
and ALLBUS 2014 used computer-assisted personal interviewing (CAPI), LINOS-1 is a
multi-mode survey that additionally involves paper-and-pencil interviewing (PAPI) and
computer-assisted web interviewing (CAWI). Table 2 presents additional information
about the specific characteristics of the surveys. Finally, we created a pooled dataset based
on the three datasets with a pooled version of the BSJO scale and comparable variables
(see Sects. 5.3, 5.4 for more details).

4.2 Operationalization

This section describes the operationalizations of the BSJO scale and notes differences
between the three datasets. We use LINOS-1 as a reference because it served as data basis
for the development of the short eight-item version of the BSJO scale, although the data
collection for the SOEP-IS 2012 version was carried out first. In LINOS-1 and SOEP-IS
2012, the BSJO scale consists of twelve items (three per dimension). The ALLBUS 2014
version of the BSJO scale is based on analyses of LINOS-1 and SOEP-IS and represents
the short eight-item version of the BSJO scale (without items A, F, G, H). Table 1 provides
an overview of the formulations of the twelve items in LINOS-1 and of how they relate to
each BSJO dimension.4 The items were presented in alphabetical order to the respondents,
meaning that item A was presented first and item L was presented last. This order is based
on an originally random selection that was retained for all surveys.5
The introductory text in LINOS-1 and ALLBUS 2014 reads as follows: ‘‘There are
different ideas about how a society can be fair and just. What is your personal opinion
about this?’’ The introductory text in the SOEP-IS 2012 differs slightly by framing the
question around ‘‘our society’’. This has the disadvantage that it unnecessarily limits the
scope of the scale. Contrary to that, the more recent formulation in LINOS-1 and ALLBUS
2014 asks the more general question of when ‘‘a society’’ is just. The respondents were
then asked to state to what extent they agree or disagree with each of the items. The five-
point Likert scale distinguishes the following answer categories in LINOS-1 and ALLBUS
2014: ‘‘strongly agree,’’ ‘‘somewhat agree,’’ ‘‘neither agree nor disagree,’’ ‘‘somewhat
disagree,’’ and ‘‘strongly disagree’’ (see Sauer and Valet 2014, pp. 25–28). All categories

4
The German version of the instrument used in LINOS-1 is provided in Table 14 in the Appendix.
5
LINOS-1 includes the PAPI mode, in which a randomization of items could not be implemented. To be
comparable across modes, the same order was applied.

123
Measuring Attitudes Toward Distributive Justice: The Basic… 671

Table 2 Overview of the dataset-specific characteristics for LINOS-1, SOEP-IS 2012, and ALLBUS 2014
LINOS-1 SOEP-IS 2012 ALLBUS 2014

Survey name First wave of the panel Innovation Sample of the German General Social
‘‘Legitimation of Socio-Economic Panel Survey (Allgemeine
Inequality Over the Life (SOEP Innovations- Bevölkerungsumfrage der
Span’’ (LINOS-1) Stichprobe [SOEP-IS]) Sozialwissenschaften
[ALLBUS])
Country Germany Germany Germany
Fieldwork 2012/13 2012/13 2014
Mode CAPI, PAPI, CAWI CAPI CAPI
Target Members of the resident Members of the resident Members of the resident
population population of Germany population of Germany population of Germany
who were subject to living in private who were living in private
social security households and who were households and who were
contributions (employees) at least 16 years old at least 18 years old
on December 31, 2011,
and who were between 18
and 57 years old
Sampling Two samples drawn from Random-route sampling in Two-step disproportionately
official data provided by combination with separate stratified random
the Federal Employment address random sampling sampling:
Agency (Bundesagentur First step: communities in
für Arbeit [BA]) East and West Germany
PAPI/CAWI-sample: drawn with a probability
nationwide random in proportion to the size of
sample their adult population
CAPI-sample: two-step Second step: respondents
selection process. First drawn randomly from
step: 60 sampling points among the residents
randomly drawn from 178 included in the resident
BA districts. Second step: registers
Random sampling of
employees from these
sampling points in
proportion to the number
of employees in each
sampling point
Specifics Oversampling of younger Several questionnaire splits Split-ballot design that
individuals and that reduced the sample reduced the sample size:
individuals with shorter size: BSJO scale BSJO scale presented in
job tenure due to a presented only for a only half of the cases
specific research agenda. subsample of respondents
This design effect can be within the ‘‘innovation
corrected when applying module’’
respective weights
N (survey) 4731 3696 3471
N (BSJO scale 4731 1644 1738
asked)
DOI for used 10.4119/unibi/sfb882.2014. 10.5684/soep.is.2013 10.4232/1.12288
dataset 9
References for Liebig et al. (2014), Sauer Richter and Schupp (2012), GESIS (2015a, b)
dataset and and Valet (2014), Valet SOEP (2014), DIW
questionnaire et al. (2014) (2015)

CAPI computer-assisted personal interviewing, CAWI computer-assisted web interviewing, PAPI paper-and-
pencil interviewing

123
672 S. Hülle et al.

are explicated, and the scale is bipolar (agreement vs. disagreement). In contrast, the BSJO
scale in SOEP-IS 2012 has a 7-point Likert answer scale ranging from 1 ‘‘1 do not agree at
all’’ to 7 ‘‘7 agree completely’’ (only these two categories are explicated). The 7-point
Likert scale has the disadvantage that it is unipolar and does not capture disagreement with
particular distributive principles.
Table 3 gives an overview of the operationalization by dataset—differences from
LINOS-1 are highlighted in italics. The three versions of the BSJO scale differ in the
wording of some items: First, in the SOEP-IS 2012, the items B and F start with ‘‘It is just
that…’’ instead of ‘‘It is just if…’’. This formulation might be interpreted as referring to the
actual situation in ‘‘our society’’ (as in the introductory text) rather than a just society in
general. This potential source of error was eliminated in the LINOS-1 and ALLBUS 2014
versions. Second, item B addresses the equity principle and reads in LINOS-1 as follows:
‘‘It is just if hard working people earn more than others.’’ Following the considerations
after the ALLBUS 2014 pretest, the formulation was changed in order to avoid cross-
loadings on the dimension of the need principle: ‘‘It is just if people who perform well in
their job earn more than others.’’ However, this problem is not entirely solved by the
modification of item B in the ALLBUS 2014. Moreover, item B in the ALLBUS 2014 is
the only item that specifies the context as an occupational setting, which brings into
question the comparability between the items. Therefore, our recommendation is to use the
initial LINOS-1 wording of item B.
For the BSJO scale with 8 items, the median response time in the ALLBUS 2014 is
105 s, i.e., approximately 13 s per question.6
In the following analyses, we had to address the differences in operationalization
between the datasets. To ensure comparability of the values of the BSJO scale across
datasets, and to facilitate an intuitive interpretation, we inverted the scale for LINOS-1 and
ALLBUS 2014 for all analyses so that higher values now reflect a greater degree of
agreement (in the case of SOEP-IS 2012, this was not necessary). In all those cases in
which factor scores were used for calculations, the scale was additionally z-standardized,
since the scaling varies between the datasets (5-point scale vs. 7-point scale).

5 Quality of the BSJO Scale

This section investigates the quality of the BSJO scale following the recommendations for
quality standards of survey instruments given by the German Data Forum (Rat für Sozial-
und Wirtschaftsdaten [RatSWD]) (see Rammstedt et al. 2015). We refer to several stan-
dards regarding objectivity, reliability, and validity and pay particular attention to the
factorial validity and construct validity of the BSJO scale in order to demonstrate its
research potential.

5.1 Objectivity

In each case, the BSJO scale was administered under standardized conditions, which is
taken as an indication of implementation objectivity. In addition, only closed-ended
questions were used, which is taken as an indication of evaluation objectivity. That
descriptions of each dimension of the scale were provided along with reference values in

6
This calculation is based on additional data for the ALLBUS 2014 (justice items) that were provided by
the ALLBUS Research Data Center at GESIS.

