Income inequality is associated with heightened test anxiety and lower academic achievement: A cross-national study in 51 countries

Abstract

Background
Research on predictors of test anxiety has focused primarily on the role of psychological factors and the proximal environment. However, the role of the broader socio-ecological context, specifically, national income inequality, is seldom explored.

Aims
The present study aimed to test whether national income inequality is associated with greater test anxiety and whether test anxiety is associated with lower academic achievement.

Data
We analyzed data from the 2015 Program for International Student Assessment (PISA), drawing on responses from 389,215 students nested in 51 countries.

Methods
Multi-level structural equation modeling was used.

Results
Results indicated that students in more unequal countries experienced greater test anxiety and had lower levels of achievement. Test anxiety, in turn, was associated with lower academic achievement in reading, math, and science. However, test anxiety did not mediate the effects of income inequality on achievement nor did income inequality moderate the relationship between test anxiety and achievement.

Conclusion
Taken together, the results of this study demonstrate the importance of taking socio-ecological factors such as income inequality into account when examining anxiety and achievement in academic settings.

Full text

Income inequality is associated with heightened test anxiety and lower academic
achievement: A cross-national study in 51 countries
© 2023 by Ronnel King, Yuyang Cai, and Andrew Elliot is licensed under CC BY-NC 4.0. To
view a copy of this license, visit http://creativecommons.org/licenses/by-nc/4.0/
Citation:
King, R. B., Cai, Y., & Elliot, A. J. (2024). Income inequality is associated with heightened
test anxiety and lower academic achievement: A cross-national study in 51
countries. Learning and Instruction, 89, 101825.
Link:
https://www.sciencedirect.com/science/article/pii/S0959475223000944
Author's personal copy - R.B. King

Learning and Instruction 89 (2024) 101825
Available online 9 September 2023
0959-4752/© 2023 Elsevier Ltd. All rights reserved.
Income inequality is associated with heightened test anxiety and lower
academic achievement: A cross-national study in 51 countries
Ronnel B. King
a
, Yuyang Cai
b
,
*
, Andrew J. Elliot
c
a
Department of Curriculum and Instruction, The Chinese University of Hong Kong, Hong Kong, China
b
School of Languages & Centre for Language Education and Assessment Research (CLEAR), Shanghai University of International Business and Economics, China
c
Department of Psychology, University of Rochester, USA
ARTICLE INFO
Keywords:
Income inequality
Academic achievement
Test anxiety
School anxiety
Cross-national
ABSTRACT
Background: Research on predictors of test anxiety has focused primarily on the role of psychological factors and
the proximal environment. However, the role of the broader socio-ecological context, specically, national in-
come inequality, is seldom explored.
Aims: The present study aimed to test whether national income inequality is associated with greater test anxiety
and whether test anxiety is associated with lower academic achievement.
Data: We analyzed data from the 2015 Program for International Student Assessment (PISA), drawing on re-
sponses from 389,215 students nested in 51 countries.
Methods: Multi-level structural equation modeling was used.
Results: Results indicated that students in more unequal countries experienced greater test anxiety and had lower
levels of achievement. Test anxiety, in turn, was associated with lower academic achievement in reading, math,
and science. However, test anxiety did not mediate the effects of income inequality on achievement nor did
income inequality moderate the relationship between test anxiety and achievement.
Conclusion: Taken together, the results of this study demonstrate the importance of taking socio-ecological factors
such as income inequality into account when examining anxiety and achievement in academic settings.
1. Introduction
Test anxiety reects the extent to which students nd examinations
threatening (von der Embse et al., 2018). It is associated with a wide
range of maladaptive outcomes including increased risk for poor grades,
mental health problems, and difculties in learning (Segool et al., 2013;
Putwain, Gallard, et al., 2021, Putwain, Stockinger, et al., 2021; von der
Embse et al., 2018). An international report indicated that 59% of stu-
dents often worry about taking tests, 66% are anxious about getting poor
grades, and 55% are very anxious about a test even if they are well
prepared (OECD, 2015).
Research on test anxiety has a long history (Sarason & Mandler,
1952). Much of the research on predictors of test anxiety has focused on
internal psychological factors. For example, a meta-analysis by Hembree
(1988) focused on factors such as fear of negative evaluation, poor study
skills, low self-concept, and an inclination to assign blame to others. A
more recent meta-analysis by von der Embse et al. (2018) highlighted
the role of other internal psychological factors such as self-concept,
motivation, goals, and personality as predictors of test anxiety.
Increasingly, albeit still to a lesser degree, research has examined the
role of environmental factors in predicting text anxiety. Some studies
have linked test anxiety to perceived threats in the environment. For
example, Segool et al. (2013) found that students had higher test anxiety
when they had to take high-stakes tests compared to low-stakes tests.
Furthermore, when students perceive the learning task as important but
have low levels of self-efcacy, they are more likely to experience test
anxiety compared to when they perceived the task as less important (Nie
et al., 2011).
Other studies have documented the role played by other environ-
mental factors such as families and peers. For example, Peleg-Popko
(2002) found that students had lower levels of test anxiety when they
had positive family relationships. Other researchers found that positive
relationships with teachers (Hoferichter et al., 2014) and peers (Lei
et al., 2021) were associated with lower test anxiety.
Despite increasing attention to environmental factors, the extant
literature has focused mostly on the proximal environment. Recent
* Corresponding author. School of Languages & CLEAR, Shanghai University of International Business and Economics Shanghai, China.
E-mail address: sailor_cai@hotmail.com (Y. Cai).
Contents lists available at ScienceDirect
Learning and Instruction
journal homepage: www.elsevier.com/locate/learninstruc
https://doi.org/10.1016/j.learninstruc.2023.101825
Received 4 May 2022; Received in revised form 9 August 2023; Accepted 18 August 2023
Author's personal copy - R.B. King

