Content uploaded by Jeremy D.W. Clifton
Author content
All content in this area was uploaded by Jeremy D.W. Clifton on Jul 04, 2021
Content may be subject to copyright.
Psychological Assessment
Brief Measures of the Four Highest-Order Primal World Beliefs
Jeremy D. W. Clifton and David B. Yaden
Online First Publication, July 1, 2021. http://dx.doi.org/10.1037/pas0001055
CITATION
Clifton, J. D. W., & Yaden, D. B. (2021, July 1). Brief Measures of the Four Highest-Order Primal World Beliefs. Psychological
Assessment. Advance online publication. http://dx.doi.org/10.1037/pas0001055
BRIEF REPORT
Brief Measures of the Four Highest-Order Primal World Beliefs
Jeremy D. W. Clifton
1
and David B. Yaden
1, 2
1
Department of Psychology, University of Pennsylvania
2
Department of Psychiatry and Behavioral Sciences, Johns Hopkins University School of Medicine
Primal world beliefs (“primals”) are beliefs about the basic character of the world (e.g., “the world is an
abundant place”). The first effort to empirically map primals identified over two dozen such beliefs. The four
highest-order beliefs––the overall belief that the world is Go od (vs. bad), followed by Good’s three dimensions
of Safe (vs. dangerous), Enticing (vs. dull), and Alive (vs. mechanistic)—were novel and strongly correlated to
many theoretically relevant outcomes such as depression. However, measuring these four beliefs currently
requires administeringthe 99-item Primals Inventory (PI-99)and computing lengthy subscales (71, 29,28, and
14 items). This article validates briefer measures. Study 1 (N=459) and Study 2 (N=5,171) examines the
dimensionality, internal reliability, and test–retest reliability of scores on an 18-item measure of Good,Safe,
Enticing,andAlive (PI-18). Study 3 (N=3,947) does the same for a briefer 6-item measure of overall Good
world belief (PI-6). Study 4 (N=5,794) compares both versionsto the PI-99 (thegold standard) and 14 of its
correlates, including depression and life satisfaction. We conclude by recommending the PI-6 and PI-18
for most research and clinical uses and note that correspondence of three parallel forms implies not only
scale accuracy but also robustness of the latent phenomena.
Public Significance Statement
Scientists have found that people vary on four main beliefs about the world that are highly correlated to
personality and mental health: The beliefs that the world is Good,Safe,Enticing, and Alive. But
currently, the only way to measure these beliefs is with a long 99-question survey. This article helps
researchers and clinicians by creating two shorter measures of these important beliefs.
Keywords: measurement, Primals Inventory, primal world beliefs, short-form scale, Good world belief
Supplemental materials: https://doi.org/10.1037/pas0001055.supp
Rapunzel: Why can’t I go outside?
Witch: The outside world is a dangerous place.
People frequently make extremely broad statements about the
world’s basic qualities, such as “The world is a shithole, overflowing
with garbage and disease”(2010 tweet) and “The world is beautiful,
youjusthavetolookaround”(2021 tweet). Some statements portray
world beliefs as developing, (e.g., a 2021 tweet “The older I’ve gotten
the more I’ve realized the world is dangerous, you gotta be safe and
lookout for yourself”) while others aim to manipulate world beliefs to
gain advantage (e.g., the fictional witch teaching Rapunzel the world
is dangerous to make her afraid to leave her tower). While many of
these statements are likely mere expressions, might they sometimes
point to something deeper? This article aims to validate brief ways of
measuring some of these primitive-sounding world beliefs.
Such world beliefs are understudied. Beck (e.g., 1979) organized
depression-inducing beliefs into beliefs about the self, the self’s
future, and the self’s world (i.e., the Cognitive Triad). But in
practice, consistent with Beck’s intention (personal communication,
March 1st, 2019), world here concerns people in one’s immediate
social environment (e.g., “My boss hates me”). Traumatologist
Janoff-Bulman (1989) suggested humans have hyper-globalized
world schemas that influence how ambiguity is interpreted across
domains. But the handful of beliefs Janoff-Bulman identified a priori
are conceptually similar and difficult to distinguish empirically
(e.g., Kaler, 2009). In a seminal review of the worldview literature,
Koltko-Rivera (2004) discusses dozens of beliefs about freewill,
God, and so forth, but only one such primitive-sounding belief
concerning overall “world nature,”called belief in a just world
(BJW) or just world belief.
By far the most-studied world belief, BJW is the view that the world
is a “karmic”place where individuals get what they deserve and
deserve what they get (Montada & Lerner, 1998;Nesbit et al., 2012).
This belief is thought to have a cascading causal influence across
personality and well-being domains (Bartholomaeus & Strelan, 2019).