123
 Table 3 Operationalization of BSJO scale in LINOS-1, SOEP-IS 2012, and ALLBUS 2014
LINOS-1 SOEP-IS 2012 ALLBUS 2014

Number of items 12 12 8 (without A, F, G, H)

Variable names in dataset v2a–v2l SIM0401–SIM0412 V217–V224
Question order Alphabetical, A to L Same order as in LINOS-1 Same order as in LINOS-1 (without A, F, G, H)
Coding 1 ‘‘strongly agree’’ 1 ‘‘1 do not agree at all’’ 1 ‘‘strongly agree’’
2 ‘‘somewhat agree’’ 2 ‘‘2’’ 2 ‘‘somewhat agree’’
3 ‘‘neither agree nor disagree’’ 3 ‘‘3’’ 3 ‘‘neither agree nor disagree’’
4 ‘‘somewhat disagree’’ 4 ‘‘4’’ 4 ‘‘somewhat disagree’’
5 ‘‘strongly disagree’’ 5 ‘‘5’’ 5 ‘‘strongly disagree’’
6 ‘‘6’’
7 ‘‘7 agree completely’’
Measuring Attitudes Toward Distributive Justice: The Basic…

All categories explicated Yes No Yes

Scaling 5-point 7-point 5-point
Polarity Bipolar (agreement vs. disagreement) Unipolar (agreement) Bipolar (agreement vs. disagreement)
Scale inverted Yes No Yes
for analyses in
this article (so that
higher values
correspond with higher agreement)
673

123
 Table 3 continued
674

Wording differences in BSJO scale compared to LINOS-1

123
Introductory ‘‘There are different ideas ‘‘Now we come to another topic. There are different views when our ‘‘There are different ideas about how a society
text about how a society can be society is just. I am now going to read out a series of statements to can be fair and just. What is your personal
fair and just. you. In each case, please tell me your personal opinion […]’’ opinion about this?’’
What is your personal
opinion about this?’’
Instruction ‘‘Please indicate for each of the ‘‘Please answer using the following scale 1 to 7. A value ‘‘Please answer using the card’’
following statements whether of 1 means ‘‘strongly disagree’’, the value of 7 ‘‘strongly [The interviewer displays card 74 with the
you strongly agree‘‘. With the values in between you can make your estimate’’ answer scale]
agree, somewhat agree, [The interviewer displays list 24569 with the answer scale]
neither agree nor disagree,
somewhat disagree,
or strongly disagree’’
[The interviewer displays list 2
with the answer scale]
Item B ‘‘It is just if hard working ‘‘It is just that hard working people earn more than others’’a ‘‘It is just if people who perform well in their
people earn more than job earn more than others’’c
others’’
Item F ‘‘It is just if people who have ‘‘It is just that people who have achieved good [not asked]
achieved good reputation and reputation and wealth profit from this later in life’’a
wealth
profit from this later in life’’
Item G ‘‘A society is just if there are ‘‘A society is just if there are only minor income [not asked]
only minor income disparities […]’’b
disparities between people’’
S. Hülle et al.
 Table 3 continued

Wording differences in BSJO scale compared to LINOS-1

Item K ‘‘It is just if income and wealth (Only small difference between German
are equally distributed versions that does not exist in English
among the members of our version)d
society’’

Italic = differences in operationalization of the BSJO scale compared to LINOS-1

a
Original German items B and F in LINOS-1: ‘‘Es ist gerecht, wenn’’ versus SOEP-IS 2012: ‘‘Es ist gerecht, dass’’
b
Original German item G in LINOS-1 ‘‘Eine Gesellschaft ist gerecht, wenn die Einkommensunterschiede zwischen den Menschen gering sind.’’ versus SOEP-IS 2012: ‘‘Eine
Gesellschaft ist gerecht, wenn […] Einkommensunterschiede […] gering sind’’
c
Original German item B in LINOS-1 ‘‘Es ist gerecht, wenn Personen, die hart arbeiten, mehr verdienen als andere.’’ versus ALLBUS 2014: ‘‘Es ist gerecht, wenn Personen,
die im Beruf viel leisten, mehr verdienen als andere’’
d
Original German item K in LINOS-1 ‘‘Es ist gerecht, wenn Einkommen und Vermögen in unserer Gesellschaft an alle Personen gleich verteilt werden.’’ versus ALLBUS
2014: ‘‘Es ist gerecht, wenn Einkommen und Vermögen in unserer Gesellschaft an alle Personen gleich verteilt sind’’
Measuring Attitudes Toward Distributive Justice: The Basic…
675

123
676 S. Hülle et al.

the form of means and standard deviations is taken as an indication that interpretation
objectivity was achieved to a large extent (see Rammstedt 2010).

5.2 Reliability

The reliability—or measuring accuracy—of a scale indicates the accuracy with which it
measures a certain personality trait or behavioral characteristic (Lienert and Raatz 1998,
p. 9). We measured the reliability with Cronbach’s alpha (Cronbach 1951), which is
commonly used to assess internal consistency, i.e., the overall degree to which the items of
a scale are intercorrelated.
Table 4 provides an overview of the Cronbach’s alpha values for the four dimensions by
dataset. The highest alpha values in all datasets can be found for the equality dimension.
An exception is SOEP-IS 2012, where the highest alpha exists for the need dimension. The
lowest alpha values appear with the equity dimension. All values are below the frequently
cited value of .7 (Nunnally 1978), which is considered by many authors to be acceptable.
However, this threshold has been called into question, as it can be considered arbitrary.
Cronbach’s alpha, as a measure of internal consistency, has also been questioned. A major
problem is that alpha depends, beside other factors, on the number of items. Hence, the
higher the number of items, the higher is the alpha (Boyle 1991; Clark and Watson 1995).
For this reason, one should expect lower values for a scale if it consists of only the
minimum number of two items, as is the case here with the dimensions (or subscales) of
the short version of the BSJO scale.
Therefore, Clark and Watson (1995) recommend ‘‘the average inter-item correlation
(which is a straightforward measure of internal consistency) [and] a much more useful
index than coefficient alpha per se (which is not)’’ (Clark and Watson 1995, p. 323).
According to Clark and Watson (1995, p. 323), the average inter-item correlation should be
between .15 and .50 for all items. Table 4 provides evidence that all average inter-item
correlations fall within this acceptable range, the only exception being the equality

Table 4 Internal consistency of

Equality Need Equity Entitlement
the short version of the BSJO
scale (eight items) by dataset
LINOS-1
a 0.609 0.407 0.350 0.473
Correlation (AIIC) 0.438 0.255 0.212 0.310
N 4509
ALLBUS 2014
a 0.533 0.458 0.236 0.421
Correlation (AIIC) 0.363 0.297 0.134 0.267
N 1036
SOEP-IS 2012
a 0.435 0.505 0.313 0.437
Correlation (AIIC) 0.278 0.338 0.185 0.280
N 742
a = standardized Cronbach’s a; Pooled dataset
AIIC = average inter-item
correlation; N = sample size; a 0.575 0.443 0.318 0.477
listwise deletion as in the cases of Correlation (AIIC) 0.404 0.284 0.189 0.313
the respective factor analyses N 7748
(see Sect. 5.3)

123
Measuring Attitudes Toward Distributive Justice: The Basic… 677

dimension in the ALLBUS 2014, with an average inter-item correlation of .13. Regarding
the relative size of the values, the same pattern exists as for Cronbach’s alpha: The highest
average inter-item correlations are for equality (except for the need dimension in the
SOEP-IS 2012) and the lowest values are for equity.
Because a theory-driven assessment aims at measuring a single construct systematically,
the ultimate goal is homogeneity or unidimensionality rather than internal consistency.
Moreover, internal consistency is a necessary but insufficient condition for unidimen-
sionality, i.e., a scale can have a high internal consistency but can load on more than one
factor in a factor analysis (Clark and Watson 1995, pp. 322–323). Thus, we turn in the
following to the dimensionality of the BSJO scale and the question of whether the items of
the four dimensions (or subscales) each have a single underlying construct or factor.