Learning and Instruction 89 (2024) 101825
2
trends in socio-ecological psychology, however, have underscored the
importance of examining the broader social ecology, including the
economic environment (Oishi, 2014). The present study advances the
test anxiety literature by examining the role of a critical socio-ecological
variable, national income inequality (which we refer to as income
inequality for shorthand), as a predictor of test anxiety. More generally,
the aim of this study was to examine the associations among income
inequality, test anxiety, and achievement in a large cross-national
sample of adolescent students.
1.1. Income inequality and test anxiety
Income inequality pertains to disparities in income between the rich
and the poor and is recognized as one of the world’s most serious social
problems (Pickett & Wilkinson, 2015). It has been found to be associated
with less sustainable economic growth, lower civic engagement, more
health problems, and lower psychological well-being (Oishi, 2014;
Pickett & Wilkinson, 2015).
1
Studies have found that the effects of
inequality remain robust after controlling for income levels, suggesting
that its effects are distinct from absolute income. In the educational
context, inequality has most often been explored in relation to
achievement outcomes, and past studies have shown a negative associ-
ation between the two (Chiu, 2015; Condron, 2011). However, less
attention has been paid to affective outcomes such as test anxiety.
Although we are not aware of any previous empirical study that has
linked income inequality to test anxiety, income inequality “may serve
as a contextual stressor” (Jiang & Probst, 2017, p. 673) and make test
anxiety more prevalent. Indeed, studies have shown that income
inequality is perceived as threatening and stressful (Pickett & Wilkinson,
2007).
Indirect empirical evidence for the potential linkage between income
inequality and test anxiety can also be found in two interrelated yet
distinct strands of literature. The rst strand is from epidemiological
research. Epidemiological studies have shown that people living in areas
with higher levels of income inequality have worse mental health out-
comes, including higher levels of depression and anxiety. This pattern
has been found both within and across countries, and applies to both
those of lower and higher socioeconomic status (SES) (Du et al., 2019a,
2019b; Messias et al., 2011; Pickett & Wilkinson, 2015; Wilkinson &
Pickett, 2010, 2019; however, see also Ngamaba et al., 2018; Sommet
et al., 2022). Research has also shown that students who have high test
anxiety are more likely to suffer from mental health and socio-emotional
problems (Cassady et al., 2019; Owens et al., 2012).
The second strand of work comes from sociological literature which
focuses on income inequality and status anxiety (Layte & Whelan,
2014). In highly unequal societies, individuals can gain more material
and social resources by doing better than their peers (Layte, 2012).
Inequality also makes one’s position in the social hierarchy more
important and salient (Kraus et al., 2013). Thus, people become stressed
and anxious about their relative social position and fearful of being left
behind by their peers (Kraus et al., 2013). Although test anxiety is
distinct from status anxiety, these different forms of anxiety nevertheless
share a common conceptual dimension (Hill et al., 2016). Furthermore,
different forms of anxiety have typically been found to be positively
correlated to each other and share a common underlying core (Norton &
Paulus, 2017; Sharp et al., 2015). For example, Lowe et al. (2011) found
a correlation of r =.70 between test anxiety and general anxiety, while
Xie et al. (2019) found a positive correlation of r =0.64. Hence, previous
ndings on the role of income inequality in status anxiety might also be
potentially applicable to test anxiety.
Given the importance of academic achievement for upward mobility,
performing well in school might be highly important in unequal soci-
eties, making students even more anxious. Research has shown that