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Jeremy D. W. Clifton https://orcid.org/0000-0003-3185-3105
David B. Yaden https://orcid.org/0000-0002-9604-6227
No conflict of interest to disclose.
Correspondence concerning this article should be addressed to Jeremy
D. W. Clifton, Department of Psychology, University of Pennsylvania,
Philadelphia, PA 19104, United States. Email: cliftonj@sas.upenn.edu
Psychological Assessment
© 2021 American Psychological Association
ISSN: 1040-3590 https://doi.org/10.1037/pas0001055
1
Correlational (and some experimental) research from a few hundred
studies tie high BJW to (a) working harder, presumably because the
world is expected to reward hard work (e.g., higher-gross domestic
product, Furnham, 1993; higher grades, Dalbert & Stoeber, 2005);
(b) being nicer, presumably because the world rewards kindness and
playing by the rules (e.g., Correia & Dalbert, 2008); (c) being
happier and more successful, presumably because they worked
harder, were nicer, just contexts are pleasant; and just perceptions
provide post hoc rationalizations for success (e.g., increased life
satisfaction Otto et al., 2009); and (d) blaming the unfortunate such
as the sick, presumably because the world punishes fairly (e.g.,
Sakallı-Uğurlu et al., 2007). Could other world beliefs exert a
similarly broad influence? For example, maybe some optimists
are simply those that happen to think the world is objectively a
good place. Maybe neuroticism is partly driven by the belief the
world is dangerous.
Clifton et al. (2019) made the first broad-based effort to empiri-
cally map world beliefs. They labeled them primal world beliefs
(“primals”) to distinguish simple, adjectival, goal-relevant world
beliefs (e.g., “the world is dangerous”) from metaphysical, inciden-
tal, or historical world beliefs (e.g., “the world is composed of 118
chemical elements”). Like other descriptive taxonomic efforts (e.g.,
identifying Big Five personality traits), work was pursued with no
particular dependent variables in mind or strong dimensionality
expectations. Ten initial projects sought to identify all major
candidate primal world beliefs. Example projects included the
analysis of over 80,000 tweets beginning with phrases like “the
world is”; the 840 most-frequently used adjectives derived from
190,000 texts (450 million words); and over 1,700 descriptions of
the world gleaned from 385 of history’s most influential texts,
including: philosophical treatises, religious scriptures, novels, polit-
ical speeches, and films. This led to the identification of 234 items
representing a reasonably exhaustive list of candidate primals which
were subjected to three rounds of exploratory and confirmatory
factor analyses.
Dimensionality reduction analyses identified 26 hierarchically
arranged dimensions (Figure 1): 22 dimensions at the bottom (tertiary
primals, including just world belief), 17 of which form three clusters
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Figure 1
Definitions and Structure of Primal World Beliefs
Note. 26 primals (22 tertiary, 3 secondary, and 1 primal) were identified by Clifton et al. (2019);figure from Clifton & Kim, 2020).
2CLIFTON AND YADEN
(secondary primals, informally called “Big Three”), which in turn
form a supercluster (the primary primal). The primary primal is the
overall belief that the world is a Good (vs. bad) place. The three
secondary primals—essentially the main reasons to see the world as
more or less good—are Safe (vs. dangerous), Enticing (vs. dull), and
Alive (vs. mechanistic)—with Alive being less central. All are nor-
mally distributed. Five test–retests now suggest high stability over
time (even during the COVID19 pandemic, Ludwig et al., 2021).
Arguably the only two primals that have received previous research
attention are Just and Progressing, both under the Safe cluster.
Promising research directions have emerged from mapping these
beliefs. Based on correlations with theoretically related outcomes
such as depression, suicide ideation, curiosity, neuroticism, trust,
and optimism, the four higher-order primals (Good,Safe,Enticing,
and Alive) appear most clinically relevant (e.g., optimism and Good
correlated at r=.67). Yet currently the only way to measure these
four beliefs is to administer very lengthy subscales from the 99-item
Primals Inventory (PI-99) of 71, 29, 28, and 14 items, respectively
(Clifton et al., 2019). This lengthiness stymies researchers, clin-
icians, and the general public when measuring the very primals most
worth measuring. Shorter measures are needed.
The goal of this brief report is to validate two short Primals
Inventories to measure higher-order primals Good,Safe,Enticing,
and Alive.Thefirst (PI-6) is a unidimensional 6-item measure of Good
for use when the highest-order primal is the chief interest and space is
limited. The second (PI-18) is a multidimensional 18-item measure of
secondary primals Safe,Enticing,andAlive where pooled items also
measure higher-order Good world belief, as done in the original PI-99
(Clifton et al., 2019) and common among measures of higher- and
lower-order factors (e.g., CES-D, Radloff, 1977; PERMA Profiler,
Kern & Butler, 2016; NEO-PI-R, Costa & McCrae, 2008). To create
these two scales, Study 1 and 2 examines the dimensionality, internal
reliability, and test–retest reliability of PI-18 scores; Study 3 does the
same for the PI-6; and Study 4 compares both to the PI-99 as the
gold standard through direct score comparison and comparison of
correlational relationships to 14 mental health and demographic
variables in the wider nomological net (e.g., depression).