5.3 Factorial Validity

The factorial validity can be examined by means of a dimensionality analysis, which must
confirm the existing assumptions concerning the dimensional structure of the construct in
question (see Rammstedt 2010, p. 253). In the current context, this means that empirical
evidence must be provided that the BSJO scale has a four-dimensional structure consisting
of the dimensions equality, need, equity, and entitlement. This section draws on the results
of a number of factor analyses to examine the factorial validity of the BSJO scale. The
dimensionality analyses include the three datasets (LINOS-1, SOEP IS 2012, and ALL-
BUS 2014) and a dataset that was pooled from these datasets.
The short version of the BSJO scale was constructed based on the twelve items in the
LINOS-1 dataset, which consists of four dimensions, each of which has three items. In a
first step, these twelve items were examined for their dimensionality. The factor analysis
(in each case: principal component analysis; rotation method: varimax) reveals a four-
factor solution, which confirms the assumed factor structure. However, the item loadings

Table 5 Factor structure of the justice principles on the basis of the original twelve items (LINOS-1)
Item Equality Need Equity Entitlement Uniqueness
Factor 1 Factor 3 Factor 2 Factor 4

C 0.724 0.153 0.060 -0.017 0.449

K 0.810 0.022 -0.095 0.038 0.333
G 0.716 0.136 0.005 -0.110 0.457
E 0.157 0.742 -0.099 -0.041 0.413
J 0.037 0.636 0.096 -0.052 0.582
A 0.183 0.554 0.030 -0.040 0.658
B -0.220 0.140 0.674 -0.001 0.477
I 0.219 -0.174 0.763 0.039 0.338
H -0.323 0.124 0.550 0.210 0.534
D 0.117 -0.039 -0.023 0.852 0.259
L -0.260 -0.065 0.172 0.692 0.420
F -0.239 0.311 0.356 0.394 0.564
Factor loadings of the principal component analysis; rotation method: varimax; total explained variance:
54.3 percent; data: LINOS-1; all respondents; listwise deletion; N = 4457; cells highlighted in bold: items
that are assumed to load high on the respective dimension; values in italics: items that were excluded in the
short version of the BSJO scale

123
678 S. Hülle et al.

Table 6 Factor structure of the justice principles on the basis of the short version of the BSJO scale
(LINOS-1)
Item Equality Need Equity Entitlement Uniqueness
Factor 1 Factor 3 Factor 4 Factor 2

C 0.797 0.116 0.084 -0.056 0.342

K 0.827 0.054 -0.063 0.015 0.308
E 0.171 0.764 -0.103 -0.045 0.374
J 0.009 0.791 0.096 -0.009 0.365
B -0.298 0.168 0.710 0.034 0.377
I 0.183 -0.103 0.819 0.078 0.279
D 0.139 -0.021 -0.002 0.860 0.241
L -0.302 -0.028 0.150 0.739 0.339
Factor loadings of the principal component analysis; rotation method: varimax; total explained variance:
67.2 percent; data: LINOS-1; all respondents; listwise deletion; N = 4509; cells highlighted in bold: items
that are assumed to load high on the respective dimension

Table 7 Factor structure of the justice principles on the basis of the short version of the BSJO scale (SOEP-
IS 2012)
Item Equality Need Equity Entitlement Uniqueness
Factor 2 Factor 1 Factor 4 Factor 3

C 0.690 0.190 0.297 -0.251 0.336

K 0.794 -0.010 -0.106 0.153 0.334
E 0.137 0.780 0.017 -0.133 0.356
J -0.001 0.773 -0.021 -0.074 0.396
B -0.321 0.468 0.508 0.166 0.392
I 0.074 -0.035 0.889 -0.016 0.203
D 0.417 -0.123 -0.005 0.670 0.362
L -0.130 -0.061 0.013 0.854 0.250
Factor loadings of the principal component analysis; rotation method: varimax; eigenvalues: at least 0.95;
total explained variance: 67.2 percent; data: SOEP-IS 2012; full-time and part-time employees; listwise
deletion; N = 742; cells highlighted in bold: items that are assumed to load high on the respective
dimension

on each of the theoretically derived attitude dimensions are not satisfactory in all cases (see
Table 5). This is particularly true of item F, which has a weak loading on entitlement and
cross loadings on the other dimensions. Also the factor loadings of items H (equity) and A
(need) are relatively low. In a second step, the number of items was reduced to eight, with
only two items per dimension, to construct the short version of the BSJO scale. For the
item selection, factorial validity was used as a criterion for exclusion and was considered
an indicator of the degree to which a simple structure of the factor loading matrices exists.
A number of stepwise factor analyses were run, and items that had either a weak loading on
the respective construct or strong cross loadings were excluded in this process (in Table 5,
these excluded items are indicated in italics).
In the case of LINOS-1, the remaining eight items of the BSJO scale reveal a four-factor
solution, which confirms the assumed simple structure (see Table 6): each item shows a

123
Measuring Attitudes Toward Distributive Justice: The Basic… 679

Table 8 Factor structure of the justice principles on the basis of the short version of the BSJO scale
(ALLBUS 2014)
Item Equality Need Equity Entitlement Uniqueness
Factor 1 Factor 2 Factor 4 Factor 3 Factor 1

C 0.698 0.230 0.009 -0.012 0.460

K 0.821 0.017 0.076 0.041 0.318
E 0.089 0.790 -0.172 -0.066 0.334
J 0.031 0.752 0.125 -0.053 0.415
B -0.474 0.312 0.407 0.146 0.492
I 0.049 -0.049 0.932 0.033 0.126
D 0.274 -0.082 0.048 0.809 0.262
L -0.322 -0.015 0.030 0.770 0.302
Factor loadings of the principal component analysis; rotation method: varimax; eigenvalues: at least 0.96;
total explained variance: 66.2 percent; data: ALLBUS 2014; full-time and part-time employees; listwise
deletion; N = 1036; cells highlighted in bold: items that are assumed to load high on the respective
dimension

Table 9 Factor structure of the justice principles on the basis of the short version of the BSJO scale (pooled
dataset)
Item Equality Need Equity Entitlement Uniqueness
Factor 1 Factor 3 Factor 4 Factor 2

C 0.769 0.161 0.106 -0.076 0.366

K 0.823 0.025 -0.031 0.057 0.318
E 0.146 0.786 -0.077 -0.068 0.350
J 0.011 0.775 0.097 -0.025 0.389
B -0.351 0.251 0.635 0.063 0.407
I 0.153 -0.077 0.860 0.041 0.230
D 0.201 -0.046 0.000 0.830 0.268
L -0.263 -0.037 0.095 0.781 0.310

Factor loadings of the principal component analysis; rotation method: varimax; eigenvalues: at least 0.97;
variables were z-standardized prior to analysis; total explained variance: 67.0 percent; data: pooled dataset
from LINOS-1, ALLBUS 2014, and SOEP-IS 2012; all respondents; listwise deletion; N = 7748; cells
highlighted in bold: items that are assumed to load high on the respective dimension

strong loading only on the respective theoretically assumed construct and only weak cross
loadings. As expected, items C and K load on equality, items E and J on need, items B and
I on equity, and items D and L on entitlement. This four-dimensional factor structure can
also be found in SOEP-IS 2012 (see Table 7), ALLBUS 2014 (see Table 8), and the pooled
dataset (see Table 9). Regarding the SOEP-IS 2012, it should be noted that item D loads
the strongest on entitlement, as expected, but it also loads on equality. Additionally, item B
loads strongest on equity, as expected, but it also loads positively on need and negatively
on equality. This additional negative loading of item B on equality can be found to a
greater or lesser extent in all considered datasets. This negative association is plausible if
one considers that people in favor of a form of (income) distribution that is based on

123
680 S. Hülle et al.

individual effort (equity) are likely to be skeptical about an unconditional equal distri-
bution of income (equality).
It is worth mentioning that the BSJO scale can be used with different interviewing
modes. Using the LINOS-1 dataset, we conducted separate factor analyses by mode for
PAPI, CAWI, and PAPI. The results confirmed the factor structure described above (see
Electronic Supplementary Material for the results).
To conclude, the factor analytical results reported in this section confirmed the assumed
factor structure of the short version of the BSJO scale with eight items. The four-dimensional
structure with equality, need, equity, and entitlement can be found consistently in all con-
sidered datasets and interview modes. Therefore, we regard the factorial validity as given.