students become more anxious when reminded of the importance of
examinations and the dire consequences of doing poorly (Putwain &
Best, 2011). In unequal societies, students might be judged more harshly
for poorer performance. They are also more likely to have lower social
mobility in the future (Jerrim & Macmillan, 2015). This lower mobility
applies to everyone, including those from higher and lower-SES back-
grounds (Andrews & Leigh, 2009; Kuo & Kawachi, 2023). Given the
importance of education for upward mobility, school success is critical,
and this might increase students’ test anxiety.
Students in highly unequal societies might also develop a fear of
being left behind by their peers if they do not do well enough. This is
because unequal societies make status differences more salient (Layte,
2012). Indeed, studies have shown that higher inequality lowers sense of
belonging, fosters greater competitiveness, and makes students more
sensitive to interpersonal comparisons (King et al., 2022; Sommet et al.,
2018, 2022). Given these ndings, we posited the following hypothesis.
Hypothesis 1. Income inequality is positively associated with test
anxiety.
1.2. Test anxiety and achievement
Test anxiety has deleterious consequences for student achievement
and cognitive performance (Maloney et al., 2014). Numerous studies
have shown that greater test anxiety is related to lower school
achievement (von der Embse et al., 2018). Longitudinal studies have
also indicated that test anxiety negatively predicts performance, even
after controlling for prior cognitive ability or prior academic attainment
(Pekrun, 1992; Putwain et al., 2013, 2016).
Meta-analytic investigations also converged on the same ndings.
For example, Hembree (1988) found that test anxiety was negatively
correlated with performance on standardized achievement tests (r =
−0.29). A subsequent meta-analysis by Seipp (1991) found that test
anxiety was negatively correlated with academic performance (r =
−0.23). A meta-analytic study conducted by von der Embse et al. (2018)
synthesized ndings from 238 studies and found that the negative as-
sociation between test anxiety and achievement held across primary
school, middle school, secondary school, and college (rs ranging from
−0.16 to −0.27). Two recent meta-analyses further conrmed these
ndings, with the researchers nding a negative correlation between
test anxiety and academic achievement (rs ranging from −0.20 to
−0.23) (Caviola et al., 2022; Robson et al., 2023).
Cognitive interference theories posit that anxiety in evaluative con-
texts places a heavy burden on the cognitive system (e.g., working
memory) which interferes with key thinking processes (Caviola et al.,
2022; Maloney et al., 2014). Test anxiety disrupts the performance of
students because worrying uses up valuable cognitive resources. Hence,
less working memory is available for the task at hand (Owens et al.,
2008; 2012a; 2012b). Aside from having a detrimental effect on
achievement, test anxiety also predicts a host of other maladaptive
outcomes, including lower expectancies of success, motivation, and
well-being (Fr´
echette-Simard et al., 2022; Putwain & Symes, 2020; von
der Embse et al., 2018). These considerations lead to the second
hypothesis.
Hypothesis 2. Test anxiety is negatively associated with achievement.
Although most of the existing studies on test anxiety and achieve-
ment have looked at linear relationships, there is also a smaller body of
work arguing that the relationship might be curvilinear; perhaps mod-
erate levels of anxiety might have a positive motivating function. For
example, Eysenck and colleagues (2007) theorized that anxiety may not
1
This is the prevailing view in the literature. However, it should be noted
that there is emerging scholarship suggesting a more nuanced view of the
relationship between income inequality and psychological functioning (e.g.,
that the overall inuence of inequality is actually negligible in some instances
because it can evoke positive as well as negative processes; see Ngamaba et al.,
2018; Sommet et al., 2019).
R.B. King et al.
Author's personal copy - R.B. King