Our scale-building philosophy for this effort has been called
construction-minded scale-building (Clifton, 2020b). This means
that instead of exploring the shape of latent phenomena, we assume
it is already mapped (by the PI-99) and we must merely construct a
mimicked measured variable with fewer items. A secondary goal
was to retain a few items that tap secondary or tertiary primals for
practical research purposes. For example, if a survey only has room
for the PI-6, it would be helpful if the PI-6 included an item that is
both an adequate indicator of Good and Enticing, so item-level
analysis can suggest relevant secondary primals. Though useful for
research and clinical follow-up, and though this variance should
mostly cancel out in pooled scale scores, intentionally retaining
systematic signal (across subjects not items) not included in the
measurement model will also slightly worsen fit statistics, rendering
confirmatory factor analysis (CFA) indices difficult to interpret. For
this reason, studies below report but do not interpret CFA fit
statistics, leaving the reader to judge how close is close enough.
Otherwise, item-selection decisions mirrored PI-99 decisions. We
selected items from the initial pool of 234 items (Clifton et al., 2019;
Study 1; N=930; pp. 210–217 of their supplement) based on factor
loadings balanced against four additional considerations: language
variety, item response characteristics, retaining reverse-scored
items, and key scale-specific concerns. As for the PI-99, retaining
at least one opposite-scored item was considered essential for
primals measurement, allowing continuum specification (Tay &
Jebb, 2018). The major scale-specific concern was skew; only
affecting Enticing. Retaining top-loading items from a large item
pool resulted in (a) retaining some items in the PI-18 and PI-6 not in
the PI-99, where the target had been tertiary variance, and (b) not
including all PI-18 items in how Good is calculated, as in the PI-99.
See Supplemental Material for methodological details.
Study 1
Starting with the PI-18, Study 1 seeks to confirm the adequacy of
the intended measurement model (three separate subscales for Safe,
Enticing, and Alive and 1 pooled scale for Good) and more generally
test how items behave when greatly reducing items administered.
Method
Of 459 Americans recruited via mTurk in June 2016 (M
age
=
36.5 years, SD
age
=11.7 years), 63% were female, 81% white,
40% currently married, and 67% college graduates. Three subjects
had missing demographic information. We administered the five-
item Satisfaction With Life Scale (SWLS; Diener et al., 1985)asan
early convergent validity check, comparing relationships to previ-
ous mTurk samples completing the PI-99 (Clifton et al., 2019).
Internal reliability was assessed with standardized Cronbach’sα.To
test the measurement model, we conducted scree plot analysis,
minimum average partial analysis, parallel analysis, exploratory
factor analysis, confirmatory factor analysis, and reliability analysis.
All studies received IRB approval.
Results and Discussion
The intended measurement model was the best way to interpret the
data (pp. 5–10 of Supplemental Material). Only three factors ex-
plained >1 eigenvalue. Scree analysis suggested 3–4 factors. Mini-
mum average partial analysis suggested three factors. Parallel analysis
suggested four maximum. A 3-factor exploratory factor solution
(PROMAX rotation, EQUAMAX prerotation, k=3) produced sim-
ple structure (salience =.295), explaining just over 100% of item
variance (possiblewhen examining squared multiple correlations).An
inferior four-factor solution produced factors reflecting Safe,Enticing,
and Alive, and a fourth less meaningful factor involving six items (four
multiloading) comparatively more associated with Good world belief
(e.g., the top-loading item on the fourth factor was “Most things in the
world are good”at .59). CFA indicated the intended model (CFI =
.96, RMSEA =.06) was superior to a model measuring only Safe,
Enticing,andAlive (i.e., excluding higher-order Good;CFI=.83,
RMSEA =.11). In both models, all parameters were significant per
Wald tests. A one-factor exploratory factor solution had one non-
loading item (“The universe doesn’t care :::”). Compared to previ-
ous mTurk PI-99 studies (Clifton et al., 2019), primals correlated
similarly with sex, age, and life satisfaction, but Enticing items were
more skewed. Internal reliability was adequate at .78 (Safe), .80
(Enticing), .79 (Alive), and .81 (Good). Two items damaged the
reliability of subscale scores (“It takes a lot for things to fall apart”
and “The universe doesn’t care ::: ”) and also loaded lowest on their
respective factors (.45 and .54). The highest item-total correlation
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
BRIEF MEASURES OF HIGHEST-ORDER PRIMALS 3
across subscales was .71. Five item-total correlations <.40 were:
“It takes a lot for things to fall apart”on Safe and Good,and“The
world is a somewhat dull ::: ,”“It feels like dangers ::: ,”and
“The universe doesn’t care ::: ”on Good, yet these items marginally
damaged reliability.