5.4 Construct Validity

A construct is valid if a measured value of a scale is a suitable indicator of the charac-

teristic to be measured with the scale. The degree of suitability can be evaluated on the
basis of how well the method stands up to empirical scrutiny during the testing of
hypotheses that have been derived from the model of the characteristic that is to be
measured. Construct validation is based on theoretical considerations that are considered to
postulate positive, neutral, and negative relationships between the construct or scale that is
to be validated and certain other constructs (nomological network). The next step is to
examine whether the predicted pattern can be confirmed empirically (Krohne and Hock
2007, p. 70; Kovaleva et al. 2012, p. 15).
This can be done by comparing whether known patterns of attitude formation reported in
previous studies can be found as well in one’s own data (Sect. 5.4.1). Another approach to
construct validation is to compare relationships between constructs that measure the same or
similar constructs (Rammstedt 2010, pp. 253–255). We chose the justice ideologies (We-
gener and Liebig 1995; Stark et al. 2000) because they also measure preferences for dis-
tributive principles and were implemented in the LINOS-1 questionnaire (Sect. 5.4.2).

5.4.1 Order-Related Justice Attitudes Among Different Population Groups

Justice attitudes differ among socio-demographic groups. Therefore, we assess the con-
struct validity of the BSJO scale by testing a number of assumptions concerning certain
relationships among order-related justice attitudes measured via the BSJO scale and a list
of socio-demographic characteristics (sex, region, age, education, and income). In doing
so, we will draw from previously published research results to formulate our expectations
regarding the assumed relationships. Then, we will compare these results with or own
findings using the BSJO scale to determine whether we will arrive at the same or similar
conclusions.
The results we report in this section are based on the LINOS-1 dataset. The associations
are basically the same when using instead a pooled dataset (with standardized mean indices
comprising LINOS-1, SOEP-IS 2012, and ALLBUS 2014) that leads to the same con-
clusions. When conducting separate analyses with either the SOEP-IS 2012 or the ALL-
BUS 2014, the associations are also very similar (i.e., in the size and direction of mean
differences or correlations), but they do not reach significance in some cases compared to
LINOS-1. This is not surprising, since the sample size for these datasets is much smaller
compared to LINOS-1. Differences in results when using a different dataset instead are
reported in footnotes.

123
Measuring Attitudes Toward Distributive Justice: The Basic… 681

Table 10 Descriptive statistics of the dimensions of short version of the BSJO scale by socio-demographic
characteristic (LINOS-1)
N Equality Need Equity Entitlement

M SD M SD M SD M SD

Region
West 3588 2.74 1.07 4.59 0.50 4.01 0.76 1.97 0.85
East 875 2.89 1.13 4.56 0.51 4.08 0.77 1.90 0.81
Sex
Male 2151 2.68 1.09 4.56 0.52 4.09 0.74 2.05 0.87
Female 2312 2.87 1.07 4.60 0.48 3.97 0.77 1.87 0.81
Age
B29 years 1874 2.80 1.06 4.54 0.51 4.06 0.74 1.96 0.83
30–44 years 1398 2.74 1.10 4.59 0.50 3.96 0.76 1.99 0.86
C45 years 1191 2.77 1.11 4.63 0.49 4.04 0.77 1.91 0.85
Total 4463 2.77 1.09 4.58 0.50 4.02 0.76 1.95 0.84

N = sample size; M = mean; SD = standard deviation; data: LINOS-1; all respondents; listwise deletion
with the additional consideration of education; N = 4463; unweighted values; inverted scale ranging from
(1) ‘‘strongly disagree’’ to (5) ‘‘strongly agree’’

Table 10 provides means and standard deviations of the BSJO scale for each of the four
dimensions equality, need, equity, and entitlement in the LINOS-1 dataset.7 As described
in Sect. 4.2, the minimum value of the BSJO scale is 0 and the maximum is 5. Therefore,
the higher the mean, the higher is the average level of support for the respective justice
principle. The bottom row of the table shows the means for the sample as a whole, which
indicates the average degree of agreement or disagreement with the four justice principles
across Germany. The results reveal that the need and equity principles are preferred most
among the population (M = 4.58 and 4.02, respectively) and that entitlement is the least-
preferred principle (M = 1.95). When asked about the equality principle, the majority of
respondents expressed an indifferent or somewhat negative attitude (M = 2.77), with the
standard deviation (SD = 1.09) indicating the largest amount of variation in the degree of
agreement and disagreement with any of the justice principles among the population in this
study. The findings we report next can be found in Table 10 as well.
Sex A finding often reported in the literature is that women are more likely than men to
support the equality principle (Liebig and Krause 2006; Forsé and Parodi 2009), whereas
men are more likely than women to prefer the equity principle (e.g., Wegener and Liebig
2000). These findings are supported by the analytical results based on LINOS-1 data (see
Table 10). Men are significantly more in favor of distributive principles that legitimize
inequalities: they have a higher preference for the equity principle (M = 4.09, SD = 0.74)
compared to women (M = 3.97, SD = 0.77), t(4461) = 5.37, p = 0.000. Men are also
significantly more in favor of the entitlement principle (M = 2.05, SD = 0.87) than are
women (M = 1.87, SD = 0.81), t(4461) = 7.17, p = 0.000. In contrast, women are more
in favor of justice principles that aim to reduce inequalities. Their preference for the

7
The Electronic Supplementary Material provides reference values (means and standard deviations) for all
three datasets by these socio-demographic characteristics. Aside from a few minor deviations, the results are
largely the same for all three datasets.

123
682 S. Hülle et al.

Table 11 Correlations between the dimensions of the short version of the BSJO scale, education, and
income
Education (CASMIN) Net income (ln) Net household income (ln)

BSJO dimension
Equality -0.214*** -0.222*** -0.227***
Need -0.027 -0.051** -0.024
Equity -0.219*** -0.050** -0.009
Entitlement -0.029 0.099*** 0.044*

Data: LINOS-1; listwise deletion; N = 2918

* p \ 0.05; ** p \ 0.01; *** p \ 0.001

equality principle (M = 2.87, SD = 1.07) is higher compared to men (M = 2.68,

SD = 1.09), t(4461) = - 5.87, p = 0.000. Women are also significantly more in favor of
the need principle (M = 4.60, SD = 0.48) than are men (M = 4.56, SD = 0.52),
t(4350) = - 2.20, p = 0.014.8
Region Wegener and Liebig (1995, 2010) showed that, on the whole, people living in
East Germany (former German Democratic Republic, GDR) are more likely to lean more
toward the equality principle than people living in West Germany. This finding has stood
the test of time, despite a slightly increasing convergence between East and West Germany
between 1990 and 2010 (Wegener and Liebig 2010; Gerlitz et al. 2012). Our analyses
using the LINOS-1 dataset confirmed this particular difference in attitudes between the two
parts of Germany: individuals from the new federal states had a significantly greater
preference for equality (M = 2.89, SD = 1.13) than did individuals from the old federal
states (M = 2.74, SD = 1.07), t(4461) = -3.64, p = 0.000 (see Table 10). This differ-
ence is explained by the fact that socio-structural conditions have an influence on attitude
formation and attitude change. It can be expected that the more the conditions in East and
West Germany converge in the future, the sooner the attitudes will converge as well
(Wegener and Liebig 2010). Another finding is that people living in West Germany have a
higher preference for the entitlement principle (M = 1.97, SD = 0.85) compared to people
who live in East Germany (M = 1.90, SD = 0.81), t(4461) = , p = 0.021.
Age Since older people are more likely to have health problems, they are more likely to
be in need. Therefore, we expect that support for the need principle increases with age. A
one-way ANOVA revealed significant differences in support for need between our three
considered age groups [F(2,4460) = 11.88, p = 0.000]. Indeed a positive age effect exists:
a Tukey post hoc test provided evidence that the youngest age group (B29 years) has a
significantly lower preference for need (M = 4.54, SD = 0.51) compared to the inter-
mediate age group (30–44 years) (M = 4.59, SD = 0.50, p = 0.019) and the oldest age
group (C34 years) (M = 4.63, SD = 0.49, p = 0.000).9

8
Conducting the analysis instead with the pooled dataset, we come to the same conclusions. However,
using ALLBUS 2014 data, the mean difference for the need principle is insignificant. Using SOEP-IS 2012
data, only the mean difference for the equality principle is significant. However, all differences point in the
same directions as with LINOS-1 data. ALLBUS 2014 and SOEP-IS 2012 have fewer cases, which may
explain these insignificant mean differences.
9
Separate analyses for ALLBUS 2014 and SOEP 2012 reveal only significant differences between the
youngest and the oldest age group.