Learning and Instruction 89 (2024) 101825
3
lead to impaired performance if it facilitates the use of positive strategies
such as enhanced effort and cognitive processing. Indeed, a recent study
by Cassady and Finch (2020) documented a curvilinear relationship
between test anxiety and other important learning-related outcomes.
Hence, we also tested the possibility of a curvilinear association as part
of the supplementary analyses.
1.3. Income inequality and academic achievement
Income inequality might also be directly associated with achieve-
ment. One of the earliest studies on this topic was conducted by Pickett
and Wilkinson (2007) who found that higher income inequality was
associated with lower math achievement. A more recent study by Con-
dron (2011) revealed that income inequality was negatively associated
with achievement, despite controlling for country afuence. Both
studies, however, were conned to using the country as the unit of
analysis.
Studies that used multi-level approaches also found the same pattern.
For example, Chiu (2015) found that income inequality negatively
predicted academic achievement. In another study, researchers found
that primary school students’ reading achievement was negatively
associated with income inequality (Chiu & Chow, 2015).
The weight of the empirical evidence seems to favor the position that
income inequality is detrimental to achievement. In highly unequal
societies, students from disadvantaged backgrounds may nd them-
selves disengaged from schoolwork because they perceive that economic
success is out of reach. Students born into higher-income families have
greater opportunities for success not available to those from more
disadvantaged families (Browman et al., 2019). Individuals who become
aware of the state of inequality become less hopeful that they can
improve their circumstances. When this happens, students might
decrease their engagement in academic tasks which might lead to lower
levels of achievement (Browman et al., 2019). This leads us to the
following hypothesis.
Hypothesis 3. Income inequality is negatively associated with
achievement.
Aside from the direct effects of income inequality on achievement, it
is possible that income inequality’s detrimental effects on achievement
will be partially mediated by test anxiety. In previous research on
inequality and achievement outcomes, the most common mechanisms
that were investigated pertained to economic resources. For example, it
has been shown that unequal countries more often have schools with
scarce learning resources (Chiu, 2015). Furthermore, learning resources
in unequal societies are more often devoted to students in more afuent
schools (Chiu, 2015). In the present research we focused on test anxiety
as a mediator of the effects of income inequality on achievement.
Hypothesis 4. The effect of income inequality on achievement is
mediated by test anxiety.
1.4. Income inequality as a moderator
We also examined whether national income inequality, a country-
specic feature, could have an impact on the relationship between test
anxiety and academic achievement for an individual. This would
represent a cross-level interaction wherein a higher-order variable (i.e.,
national income inequality) changes the nature of the association among
lower-level variables (i.e., test anxiety and achievement for an individ-
ual). Although we are not aware of any empirical research examining
whether income inequality functions as a cross-level moderator of the
inuence of test anxiety on achievement outcomes, past studies have
shown that income inequality could moderate the relationships among
individual-level psychological constructs (e.g., Jiang & Probst, 2017).
It is possible that in more unequal contexts test anxiety’s deleterious
effects on achievement become even more harmful because the
educational stakes are higher. Income inequality makes disparities be-
tween the “haves” and “have nots” greater. Societies with high income
inequality have fewer good job opportunities for students after they
graduate, schools are less well-resourced, and students face stiffer
competition for fewer good opportunities. This might be especially true
for students from more disadvantaged backgrounds, who might suffer
more from lesser social mobility (Bartram, 2022; Kerney & Levine,
2016). In unequal societies, individuals who have high levels of test
anxiety not only contend with the threat of losing good opportunities in
school but also in later life, thereby exacerbating the high stakes nature
of schooling. This leads us to the following hypothesis.
Hypothesis 5. Income inequality strengthens the negative association
between test anxiety and achievement.
1.5. Accounting for alternative explanations
Though several past studies support the argument that income
inequality is generally harmful to learning-related outcomes, there is
also a smaller body of work showing that income inequality’s effects
disappear when other more critical factors such as country afuence or
socioeconomic status are accounted for. For example, a study by Con-
dron (2011) examined how income inequality was associated with
achievement outcomes. When income inequality was the only
country-level predictor, it was associated with achievement, but once
country afuence was added into the model the effect of income
inequality disappeared. Another study by Zagorski et al. (2014) found
that income inequality was negatively related to well-being. However,
once country afuence was included in the model, income inequality
became non-signicant.
Other studies have argued that individual-level factors such as so-
cioeconomic status take precedence over country-level factors such as
national income inequality. For example, a study by Sommet et al.
(2018) explored how income inequality was associated with happiness.
They found that income inequality was only harmful to happiness
among economically disadvantaged individuals but not among the more
economically advantaged populations. Although these studies did not
directly examine test anxiety and academic achievement, they indicate
that there exists a certain ambiguity in the role of income inequality and
the possibility that inequality’s effects might be more fragile than might
be expected from the existing literature (e.g., Ngamaba et al., 2018).
Hence, to provide more robust evidence of the role played by inequality,
we controlled for the effects of country afuence, socioeconomic status,
and gender.
1.6. The present study
The present research aims to test the links between national income
inequality and students’ test anxiety and achievement. In sum, ve main
hypotheses were tested.
•H1: Income inequality is positively associated with test anxiety.
•H2: Test anxiety is negatively associated with achievement.
•H3: Income inequality is negatively associated with achievement.
•H4: The effect of income inequality on achievement is mediated by test
anxiety.
•H5: Income inequality strengthens the negative association between test
anxiety and achievement.
To test these hypotheses, we analyzed data from the OECD Program
for International Student Assessment (PISA). We choose to work with
PISA 2015 because it is the latest and only PISA study that assesses test
anxiety in school.
R.B. King et al.
Author's personal copy - R.B. King