In sum, Study 1 showed that the intended measurement model
was the best way of understanding item variance: there is an overall
factor and three subfactors. However, not all items performed as
expected in the new item context and there was room for improved
internal reliability. Thus, we chose to exchange a handful of under-
performing items for, in most cases, two items with one aimed at
capturing a third most-related tertiary primal.
Study 2
Study 2 reexamines in a larger sample the dimensionality, internal
reliability, and test–retest reliability of scores on the finalized PI-18.
Method
AuthenticHappiness.Org allows members of the public to partic-
ipate in studies and see their scores. Of 5,171 unique respondents
from Feb. 2019 to Dec. 2020, 62% were male, 56% younger than 35,
54% college graduates, and 103 countries were represented, with
most (69%) in the USA. Of these, 322 took the survey additional
times, allowing assessment of test–retest reliability via pairwise
correlation. Other analyses are same as Study 1.
Results and Discussion
The PI-18 behaved similarly as it did in Study 1, with some
improvements (pp. 11–14 of Supplement Material). Again, only
three factors explained >1 eigenvalue, scree analysis suggested 3–4
factors, minimum average partial analysis suggested three, parallel
analysis suggested four max. A three-factor exploratory factor
solution (PROMAX rotation, EQUAMAX prerotation, k=3) pro-
duced simple structure (salience =.295) save a single multiloader
(“Most things have a habit of getting worse”) loading on Safe (.35)
and Enticing (.40). Three factors again explained just over 100% of
item variance. An improbable four-factor solution was considered
anyway, producing an inadequate structure reflecting Alive,Safe,
reverse-scored Enticing, and forward-scored Enticing, the latter two
sharing five multiloaders. CFA suggested the intended model
(CFI =.90, RMSEA =.08) was superior to the model excluding
Good (CFI =.80, RMSEA =.11). All parameters were significant
per Wald tests and loaded on a one-factor exploratory factor
solution (salience =.295). Internal reliability of scores rose for
all subscales, to α=.88 for Good,toα=.83 for Safe,toα=.83
for Enticing, and to α=.85 for Alive. The only item with an item-
total correlation <.40 was also the only item that damaged—albeit
negligibly—any subscales’internal reliability of scores (“While
some things are worth checking out ::: ”), yet it was still the third
highest-loading item on its intended dimension (.65; Enticing).
The highest item-total correlation across subscales was r=.75
(“Everything happens for a reason and on purpose”on Alive).
Test–retest correlations among 322 subjects taking the survey on
average 14.3 days apart ranged from r=.90 to r=.89. However,
because 74% took surveys the same day, we reran analysis on the
26% (n=84) who took surveys at least a day apart (mean 55 days
apart, median 37 days apart), which was r=.76. In sum, in a
sample over 10 times larger than Study 1, the intended measure-
ment model was confirmed, internal reliability increased, and test–
retest reliability was found.
Study 3
Study 3 involves the same analysis in Study 2, but this time
examining the PI-6.
Method
The PI-6 was also placed on AuthenticHappiness.Org. Of 3,957
unique respondents, 61% were male, 54% younger than 35, 55%
college graduates, and in 93 countries, with most (63%) in the USA.
Of these, 179 took the PI-6 additional times, allowing test–retest
reliability assessment. All analyses are same as Study 2.
Results and Discussion
The intended one-factor model for the PI-6 performed best (pp.
15–17 of Supplement Material). Only one factor explained >1
eigenvalue, scree analysis suggested one factor, minimum average
partial analysis suggested one, and parallel analysis suggested two
maximum. A one-factor exploratory factor solution produced simple
structure, even at salience =.50, and explained just over 100% of
item variance. A two-factor solution was considered, but produced
less meaningful factors of entirely forward- and reverse-scored
items. CFA results were CFI =.91 and RMSEA =.17. All param-
eters were significant per Wald tests. Internal reliability was α=.86
with all items contributing. Item-total correlations ranged from .52
to .71. Test–retest reliability was r=.88 among 179 subjects who
took the survey on average 20 days apart and r=.78 among the 53
who took the surveys on different days on average 68 days apart. In
sum, Study 3 confirmed the unidimensional model, with good
internal reliability and test–retest reliability.
Study 4
Study 4 compares the PI-6 and PI-18 to each other and the PI-99
as the gold standard.
Method
The PI-99 was also placed on AuthenticHappiness.Org. Of 5,794
unique respondents, 63% were male, 50% younger than 35, 61%
college graduates, and in 105 countries, with most (69%) in the USA.