123
Measuring Attitudes Toward Distributive Justice: The Basic… 683

Income According to the self-interest hypothesis, an individual’s preference for partic-

ular justice principles is the result of that individual’s position within the inequality
structure of a society (Kunovich and Slomczynski 2007; Gerlitz et al. 2012). This
hypothesis suggests that individuals tend to prefer those justice principles that will be most
likely to increase their own personal gains, be they monetary or otherwise, when it comes
to the distribution of benefits and burdens within the society in which they live. One would
therefore expect to find that individuals in lower-status positions in the income hierarchy
lean more toward the equality and need principles, because they believe their implemen-
tation will lead to an increasing redistribution of benefits from top to bottom, whereas one
would expect to find that respondents in higher-status positions would lean more toward
the entitlement and equity principles.
As expected, the empirical analysis (see Table 11) provides negative relationships
between income and equality (net income: r = - 0.222, p \ 0.001; net household income:
r = - 0.227, p \ 0.001) and need (net income: -0.051, p \ 0.01), whereas a positive
relationship exists for entitlement (net income: r = 0.099, p \ 0.001; net household
income: r = 0.044, p \ 0.05).10 Thus, the results support the self-interest hypothesis: the
worse a respondent’s individual financial situation, the stronger their preference for
equality and need and the lower their preference for entitlement.
Education The use of education as a proxy for an individual’s social situation is common
practice. As a result, education has become a key demographic characteristic that is used in
almost every survey-based study. With regard to preferences for particular justice prin-
ciples, it seems reasonable to assume that education serves as a proxy for an individual’s
chances of reaching a favorable position in the social structure. One would therefore expect
that individuals who have little chance of reaching an advantageous position will be more
in favor of principles whose implementation will compensate for, or at least mitigate, the
effects of their unequal life chances and of their resulting position in the social structure.
For this reason, we expect to find a stronger preference for the equality and need
principles among respondents with a lower education and, thus, a negative relationship
between education and a preference for the equality principle. This expectation is sup-
ported by our finding that there is a negative relationship between education and a pref-
erence for the equality principle (r = - 0.214, p \ 0.001) (see Table 11). However, the
results also show a negative relationship between education and a preference for the equity
principle: individuals with a lower level of education lean more heavily toward the equity
principle than do individuals with a higher level of education. A similar pattern can be
observed in Wegener and Liebig (2010), although the negative effect of education is not
significant at the 5%-level.

5.4.2 Relationship Between the BSJO Scale and the Justice Ideology Scale

To further explore the construct validity of the BSJO scale, we compare it with the justice
ideology scale (Wegener and Liebig 1995; Stark et al. 2000), which measures similar

10
An unexpected small negative relationship also exists between equity and net income (r = 0.050,
p \ 0.01). However, this relationship does not exist if the pooled dataset is used (the other relationships
show the same directions, leading to the same conclusions). Conducting the analyses with SOEP-IS 2012 or
ALLBUS 2014 data, we also come to the same conclusions as we do with LINOS-1 data (the same
correlations are significant, pointing in the same directions). The only exception is the small negative
correlation (r = 0.053, p \ 0.05) for net household income in the ALLBUS 2014 dataset.

123
684 S. Hülle et al.

Table 12 Correlations between the dimensions of the short version of the BSJO scale and the justice
ideologies
Justice ideology

Egalitarianism Individualism Ascriptivism

BSJO dimension
Equality 0.380*** -0.208*** -0.259***
Need 0.213*** -0.062*** 0.004
Equity 0.002 0.279*** 0.234***
Entitlement -0.101*** 0.299*** 0.217***
Mean indices of both scales; data: LINOS-1; listwise deletion; N = 4342
* p \ 0.05; ** p \ 0.01; *** p \ 0.001

constructs using the same sample. The general assumption is that from a theoretical per-
spective, similar constructs should have a positive correlation, whereas conflicting con-
structs should have a negative correlation. Looking at the relationships between the
dimensions of the two scales, we postulate positive and negative relationships, which we
test empirically with LINOS-1 data. Based on mean indices of the constructs, Table 12
provides information on the correlations between the constructs of the two scales.
We predict positive correlations for the following pairs of constructs:
• Equality and egalitarianism: both constructs aim at capturing a preference for equality
as a distributive criterion, although egalitarianism differs in emphasizing the state as
allocator. This correlation should be the strongest because of the high construct
similarity.
• Need and egalitarianism: the need principle is confounded with the justice ideology
egalitarianism. The latter regards the state as a responsible allocator that also has to
consider individual needs as a distributive criterion: ‘‘The government should
guarantee everyone a minimum standard of living.’’
• Equity and individualism: both regard individual efforts as a just distributive criterion.
• Entitlement and ascriptivism: both consider ascriptive characteristics or status
characteristics as relevant criteria when distributing resources.
• Entitlement and individualism: according to the entitlement principle, individual status
is a legitimate justification for being entitled to receive an allocated good. This status
might be gained through individual effort in the past and is, in this sense, related to the
equity principle, which stresses individual efforts and contributions.
• Equity and ascriptivism: We expect this correlation to be less significant than the
correlation between entitlement and individualism. Since entitlement and ascriptivism
are related, the same argument for the latter correlation might also apply here.
We predict negative correlations for the following pairs of constructs:
• Equality and ascriptivism: the distributive principle of equality is in opposition to
principles that emphasize individual characteristics as distributive criteria. Equality
neglects individual differences and aims at treating people equally. In contrast,
ascriptivism emphasizes individual differences in ascriptive characteristics and status.
Hence, we predict a negative correlation with ascriptivism.
• Equality and individualism: this argument applies also for individualism. This justice
ideology emphasizes individual differences as efforts as well and is opposed to equality.

123
Measuring Attitudes Toward Distributive Justice: The Basic… 685

• Entitlement and egalitarianism: Being entitled because of status or ascriptive attributes

is contrary to the incorporated equality principle in egalitarianism. This is why we
expect a negative correlation between entitlement and egalitarianism.
The results in Table 12 of the aforementioned relationships confirm all our predictions and
are all highly significant (p \ 0.001). The correlation between equality and egalitarianism
turns out to be the strongest relationship. The correlation between equity and individualism is
also high, although the correlation between entitlement and individualism is even higher.
Therefore, we have consistent and highly significant results that support all our predicted
positive and negative associations between the constructs of the two scales. To conclude, this
analysis provides further evidence for the construct validity of the BSJO scale.

6 Practical Notes for the Use of the BSJO Scale

Drawing on our experience with the BSJO scale, our statements in Sect. 4, and the analyses
in Sect. 5, we suggest the following version of the BSJO scale to potential users who want
to implement it in their own surveys: We recommend the version as implemented in
LINOS-1 (see Sect. 4), with the modification that only the eight items of the short scale be
considered (see Tables 1, 3).11 We recommend random item presentation if applicable.
We provide reference values of the BSJO short scale to enable potential users of the
scale to make comparisons with their own data. Descriptive statistics (mean and standard
deviation) on the level of items by dataset are provided in Table 15 in the Appendix.
Descriptive statistics on the level of dimensions (equality, need, equity, entitlement) by
socio-demographic characteristics (region, sex, age, education, and income) by dataset are
provided in the Electronic Supplementary Material.

7 Conclusion

The perceived justice in the way goods and burdens are distributed in a society is crucial
for the observance of societal rules and norms, as well as for trust in societal institutions
and a commitment to societal, political, and economic goals (Rothmund et al. 2016;
Sachweh 2016). Citizens consider inequalities to be legitimate if, among other things, they
are in line with their basic normative orientations, i.e., their attitudes toward distributive
justice. These order-related justice attitudes address the rules or norms people think should
guide the allocation and (re)distribution of goods and burdens within a society. They differ
not only within and between societies but also over time, and they vary depending on the
generosity of welfare regimes (Arts and Gelissen 2001; Van Oorschot et al. 2012). This
becomes especially important since many societies have to face distributional conflicts,
such as the increase in inequalities in wealth and income (Nolan et al. 2014; Piketty 2014;
OECD 2015), financial problems in expanding welfare states with overaging populations,
or the obligation to ensure ecological and social sustainability. To respond to such chal-
lenges with appropriate political measures and to ensure their legitimacy, empirically
grounded knowledge about individuals’ attitudes toward distributive justice is necessary.

