ArticlePDF Available

The Development and Validation of Brief and Ultrabrief Measures of Values

Authors:

Abstract and Figures

Values are a central personality construct and the importance of studying them has been well established. To encourage researchers to integrate measures of values into their studies, brief and ultrabrief instruments were developed to recapture the 10 values measured by the 40-item Portrait Values Questionnaire (PVQ; Schwartz, 2003 Schwartz, S. H. (2003). A proposal for measuring value orientations across nations. In Questionnaire development report of the European Social Survey (pp. 259–319). Retrieved from http://www.europeansocialsurvey.org/index.php?option=com_docman&task=doc_view&gid=126&Itemid=80). Rigorous psychometric procedures based on separate derivation (N = 38,049) and evaluation (N = 29,143) samples yielded 10- and 20-item measures of values, which proved to be successful at capturing the patterns and magnitude of correlations associated with the original PVQ. These instruments should be useful to researchers who would like to incorporate a values scale into their study but do not have the space to administer a longer measure.
Content may be subject to copyright.
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=hjpa20
Download by: [Carson Sandy] Date: 21 October 2016, At: 15:42
Journal of Personality Assessment
ISSN: 0022-3891 (Print) 1532-7752 (Online) Journal homepage: http://www.tandfonline.com/loi/hjpa20
The Development and Validation of Brief and
Ultrabrief Measures of Values
Carson J. Sandy, Samuel D. Gosling, Shalom H. Schwartz & Tim Koelkebeck
To cite this article: Carson J. Sandy, Samuel D. Gosling, Shalom H. Schwartz & Tim Koelkebeck
(2016): The Development and Validation of Brief and Ultrabrief Measures of Values, Journal of
Personality Assessment, DOI: 10.1080/00223891.2016.1231115
To link to this article: http://dx.doi.org/10.1080/00223891.2016.1231115
View supplementary material
Published online: 21 Oct 2016.
Submit your article to this journal
View related articles
View Crossmark data
The Development and Validation of Brief and Ultrabrief Measures of Values
Carson J. Sandy,
1
Samuel D. Gosling,
1
,
2
Shalom H. Schwartz,
3
,
4
and Tim Koelkebeck
5
1
Department of Psychology, The University of Texas at Austin;
2
Music, Mind and Wellbeing, The University of Melbourne, Parkville, Victoria, Australia;
3
Department of Psychology, The Hebrew University of Jerusalem, Israel;
4
Expert Institute and International Research and Teaching Laboratory for Socio-
Cultural Research, National Research University Higher School of Economics, Moscow, Russia;
5
MyType, Inc., San Francisco, California
ARTICLE HISTORY
Received 30 November 2015
Revised 29 July 2016
ABSTRACT
Values are a central personality construct and the importance of studying them has been well established.
To encourage researchers to integrate measures of values into their studies, brief and ultrabrief
instruments were developed to recapture the 10 values measured by the 40-item Portrait Values
Questionnaire (PVQ; Schwartz, 2003). Rigorous psychometric procedures based on separate derivation (N
D38,049) and evaluation (ND29,143) samples yielded 10- and 20-item measures of values, which proved
to be successful at capturing the patterns and magnitude of correlations associated with the original PVQ.
These instruments should be useful to researchers who would like to incorporate a values scale into their
study but do not have the space to administer a longer measure.
There is now strong evidence for a near-universal structure of
basic human values (e.g., Bilsky, Janik, & Schwartz, 2011;
Schwartz, 1992; Schwartz & Bardi, 2001). For example, the
structure of Schwartzs 10 basic values has been supported in
210 samples from 67 countries (Schwartz & Boehnke, 2004).
These values (universalism, self-direction, stimulation, hedo-
nism, achievement, power, security, conformity, tradition, and
benevolence) relate to one another in a quasi-circumplex struc-
ture such that values that are adjacent to one another are more
positively correlated than values more distant on the circular
structure. In addition, the relative ranking of the importance of
the 10 values shows substantial consistency across different
samples and geographical contexts. This consistency suggests
that a pancultural hierarchy fosters successful societal function-
ing (Schwartz & Bardi, 2001).
The basic values have numerous established external correlates
(e.g., age, gender, education level, political orientation, religiosity;
Schwartz, 2006) and predict a broad range of meaningful decisions
and behaviors (e.g., Bardi & Schwartz, 2003; Sagiv & Schwartz,
2004). In previous research, values have predicted such diverse
behaviors as alcohol consumption (Dollinger & Kobayashi, 2003),
worrying (Schwartz, Sagiv, & Boehnke, 2000), social attraction
(Boer et al., 2011), proenvironmental behavior (Grunert & Juhl,
1995; Schultz & Zelezny, 1998), and political attitudes and voting
behavior (e.g., Schwartz, Caprara, & Vecchione, 2010).
Despite such promising ndings, the quantity of values-
related research has lagged behind that of other individual-dif-
ference constructs, such as traits. A keyword search of articles
published in the top (based on Impact Factor from the 2012
Journal Citation ReportsÒSocial Sciences Edition [Thomas
Reuters, 2013]) personality journals (Personality and Social Psy-
chology Review, Journal of Personality and Social Psychology,
Journal of Personality, European Journal of Personality, Person-
ality and Social Psychology Bulletin, Journal of Research in Per-
sonality, Journal of Personality Assessment, Personality and
Individual Differences) over the past 20 years yielded 2,413
articles using the keyword traits compared to 411 articles using
the keyword values.
1
Some researchers have essentially equated traits with personality
(Buss, 1989;Hofstee,1984) but theorists have been quick to point
out that traits alone offer only a limited understanding of personal-
ity (McAdams, 1995). In attempts to push more integrative models
of personality, researchers have proposed values as a candidate for
offering a more comprehensive understanding of the person (e.g.,
McAdams & Pals, 2006;Parks-Leduc,Feldman,&Bardi,2015).
The incremental contribution of values over traits is supported by
their differing patterns and strengths of correlations with outcomes
and constructs, such as religiosity (Roccas, Sagiv, Schwartz, and
Knafo, 2002), subjective well-being (e.g., Haslam, Whelan, & Bas-
tian, 2009), and voting preferences (e.g., Caprara, Vecchione, &
Schwartz, 2009).
A recent meta-analysis of 60 studies reporting correlations
between values and the Big Five personality traits suggested
that the two types of constructs are distinct (Parks-Leduc et al.,
2015); such ndings provide an impetus for researchers to mea-
sure both values and traits when studying the effects of individ-
ual differences. This more comprehensive approach will
provide scholars with a more holistic view of the person and
CONTACT Carson J. Sandy cjosandy@gmail.com MotiveMetrics, 410 Sherman Avenue, Palo Alto, CA 94306.
Carson J. Sandy is now at MotiveMetrics in Palo Alto, CA.
Supplemental data for this article can be accessed on the publishers website.
1
The search result for values overestimates the number of studies examining basic psychological values because the search term also captures additional constructs such
as social value orientation, truth value, and mate value.
© 2016 Taylor & Francis
JOURNAL OF PERSONALITY ASSESSMENT
http://dx.doi.org/10.1080/00223891.2016.1231115
also contribute to a better understanding of how traits and val-
ues differentially affect and predict behavior.
One of the main reasons for the relative scarcity of values
research (in comparison to research on personality traits) could be
the costs in time and effort of measuring values, especially for
investigators for whom values are not their primary research inter-
est. Even the already vibrant eld of personality trait research saw
an upsurge in studies measuring personality with the availability of
very brief, psychometrically sound personality trait measures (e.g.,
Gosling, Rentfrow, & Swann, 2003). Here, we aim to develop a
short measure of values with the hope that it will similarly facilitate
a wider integration of values measures into ongoing research. Natu-
rally, researchers whose central interests are in values will continue
to use the existing longer instruments.
Background of the empirical study of values
Building on the seminal work of Rokeach (1973), Schwartz and
Bilsky (1987,1990) sought to develop a theory of basic human
values that applies across cultural contexts and is grounded in
human social nature. Early efforts specied three facets of every
value: goal type (terminal vs. instrumental), interests served
(collective vs. individual), and motivational domain (e.g., hedo-
nism). Subsequently, Schwartz reduced this to two facets, inter-
ests served (social vs. personal) and motivational domain. For
example, the value excitement serves a personal interest in the
stimulation motivational domain. The development of these
facets resulted in a systemic theory of the content and organiza-
tion of the value systems of individuals that has been validated
empirically (Schwartz, 1992; Smith & Schwartz, 1997).
Schwartz (1992) settled on 10 basic human values that are orga-
nized on a circular motivational continuum such that adjacent val-
ues(valuesthatareclosertogetherinthecircle)areconceptually
more closely related. Values include universalism, self-direction,
stimulation, hedonism, achievement, power, security, conformity,
tradition, and benevolence values. The circular structure has been
supported in adolescent, student, adult, and representative samples
from more than 80 countries (Schwartz, 2006,2011). In addition to
replicating the circular structure with self-reported values, Bardi
and Schwartz (2003) found that behaviors chosen as likely to
express each value (e.g., obey my parents[conformity], Use
environmentally friendly products[universalism]) also exhibited
a circular structure of relations that corresponded to the circular
motivational structure of values. Thus, value-expressive behaviors
validate the motivational structure, too.
Measurement of the Schwartz values model
The 57-item Schwartz Value Survey (SVS; Schwartz, 1992) was
created to evaluate these 10 basic human values. Participants
endorse the appropriate level of personal importance (from ¡1
to 7) of each value item. For example, to measure power, a
respondent indicates how important social powercontrol
over others, dominanceis as a guiding principle in his or her
life. The SVS has strong psychometric validity and remains one
of the most pervasive measures of values.
One issue with the SVS was that the values circumplex
2
failed to replicate in about 5% of cross-cultural samples, mainly
in Africa (Schwartz et al., 2001). A major reason for this failure
is that the SVS requires respondents to have an abstract con-
ception of the values they endorse. The SVS provides no infor-
mation to contextualize the values in a real-world setting. Some
participants, especially in societies with a non-Western educa-
tion system, might have had difculty interpreting the abstract
items of the SVS.
To circumvent the problem of using abstract value concepts,
Schwartz developed the Portrait Values Questionnaire (PVQ
40; Schwartz, 2003; Schwartz et al., 2001). This 40-item instru-