Sizeable subsamples ranging from 1,848 to 5,793 (Tables S16)took
multiple versions of the Primals Inventory or nonprimal measures.
We compared scores in three ways. First, we correlated primals to
themselves across versions (e.g., comparing PI-6 Good to PI-99
Good among the 1,848 subjects who took both). Second, we com-
pared correlations to the 22 tertiary primals not measured by the PI-6
or PI-18 (e.g., PI-6 Good and PI-99 Good correlations to PI-99
Hierarchical). Third, we compared correlations to 14 demographic
and mental health variables in the wider nomological net. Single items
measured age, gender, and education. Depression was measured
by Radloff (1977) 20-item Center for Epidemiological Studies-
Depression (CES-D) scale, which concerns experiences over the
past week, uses a 4-point likert scale, and is validated for nonclinical
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
4CLIFTON AND YADEN
samples (example item: “I had crying spells”). Life satisfaction was
again measured by the SWLS. The remaining variables were mea-
sured by Butler and Kern’s (2016) PERMA Profiler. The PERMA
Profiler’s overall psychological flourishing score pools five equally
weighted criteria: positive emotion, engagement, relationships, mean-
ing, and accomplishment. These five subscales are complemented by
additional subscales for health, and negative emotion, which are all
three items each. Loneliness is one item. An example health item is:
“Compared to others of your same age and sex, how is your health?”
All involve 11 response options (e.g., terrible to excellent).
Results and Discussion
To be a valid measure of the same construct, short-form scales
must strongly correlate with parent scales (criterion validity) and
mimic the parent’s correlational relationships with other variables
(convergent and divergent validity). Study 4 found strong evidence
of both for both new scales. In samples nearing or above 2,000
subjects, Good,Safe,Enticing, and Alive correlated with themselves
across versions between .82 and .88 (Table 1) and mimicked each
other in relation to 22 other primals (Tables S17, S18, and S19 of
Supplement Material) and 14 other variables in the broader nomo-
logical net (Table 2;Tables S20, S21, S22, S23, and S24). Correla-
tions to these 14 other variables were only |r
mean
|=.03 different and
never (of 84 possibilities) more than |r|=.09 different. For exam-
ple, PI-18 Alive and PI-99 Alive correlated identically with flourish-
ing (both at r=.44) and life satisfaction (both at r=.38). This
convergence was notable given comparisons used subsamples with
at least 50% different subjects, scale administration was not con-
current, and scales were administered in different orders.
General Discussion
Cognitive behavior therapy—arguably the most evidence-based
form of psychotherapy—was designed “based on the underlying
theoretical rationale that an individual’s affect and behavior are
largely determined by the way in which he structures his world”
through various often primitive-sounding beliefs (Beck et al., 1979,
p. 3). By impacting the interpretations of ambiguity across situa-
tions, Just world belief has been thought for decades to have a
cascading influence on clinical outcomes like depression and behav-
ior patterns such as perseverance (e.g., Bartholomaeus & Strelan,
2019). Yet Clifton et al.’s (2019) effort to map world beliefs found
Just to be unexceptional among 26 world beliefs, being one of
seven facets of Safe world belief, itself a facet of Good world
belief. Moreover, after examining pairwise relationships with over
100 personality, clinical, well-being, religious, political, and demo-
graphic variables in the broader nomological net, many primals
(especially higher-order primals Good,Safe,Enticing, and Alive)
were found to be more highly correlated with theoretically influ-
enced outcomes—even seemingly Just-specific outcomes like per-
severance. Despite the obvious need for further study on the four
neglected higher-order primals (though not entirely unstudied, e.g.,
Janoff-Bulman, 1989), efforts have been hindered by scale length.
Indeed, the PI-99’sGood subscale, at 71 items long, is ten times
longer than the average scale in psychology (Cortina et al., 2020).
This article validated two shorter measures that do not trade much
accuracy for brevity. This was possible because the original PI-99’s
length was driven by the need to measure 22 tertiary primals.
Numerous items were included in higher-order subscales because
they were available, but above studies confirmonlyafractionare
actually needed. Studies 1 and 2 established the internal reliability,
test–retest reliability, and dimensionality of scores on an 18-item
Primals Inventory (PI-18) measuring Good,Safe,Enticing,andAlive.
Study 3 did likewise for a unidimensional 6-item measure of Good
(PI-6). Study 4 then compared new versions to the PI-99, establishing
criterion, concurrent, divergent, and convergent validity. A strength
of our approach was parallel validation of multiple scale versions in
sizeable samples, allowing direct comparison of scores.
A limitation was that test–retest reliability samples were relatively
small. Because results are consistent with other PI-18 test–retests
(Ludwig et al., 2021), additional PI-6 test–retests would be helpful.