11
This version is the same as the ALLBUS 2014 version except the following two deviations: first, the
instruction differs (in ALLBUS 2014, it is specified to CAPI and is thus inappropriate for other modes);
second, item B is modified, as discussed in Sect. 4.2.

123
686 S. Hülle et al.

The BSJO scale introduced in this article provides a useful instrument to measure order-
related justice attitudes in a more comprehensive way than prior measures. It is directly
compatible with psychological and sociological justice (Gollwitzer and van Prooijen 2016;
Liebig and Sauer 2016) and welfare state research and with the relevant justice principles
that are the subject of studies in those research areas. Consisting of only eight items, it can
differentiate between a person’s preferences for the four basic distributive principles of
equality, need, equity, and entitlement.
We provided several tests of the scale following the recommendations for quality
standards of survey instruments issued by the German Data Forum (Rammstedt et al. 2015).
We assessed the quality of the BSJO scale using three datasets: LINOS-1, SOEP-IS 2012,
and ALLBUS 2014. The analyses conducted for this purpose largely confirmed the factorial
validity of the BSJO scale: The dimensionality analyses revealed the expected four-factor
structure (equality, need, equity, and entitlement) in all three datasets (factorial validity).
The analyses of empirical relationships between the BSJO scale and other constructs
confirmed the construct validity of the BSJO scale: We could replicate many patterns of
order-related justice attitudes among demographic groups that were reported in other
studies. We also found the expected positive and negative correlations between constructs
of the BSJO scale and the justice ideology scale (Wegener and Liebig 1995; Stark et al.
2000). In sum, our assessment of the Basic Social Justice Orientations scale provides
evidence that it is an appropriately validated instrument for measuring preference for the
four basic justice principles of equality, need, equity, and entitlement. It is a short and thus
time-efficient instrument and can be implemented in national and international large-scale
population surveys.
How can the BSJO scale be used in further research, i.e., what is its research potential?
1. On an aggregate level, the BSJO scale can be applied to provide deeper insights into
the attitude patterns toward distributive justice among different population groups
within a society (see Sect. 5.4.1).
2. On an individual level, the BSJO scale provides the opportunity to gain deeper insights
into how an individual’s order-related justice attitudes overlap, i.e., how they
complement or contradict each other (see split-consciousness theory, Kluegel and
Smith 1986).
3. The dimensions of the scale (i.e., mean indices or factor scores12) can be used as
predictors for outcomes of interest, such as voting behavior or policies that cover
issues of (re)distribution and distributive justice. For example, using the pooled
dataset described above, Liebig et al. (2016) used the dimensions of the BSJO scale
as predictors for an individual’s voting intention or party preference regarding the
five major political parties in Germany. Most of the theoretically assumed effects
could be detected empirically. The inclusion of the BSJO predictors nearly doubled
the share of explained variance compared to the basic regression model that contained
socio-demographic characteristics. Hence, the BSJO scale—and, therewith, order-
related justice attitudes—can make a substantial contribution to explaining voting
behavior.
4. The BSJO scale has the potential to be applied for country comparisons: The fact that
the justice principles implemented in specific policies or institutions (e.g., unemploy-
ment insurance) typically differ between countries makes cross-country comparisons

12
For simple descriptive analyses, the use of mean indices might suffice (see, e.g., Section 5.4.1). However,
for multivariate analyses, the use of factor scores is preferable.

123
Measuring Attitudes Toward Distributive Justice: The Basic… 687

generally difficult. Therefore, the BSJO scale does not capture such nations-specific
characteristics. Instead, the BSJO items ask for agreement on the four basic
distributive principles without making the context too specific. Everyone can agree
or disagree with these principles, regardless of whether these principles are
institutionalized in the respective country or not. The comparative welfare state
research considers differences in institutional contexts between countries. To a
varying extent, and in different combinations, the four basic justice principles that
are measured by the BSJO scale can be identified as guiding norms for allocating
and distributing goods and burdens within the societal institutions of welfare states
(e.g., Clasen and van Oorschot 2002; Taylor-Gooby 2005; Ullrich 2008). Therefore,
the BSJO scale is predestined for application in (comparative) welfare state research,
for examining cross-national differences in preferences toward distributive justice,
and for predicting support for policies, especially those that cover issues of
(re)distribution and distributive justice in different countries with different institu-
tional contexts.
5. Finally, when following a longitudinal perspective, the BSJO can be used, on the one
hand, to capture aggregate-level changes in justice attitude patterns within and
between demographic groups and countries. On the other hand, individual-level
changes can be studied: So far, it is unclear how and why an individual’s order-related
justice attitudes change over the life-course. The LINOS panel (Liebig et al. 2014;
Sauer and Valet 2014; Valet et al. 2014) aims at investigating such questions. As a
determinant of changes (or stability) of order-related justice attitudes, the (varying)
structural conditions of the social contexts (e.g., the respondent’s household, firm, or
social network) within which an individual is embedded are considered.

Acknowledgements The authors received funding from the German Research Foundation (DFG) for their
Subproject A6 ‘‘The Legitimation of Inequalities—Structural Conditions of Justice Attitudes over the Life-
span’’ of the DFG Research Center (SFB) 882 ‘‘From Heterogeneities to Inequalities’’.

Appendix

See Tables 13, 14, 15.

Table 13 Operationalization of the ideology scale (Stark et al. 2000)

Justice ideology Item text
Introductory text

Egalitarianism
I will now read out some statements that have been The government should guarantee everyone a
made about the role of government in [country]. minimum standard of living
Using one of the phrases on this card, please tell The government should provide a job for everyone
me how much you agree or disagree with each who wants one
statement
Individualism
Now some questions about incomes in … (country). There is an incentive for individual effort only if
For each statement I read, tell me to what extent differences in income are large enough
you agree or disagree with each, using the phrases It is all right if businessmen make good profits
on this card because everyone benefits in the end

123
688 S. Hülle et al.

Table 13 continued

Justice ideology Item text

Introductory text

Fatalism
Here are some more statements that are sometimes There is no point arguing about social justice since it
made about what is just and unjust in … (country). is impossible to change things
Using the same card tell me how much you agree The way things are these days, it is hard to know
or disagree with each what is just anymore
Ascriptivism
I will read out some statements about wealth and People are entitled to keep what they have earned,
income. For each statement please tell me how even if this means some people will be wealthier
much you agree with each, using a phrase from than others
this card People are entitled to pass on their wealth to their
children

5-point scale for all items with the following categories: 1 = ‘‘strongly agree’’; 2 = ‘‘somewhat agree’’;
3 = ‘‘neither agree nor disagree’’; 4 = ‘‘somewhat disagree’’; 5 = ‘‘strongly disagree’’
Source: Stark et al. (2000)

Table 14 Operationalization of the German version of the BSJO scale in LINOS-1 by justice principle
(Sauer and Valet 2014)
Justice principle Item Item text

Equality C Gerecht ist, wenn alle die gleichen Lebensbedingungen haben

(Gleichheit)
K Es ist gerecht, wenn Einkommen und Vermögen in unserer Gesellschaft an alle
Personen gleich verteilt werden
G Eine Gesellschaft ist gerecht, wenn die Einkommensunterschiede zwischen den
Menschen gering sind
Need (Bedarf) E Eine Gesellschaft ist gerecht, wenn sie sich um die Schwachen und
Hilfsbedürftigen kümmert
J Es ist gerecht, wenn Personen, die Kinder oder pflegebedürftige Angehörige zu
versorgen haben, besondere Unterstützung und Vergünstigungen erhalten
A Eine Gesellschaft ist gerecht, wenn jede Person ausreichend Essen, Wohnraum,
Kleidung sowie Zugang zu Bildung und medizinischer Versorgung hat
Equity B Es ist gerecht, wenn Personen, die hart arbeiten, mehr verdienen als andere
(Leistung)
I Gerecht ist, wenn jede Person nur das bekommt, was sie sich durch eigene
Anstrengungen erarbeitet hat
H Eine Gesellschaft ist gerecht, wenn Unterschiede in Einkommen und Vermögen
die Leistungsunterschiede zwischen Menschen widerspiegeln
Entitlement D Es ist gerecht, wenn Personen, die aus angesehenen Familien stammen, dadurch
(Anrecht) Vorteile im Leben haben
L Es ist gerecht, wenn diejenigen, die in einer Gesellschaft oben stehen, bessere
Lebensbedingungen haben als diejenigen, die unten stehen