ment presents participants with short portraits of gender-
matched individuals. Respondents rate how much the person
described in the portrait is like himself or herself. For example,
a PVQ power item states, It is important to him to be in
charge and tell others what to do. He wants people to do what
he says.Respondentsown values are inferred from their self-
reports of how similar they are to the individuals described in
the portraits. With its greater concreteness, the PVQ was more
successful in conrming the value circumplex in countries
where the SVS had not (Schwartz, 2006). The goal of this study
is to develop brief and ultrabrief scales, based on the PVQ40,
thereby providing researchers with more options to incorporate
values measures into their research.
Why is a shorter instrument needed?
Researchers in the eld of personality assessment have identi-
ed numerous contexts where short scales are advantageous
(e.g., Gosling et al., 2003; Paulhus & Bruce, 1992; Robins,
Trzesniewski, Tracy, Gosling, & Potter, 2002). Some of these
contexts include longitudinal studies (tracking individuals on
many constructs over time), personality-rating studies (rating
individuals on multiple personality variables), large-scale Inter-
net studies (where participants might lack the patience to take
long questionnaires), and prescreening (researchers desire to
identify a number of traits before moving forward with a full
study).
In response to the demand for short measures, a wide range
of very brief scales have been developed and validated. These
address such topics as relationships (Wei, Russell, Mallinck-
rodt, Vogel, 2007), personality traits (Gosling et al., 2003),
intelligence (Minshew, Turner, & Goldstein, 2005), and self-
esteem (Robins, Hendin, & Trzesniewski, 2001). Typically,
these scales demonstrate convergence with the full scale and
correlate with external variables as predicted. The short scales
tend to have lower reliabilities (e.g., internal consistency, test
retest) than the full scales but remain satisfactory (Ziegler,
Kemper, & Kruyen, 2014).
Abbreviating the PVQ40
In abbreviating the PVQ40, we had two objectives. First, we
wished to create brief (two items per value) and ultrabrief (one
item per value) questionnaires to allow researchers to incorpo-
rate values into studies that cannot accommodate the length of
2
Strictly speaking, the circular motivational structure of values is a quasi-circum-
plex (Schwartz & Boehnke, 2004). For simplicity, we refer to it as a circumplex.
2SANDY, GOSLING, SCHWARTZ, KOELKEBECK
the standard measures. Second, to promote incremental
research, we wanted to construct the scales in a way that would
readily facilitate comparisons across the scales of different
lengths. Third, we wished to provide a record of all relevant
selection and validation psychometrics.
A brief 21-item measure of values, derived from the PVQ,
already exists, the PVQ21 (Schwartz, 2003). Schwartz devel-
oped this abbreviated scale for use in the European Social Sur-
vey (ESS; Schwartz, 2003). A number of issues associated with
the development and publication of the PVQ21 prompted us
to create a new scale. First, several of its items were modied
versions of PVQ40 items, making direct psychometric com-
parisons between the 21-item and 40-item versions problem-
atic. Second, few details were reported regarding the sample of
444 individuals used for the item-derivation sample other than
that they originated from the United Kingdom and Nether-
lands. Third, no external variables were measured in the scale-
development process; evaluating the pattern of relationships
with values and other constructs is a key step in establishing
the validity of a new measure (e.g., Gosling et al. 2003). Fourth,
convergent validity scores are not reported. An abbreviated
scales correlation with the original scale is an important step in
demonstrating that it can effectively recapture the psychomet-
ric properties of the long-form version. Finally, no testretest
data are available; this form of reliability is particularly impor-
tant because, as we discuss later, internal consistency indexes
are vulnerable to underestimating the reliability of short scales.
The current scale-development efforts sought to address
these drawbacks of the PVQ21. First, to facilitate comparabil-
ity across scales, we drew on PVQ40 items to construct both
the 10- and 20-item scales, using identical methods and equiva-
lent samples. This procedure guaranteed that the items of the
10-item scale were a subset of the items on the 20-item scale.
Because the longer instruments perfectly subsume the shorter
instruments, researchers can directly compare ndings across
studies that used either of the two new scales or the existing
PVQ40. We also compared the performance of our new 10-
and 20-item scales with that of the existing 21-item PVQ. Sec-
ond, we present detailed demographic information on partici-
pants and documentation of the data collection procedures.
Third, a number of external variables were collected, allowing
us to examine their expected relationships with the values
scales, furnishing further evidence regarding the validity of the
new scales. Fourth, convergent validity scores with respect to
the original PVQ40 were estimated and used in the item-selec-
tion and validation phases. Finally, we estimated testretest
reliability data.
Overview of this research
Phase 1 of the study involved the selection of items for the new
scales. To derive the best performing 10-item and 20-item
measures, the following psychometric criteria were used: con-
vergence with the full scale, internal consistency, and patterns
of predicted external correlates. Phase 2 then evaluated the
abbreviated scales in a new sample. Additionally, two follow-up
studies were performed. The rst study aimed to identify the
time savings a researcher might gain from using the new
shorter scales The second study estimated testretest reliability.
A number of automated methods, such as those based on
item response theory and genetic algorithms, can also be used
to select the subsets of items that best represent the longer
scales from which the items are drawn. Elsewhere we have
empirically compared the results of the traditional and auto-
mated methods and found the results to be virtually identical
(Sandy, Gosling, & Koelkebeck, 2014). Here we use the tradi-
tional method, which is more widely understood.
Method
Participants
The procedures required a large sample to detect the subtle psy-
chometric effects of measurement differences, and one that was
reasonably diverse to establish the generalizability of the nd-
ings. It was also necessary to obtain responses to a number of
other variables to evaluate patterns of external correlates. Col-
lecting the data online facilitated meeting these criteria (Gos-
ling & Mason, 2015). Specically, two samples of volunteers
responded to items hosted on an application (MyType) run-
ning on Facebook. Respondents received feedback on their
scores that they could post to their walls if they wished. Sample
1 (derivation sample) consisted of 38,049 individuals (63%
female; ages 1894, MD26.42, SD D10.01) who responded
between December 2009 and November 2010. Sample 2 (evalu-
ation sample) consisted of 29,143 individuals (62% female; ages
1892, MD27.45, SD D9.52) who responded between August
2010 and February 2011 while Sample 1 was being analyzed.
Sample 1 and Sample 2 were both pulled from a larger database
of participants and are completely nonoverlapping (meaning
that no participants appear in both data sets). The nationality
of the participants was as follows: United States (72%), Singa-
pore (8%), Canada (3%), Australia (3%), and Great Britain
(3%). Twelve percent of participants were from various other
countries and 11% did not report a location.
Measures
The data were collected via a series of questionnaires hosted by
MyType.com, an application developed for Facebook. A pri-
mary feature of this application is to provide users the opportu-
nity to respond to and get feedback on psychometrically valid
personality scales. Participants arrived at MyType.com through
embedded links in other Web sites, online search engines, and
word of mouth. Two broad classes of measures were collected.
First were the values measures from which the short scales
were derived. Second were variables that could be used to com-
pare the patterns of external correlates of the old and new
instruments. These were demographics, Big Five, political ori-
entation, and religiosity, which MyType happened to be
collecting.
Portrait Values Questionnaire 40-item
One survey available through MyType was the PVQ (PVQ40;
Schwartz, 2003; Schwartz et al., 2001), a 40-item questionnaire
that includes descriptions of people who endorse certain values.
Respondents use a 6-point scale ranging from 1 (not like me at
all)to6(very much like me) to describe the extent to which the
SHORT VALUES MEASURE 3
individuals portrayed in 40 brief descriptions are similar to
themselves (see online supplemental Appendix A for sample
items). Respondentsown values are inferred from these self-
reported similarity estimates. Reliability reached acceptable lev-
els (typically greater than .7; Nunnally, 1978) in both samples
with one exception (tradition). Cronbachs alphas for the selec-
tion and evaluation samples were, respectively, conformity D
.73/.73, tradition D.51/.51, benevolence D.85/.86, universalism
D.84/.85, self-direction D.88/.89, stimulation D.77/.78, hedo-
nism D.80/.81, achievement D.84/.84, power D.69/.69, and
security D.65/.64.
Portrait Values Questionnaire 21-item
Schwartz created a 21-item version of the PVQ (PVQ21;
Schwartz, 2003) for inclusion in the ESS (Schwartz, 2003). To
evaluate whether the new 10- and 20-item instruments offer
any improvement over this version, we also calculated scores
for a close proxy of the PVQ21. It was necessary to use a proxy
for the PVQ21 because eight of its items were slightly modi-
ed from the PVQ40 items. Our data set only included items
from the PVQ40, so we substituted the items from the PVQ
40 that matched the modied PVQ21 items most closely for
our proxy. The differences between the original PVQ40 and
modied PVQ21 items are quite minor. For example, Enjoy-
ing lifes pleasures is important to him. He likes to spoilhim-
selfwas modied to Having a good time is important to him.
He likes to spoilhimselfin the PVQ21.
Nonetheless, to evaluate the extent to which these minor
edits had an impact on responses to the items, we conducted a
series of tests that examined the equivalences between the
items. Data were collected via Amazons Mechanical Turk (see
Buhrmester, Kwang, & Gosling, 2011, for a description of
Mechanical Turk and support for the reliability of data col-
lected on that platform). With a sample of 307 participants, we
evaluated correlations between the PVQ21 items and their
PVQ40 equivalents. The correlations ranged from .43 to .77
(MD.64). Combining these estimates across studies and cor-
recting them for attenuation (using the testretest correlations
derived from perfectly identical items) yielded item-equivalence
estimates that ranged from .53 to .91 (MD.74, SD D.10).
Based on these estimates, we concluded that the item-equiva-
lences were sufciently strong to proceed with our proxy mea-
sure of the PVQ21. However, readers should keep in mind
that our proxy is not identical to the PVQ21. Cronbachs
alphas for the selection and evaluation samples for the PVQ21
were, respectively, conformity D.67/.68, tradition D.18/.16,