Learning and Instruction 89 (2024) 101825
4
2. Method
2.1. Data and measures
Our study used OECD PISA 2015 data (OECD, 2016) which included
responses from 389,215 15-year-old adolescent students (Mean age =
15.80, SD =0.29) from 51 countries,
2
each occupying 0.4%–8.2% of the
total sample. The gender ratio was nearly equal: males =193,818
(49.8%), females =195,317 (50.2%). These countries are shown in
Table 2.
Income inequality. The Gini index provided by the Standardized
World Income Inequality Database (Solt, 2016) was used to represent
national income inequality. The Gini index ranges from 0 (all people have
equal income) to 100 (one person has all of the income and others have none)
(Solt, 2016). The Gini values of the 51 countries for the year 2014 were
used; Mean =33.60 (SD =6.48).
Test anxiety. Test anxiety was measured using the index of school-
work anxiety provided in the 2015 PISA data set (OECD, 2017). This
index comprises 5 items and measures students’ cognitive and emotional
reactions to test taking using a 4-point scale from 1 (strongly disagree) to
4 (strongly agree). This scale was adapted and developed by OECD partly
based on previously published questionnaires measuring test anxiety (e.
g., Cassady & Johnson, 2002; Spielberger, 1980). A sample item is, “I
often worry that it will be difcult for me taking a test.” Cronbach’s
alpha for the anxiety scale was 0.82. We used the Rasch calibrated test
anxiety score supplied by OECD which has a mean of 0 and a standard
deviation of 1 (OECD, 2017).
Achievement. The PISA dataset contained achievement scores in
three subjects: reading, math, and science. OECD used the Rasch
modeling approach to estimate the PISA achievement scores (OECD,
2017). A Rasch model species that the probability with which an
examinee answers an item correctly depends on the difference between
the ability of the examinee and the difculty of the item (Bond & Fox,
2015). The particular Rasch model that OECD developed can be applied
to multiple populations by assuming one population for each partici-
pating country (OECD, 2017). The mean for reading was 484.79 (SD =
99.61), the mean for mathematics was 479.89 (SD =99.07), and the
mean for science was 485.08 (SD =99.82). The reliability of
Rasch-calibrated scores for reading, math, and science across groups
ranged from 0.80 to 0.85 (OECD, 2017).
Covariates. At the student level, gender (female =0 and male =1)
and SES were included as covariates. PISA uses the variable economic
and social cultural status to represent SES. The economic and social
cultural status variable contained information about students’ family
background such as the number of books at home and their parents’
education and occupation among others (OECD, 2016b). SES was a
standardized score and had a mean of −0.20 (SD =1.07). At the country
level, country afuence was indexed by the Gross Domestic Product
(GDP) per capita, which was log transformed (World Bank, 2018).
2.2. Data analysis
A multi-level approach was required given that students were nested
within countries. The data were analyzed using multi-level structural
equation modeling (SEM) in Mplus 8.2 (Muth´
en & Muth´
en, 1998–2018).
The multi-level SEM approach has considerable advantages over con-
ventional multi-level modeling procedures (e.g., hierarchical linear
modeling), as it allows the simultaneous estimation of complex models
with multiple mediators and/or outcome variables and both direct and
indirect effects (Preacher et al., 2016).
The present study tested a multi-level model, wherein income
inequality was measured at the country level (Level 3), while test anx-
iety and achievement were measured at the student level (Level 1). We
also accounted for the school-level effects (Level 2) but did not measure