123
Measuring Attitudes Toward Distributive Justice: The Basic… 689

Table 14 continued

Justice principle Item Item text

F Es ist gerecht, wenn Personen, die es im Laufe ihres Lebens zu Ansehen und
Wohlstand gebracht haben, auch im Alter davon profitieren
Items were asked in alphabetical order. Items in italics are items that are not included in the short eight-item
BSJO scale. The German introductory text read: ‘‘Es gibt unterschiedliche Vorstellungen darüber, wann eine
Gesellschaft gerecht ist. Wie ist Ihre persönliche Meinung dazu? Bitte geben Sie an, ob Sie jeweils: voll
zustimmen, etwas zustimmen, weder zustimmen noch ablehnen, etwas ablehnen oder ganz ablehnen’’
Source: Question 2 in the LINOS-1 questionnaire (Sauer and Valet 2014, p. 188)

Table 15 Descriptive statistics of items and dimensions of the short BSJO scale by dataset
Item LINOS-1 ALLBUS 2014 SOEP IS 2012

Subscale M SD Min Max M SD Min Max M SD Min Max

Equality C 3.30 1.30 1 5 3.34 1.27 1 5 4.60 2.03 1 7

K 2.26 1.26 1 5 2.41 1.31 1 5 2.91 1.91 1 7
Subscale 2.78 1.09 1 5 2.88 1.07 1 5 3.75 1.60 1 7
Need E 4.56 0.64 1 5 4.66 0.60 1 5 6.20 1.17 1 7
J 4.60 0.63 1 5 4.75 0.50 1 5 6.29 1.04 1 7
Subscale 4.58 0.50 1 5 4.71 0.44 2.5 5 6.24 0.91 1 7
Equity B 4.48 0.74 1 5 4.59 0.72 1 5 6.20 1.20 1 7
I 3.57 1.17 1 5 3.69 1.14 1 5 5.21 1.72 1 7
Subscale 4.03 0.76 1 5 4.14 0.71 1 5 5.70 1.13 1 7
Entitlement D 1.62 0.92 1 5 1.65 0.90 1 5 2.06 1.48 1 7
L 2.29 1.16 1 5 2.34 1.22 1 5 2.87 1.75 1 7
Subscale 1.95 0.85 1 5 2.00 0.86 1 5 2.47 1.31 1 7

M = mean; SD = standard deviation; Min = minimum; Max = maximum; listwise deletion in all cases; all
respondents; LINOS-1: inverted scale ranging from (1) ‘‘strongly disagree’’ to (5) ‘‘strongly agree,’’
N = 4509; ALLBUS 2014: inverted scale ranging from (1) ‘‘strongly disagree’’ to (5) ‘‘strongly agree,’’
N = 1682; SOEP-IS 2012: scale ranging from (1) ‘‘do not agree at all’’ to (7) ‘‘agree completely,’’ N = 1557

References
Adams, J. S. (1963). Towards an understanding of inequity. The Journal of Abnormal and Social Psy-
chology, 67(5), 422–436.
Arts, W., & Gelissen, J. (2001). Welfare states, solidarity and justice principles: Does the type really matter?
Acta Sociologica, 44(4), 283–299.
Atkinson, A. B., & Bourguignon, F. (2015). Handbook of income distribution (Vol. 2A). Amsterdam: North
Holland.
Boyle, G. J. (1991). Does item homogeneity indicate internal consistency or item redundancy in psycho-
metric scales? Personality and Individual Differences, 12(3), 291–294.
Clark, A. E., & d’Ambrosio, C. (2015). Attitudes to income inequality: Experimental and survey evidence.
In A. B. Atkinson & F. Bourguignon (Eds.), Handbook of income distribution (Vol. 2A,
pp. 1147–1208). Amsterdam: North Holland.

123
690 S. Hülle et al.

Clark, L. A., & Watson, D. (1995). Constructing validity: Basic issues in objective scale development.
Psychological Assessment, 7(3), 309.
Clasen, J., & Van Oorschot, W. (2002). Changing principles in European social security. European Journal
of Social Security, 4(2), 89–116.
Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika, 16(3), 297–334.
Deutsch, M. (1975). Equity, equality, and need: What determines which value will be used as the basis of
distributive justice? Journal of Social Issues, 31(3), 137–149.
Deutsch, M. (1985). Distributive justice: A social-psychological perspective. New Haven: Yale University
Press.
DIW. (2015). SOEP innovation sample (SOEP-IS), data from 1998–2013. Berlin: German Institute for
Economic Research, Research Data Center SOEP. doi:10.5684/soep.is.2013.
Douglas, M. (1982). In the active voice. London: Routledge.
Fiske, A. P. (1992). The four elementary forms of sociality: Framework for a unified theory of social
relations. Psychological Review, 99(4), 689–723.
Forsé, M., & Parodi, M. (2009). Distributive justice: An ordering of priorities. A comparative analysis of
European opinions. International Review of Sociology, 19(2), 205–225.
Gerlitz, J.-Y., Mühleck, K., & Scheller, P. (2007). Zur Erhebung von Gerechtigkeitsideologien mit Bezug
zur Altersvorsorge. Dokumentation der Instrumentenentwicklung für das ISJP 2006. ISJP Arbeits-
bericht No. 117, Berlin: Humboldt University of Berlin.
Gerlitz, J.-Y., Mühleck, K., Scheller, P., & Schrenker, M. (2012). Justice perception in times of transition:
Trends in Germany, 1991–2006. European Sociological Review, 28(2), 263–282.
GESIS. (2015a). ALLBUS/GGSS 2014 (Allgemeine Bevölkerungsumfrage der Sozialwissenschaften/Ger-
man General Social Survey 2014). Cologne: GESIS Data Archive. ZA5240 Data file Version 2.1.0.
doi:10.4232/1.12288.
GESIS. (2015b). ALLBUS Allgemeine Bevölkerungsumfrage der Sozialwissenschaften: ALLBUS 2014—
Variable report: Study no. 5240. GESIS variable reports no. 2015|30. Cologne: GESIS—Leibniz
Institute for the Social Sciences.
Gollwitzer, M., & van Prooijen, J. W. (2016). Psychology of justice. In C. Sabbagh & M. Schmitt (Eds.),
Handbook of social justice theory and research (pp. 61–82). New York, NY: Springer.
Hadler, M. (2005). Why do people accept different income ratios? A multi-level comparison of thirty
countries. Acta Sociologica, 48(2), 131–154.
Harkness, J. A. (2008). Comparative survey research: Goal and challenges. In E. D. de Leeuw, J. J. Hox, &
D. A. Dillman (Eds.), International handbook of survey methodology (pp. 56–77). New York: Psy-
chology Press.
Hegtvedt, K. A., & Isom, D. (2014). Inequality: A matter of justice? In J. D. McLeod, E. J. Lawler, & M.
Schwalbe (Eds.), Handbook of the social psychology of inequality (pp. 65–94). Dordrecht: Springer.
Jäckle, N. (2002). Entwicklung eines Fragebogens zur Erfassung der Gerechtigkeitsideologien Egalitaris-
mus, Askriptivismus, Individualismus und Fatalismus. ISGF work report no. 38. Berlin: Humboldt
University of Berlin.
Jasso, G. (2015). Thinking, saying, doing in the world of distributive justice. Social Justice Research, 28,
435–478.
Kluegel, J. R., & Smith, E. R. (1986). Beliefs about inequality: Americans’ views of what is and what ought
to be. New York, NY: Aldine de Gruyter.
Konow, J. (2003). Which is the fairest one of all? A positive analysis of justice theories. Journal of
Economic Literature, 41(4), 1188–1239.
Kovaleva, A., Beierlein, C., Kemper, C. J., & Rammstedt, B. (2012). Eine Kurzskala zur Messung von
Kontrollüberzeugung: Die Skala Internale-Externale-Kontrollüberzeugung-4 (IE-4). GESIS working
papers no. 2012|19. Cologne: GESIS.
Krohne, H. W., & Hock, M. (2007). Psychologische Diagnostik: Grundlagen und Anwendungsfelder.
Stuttgart: Kohlhammer.
Kunovich, S., & Slomczynski, K. M. (2007). Systems of distribution and a sense of equity: A multilevel
analysis of meritocratic attitudes in post-industrial societies. European Sociological Review, 23(5),
649–663.
Legewie, J., Gerlitz, J.-Y., Mühleck, K., Scheller, P., & Schrenker, M. (2007). Dokumentation des inter-
national social justice project 2006 für Deutschland. ISJP technical report no. 118, Berlin: Humboldt
University of Berlin.
Liebig, S. (2001). Lessons from philosophy? Interdisciplinary justice research and two classes of justice
judgments. Social Justice Research, 14(3), 265–287.