benevolence D.83/.85, universalism D.73/.75, self-direction D
.77/.79, stimulation D.70/.70, hedonism D.67/.67, achieve-
ment D.73/.73, power D.55/.55, and security D.40/.39.
The Big Five Inventory
The Big Five Inventory (BFI; John, Donahue, & Kentle, 1991)is
designed to measure the factors of the Big Five (Extraversion
[E], Agreeableness [A], Conscientiousness [C], Neuroticism
[N], Openness [O]). It consists of 44 short items that respond-
ents rate on a scale ranging from 1 (strongly disagree)to5
(strongly agree). Cronbachs alpha reached acceptable levels in
both the selection and validation samples, respectively (E D
.77/.77, A D.75/.70, C D.78/.77, N D.81/.77, O D.71/.67).
Demographic and attitudinal variables
These variables were age (measured in years), gender, income,
education level, political conservatism, and religiosity. Income
was self-reported household income (seven intervals ranging
from less than $25,000 to $200,000 or more). Political conser-
vatism was measured with a single-item scale ranging from 1
(extremely liberal)to5(extremely conservative). Religiosity was
measured with a single-item scale ranging from 1 (not religious
at all)to5(devoutly religious).
Phase 1: Item selection
We based item selection on three factorsreliability, patterns
of predicted external correlations, and mean external correla-
tions. We measured reliability via Cronbachs alpha, which esti-
mates internal consistency, or the degree to which the items tap
the same construct. For scales that are designed to estimate uni-
tary constructs, as the value scales are, items that yield higher
alpha levels are preferred. In the case of the 10-item measure,
internal consistency could not be evaluated because only one
item measured each value. It is important to note that although
we used Cronbachs alpha as one of the criteria for selecting
items, this index can underestimate reliability, especially when
applied to very short measures of broad constructs. This bias
results from the fact that alpha is a function of the number of
items on a scale and average correlation among the items. Typi-
cally the goal of a brief measure is to capture the full breadth of
a construct (i.e., content validity) with as few items as possible.
To capture the breadth of the construct, maximally different
items could be chosen, thereby reducing the mean interitem
correlations. To keep the scale short, few items are selected.
Together both these factors (low interitem correlations, few
items) will tend to diminish the alpha coefcients.
If the new value instruments are to serve as useful alterna-
tives to the PVQ40, the values they measure should show pat-
terns of correlation with external variables that are similar to
those found with the PVQ40. To evaluate the degree to which
this is the case, we performed a three-step analysis. First, we
centered each participants value scores around his or her own
scale mean. This is an important step when evaluating external
correlations because it controls for response biases such as
acquiescence or social desirability (e.g., Schwartz, 2004). Next,
we computed the correlations of the values measured with the
PVQ40 with 11 external variables (e.g., personality traits, soci-
odemographic variables). Finally, we computed the correlations
of the values measured with the new instruments with the same
11 external variables. For this analysis, we are concerned with
the similarity between the patterns of value correlations of the
PVQ40 and the abbreviated scales. Recapturing the pattern of
correlations supports the construct validity of a measureor
its expected relationships with other constructs (whether they
be related or unrelated to the construct of interest; Cronbach &
Meehl, 1955).
Ideally, the value correlations of the new instruments should
demonstrate both convergent and discriminant validity. So, in
the second step, we computed column-vector correlations
between the sets of correlations (i.e., between the correlations
of the PVQ40 and of the abbreviated scales with the external
variables). We rst transformed the correlations to a linear
4SANDY, GOSLING, SCHWARTZ, KOELKEBECK
scale using Fishersr-to-zformula. Strong vector correlations
would indicate that an abbreviated scale captures external pre-
dictions in a manner similar to the full PVQ40.
In addition to reproducing the original pattern of correla-
tions, it is also important that the correlations with the new
instrument are of an approximately similar magnitude as the
correlations with the original instrument; of course, some
reduction in magnitude usually results as a consequence of the
somewhat lower reliability associated with short scales. To
examine whether the new scales recaptured the magnitude of
correlations found with the original scale, the mean of the abso-
lute correlations with external variables was computed for the
original and new values scales.
The 10-item and 20-item scales that optimized these three
criteria (reliability, patterns of external correlates, magnitude of
external correlations) were selected for further evaluation (in
Stage 2). The 10-item and 20-item scale are henceforth referred
to as the Ten Item Value Inventory (TIVI) and the Twenty
Item Value Inventory (TwIVI), respectively.
Results
Table 1 presents the scale reliabilities for the TwIVI, the PVQ40,
and the PVQ21. Scale reliabilities could not be estimated for the
TIVI because it consists of only one item per scale. Reliabilities for
the TwIVI, with two items per value, ranged from .33 to .91 (MD
.71). Reliabilities for the PVQ21 proxy ranged from .18 to .83 (M
D.62). Reliabilities for the full 40-item PVQ ranged from .51 to .88
(MD.76). Two values (security and tradition) had consistently
low reliabilities, even in the full PVQ40. Not surprisingly, with
three to six items per scale, alpha levels for the PVQ40 were higher
on average than for any of the short scales.
The vector correlations revealed that the values measured
with the new, abbreviated scales had patterns of correlation
with the 11 external variables that were highly similar to the
patterns for the PVQ40. Vector correlations ranged from .78
to .98 (MD.91) for the 10-item TIVI, from .85 to .99 (MD
.93) for the 20-item TwIVI, and from .65 to .99 (MD.91) for
the PVQ21 proxy (for full results, see online supplemental
Table S.1).
Mean external correlates ranged from .09 to .14 (MD.11)
for the one-item scales of the TIVI, from .09 to .14 (MD.12)
for the two-item scales of the TwIVI, and from .08 to .13 (MD
.11) for the PVQ21 (for full results, see online supplemental
Table S.2). The abbreviated scales almost fully reproduced the
magnitude of external correlates seen in the scales of the full
PVQ40 scale, which ranged from .10 to .15 (MD.13). A full
summary of external correlates can be found in Table 2. The
compositions of the different scales, including the degree to
which the derived items overlap with the PVQ21, are shown
in Table 3.
Phase 2: Validation
To ensure that the psychometric criteria derived in the original
sample are not capitalizing on correlated error, it is necessary
to compute internal consistency reliability, vector correlations,
and mean correlations in a new, independent sample. Thus, the
goal of Phase 2 was to undertake an evaluation of the scales in
a new sample. To evaluate the degree to which each new scale
recaptured the full original scale, we also assessed the conver-
gent and discriminant validity between the new short scales
and the scales of the original 40-item instrument. Additionally,
we tested the ability of the abbreviated scales to recapture the
value hierarchy (or the mean ranking of the 10 values) of the
PVQ40.
Measures
The same variables were measured in Phase 2 as in Phase 1.
These variables were also measured in the same manner (i.e.,
the same questionnaires were used).
Validation criteria
It is important to establish that the measures are reliable in an
independent sample to ensure that they are sufciently precise
for use in research. We evaluated reliability and external corre-
lations in the same manner as in Phase 1. Predicting the pattern
of external correlation (both the magnitude and pattern) is an
important step in establishing a nomological network, thereby
supporting the construct validity of the instruments (Cronbach
& Meehl, 1955).
In addition to examining the pattern and magnitude of the
external correlates, we also tested the predicted relationships
that the values would have with external variables based on pre-
vious literature. For each value dimension, we formed predic-
tions about the directionality of the relationship the dimension
would have with the external variables. These predictions were
made without reference to the ndings in this data set.
Informed predictions were made only for external variables
with sufcient background research to support the hypothesis.
Hypotheses for values and the Big Five are based on meta-anal-
yses performed by Fischer and Boer (2015), Parks (2007),
Parks-Leduc et al. (2015), and Roccas et al. (2002). Religiosity
hypotheses are based on meta-analyses by Saroglou, Delpierre,
and Dernelle (2004) and Schwartz and Huismans (1995). The
predictions for sociodemographic variables (e.g., age, gender)
are based on Schwartz and Rubel (2005) and on representative
Table 1. Scale alpha reliabilities in the selection and validation samples.
TwIVI PVQ40 PVQ21.
Value Selection Validation Selection Validation Selection Validation
Conformity .61 .60 .73 .73 .67 .68
Tradition .50 .50 .51 .51 .18 .16
Benevolence .91 .91 .85 .86 .83 .85
Universalism .76 .77 .84 .85 .73 .75
Self-direction .81 .82 .88 .89 .77 .79
Stimulation .70 .70 .77 .78 .70 .70
Hedonism .85 .85 .80 .81 .67 .67
Achievement .79 .79 .84 .84 .73 .73
Power .80 .80 .69 .69 .55 .55
Security .33 .33 .65 .64 .40 .39
M.71 .71 .76 .76 .62 .63
Note. TwIVI DTwenty Item Values Inventory; PVQ40 D40-item Portrait Values
Questionnaire; PVQ21 D21-item Portrait Values Questionnaire. The Ten Item
Values Inventory (TIVI) is not represented in this table because internal consis-
tency cannot be estimated for a scale that only uses one item per dimension.
Reliabilities from both the selection phase (Sample 1) and validation phase (Sam-
ple 2) are presented side by side.
SHORT VALUES MEASURE 5
national data from 27 countries in rounds one and two of the
ESS. Hypotheses for values and political conservatism were
formed based on the work of Piurko, Schwartz, and Davidov
(2011) and Schwartz et al. (2010). Table 4 displays the full
matrix of predictions.
In addition, we evaluated the convergent and discriminant
validity of the abbreviated scales by correlating the 10 values
measured by each scale with the 10 values measured by the
PVQ40. High correlations on the diagonal of the intercorrela-
tion matrices would support convergent validity. Traditionally,
near-zero correlations on the off-diagonals support discrimi-
nant validity. However, the values form a circumplex, so a dif-
ferent test of discriminant validity is appropriate. Values
adjacent in the circumplex (e.g., power and achievement)
should correlate positively, opposing values (e.g., power and
benevolence) should correlate negatively, and only orthogonal
values (e.g., power and stimulation) should correlate near zero.
This pattern is found with the PVQ40 values.
Finally, we evaluated the ability of the new scales to recapture
the value hierarchy of the PVQ40. The value hierarchy, or ranking
Table 2. External correlations for selection sample.
CO TR BE UN SD ST HE AC PO SE
PVQ40
Extraversion ¡.10