specic school-level variables, as our hypotheses were focused on vari-
ables at the country and student levels.
The primary analyses proceeded in three steps moving from simpler
models to more complex models. In Model 1, we tested an unconditional
model to determine the appropriateness of using multi-level analyses. In
Model 2, we tested a multi-level SEM model that tested the linkages
among the focal variables. In Model 3, we added covariates to account
for alternative explanations and tested H1, H2, H3, and H4.
Model 4 added a random slope component to Model 3 and enabled us
to test H5. H5 involves a cross-level interaction which we tested using
multi-level SEM with random slopes. Multi-level SEM with random
slopes has two key assumptions: a dependent variable (e.g., reading,
mathematics, or science achievement) can be predicted by independent
variables at two levels (test anxiety at the student level; national income
inequality and country afuence at the country level). The effect of the
independent variable at the lower level depends on the value of the
independent variable(s) at the higher level. In the current study, we
tested whether a contextual characteristic (i.e., national income
inequality at Level 3) moderates the strength of a lower-level relation-
ship (i.e., the relationship between test anxiety and achievement at Level
1). Psychometricians have argued that multi-level models involving
cross-level interactions should include the random slope component, as
failure to do so may lead to t-ratios that are too high, condence in-
tervals that are too narrow, and standard errors and p values that are too
low (Heisig & Schaeffer, 2019). Hence, from a psychometric perspective,
using multi-level SEM with random slopes is the optimal approach for
testing cross-level interactions. The variables that were included at both
levels of analysis were group-mean centered.
Supplementary data analyses. We also conducted supplementary an-
alyses designed to further test the nature of the relationship between test
anxiety and academic achievement. Though not part of our main hy-
potheses, we also tested the possibility of a curvilinear relationship.
3. Results
3.1. Preliminary analyses
At the country level, income inequality was positively correlated
with test anxiety but negatively correlated with achievement in reading,
science, and math (Table 1). The scatterplots show how income
inequality is associated with test anxiety (Fig. 1) and achievement in
reading, math, and science (Fig. 2).
At the student level, test anxiety was negatively associated with
achievement in reading, math, and science. The overall correlations for
the whole sample are shown in Table 1 and the correlations within each
country can be found in Table 2.
3.2. Primary results
The results of the primary analyses are described below (see Fig. 3).
3.2.1. Model 1: unconditional model
We rst tested whether multi-level modeling was necessary by
calculating ICC (the contribution of between-group variance to the total
variance) for all outcome variables. The ICC of each outcome variable
was larger than zero: 0.13 for reading achievement, 0.20 for math
achievement, and 0.16 for science achievement. Researchers have sug-
gested that multi-level modeling is justied when ICC values are close to
or exceed 0.10 (Heck & Thomas, 2015, 2020; 2020).
3.2.2. Model 2: linkages among inequality, anxiety, and achievement
Next, we tested a multi-level SEM model that focused on testing the
linkages among income inequality, test anxiety, and academic
2
We use the term country for shorthand but note that some of the contexts
included in the PISA dataset are more appropriately classied as cities, juris-
dictions, or regions (e.g., Hong Kong).
R.B. King et al.
Author's personal copy - R.B. King