Another concern is marginal CFA fit statistics (CFI twice approached
.95 and once exceeded it). As noted, some decline was expected
because systematic item-level variance was purposefully retained but
not included in the measurement model. Tables S23, S24, S25, and
S26 of supplement indicate these retained items are indeed capable of
indicating more granular primals, but marginal CFIs nonetheless limit
the PI-6 and PI-18 compared to the PI-99. Given high internal
reliability and convergence across scale versions, we see this as a
reasonable trade-off, allowing more information to be gathered with
fewer items. Another limitation was fixed order administration (in
Studies 2–4). Split-half reliability, invariance, and performance in
pen-and-pencil contexts remain unexamined.
In the process of validating measures, some light was also shed on
the latent phenomena itself. Five years ago, it was not apparent if
primals existed or involved robust signal—perhaps too much was
being made of hyperbolic, affect-driven, state-like overgeneralizations
(such as those in the opening paragraph). Clifton et al.’s(2019)
validation of the PI-99 established the robustness of tertiary primals
especially, but legitimate questions remained about the newly identi-
fied higher-order primals, especially the 71-item Good subscale,
which concerns not only the most general of environments (the
world) but also the most general of evaluations (goodness)—and α
is known to become a useless indicator for scales half as long
(Cortina, 1993). Maybe tertiary primals are robust and can be
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Table 1
Correlating Primals to Themselves as Measured by the PI-6, PI-18, and PI-99
PI-6 PI-18
Variable Good Good Safe Enticing Alive
n1,848 1,947 1,947 1,947 1,947
Same belief measured by the PI-99 .82*.88*.86*.83*.84*
Note. Likewise, PI-18 Good and PI-6 Good correlated at r=.86 (n=2,369).
*p<.0001.
BRIEF MEASURES OF HIGHEST-ORDER PRIMALS 5
measured with 4–5 items sprinkled across 95 other items, but Good
world belief might express a mere artifact of dispositional optimism.
However, the above close correspondence of three alternative Good
subscales (i.e., parallel form reliability), increases confidence that
primals exist, are measurable, and are being accurately measured—
Good especially. The six-item PI-6 correlated with the PI-99’s71-
item subscale at r=.82 and correlated similarly to all 14 mental
health and demographic variables (rvalues were never >.06 different;
M
diff.
=.03). Depression, for example, correlated with PI-99 Good at
r=−.54 and PI-6 Good almost the same at r=−.51. Some diver-
gence is also due to mere attenuation: more internally reliable scales
inevitably tend to correlate more with other variables and PI-99 Good
is more internally reliable and indeed slightly more correlated to 10of
the 14 variables. In short, the new six-item Good scale appears
essentially equivalent to one nearly 12 times longer.
Though self-report may be unavoidable when measuring primals,
future research could further determine the robustness of primals by
testing more dissimilar self-report approaches. For example, an
ultra-brief single-item scale with 11-reponse options anchored by
opposing adjectives (e.g., very good to very bad) may be sufficient
to capture overall Good world belief. Indeed, that this sort of
barebones approach is even plausible suggests much has been
learned in recent years about the robustness of primal world beliefs.
As primals measurement progresses, we recommend the PI-6 and
PI-18 for research and clinical use. For clinical purposes, the PI-6,
PI-18, and PI-99 remain publicly available at www.AuthenticHa
ppiness.org where individuals can identify their primals and how
they differ from the general population. For research uses, a guide to
administering the Primals Inventory and selecting among versions is
available at http://dx.doi.org/10.13140/RG.2.2.26716.72320. When
length allows, we recommend starting research efforts with the
PI-99 because intuitions have been wrong about which primals are
relevant (e.g., dangerous world belief appears uncorrelated to polit-
ical conservatism, Clifton, 2020a). When length precludes the
PI-99, the PI-18 is likely the best balance of brevity and
granularity. For many clinical purposes, however, the PI-6 may
be the only version many researchers ever need to use.
References
Bartholomaeus, J., & Strelan, P. (2019). The adaptive, approach-oriented
correlates of belief in a just world for the self: A review of the research.
Personality and Individual Differences,151(May), Article 109485.
https://doi.org/10.1016/j.paid.2019.06.028
Beck, A. T., Rush, A. J., Shaw, B. F., & Emery, G. (1979). Cognitive therapy
of depression. Guilford Press.
Butler, J., & Kern, M. L. (2016). The PERMA-Profiler: A brief multi-
dimensional measure of flourishing. International Journal of Wellbeing,
6(3), 1–48. https://doi.org/10.5502/ijw.v6i3.1
Clifton, J. D. W. (2020a). Leveraging the first comprehensive measure of
primal world beliefs to further discussions in political, developmental, and
positive psychology (Publication No. 28149825). [Doctoral Dissertation].