¡.12

.19

.05

.20

.39

.29

.30

.43

.06

Agreeableness .26

.24

.57

.35

¡.07

.02
¡.02

¡.24

¡.39

¡.01
Conscientiousness .29

.22

.06

¡.02
¡.02
¡.14

¡.15

.16

.17

.45

Emotional Stability .01 .04

.11

.04

.11

.15

.01 ¡.04

¡.02

¡.02

Openness ¡.18

¡.14

.14

.26

.38

.23

.02

.06

.01 ¡.11

Conservatism .23

.32

.01 ¡.24

¡.09

¡.08

¡.07

¡.03
.01 .18

Religiosity .28

.47

.23

.05

¡.07

.01 ¡.08

¡.06

¡.04

.13

Income ¡.01 ¡.07

¡.04

¡.06

.01 ¡.02
¡.01 .08

.13

.05

Education .02

¡.01 ¡.01 .01 .01 ¡.03

¡.04

¡.02

.02

.04

Gender .03

.03

.03

.05

¡.01
.01 .01 ¡.02

¡.05

.03

Age .01 ¡.01 .05

.08

.06

¡.09

¡.07

¡.16

¡.05

.12

PVQ21
Extraversion ¡.11

¡.23

.18

.06

.20

.37

.28

.25

.30

¡.04

Agreeableness .14

.19

.49

.34

¡.10

¡.01 ¡.05

¡.20

¡.39

.00
Conscientiousness .29

.23

.08

¡.02
¡.10

¡.11

¡.15

.07

.12

.25

Emotional Stability ¡.04

¡.02 .01 .06

.06

.17

.00 ¡.10

¡.07

¡.12

Openness ¡.20

¡.17

.11

.27

.36

.26

.01 .03

¡.06

¡.16

Conservatism .25

.22

.00 ¡.21

¡.09

¡.08

¡.08

¡.03
.02 .20

Religiosity .24

.23

.17

.03

¡.05

.00 ¡.10

¡.05

¡.07

.11

Income level .01 ¡.04

¡.04

¡.06

.00 ¡.01 .00 .07

.12

.03

Education level .02
¡.02

¡.02
.02
¡.01
¡.02
¡.02

¡.02
.01 .01
Gender .02

.01
.03

.05

¡.01

¡.01
.00 ¡.00 ¡.05

.05

Age ¡.01 .00 .03

.08

.05

¡.07

¡.04

¡.14

¡.07

.08

TwIVI
Extraversion ¡.06

.03

.19

.05

.18

.37

.31

.24

.49

.08

Agreeableness .21

.13

.56

.33

.01 ¡.01 .04

¡.27

¡.34

¡.03

Conscientiousness .27

.22

.06

¡.04

¡.10

¡.11

¡.13

.17

.20

.50

Emotional Stability ¡.00 .00 .01 .00 .07

.17

.05

¡.04

.02
.00
Openness ¡.18

¡.13

.13

.18

.45

.26

.03

¡.01 .06

¡.07

Conservatism .23

.38

¡.02 ¡.26

¡.10

¡.08

¡.06

¡.01 .02 .15

Religiosity .28

.56

.19

.03

¡.02
.00 ¡.05

¡.07

.01 .11

Income level .00 ¡.02
¡.03

¡.06

¡.01 ¡.01 ¡.03

.08

.11

.05

Education level .01
.02

.00 ¡.01 .00 ¡.02
¡.06

¡.03

.03

.05

Gender .03

.05

.05

.05

¡.02

¡.01
.00 ¡.04

¡.03

.02

Age ¡.01 .03

.07

.05

.04

¡.07

¡.09

¡.20

¡.01 .11

TIVI
Extraversion .04

.06

.19

.04

.13

.36

.31

.22

.39

.06

Agreeableness .20

.16

.52

.29

.04

¡.06

.02

¡.21

¡.36

¡.03

Conscientiousness .24

.13

.05

¡.03

¡.03

¡.10

¡.14

.10

.17

.58

Emotional Stability .06

.04

.02
¡.00 .09

.17

.05

¡.07

¡.04

.00
Openness ¡.12

¡.06

.12

.13

.34

.20

.03

¡.03

¡.00 ¡.06

Conservatism .22

.37

¡.03
¡.19

¡.08

¡.04

¡.05

¡.01 .04

.10

Religiosity .28

.61

.19

.02

¡.02

.01 ¡.05

¡.06

.01 .09

Income level ¡.00 ¡.03

¡.03

¡.06

.00 ¡.01 ¡.03

.07

.09

.03

Education level .02
.03

¡.00 ¡.01 .03

¡.04

¡.06

¡.03

.02

.06

Gender .03

.04

.05

.05

¡.01
¡.04

.00 ¡.02

¡.02

.05

Age .02

.02

.05

.04

.05

¡.09

¡.08

¡.20

¡.03

.12

Note.CODConformity; TR DTradition; BE DBenevolence; UN DUniversalism; SD DSelf-direction; ST DStimulation; HE DHedonism; AC DAchievement; PO DPower;
SE DSecurity; PVQ40 D40-item Portrait Values Questionnaire; PVQ21 D21-item Portrait Values Questionnaire; TwIVI DTwenty Item Values Inventory; TIVI DTen
Item Values Inventory.
p<.01.

p<.001.
6SANDY, GOSLING, SCHWARTZ, KOELKEBECK
of values within a sample (as established through the sample
means), describes the relative value priorities of a given population.
There is considerable convergence regarding the relative impor-
tance (and unimportance) of the values across countries. Benevo-
lence, self-direction, and universalism typically emerge as the most
important values, and power, tradition, and stimulation as the least
important (Schwartz & Bardi, 2001). However, the value priorities
of different populations tend to vary around this pancultural base-
line. Given the importance of the value hierarchy for understanding
particular populations, it is important for any short scale of values
to be able to effectively capture this hierarchy. We therefore com-
pared the mean value ranking obtained with each shorter scale to
that obtained with the PVQ40.
Results
Reliability
The reliability estimates in the validation sample were generally
similar to those in the selection sample. Table 1 reports coefcient
alphas between .33 and .91 (MD.71) for the values measured by
the TwIVI. For the PVQ21, reliabilities ranged from .16 to .85 (M
D.63) and for the PVQ40 from .51 to .89 (MD.76). As previ-
ously noted, reliability could not be estimated for the TIVI because
the scales consisted of only one item each.
Correlations with external variables
Each computed scale (PVQ40, PVQ21,TwIVI,andTIVI)was
correlated with a set of external variables (e.g., Big Five, gender).
We compared these correlations with the predicted relationships
that we posited based on past research (predictions can be seen in
Table 4). As shown in Table 5, the relationships between the full
PVQ40 and the external variables closely matched the predicted
relationships in direction and signicance. Overall, 82% of the 90
predictions matched both in the predicted direction and signi-
cance level. Results for the external correlations of the PVQ21,
TwIVI, and TIVI are also found in Table 5. The results revealed
that the TwIVI matched 85% of the initial predictions. The PVQ
21 and the TIVI both matched 83% of predictions. As can be seen
in Table 5, many of the predictions that failed were in the proper
predicted direction but only met marginal signicance.
The full pattern of external correlations was also similar to those
in the selection sample. Online supplemental Table S.1 reports vec-
tor correlations for the TwIVI of .95 to .99 (MD.97). Vector corre-
lations for the TIVI ranged from .82 to .99 (MD.91). Vector
correlations for the PVQ21 ranged from .79 to .99 (MD.93).
The correlations of values with external variables were gen-
erally stronger for the longer, more reliable measures, as psy-
chometric principles would predict. Online supplemental
Table S.2 reports absolute mean correlations for the PVQ40 of
.09 to .13 (MD.11), for the 20-item TwIVI of .09 to .13 (MD
.11), for the 10-item TIVI scale of .07 to .12 (MD.10), and for
the 21-item PVQ21 of .08 to .12 (MD.10). Table 5 provides
full details of the correlations with external variables.
Convergence across measures
The correlations along the diagonals of Table 6 reveal generally
strong convergence between the values measured by the abbrevi-
ated measures and by the PVQ40. As expected, the highest con-
vergence emerged for the two-item measures of the TwIVI (MrD
.91). Convergent validity for the PVQ21 values was also strong
(MrD.91).TheTIVIalsoperformedwellwithameanrof .81.
For the reasons discussed earlier, we assessed the discrimi-
nant validity of the scales by computing the intervalue correla-
tions of each abbreviated scale with the intervalue correlations
Table 3. Overlap of items across measures.
Item TIVI TwIVI PVQ21
1SDSD
2PO
3UNUNUN
4AC
5SE
6STST
7CO
8UN
9
10 HE HE HE
11 SD
12 BE BE BE
13 AC AC AC
14 SE
15 ST ST ST
16 CO CO
17 PO PO PO
18 BE
19 UN
20 TR TR
21 SE SE
22 SD SD
23 UN
24
25 TR TR
26 HE
27 BE
28 CO CO
29
30
31
32
33 AC
34
35 SE
36
37 ST
38 TR
39 PO
40
Note. TIVI DTen Item Values Inventory; TwIVI DTwenty Item Values Inventory;
PVQ21 D21-item Portrait Values Questionnaire. The letter abbreviations indi-
cate which of the 10 values the item was used to measure: CO DConformity; TR
DTradition; BE DBenevolence; UN DUniversalism; SD DSelf-direction; ST D
Stimulation; HE DHedonism; AC DAchievement; PO DPower; SE DSecurity.
Table 4. External correlate predictions for validation sample.
Value E A C N O Gender Age Religiosity Conservatism
Conformity x CCC¡ // CC
Tradition ¡CCC¡ // CC
Benevolence CCC xCC C //
Universalism x CC xC/C/¡
Self-direction CxC¡C // ¡¡
Stimulation Cx¡¡C /¡¡ /
Hedonism Cx¡¡C /¡¡ ¡
Achievement C¡C xx / / / /
Power x¡¡ ¡ // /
Security x x CC¡ /C/C
Note.EDExtraversion; A DAgreeableness; C DConscientiousness; N DNeuroti-
cism; O DOpenness; Dnegative correlation; CDpositive correlation; / D
insufcient literature to make a prediction; x Dno predicted relationship.
SHORT VALUES MEASURE 7
of the PVQ40 scales. The patterns of correlations for the short
scales very closely matched those of the criterion correlation
matrix of the PVQ40. Correlations were .92 (TwIVI), .93
(TIVI), and .92 (PVQ21).
Value hierarchy
The ranking of the mean score for each abbreviated scale was com-
pared to the ranking of the scores for the PVQ40 scales. The
TwIVI came the closest to recapturing the value hierarchy of the
full scale, with only 4 (out of 50 possible) deviations. The PVQ21
and the TIVI each deviated 8 ranks. Rank-order correlations with
the PVQ40 were .98 for the TwIVI, .94 for the PVQ21, and .91
for the TIVI. A full report of the ranks and mean differences for
each of the abbreviated scales canbeseeninonlinesupplemental
Tables S.3, S.4, and S.5.
Testretest reliability
Testretest reliability statistics are particularly important met-
rics to gather when evaluating short measures because, unlike
Cronbachs alpha, they are not deated by the low number of
items (e.g., Gosling et al., 2003). We therefore conducted a
Table 5. External correlations validation sample.
CO TR BE UN SD ST HE AC PO SE
PVQ40
Extraversion ¡.08