123
Measuring Attitudes Toward Distributive Justice: The Basic… 691

Liebig, S., Hülle, S., & May, M. (2016). Principles of the just distribution of benefits and burdens: The
‘‘Basic Social Justice Orientations’’ scale for measuring order-related social justice attitudes. SOEP
paper on multidisciplinary panel data research, no. 831. Berlin: DIW Berlin.
Liebig, S., & Krause, A. (2006). Soziale Einstellungen in der Organisationsgesellschaft. Betriebliche
Strukturen und die gerechte Verteilungsordnung der Gesellschaft. Zeitschrift für Arbeitsmarkt-
forschung, 39(2), 255–276.
Liebig, S., May, M., Sauer, C., Schneider, S., & Valet, P. (2014). Erwartungen an Wirtschaft und Gesell-
schaft. DFG Research Center (SFB) 882 ‘‘From Heterogeneities to Inequalities’’ doi:10.4119/unibi/
sfb882.2014.9.
Liebig, S., & Sauer, C. (2016). Sociology of justice. In C. Sabbagh & M. Schmitt (Eds.), Handbook of social
justice theory and research (pp. 37–59). New York, NY: Springer.
Liebig, S., & Schlothfeldt, S. (2002). Das Grid-Group-Paradigma und sein Beitrag für die interdisziplinäre
soziale Gerechtigkeitsforschung. In S. Liebig & H. Lengfeld (Eds.), Interdisziplinäre
Gerechtigkeitsforschung: Zur Verknüpfung empirischer und normativer Perspektiven (pp. 219–242).
Frankfurt/Main: Campus.
Lienert, G. A., & Raatz, U. (1998). Testaufbau und Testanalyse (6th ed.). Weinheim: Beltz Psychologie
Verlags Union.
Miller, D. (1976). Social justice. Oxford: Clarendon Press.
Miller, D. (1999). Principles of social justice. Cambridge, MA: Harvard University Press.
Mohler, P., Dorer, B., de Jong, J., & Hu, M. (2016). Translation. Guidelines for best practice in cross-
cultural surveys. Ann Arbor, MI: Survey Research Center, Institute for Social Research, University of
Michigan. http://www.ccsg.isr.umich.edu/images/PDFs/CCSG_Full_Guidelines_2016_Version.pdf.
Accessed January 10, 2017.
Nielsen, F. (2017). Inequality and inequity. Social Science Research, 62, 29–35.
Nolan, B., Salverda, W., Checchi, D., Marx, I., McKnight, A., Tóth, I. G., et al. (2014). Changing
inequalities and societal impacts in rich countries: Thirty countries’ experiences. Oxford: Oxford
University Press.
Nunnally, J. C. (1978). Psychometric theory (2nd ed.). New York: McGraw-Hill.
OECD. (2015). In it together: Why less inequality benefits all. Paris: Organisation for Economic Co-
operation and Development.
Piketty, T. (2014). Capital in the twenty-first century. Cambridge, MA: Harvard University Press.
Rammstedt, B. (2010). Reliabilität, Validität, Objektivität. In C. Wolf & H. Best (Eds.), Handbuch der
sozialwissenschaftlichen Datenanalyse (pp. 239–258). Wiesbaden: VS Verlag für
Sozialwissenschaften.
Rammstedt, B., Beierlein, C., Brähler, E., Eid, M., Harting, J., Kersting, M., et al. (2015). Quality standards
for the development, application, and evaluation of measurement instruments in social science survey
research. RatSWD working paper no. 245. Berlin: German Council for Social and Economic Data
(RatSWD).
Richter, D., & Schupp, J. (2012). SOEP innovation sample (SOEP-IS)—Description, structure and docu-
mentation. SOEP paper of multidisciplinary panel data research no. 463. Berlin: German Institute for
Economic Research, Research Data Center SOEP.
Roller, E. (1995). The welfare state: The equality dimension. In O. Borre & E. Scarbrough (Eds.), The scope
of government (pp. 165–197). Oxford: Oxford University Press.
Rothmund, T., Becker, J. C., & Jost, J. T. (2016). The psychology of social justice in political thought and
action. In C. Sabbagh & M. Schmitt (Eds.), Handbook of social justice theory and research (pp.
275–291). New York, NY: Springer.
Sachweh, P. (2016). Social justice and the welfare state: Institutions, outcomes, and attitudes in comparative
perspective. In C. Sabbagh & M. Schmitt (Eds.), Handbook of social justice theory and research (pp.
293–313). New York, NY: Springer.
Sauer, C., & Valet, P. (2014). LINOS-1: Legitimation of inequality over the life-span. SFB 882 technical
report no. 13. Bielefeld: DFG Research Center (SFB) 882 ‘‘From Heterogeneities to Inequalities’’.
Sauer, C., Valet, P., & Meyer, L. (2014). Expectations towards economy and society: Codebook of the
employee survey. SFB 882 technical report no. 11. Bielefeld: DFG Research Center (SFB) 882 ‘‘From
Heterogeneities to Inequalities’’.
SOEP. (2014). SOEP-IS 2012—Methodenbericht zum Befragungsjahr 2012/13 des SOEP-Innovation-
ssamples. SOEP survey paper no. 179. Berlin: German Institute for Economic Research.
Stark, G., Liebig, S., & Wegener, B. (2000). Gerechtigkeitsideologien. Theoretische Grundlagen, Opera-
tionalisierung und Ergebnisse für die Erhebung von Einstellungen zur sozialen Gerechtigkeit.
Lieferung für das ZUMA-Informations-System Sozialwissenschaftlicher Skalen (ZIS). Arbeitsbericht
No. 14. Berlin: Nachwuchsgruppe ‘‘Interdisziplinäre Soziale Gerechtigkeitsforschung’’.

123
692 S. Hülle et al.

Taylor-Gooby, P. (2005). Ideas and welfare state reform in Western Europe. Basingstoke: Palgrave
Macmillan.
Ullrich, C. G. (2008). Die Akzeptanz des Wohlfahrtsstaates. Wiesbaden: VS Verlag für
Sozialwissenschaften.
Valet, P., May, M., Sauer, C., & Liebig, S. (2014). LINOS-1: Legitimation of inequality over the life-span.
SFB 882 technical report no. 13. Bielefeld: DFG Research Center (SFB) 882 ‘‘From Heterogeneities to
Inequalities’’.
Van Oorschot, W., Reeskens, T., & Meuleman, B. (2012). Popular perceptions of welfare state conse-
quences: A multilevel, cross-national analysis of 25 European countries. Journal of European Social
Policy, 22(2), 181–197.
Verme, P. (2011). Life Satisfaction and Income Inequality. Review of Income and Wealth, 57(1), 111–127.
Wegener, B. (1992). Gerechtigkeitsforschung und Legitimationsnormen. Zeitschrift für Soziologie, 21(4),
269–283.
Wegener, B., & Liebig, S. (1995). Hierarchical and social closure conceptions of distributive social justice:
A comparison of East and West Germany. In J. R. Kluegel, D. S. Mason, & B. Wegener (Eds.), Social
justice and political change. Political opinion in capitalist and post-communist nations (pp. 263–284).
New York, NY: De Gruyter.
Wegener, B., & Liebig, S. (2000). Is the ‘‘inner wall’’ here to stay? Justice ideologies in unified Germany.
Social Justice Research, 13(2), 177–197.
Wegener, B., & Liebig, S. (2010). Gerechtigkeitsvorstellungen in Ost- und Westdeutschland im Wandel:
Sozialisation, Interessen, Lebenslauf. In P. Krause & I. Ostner (Eds.), Leben in Ost- und West-
deutschland: Eine sozialwissenschaftliche Bilanz der deutschen Einheit (pp. 83–102). Frankfurt/Main:
Campus.

123