University of Pennsylvania, Proquest. https://doi.org/10.13140/RG.2.2
.12296.85760
Clifton, J. D. W. (2020b). Managing validity versus reliability trade-offs in
scale-building decisions. Psychological Methods,25(3), 259–270. https://
doi.org/10.1037/met0000236
Clifton, J. D. W., Baker, J. D., Park, C. L., Yaden, D. B., Clifton, A. B. W.,
Terni, P., Miller, J. L., Zeng, G., Giorgi, S., Schwartz, H. A., & Seligman,
M. E. P. (2019). Primal world beliefs. Psychological Assessment,31(1),
82–99. https://doi.org/10.1037/pas0000639
Clifton, J. D. W., & Kim, E. S. (2020). Healthy in a crummy world:
Implications of primal world beliefs for health psychology. Medical
Hypotheses,135, Article 109463. https://doi.org/10.1016/j.mehy.2019.109463
Correia, I., & Dalbert, C. (2008). School bullying: Belief in a personal just
world of bullies, victims, and defenders. European Psychologist,13,
248–254. https://doi.org/10.1027/1016-9040.13.4.248
Cortina, J. M. (1993). What is coefficient alpha? An examination of theory
and applications. Journal of Applied Psychology,78(1), 98–104. https://
doi.org/10.1037/0021-9010.78.1.98
Cortina, J. M., Sheng, Z., Keener, S. K., Keeler, K. R., Grubb, L. K., Schmitt,
N., Tonidandel, S., Summerville, K. M., Heggestad, E. D., & Banks, G. C.
(2020). From αto omega and beyond! A look at the past, present, and
(possible) future of psychometric soundness in the Journal of Applied
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Table 2
Primals Measured by the PI-6, PI-18, and PI-99 Correlate Similarly to 14 Mental Health and Demographic Variables in the Wider
Nomological Net
Good Safe Enticing Alive
Variable PI-99 PI-6 PI-18 PI-99 PI-18 PI-99 PI-18 PI-99 PI-18
Depression −.54*−.51*−.49*−.51*−.44*−.42*−.38*−.35*−.34*
Life Satisfaction .51*.46*.49*.45*.40*.44*.38*.38*.38*
Flourishing .57*.53*.56*.47*.41*.52*.49*.44*.44*
Positive Emotion .55*.54*.55*.47*.43*.49*.46*.41*.43*
Engagement .44*.39*.44*.33*.28*.46*.45*.27*.29*
Relationships .45*.44*.44*.38*.33*.39*.39*.33*.34*
Meaning .55*.50*.53*.44*.38*.50*.45*.49*.47*
Accomplishment .46*.43*.45*.38*.33*.41*.38*.37*.36*
Negative Emotion −.48*−.47*−.46*−.47*−.43*−.38*−.39*−.19*−.24*
Loneliness −.34*−.31*−.32*−.31*−.29*−.29*−.29*−.15*−.17*
Health .36*.36*.37*.34*.33*.29*.28*.22*.26*
Female −.03 −.01 −.08*−.03 −.00 −.09*−.10*−.10*−.11*
Age .25*.22*.30*.28*.34*.19*.26*−.08*−.07*
Education .19*.17*.26*.19*.29*.15*.24*−.09*−.09*
Note.nfor all correlations ranged from 2,063 to 5,793 (Table S16). All comparisons involve samples overlapping <50%, indicating robustness. Scale
administration was not concurrent, but on average weeks apart, which likely dampened Pearson’sr, but unlikely matters for comparability purposes since all
samples were similarly impacted.
*p<.0001.
6CLIFTON AND YADEN
Psychology.Journal of Applied Psychology,105(12), 1351–1381. https://
doi.org/10.1037/apl0000815
Costa, P. T., & McCrae, R. R. (2008). TheRevised NEO PersonalityInventory
(NEO-PI-R). In R. Atkinson, R. Atkinson, E. Smith, & D. Bem (Eds.),
Personality Theory and Assessment (pp. 179–198). Sage Publications.