¡.08

.19

.05

.16

.36

.28

.25

.37

.07

Agreeableness .26

.25

.47

.33

¡.04

.01 ¡.02 ¡.24

¡.36

.00
Conscientiousness .26

.17

.05

¡.02 .02 ¡.11

¡.11

.17

.17

.40

Emotional Stability .02 .03
.08

.01 .07

.12

.03
¡.03
¡.01 ¡.01
Openness ¡.14

¡.11

.13

.22

.31

.20

.02 .04

.00 ¡.09

Conservatism .22

.30

.02
¡.24

¡.10

¡.07

¡.06

¡.02
.01 .17

Religiosity .27

.46

.20

.04

¡.06

.00 ¡.09

¡.06

¡.04

.13

Income level .00 ¡.06

¡.03

¡.05

.02
¡.02 ¡.01 .07

.12

.05

Education level .02
¡.01 ¡.03

¡.02
¡.01 ¡.05

¡.05

¡.03

.03

.03

Gender .03

.03

.03

.04

¡.01 .01 .01 ¡.01
¡.04

.03

Age .00 ¡.01 .05

.08

.06

¡.09

¡.07

¡.15

¡.04

.11

PVQ21
Extraversion ¡.10

¡.19

.17

.05

.16

.34

.27

.20

.25

¡.03
Agreeableness .14

.22

.39

.31

¡.03
¡.02 ¡.05

¡.21

¡.36

.01
Conscientiousness .26

.20

.07

¡.02 ¡.05

¡.08

¡.12

.08

.12

.20

Emotional Stability ¡.04

¡.02 ¡.01 .03
.04

.13

.01 ¡.07

¡.05

¡.10

Openness ¡.17

¡.14

.10

.23

.30

.23

.01 .01 ¡.06

¡.14

Conservatism .22

.21

.01 ¡.22

¡.10

¡.07

¡.06

¡.03

.02
.19

Religiosity .23

.22

.16

.02 ¡.05

¡.01 ¡.10

¡.06

¡.06

.10

Income level .01 ¡.03

¡.03

¡.05

.01 .00 .00 .06

.11

.03

Education level .02
¡.02
¡.04

¡.01 ¡.03

¡.03

¡.03

¡.03

.02
.00
Gender .02

.01 .03

.04

¡.01 ¡.01 .01 .00 ¡.05

.05

Age ¡.01 ¡.01 .03

.07

.05

¡.07

¡.05

¡.13

¡.07

.07

TwIVI
Extraversion ¡.06

.07

.19

.03

.15

.34

.23

.19

.44

.09

Agreeableness .22

.12

.46

.30

.03
¡.02 .03

¡.26

¡.32

¡.02
Conscientiousness .23

.17

.05

¡.03
¡.06

¡.08

¡.10

.17

.19

.45

Emotional Stability .01 .01 ¡.01 ¡.02 .05

.13

.06

¡.02 .03
.02
Openness ¡.14

¡.11

.11

.15

.38

.23

.03

¡.01 .06

¡.05

Conservatism .22

.36

¡.01 ¡.26

¡.11

¡.07

¡.05

.00 .02 .13

Religiosity .27

.53

.17

.02
¡.02
¡.01 ¡.06

¡.08

.01 .11

Income level .00 ¡.02
¡.02 ¡.06

¡.01 .00 ¡.02
.08

.10

.05

Education level .01 .02
¡.02

¡.03

¡.02
¡.03

¡.08

¡.03

.03

.05

Gender .02

.04

.04

.05

¡.01
¡.01 .01 ¡.03

¡.03

.01
Age ¡.02

.03

.06

.05

.04

¡.07

¡.09

¡.18

¡.00 .11

TIVI
Extraversion .04

.10

.18

.03
.12

.33

.30

.17

.34

.06

Agreeableness .21

.13

.43

.26

.04

¡.06

.02 ¡.23

¡.34

¡.01
Conscientiousness .21

.09

.05

¡.03
.00 ¡.07

¡.11

.10

.16

.52

Emotional Stability .06

.04

.00 ¡.02 .05

.14

.05

¡.05

¡.02 ¡.00
Openness ¡.09

¡.05

.11

.11

.30

.18

.03
¡.02 .00 ¡.06

Conservatism .22

.33

¡.01 ¡.21

¡.09

¡.04

¡.04

¡.01 .03

.10

Religiosity .27

.59

.17

.01 ¡.02
¡.01 ¡.06

¡.06

.01 .10

Income level ¡.00 ¡.03

¡.03

¡.05

.01 ¡.01 ¡.03

.08

.08

.03

Education level .01 .03

¡.03

¡.03

.01 ¡.05

¡.07

¡.04

.02

.05

Gender .03

.04

.04

.04

¡.01 ¡.03

.00 ¡.02
¡.02

.04

Age .01 .03

.05

.04

.04

¡.09

¡.08

¡.18

¡.03

.11

Note.CODConformity; TR DTradition; BE DBenevolence; UN DUniversalism; SD DSelf-direction; ST DStimulation; HE DHedonism; AC DAchievement; PO DPower; SE DSecu-
rity; PVQ40 D40-item Portrait Values Questionnaire; PVQ21 D21-item Portrait Values Questionnaire; TwIVI DTwenty Item Values Inventory; TIVI DTen Item Values Inventory.
Terms shown in bold italics indicate a conict with predicted correlations. Terms shown in bold indicate consistency with predicted relationships. Nonbolded numbers indicate that
nopredictionsweremadeaboutthosecorrelations.
p<.01.

p<.001.
8SANDY, GOSLING, SCHWARTZ, KOELKEBECK
study to establish the testretest reliability of the two new pro-
posed measures (the TIVI and the TwIVI).
Participants and method
Two samples (one for the TIVI and one for the TwIVI) of partici-
pants were assessed twice, with a 2-week interval between assess-
ments. Data were again collected via Amazons Mechanical Turk.
One hundred and fteen participants took the TIVI at Time 1 and
76 of them were available to retakeitatTime2.Onehundredand
twenty-one participants took the TwIVI at Time 1 and 46 of them
were available to retake it at Time 2. Demographic variables were
not collected for this sample.
Table 6. Convergent and discriminant validity with the full PVQ40 in validation sample.
Scales CO TR BE UN SD ST HE AC PO SE
PVQ40
Conformity 1
Tradition .62

1
Benevolence .36

.27

1
Universalism .24

.15

.72

1
Selfdirection .07

¡.05

.69

.67

1
Stimulation ¡.05

¡.07

.45

.45

.59

1
Hedonism .05

¡.09

.47

.42

.57

.61

1
Achievement .15

¡.09

.30

.26

.49

.41

.45

1
Power ¡.06

¡.17

¡.26

¡.25

¡.09

.07

.09

.44

1
Security .49

.34

.30

.28

.25

.08

.22

.36

.17

1
TwIVI
Conformity .93

.59

.34

.21

.07

¡.02

.07

.17

¡.02

.47

Tradition .49

.82

.06

¡.10

¡.23

¡.14

¡.16

¡.05

.06

.28

Benevolence .33

.23

.95

.70

.66

.43

.46

.30

¡.23

.28

Universalism .20

.10

.65

.91

.60

.38

.40

.24

¡.22

.23

Selfdirection .06

¡.04

.67

.67

.94

.58

.52

.43

¡.13

.19

Stimulation ¡.08

¡.08

.40

.41

.56

.95

.53

.38

.07

.05

Hedonism .09

¡.02

.55

.49

.62

.65

.95

.43

¡.01 .22

Achievement .12

¡.10

.12

.09

.30

.29

.34

.92

.51

.33

Power ¡.03

¡.09

¡.13

¡.14

.01 .10

.07

.41

.91

.16

Security .40

.29

.14

.16

.11

.01 .07

.24

.20

.84

TIVI
Conformity .76

.54

.43

.28

.18

.11

.17

.17

¡.07

.42

Tradition .38

.71

.13

¡.04

¡.13

¡.06

¡.15

¡.07

¡.01
.19

Benevolence .32

.24

.90

.67

.62

.41

.42

.28

¡.22

.27

Universalism .19

.08

.66

.83

.62

.37

.40

.26

¡.23

.20

Selfdirection .12

.00 .68

.67

.89

.53

.51

.43

¡.15

.23

Stimulation ¡.16

¡.09

.18

.19

.31

.81

.39

.26

.16

¡.05

Hedonism .03

¡.04

.45

.39

.51

.60

.89

.36

.02

.16

Achievement .13

¡.10

.12

.08

.25

.26

.33

.85

.47

.27

Power ¡.01
¡.07

¡.25

¡.26

¡.15

¡.02
¡.03

.31

.86

.12

Security .31

.23

.12

.10

.09

¡.01 .02

.18

.15

.60

PVQ21
Conformity .86

.53

.06

¡.03

¡.19

¡.23

¡.15

.05

.07

.39

Tradition .59

.81

.26

.19

.04

¡.07

.01 .00 ¡.14

.38

Benevolence .33

.23

.95

.70

.71

.45

.49

.34

¡.21

.30

Universalism .21

.11

.74

.95

.71

.46

.44

.28

¡.26

.25

Selfdirection .03

¡.06

.64

.62

.95

.57

.56

.45

¡.07

.20

Stimulation ¡.08

¡.08

.40

.41

.56

.95

.53

.38

.07

.05

Hedonism ¡.01 ¡.12

.34

.31

.44

.54

.96

.40

.15

.17

Achievement .14

¡.11

.29

.23

.44

.35

.44

.92

.38

.29

Power ¡.05

¡.17

¡.36

¡.34

¡.22

¡.04

.04

.36

.94

.14

Security21 .45

.30

.19

.15

.11

¡.06

.16

.26

.14

.83

Note.CODConformity; TR DTradition; BE DBenevolence; UN DUniversalism; SD DSelf-direction; ST DStimulation; HE DHedonism; AC DAchievement; PO DPower;
SE DSecurity; PVQ40 D40-item Portrait Values Questionnaire; TwIVI DTwenty Item Values Inventory; TIVI DTen Item Values Inventory; PVQ21 D21-item Portrait
Values Questionnaire. The numbers along the diagonal reect convergent validityor the correlations between the PVQ40 and each other scale. The numbers on the
upper diagonal reect the correlations between the PVQ40 values and the values computed by each other scale. The lower diagonal reects discriminant validityor
the intercorrelations within each scale. Only the lower diagonal for the PVQ40 correlations is reported because the numbers mirror the upper diagonal.
p<.01.

p<.001.
Table 7. Testretest correlations for TwIVI and TIVI.
Value TwIVI TIVI
Conformity .65 .70
Tradition .58 .79
Benevolence .69 .73
Universalism .81 .60
Self-direction .77 .60
Stimulation .71 .64
Hedonism .76 .60
Achievement .51 .68
Power .66 .62
Security .53 .67
M.67 .66
Note. TwIVI DTwenty Item Values Inventory; TIVI DTen Item Values Inventory. All
correlations were signicant at p<.001.
SHORT VALUES MEASURE 9
Results
Testretest correlations for both the TwIVI and TIVI can be seen in
Table 7. Correlations for the TwIVI ranged from .51 to .81
(MD.67). Correlations for the TIVI ranged from .60 to .79 (MD
.66). These results are in line with similarly abbreviated personality
measures (e.g., Gosling et al., 2003). Additionally, in the case of the
PVQ40, Schwartz et al. (2001) reported 2-week testretest correla-
tions ranging from .66 to .88 (MD.81) and Bardi, Lee, Hofmann-
Towgh, and Soutar (2009) reported 9-month testretest correla-
tions ranging from .58 to .68 (MD.63).
Given the importance of value priorities in predicting
behavior, prole stability coefcientsor the relative stability
of the value priorities from Time 1 to Time 2were com-
puted (Schwartz & Bardi, 2001). These coefcients were cal-
culated by averaging the intraindividual correlations for the
TwIVI and TIVI. The stability coefcient for the TwIVI was
.86 (SD D.15) and slightly lower at .77 (SD D.31) for the
TIVI.
Discussion
Our goal was to develop psychometrically sound brief (20-item)
and ultrabrief (10-item) measures of values. We evaluated the
instrumentsconvergent and discriminant validities, reliabilities,
and external correlations. The TwIVI outperformed the PVQ21
in almost every area of psychometric evaluation. As expected, the
longer measures outperformed the 10-item measure (TIVI). All of
the new measures met a satisfactory standard of reliability and
validity.Additionally,allofthemeasureswereabletoalmost
completely reproduce the pattern of predicted relationships with
external variables.
We wanted to develop measures that made direct cross-
instrument comparisons possible, with the ultrabrief measure
being a subset of the brief measure, and the brief measure being
a subset of the PVQ40. Our psychometric analyses allowed us
to achieve this goal, with the TIVI being completely subsumed
by the TwIVI. Specically, in four cases (at the item-selection
phase), there were two candidate items with virtually identical
psychometric properties, so we chose the ones that would
ensure cross-instrument overlap. The online supplemental
Appendices present the TIVI and the TwIVI along with their
scoring instructions. Normative data (calculated from the
Mytype.com data) for the two new measures can be found in
online supplemental Tables S.6 and S.7.
Of the two measures, we recommend using the TwIVI over
the TIVI for six reasons. First, the TwIVI has slightly superior
psychometric properties (see Table 1 and online supplemental
Tables S.1 and S.2). Second, the TwIVI more successfully recap-
tures predicted relationships between values and external varia-
bles. Third, the TwIVI comes closer than the TIVI to
duplicating the value hierarchy of the PVQ40. Fourth, the two
items on each scale allow researchers to undertake rudimentary
checks for random responding and other rating nuisances and
to compute indexes of internal consistency (although research-
ers are warned against relying too heavily on such measures;
see McCrae, Kurtz, Yamagata, & Terracciano, 2011). Fifth, the
TwIVI offers greater content validity because there are two
items measuring each dimension instead of one. The TIVI
remains a viable option for researchers extremely short on
time, questionnaire space, or both. Our results showed that,
even with a lower internal consistency, the TIVI maintains
equally high testretest reliabilities as the TwIVI.
Limitations and future directions
A number of limitations with this study should be noted. First,
our testretest sample was relatively small and, therefore, gen-
eralizability of these ndings is limited. Second, our samples
were heavily U.S.-based and collected via a Facebook applica-
tion; future research should examine these scales in a broader
sample of countries and using a broader array of assessment
methods. Additionally, we were unable to collect nationality
information on participants in the testretest study. This limits
the generalizability of those ndings. Third, the emphasis on a
data-driven approach resulted in the selection of some items
that suffer from low face validity. For example, the single item
assessing tradition in the TIVI was, Religious belief is impor-
tant to him/her. S/he tries hard to do what his/her religion
requires,which would appear to miss nonreligious people
high on tradition. Nonetheless, we retained such items because
the analyses clearly showed them to have the strongest conver-
gent validity with the PVQ40 scales; had we abandoned our
data-driven decisions for these items, we would be compelled
to do so for all the scales, unhitching us from a key strength of
this work.
Finally, the TwIVI sometimes suffered from low internal
consistency. As previously noted, however, we view this more
as an apparent limitation than a genuine limitation, because
researchers should not rely too heavily on measures of inter-
nal consistency when it comes to evaluating the efcacy of
abbreviated measures. To attenuate for the bias in alpha reli-
abilities, they should put more weight on other forms of reli-
ability (e.g., testretest) and validity (e.g., convergent, external
correlates).
Funding
The work of Shalom H. Schwartz on this article was done within the
framework of the Basic Research Program at the National Research Uni-
versity Higher School of Economics (HSE) that was supported by the Gov-
ernment of the Russian Federation for the implementation of the Global
Competitiveness Program.
References
Bardi, A., Lee, J. A., Hofmann-Towgh, N., & Soutar, G. (2009). The struc-
ture of intraindividual value change. Journal of Personality and Social
Psychology,97, 913929.
Bardi, A., & Schwartz, S. H. (2003). Values and behavior: Strength and
structure of relations. Personality & Social Psychology Bulletin,29,
12071220.
Buss, A. H. (1989). Personality as traits. American Psychologist,44, 1378
1388.
Bilsky, W., Janik, M., & Schwartz, S. H. (2011). The structural organization
of human valuesEvidence from three rounds of the European Social
Survey (ESS). Journal of Cross-Cultural Psychology,42, 759776.
Boer, D., Fischer, R., Strack, M., Bond, M. H., Lo, E., & Lam, J. (2011). How
shared preferences in music create bonds between people: Values as the
missing link. Personality & Social Psychology Bulletin,37, 11591171.
Buhrmester, M., Kwang, T., & Gosling, S. D. (2011). Amazons Mechanical
Turk: A new source of inexpensive, yet high-quality, data? Perspectives
on Psychological Science,6,35.
10 SANDY, GOSLING, SCHWARTZ, KOELKEBECK
Caprara, G. V., Vecchione, M., & Schwartz, S. H. (2009). Mediational role
of values in linking personality traits to political orientation. Asian
Journal of Social Psychology,12,8294.
Cronbach, L. J., & Meehl, P. E. (1955). Construct validity in psychological
tests. Psychological Bulletin,52, 281302.
Dollinger, S. J., & Kobayashi, R. (2003). Value correlates of collegiate alco-
hol abuse. Psychological Reports,93, 848850.
Fischer, R., & Boer, D. (2015). Motivational basis of personality traits: A
meta-analysis of value-personality correlations. Journal of Personality,
83, 491510.
Gosling, S. D., & Mason, W. (2015). Internet research in psychology.
Annual Review of Psychology,66, 877902.
Gosling, S. D., Rentfrow, P. J., & Swann, W. B. (2003). A very brief measure
of the Big-Five personality domains. Journal of Research in Personality,
37, 504528.
Grunert, S., & Juhl, H. (1995). Values, environmental attitudes, and buying
of organic foods. Journal of Economic Psychology,16,3962.
Haslam, N., Whelan, J., & Bastian, B. (2009). Big Five traits mediate associ-
ations between values and subjective well-being. Personality and Indi-
vidual Differences,46,4042.
Hofstee, W. K. B. (1984). Whats in a trait: Reections about the inevitabil-
ity of traits, their measurement, and taxonomy. Personality Psychology
in Europe,1,7581.
John, O. P., Donahue, E. M., & Kentle, R. L. (1991). The Big Five Inven-
toryVersions 4a and 54. Berkeley, CA: University of California, Ber-
keley, Institute of Personality and Social Research.
McAdams, D. P. (1995). What do we know when we know a person? Jour-
nal of Personality,63, 365396.
McAdams, D. P., & Pals, J. L. (2006). A new Big Five: Fundamental princi-
ples for an integrative science of personality. American Psychologist,61,
204217.
McCrae, R. R., Kurtz, J. E., Yamagata, S., & Terracciano, A (2011). Internal
consistency, retest reliability, and their implications for personality
scale validity. Personality and Social Psychology Review,15,2850.
Minshew, N. J., Turner, C. A., & Goldstein, G. (2005). The application of
short forms of the Wechsler Intelligence Scales in adults and children
with high functioning autism. Journal of Autism and Developmental
Disorders,35,4552.
Nunnally, J. C. (1978). Psychometric theory (2nd ed.). New York, NY:
McGraw-Hill.
Parks, L. (2007, April). Personality and values: A meta-analysis. Paper pre-
sented at the Society of Industrial and Organizational Psychology
annual conference, New York, NY.
Parks-Leduc, L., Feldman, G., & Bardi, A. (2015). Personality traits and
personal values: A meta-analysis. Personality and Social Psychology
Review,19,329.
Paulhus, D. L., & Bruce, M. (1992). The effect of acquaintanceship on the
validity of personality impressions: A longitudinal study. Journal of
Personality and Social Psychology,63, 816824.
Piurko, Y., Schwartz, S. H., & Davidov, E. (2011). Basic personal values and
the meaning of left-right political orientations in 20 countries. Political
Psychology,32, 537561.
Robins, R. W., Hendin, H. M., & Trzesniewski, K. H. (2001). Measuring
global self-esteem: Construct validation of a single-item measure and
the Rosenberg Self-Esteem Scale. Personality and Social Psychology Bul-
letin,27, 151161.
Robins, R. W., Trzesniewski, K. H., Tracy, J. L., Gosling, S. D., & Potter, J.
(2002). Global self-esteem across the life span. Psychology and Aging,
17, 423434.
Roccas, S., Sagiv, L., Schwartz, S. H., & Knafo, A. (2002). The Big Five per-
sonality factors and personal values. Personality and Social Psychology
Bulletin,28, 789801.
Rokeach, M. (1973). The nature of human values. New York, NY: The Free
Press.
Sagiv, L., & Schwartz, S. H. (2004). Values, intelligence and client behavior
in career counseling: A eld study. European Journal of Psychology of
Education,19, 237255.
Sandy, C. J., Gosling, S. D., & Koelkebeck, T. (2014). Psychometric com-
parison of automated versus rational methods of scale abbreviation: An
illustration using a brief measure of values. Journal of Individual Differ-
ences,35, 221235.
Saroglou, V., Delpierre, V., & Dernelle, R. (2004). Values and religiosity: A
meta-analysis of studies using Schwartzs model. Personality & Individ-
ual Differences,37, 721734.
Schultz, P., & Zelezny, L. C. (1998). Values and proenvironmental behav-
ior: A ve-country survey. Journal of Cross-Cultural Psychology,29,
540558.
Schwartz, S. H. (1992). Universals in the content and structure of values:
Theoretical advantages and empirical tests in 20 countries. In M. P.
Zanna (Ed.), Advances in experimental and social psychology (Vol. 25,
pp. 165). New York, NY: Academic Press.
Schwartz, S. H. (2003). A proposal for measuring value orientations across
nations. In Questionnaire development report of the European Social
Survey (pp. 259319). Retrieved from http://www.europeansocialsur
vey.org/index.php?optionDcom_docman&taskDdoc_view&gidD126
&ItemidD80
Schwartz, S. (2004). Evaluating the structure of human values with
conrmatory factor analysis. Journal of Research in Personality,38,
230255.
Schwartz, S. H. (2006). Basic human values: An overview. Retrieved from
http://www.fmag.unict.it/Allegati/convegno%207-8-10-05/Schwartz
paper.pdf
Schwartz, S. H. (2010). Basic values: How they motivate and inhibit proso-
cial behavior. In M. Mikulincer, & P. R. Shaver (Eds.), Prosocial
motives, emotions, and behavior: The better angels of our nature (pp.
221241). Washington, DC: American Psychological Association.
Schwartz, S. H. (2011). Values: Cultural and individual. In F. J. R. van de
Vijver, A. Chasiotis, & S. M. Breugelmans (Eds.), Fundamental ques-
tions in cross-cultural psychology (pp. 463493). Cambridge, UK: Cam-
bridge University Press.
Schwartz, S. H., & Bardi, A. (2001). Value hierarchies across cultures: Taking a
similarities perspective. Journal of Cross-Cultural Psychology,32, 268290.
Schwartz, S. H., & Bilsky, W. (1987). Toward a universal psychological
structure of human values. Social Psychology,3, 550562.
Schwartz, S. H., & Bilsky, W. (1990). Toward a theory of the universal con-
tent and structure of values: Extensions and cross-cultural replications.
Journal of Personality and Social Psychology,58, 878891.
Schwartz, S. H., & Boehnke, K. (2004). Evaluating the structure of human
values with conrmatory factor analysis. Journal of Research in Person-
ality,38, 230255.
Schwartz, S. H., Caprara, G. V., & Vecchione, M. (2010). Basic personal
values, core political values, and voting: A longitudinal study. Political