Dalbert, C., & Stoeber, J. (2005). The belief in a just world and distress at
school. Social Psychology of Education,8(2), 123–135. https://doi.org/10
.1007/s11218-005-1835-2
Diener, E., Emmons, R. A., Larsen, R. J., & Griffin, S. (1985). The
satisfaction with life scale. Journal of Personality Assessment,49(1),
71–75. https://doi.org/10.1207/s15327752jpa4901_13
Furnham, A. (1993). Just world beliefs in twelve societies. The Journal of
Social Psychology,133(3), 317–329. https://doi.org/10.1080/00224545
.1993.9712149
Janoff-Bulman, R. (1989). Assumptive worlds and the stress of traumatic
events: Applications of the schema construct. Social Cognition,7(2), 113–
136. https://doi.org/10.1521/soco.1989.7.2.113
Kaler, M. E. (2009). The World Assumptions Questionnaire: Development of
a measure of the assumptive world. University of Minnesota. http://conse
rvancy.umn.edu/bitstream/55049/1/Kaler_umn_0130E_10613.pdf
Kern, M. L., & Butler, J. (2016). The PERMA-Profiler: A brief multidimen-
sional measure of flourishing. International Journal of Wellbeing,6(3),
1–48. https://doi.org/10.5502/ijw.v6i3.526
Koltko-Rivera, M. E. (2004). The psychology of worldviews. Review of
General Psychology,8(1), 3–58. https://doi.org/10.1037/1089-2680.8.1.3
Ludwig, V. U., Crone, D., Clifton, J. D. W., Rebele, R. W., Schor, J., & Platt,
M. L. (2021). Testing the resilience of primal world beliefs during the
COVID-19 pandemic [Manuscript submitted for publication]. Department
of Neuroscience, University of Pennsylvania.
Montada, L., & Lerner, M. J. (Eds.). (1998). Responses to victimizations and
belief in a just world. Plenum Press. https://doi.org/10.1007/978-1-4757-
6418-5
Nesbit, S. M., Blankenship, K. L., & Murray, R. A. (2012). The influence of
just-world beliefs on driving anger and aggressive driving intentions.
Aggressive Behavior,38(5), 389–402. https://doi.org/10.1002/ab.21439
Otto, K., Glaser, D., & Dalbert, C. (2009). Mental health, occupational trust,
and quality of working life: does belief in a just world matter? Journal of
Applied Social Psychology,39(6), 1288–1315. https://doi.org/10.1111/j
.1559-1816.2009.00482.x
Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for
research in the general population. Applied Psychological Measurement,
1(3), 385–401. https://doi.org/10.1177/014662167700100306
Sakallı-Uğurlu, N., Yalçın, Z. S., & Glick, P. (2007). Ambivalent sexism,
belief in a just world, and empathy as predictors of Turkish students’
attitudes toward rape victims. Sex Roles,57(11–12), 889–895. https://
doi.org/10.1007/s11199-007-9313-2
Tay, L., & Jebb, A. T. (2018). Establishing construct continua in construct
validation: The process of continuum specification. Advances in Methods
and Practices in Psychological Science, 375–388. Advance online publi-
cation. https://doi.org/10.1177/2515245918775707
Appendix
Final Versions of the PI-6 and PI-18
PI-6
Instructions (please bold as noted)
Below are very general statements about the world—not the
world we wish we lived in, but the actual world as it is now. Please
share your sense of agreement or disagreement. When in doubt, go
with what initially feels true of the real world. There are no wrong
answers. There’s no need to overthink.
Response Options (response score should not be visible to
respondents)
Strongly Agree (5), Agree (4), Slightly Agree (3), Slightly Dis-
agree (2), Disagree (1), Strongly Disagree (0)
Items in the Order Given in Study 3 and 4 (*indicates r-scored)
Most things in the world are good.
In life, there’s way more beauty than ugliness.
Most things have a habit of getting worse.*
On the whole, the world is an uncomfortable and unpleas-
ant place.*
Good things in the world outweigh the bad things.
Please mark this statement “slightly disagree.”(attention check
optional)
On the whole, the world is a bad place.*
PI-18
Instructions and Response Options
Same as above
Items in the Order Given in Study 2 and 4 (*=r-scored, G =
Good, S =Safe, E =Enticing, A =Alive)
In life, there’s way more beauty than ugliness.
GE
It often feels like events are happening in order to help me in
some way.
GA
I tend to see the world as pretty safe.
GS
What happens in the world is meant to happen.
A only
While some things are worth checking out or exploring further,
most things probably aren’t worth the effort.*
GE
Most things in life are kind of boring.*
GE
The world is an abundant place with tons and tons to offer.
GE
No matter where we are or what the topic might be, the world is
fascinating.
GE
The world is a somewhat dull place where plenty of things are not
that interesting.*
GE
On the whole, the world is a dangerous place.*
GS
Instead of being cooperative, the world is a cut-throat and
competitive place.*
GS
Events seem to lack any cosmic or bigger purpose.*
A only
Most things have a habit of getting worse.*
GS
The universe needs me for something important.
GA
Most things in the world are good.
GS
Please mark this statement “slightly disagree.”(attention check
optional)
Everything happens for a reason and on purpose.
A only
Most things and situations are harmless and totally safe.
GS
No matter where we are, incredible beauty is always around us.
GE
Detailed scale administration instructions are available here:
https://doi.org/10.13140/RG.2.2.26716.72320
Received February 9, 2021
Revision received June 7, 2021
Accepted June 10, 2021 ▪
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
BRIEF MEASURES OF HIGHEST-ORDER PRIMALS 7