Psychology,31, 421452.
Schwartz, S. H., & Huismans, S. (1995). Value priorities and religiosity in
four Western religions. Social Psychology Quarterly,58,88107.
Schwartz, S. H., Melech, G., Lehmann, A., Burgess, S., Harris, M., &
Owens, V. (2001). Extending the cross-cultural validity of the theory of
basic human values with a different method of measurement. Journal
of Cross-Cultural Psychology,32, 519542.
Schwartz, S. H., & Rubel, T. (2005). Sex differences in value priorities:
Cross-cultural and multi-method studies. Journal of Personality and
Social Psychology,89, 10101028.
Schwartz, S. H., Sagiv, L., & Boehnke, K. (2000). Worries and values. Jour-
nal of Personality,68, 309346.
Smith, P. B., & Schwartz, S. H. (1997). Values. In J. W. Berry, M. H. Segall,
& C. Kagitcibasi (Eds.), Handbook of cross-cultural psychology (pp. 77
118). Needham Heights, MA: Allyn & Bacon.
Wei, M., Russell, D. W., Mallinckrodt, B., & Vogel, D. L. (2007).
The Experiences in Close Relationship Scale (ECR) short form:
Reliability, validity, and factor structure. Journal of Personality
Assessment,88,187204.
Ziegler, M., Kemper, C. J., & Kruyen, P. (2014). Short scalesFive misun-
derstandings and ways to overcome them. Journal of Individual Differ-
ences,35, 185189.
SHORT VALUES MEASURE 11
... Short scales, which are quick and efficient to administer, have shown to be advantageous in numerous contexts such as large-scale studies, where long questionnaires pose the risk of introducing bias due to participant fatigue (Sandy et al., 2017;Schipolowski et al., 2014). Nonetheless, it is important that all such scales have a profile of high psychometric quality, including cross-cultural validation. ...
... Las escalas cortas, de administración rápida y eficiente, han mostrado tener ventajas en diversos contextos, como en los estudios de gran escala, en los que los cuestionarios largos conllevan el riesgo de introducir sesgos debidos a la fatiga de los participantes (Sandy et al., 2017;Schipolowski et al., 2014). Sin embargo, es importante que todas estas escalas tengan un perfil de alta calidad psicométrica, incluyendo la validación transcultural. ...
... Hence, the two value dimensions measure a preference for Openness over Conservation and for Self-Transcendence over Self-Enhancement. Although a popular scale in values research, it lacks proper validation and psychometric properties are seldom given in previous research (e.g., mostly western samples, values seem to overlap/difficult to form, see Sandy et al., 2017). Further, research has shown the scale to be measurement invariant across 15 European nations only regarding openness and self-enhancement (Cieciuch et al., 2018). ...
Article
Full-text available
Previous research on Schwartz’s theory of basic human values has mostly applied a variable-oriented approach . This study took a person-oriented approach and investigated how values co-occur and are manifested in individuals, that is, what kind of value profiles exist and how they differ in terms of subjective (i.e., life satisfaction) and social well-being (i.e., sense of belonging). In a sample of Finnish adolescents ( N = 973; women 59.6%; M age = 16.8, SD = .70), three value profiles emerged: personal-focus , growth-focus and self-protective , of which the personal-focus group scored highest on life satisfaction, the growth-focus group scored lowest on belonging to social media and the self-protective group scored highest on belonging to organizations. In all, subjective and social well-being were differently related to opposing values. The findings are discussed within a cultural and contextual framework.
... In addition, producers' perspectives relevant to calf care were explored by (1) how producers valued their male calves, (2) characteristics of calf care practices that producers prioritized (e.g., whether they are economical to implement; whether they promote calf health), and (3) which dairy industry stakeholder opinions were most influential, with a focus on how producers interacted with calf buyers. The personal values of producers were assessed using a series of questions measuring Schwartz's 10 basic values (conformity, tradition, security, power, achievement, hedonism, stimulation, self-direction, universalism, and benevolence), which have been validated across cultures for predicting decisions and behaviors (Sandy et al., 2017). Further analysis of producer values is being prepared in another manuscript. ...
Article
The primary objective of this study was to compare male and female dairy calf management practices and evaluate risk factors associated with differences in care. Secondary objectives were to understand surplus calf transportation and marketing practices and investigate incentives to motivate calf care improvements. An online survey was distributed to all dairy producers in Ontario (n = 3,367) from November 2020 to March 2021 and Atlantic Canada (n = 557) from April to June 2021. Dairy producers were identified through provincial dairy associations and contacted via e-mail and social media. Descriptive statistics were computed, and a logistic regression model was created to evaluate factors associated with using discrepant feeding practices (i.e., fed less colostrum, fed colostrum later, or fed raw, unsalable milk) for male calves compared with females. The survey had a 7.4% response rate (n = 289/3,924) and was primarily filled out by farm owners (76%). Although colostrum and milk feeding practices were similar between male and female calves, male calves received less milk while still on the dairy farm of origin compared with females. Male calves were also more likely to be fed a higher proportion of raw, unsalable milk. Female producers and those that kept their male calves beyond 10 d of age had lower odds of using poorer feeding practices for male calves. Male calves were mostly sold between 1 and 10 d (64%), primarily through direct sales to a calf-rearing facility (45%), with auctions being the next most common method (35%). A small but notable proportion of producers (18%) agreed that euthanizing male calves is a reasonable alternative when their sale price is very low; however, few producers (13%) reported that financial costs limited their male calf care. The largest proportion (43%) of producers reported that a price premium for more vigorous calves would motivate them to take better care of their male calves. Conversely, only 28% of producers reported that a price discount for calves in poor condition would be motivating. Producers placed importance on the opinion of their calf buyer, their herd veterinarian, and the Canadian Code of Practice for the Care and Handling of Dairy Cattle when considering their calf care practices, and they highly valued practices that promote calf health. Respondents to this survey reported a lower proportion of tiestall barn use and higher milk productivity compared with typical dairy herds in the region, suggesting selection bias for more progressive dairy producers. Nevertheless, our results suggest that dairy producers provide similar care between male and female calves, but some male calves experience challenges due to milk feeding and marketing practices. Feedback from calf buyers along with continued support and guidance from herd veterinarians and the Code of Practice may motivate dairy producers to improve male calf care.
... We used the Ten Item Values Inventory (TIVI) originally developed by (Sandy et al., 2017) to measure the basic human values proposed by (Schwartz, 1992(Schwartz, , 1994. We adapted the TIVI instructions so that participants were asked to indicate the extent to which partners in the relationship shared each value (e.g., "Both members of the couple like to take risks. ...
Article
Full-text available
Studies have shown that romantic partners in consensual non-monogamous (CNM) relationships are targets of stigmatization. However, little is known about the underlying mechanisms and the conditions under which such stigmatization occurs. In two experimental studies (combined N = 772), we asked participants to read the description of two partners in a relationship (monogamous vs. open relationship vs. polyamorous) and make a series of judgments about those partners. Overall results showed that CNM (vs. monogamous) partners were perceived as less trustworthy and as having more sexual health concerns (Studies 1 and 2), and as being less committed and less sexually satisfied (Study 2). Results from a conditional mediation analysis (Study 2) further showed that participants with negative attitudes toward consensual non-monogamy perceived CNM (vs. monogamous) partners as having less conservation and more openness to change values, which was then associated with more stigmatization. In contrast, participants with positive attitudes toward consensual non-monogamy perceived CNM (vs. monogamous) partners as having more openness to change values, which was then associated with less stigmatization. Taken together, these results extended the literature focused on prejudice, discrimination, and stigmatization of minority groups and highlighted key elements that can be used to buffer stigmatization.
... Personal values were measured via the Ten-item Value Inventory (TIVI; Sandy et al., 2017), an ultra-brief version of the Portrait Values Questionnaire (PVQ; Schwartz, 2003). The 10 values assessed by the TIVI are as follows: conformity, tradition, benevolence, universalism, selfdirection, stimulation, hedonism, achievement, power, and security. ...
Article
Full-text available
Microplastics are an issue of rising concern, in terms of their possible implications for both the environment and human health. A survey was distributed among a representative sample of the adult Norwegian population (N = 2720) to explore the public understanding of microplastics. Respondents were asked to report the first thing that came to mind when they read or heard the word "microplastics," based on which a coding scheme was developed that served to categorize the obtained answers into thematic clusters. Results indicate that the public seem to think of microplastics as something bad that might pollute the ocean and harm animal species. Awareness of the sources of microplastics appeared to be rather low, and few respondents mentioned potential ways to solve the problem. Responses differed across certain socio-demographic characteristics; for example, female and younger respondents were more likely to think about the spread and causes/sources of microplastics, whereas a higher educational level was associated positively with thinking of ways to solve the problem. Additional analyses indicated relationships between personal values and the identified thematic clusters; for example, endorsing self-transcendence and openness-to-change values was associated with thinking of ways to solve and of consequences of microplastics. These findings are informative to those wanting to design tailored communications and interventions aimed at reducing plastic pollution and plastic waste.
Article
Democracy was forged in the furnaces of oppression, whether combatting tyranny or affirming the rights of the individual. As democracy is under threat in many parts of the world, there has never been a more urgent need to understand political thoughts and behaviours. This lucid and accessible book brings together a global group of scholars from psychology, political science, communication, sociology, education and psychiatry. The book's structure, based on Abraham Lincoln's well-known phrase 'Of, by and for' the people, scrutinises the psychological factors experienced by politicians as representatives 'of' the electorate, the political institutions and systems devised 'by' those we elect, and the societies that influence the context 'for' us as citizens. From trust to risk, from political values to moral and religious priorities, from the personality and language of leaders to fake news and anti-democratic forces, this book provides vital new insights for researchers, politicians and citizens alike.
Article
Introduction Personal values are important for understanding psychological phenomena by serving as predictors of individuals’ attitudes and behavior. The Twenty-Item Value Inventory (TwIVI, Sandy et al., 2017) was developed to assess values with items from the Portrait Values Questionnaire (PVQ, Schwartz et al., 2001). Objective The goal of this study was to assess the reliability and validity of the French version of the TwIVI. Method Two French-Canadian samples of adults (N1 = 825; N2 = 429) were recruited in order to test the reliability and validity of the TwIVI. Results Results revealed that the French version of the TwIVI showed satisfying temporal stability, convergence validity with the PVQ, and concurrent validity with personality traits. Some issues were observed for internal consistency and structural validity, assessed with both multidimensional scaling and exploratory structural equation modeling. Conclusion Overall, these findings partially support the psychometric qualities of the French version of the TwIVI. Implications for theory and research on values are discussed.
Article
Introduction Several scales were developed to describe and assess the work values system. Although some overlap is observed, none of these scales was developed to incorporate the whole array of work values that can be identified in the literature. Objective The goal of this set of studies was to develop and validate a work values scale based on a comprehensive review of the literature, namely the Integrative Work Values Scale (IWVS). Method A panel of experts (N1 = 9) and two French-Canadian samples of adults (N2 = 229; N3 = 476) were recruited to develop the IWVS and to test its reliability and validity. Results Omega coefficients revealed that the IWVS showed satisfying reliability estimates. Results also supported its content validity, convergent validity with general values, and factor validity. Conclusion Overall, these findings support the utility of the IWVS to assess the wide array of work values that were identified in the literature. Implications for theory and research on work values are discussed.
Article
Purpose Technology-based food innovations are likely to face opposition in consumer acceptance despite certain perceived benefits. This study investigates distrust in scientists as a barrier to cultured meat acceptance and its interplay with positive drivers. For the latter, this study draws upon higher-order values (universalism) and domain-specific innovativeness. Design/methodology/approach This study proposes a conceptual model, which is empirically tested using an online survey of Austrian consumers ( n = 509) and structural equation modelling. Findings First, this study provides empirical support for the inhibiting role of distrust in scientists on the acceptance of cultured meat. Second, universalism and domain-specific innovativeness both positively impact consumers' behavioural intentions. Third, this study finds that consumers' innovativeness mitigates the negative effect of distrust in scientists on the willingness to try cultured meat, whereas the value of universalism does not moderate this negative effect. Practical implications The findings are strategically useful for companies in the innovative food sector in terms of segmentation and targeting. For both companies and policymakers, the impeding role of distrust in scientists for technology-based food innovations depicts an aspect that might be considered relevant to be addressed. Originality/value This study adds to the current body of knowledge about consumer acceptance of technology-based food innovations, such as cultured meat, by jointly examining the relevance of human values, domain-specific innovativeness and distrust in scientists as well as their interactions to explain behavioural intentions.
Article
A positive and reciprocal relationship between subjective wellbeing and pro-environmental behaviour (PEB) has been observed across a range of countries worldwide. There is good reason however to think that the nature of the PEB-wellbeing link might vary between individuals and cross-culturally. We use data obtained in Brazil, China, Denmark, India, Poland, South Africa and the UK (total n = 6969) to test a series of hypotheses using pre-registered regression models. First, we assess the relationship between PEB and wellbeing across countries, and test the ‘privilege’ hypothesis that this varies according to personal income and a country’s level of development. Second, we consider the role of individual values and motivations in relation to PEB and wellbeing. To this end, we test the ‘enhancement’ hypothesis, in which the PEB-wellbeing link is strengthened by people holding particular values and motivations. Third, we consider the role of cultural differences for the nature of the PEB-wellbeing link. We test the ‘social green’ hypothesis that public sphere behaviours (e.g. addressing environmental issues with other people) are more closely linked to wellbeing than are private sphere behaviours (e.g. product purchasing) in collectivistic cultures; in tandem, we assess whether private sphere behaviours are more closely linked to wellbeing in individualistic cultures. We obtain strong evidence for a PEB-wellbeing link across nations. There is partial evidence across countries to support the ‘social green’ hypothesis, but little evidence for the ‘privilege’ or ‘enhancement’ hypotheses. We discuss the implications of our findings for understanding the relationship between PEB and wellbeing, and consider how its promotion might feature in environmental and public health policy.
Chapter
Full-text available
During the past two decades, I have developed, validated, and applied two separate theories of value related constructs. The first concerns the basic human values on which individual people in all societies differ (e.g., security, achievement, hedonism, concern for others). Basic individual values are an aspect of personality. The second theory deals with normative value orientations on which cultures differ (e.g., hierarchy, egalitarianism, harmony). These orientations underlie and justify the functioning of societal institutions. Are two value theories really necessary? Could the same value constructs or dimensions serve at both individual and cultural levels of analysis? It would certainly be more parsimonious to do with one theory. Logically, one theory would suffice if we could assume that culture is simply personality writ large or that individual values are culture writ small. Sadly—for accepting either assumption would make life easier—I find them both unacceptable. In addressing the question of the relationship between individual and culture levels, I hold that we need separate theories of values. This chapter presents and contrasts my individual-level and culture-level theories and suggests how to apply them fruitfully together. It is structured as follows: First, I explicate each theory, specifying its constructs and the relations among them and citing evidence to support them. Next, I compare the empirical structures obtained when the same values data are analyzed at the two levels of analysis and discuss how to interpret these structures as expressions of individual personality and of societal culture. I then contrast the causes of individual differences in basic values and the causes of societal differences in cultural orientations. Next, I present and illustrate the questions that cultural orientations are suited to address and the different questions that individual values are suited to address. Finally, I discuss and illustrate how multi-level analyses that exploit both types of values together can explain national and individual differences in behavior and attitudes.
Article
Full-text available
Four studies examined the construct validity of two global self-esteem measures. In Studies 1 through 3, the Single-Item Self-Esteem Scale (SISE) and the Rosenberg Self-Esteem Scale (RSE) showed strong convergent validity for men and women, for different ethnic groups, and for both college students and community members. The SISE and the RSE had nearly identical correlations with a wide range of criterion measures, including domain-specific self-evaluations, self-evaluative biases, social desirability, personality, psychological and physical health, peer ratings of group behavior, academic outcomes, and demographic variables. Study 4 showed that the SISE had only moderate convergent validity in a sample of children. Overall, the findings support the reliability and validity of the SISE and suggest it can provide a practical alternative to the RSE in adult samples. More generally, the findings contribute to the research literature by further elaborating the nomological network of global self-esteem.
Article
Full-text available
What do we know about short psychological scales? The idea for this special issue was born out of a common interest in short scales, defined as scales measuring a specific psychological construct with fewer than, say, ten items. All editors of this special issue were recently either active in short scale construction themselves (Kemper, Beierlein, Kovaleva, & Rammstedt, 2013; Ziegler, Kemper, & Rammstedt, 2013) or evaluated the current status of short scales in the field (Kemper, Brähler, & Zenger, 2013; Kruyen, Emons, & Sijtsma, 2012, 2013a, 2013b, 2014). Based on discussions with other scientists about the challenges and caveats of short scales, they felt a strong need to share contemporary issues regarding short scales with a broader scientific audience. This special issue brings together a variety of researchers who have approached the topic of short scales from different angles. It is our ambition to resolve some misunderstandings and open questions regarding the construction, psychometric quality, and use of short scales by bringing together methodological, statistical and construction-oriented perspectives . We aim at contributing to an ongoing debate.
Article
Full-text available
Personality traits and personal values are important psychological characteristics, serving as important predictors of many outcomes. Yet, they are frequently studied separately, leaving the field with a limited understanding of their relationships. We review existing perspectives regarding the nature of the relationships between traits and values and provide a conceptual underpinning for understanding the strength of these relationships. Using 60 studies, we present a meta-analysis of the relationships between the Five Factor Model (FFM) of personality traits and the Schwartz (1992) values, and demonstrate consistent and theoretically-meaningful relationships. However, these relationships were not generally large, demonstrating that traits and values are distinct constructs. We find support for our premise that more cognitively-based traits are more strongly related to values and more emotionally-based traits are less strongly related to values. Findings also suggest that controlling for personal scale-use tendencies in values is advisable.
Article
A family of automated analyses known as Genetic Algorithms (GAs) have recently been applied to abbreviating scales in the psychological literature. This method allows long scales to be abbreviated in a matter of minutes, representing a huge potential savings of time over standard rational methods, which are based on psychometric analyses. To compare the results of the automated GA approach with those of the traditional approach, we used both methods to abbreviate a 40-item measure of psychological values (Portrait Values Questionnaire). The two methods yielded 10- and 20-item measures of values, which proved to be successful at recapturing the properties of the original measure. The abbreviated scales created from two approaches displayed strikingly similar psychometric qualities. The advantages and disadvantages of the different approaches are discussed.
Article
Today the Internet plays a role in the lives of nearly 40% of the world's population, and it is becoming increasingly entwined in daily life. This growing presence is transforming psychological science in terms of the topics studied and the methods used. We provide an overview of the literature, considering three broad domains of research: translational (implementing traditional methods online; e.g., surveys), phenomenological (topics spawned or mediated by the Internet; e.g., cyberbullying), and novel (new ways to study existing topics; e.g., rumors). We discuss issues (e.g., sampling, ethics) that arise when doing research online and point to emerging opportunities (e.g., smartphone sensing). Psychological research on the Internet comes with new challenges, but the opportunities far outweigh the costs. By integrating the Internet, psychological research has the ability to reach large, diverse samples and collect data on actual behaviors, which will ultimately increase the impact of psychological research on society. Expected final online publication date for the Annual Review of Psychology Volume 66 is November 30, 2014. Please see http://www.annualreviews.org/catalog/pubdates.aspx for revised estimates.
Article
Objective We investigated the relationships between personality traits and basic value dimensions. Furthermore, we developed novel country-level hypotheses predicting that contextual threat moderates value-personality trait relationships.Method We conducted a three-level v-known meta-analysis of correlations between big five traits and Schwartz’ (1992) ten values involving 9,935 participants from 14 countries. Variations in contextual threat (measured as resource threat, ecological threat and restrictive social institutions) were used as country-level moderator variables.ResultsWe found systematic relationships between big five traits and human values that varied across contexts. Overall, correlations between Openness traits and the Conservation value dimension and Agreeableness traits and the Transcendence value dimension were strongest across all samples. Correlations between values and all personality traits (except Extraversion) were weaker in contexts with greater financial, ecological and social threats. In contrast, stronger personality-value links are typically found in contexts with low financial and ecological threats and more democratic institutions and permissive social context. These effects explained on average more than 10% of the variability in value-personality correlations.Conclusions Our results provide strong support for systematic linkages between personality and broad value dimensions, but also point out that these relations are shaped by contextual factors.