Learning and Instruction 89 (2024) 101825
5
achievement. We considered variance at the student level (level 1),
school level (level 2), and country level (level 3) though our focal var-
iables were only located at the student (i.e., test anxiety and academic
achievement) and country-levels (i.e., national income inequality). In-
come inequality positively predicted test anxiety (β =0.10, p <.001).
Test anxiety negatively predicted achievement in reading (β = − 0.08, p
<.001), math (β = − 0.16, p <.001), and science (β = − 0.16, p <.001).
Furthermore, income inequality negatively predicted achievement in
reading (β = − 0.09, p <.001), math (β = − 0.10, p <.001), and science
(β = − 0.09, p <.001). The diagram is presented in the Supplementary
Files (Fig. S1).
3.2.3. Model 3: inclusion of covariates
To conduct a more rigorous test of our hypotheses and account for
alternative explanations, we added covariates such as gender, socio-
economic status, and country afuence to Model 2 above. Results
showed that income inequality positively predicted test anxiety (β =
0.67, p <.001), supporting H1. Test anxiety negatively predicted
achievement in reading (β = − 0.08, p <.001), math (β = − 0.10, p <
.001), and science (β = − 0.10, p <.001), supporting H2. Furthermore,
inequality negatively predicted achievement in math (β = − 0.32, p <
.001), but its effects on reading (β = − 0.15, p =.25) and science (β =
−0.20, p =.13), were not statistically signicant. This meant that in
more unequal societies, test anxiety was higher, and higher test anxiety
was associated with lower achievement. Students in more unequal so-
cieties had lower math, but not reading and science scores. Thus, H3 was
only partly supported.
H4 was a mediational hypothesis, indicating that the effects of in-
come inequality on academic achievement would be mediated by test
anxiety. The indirect effects were non-signicant for all three domains of
reading (indirect effect = − 0.04, p =.55, 90% CI =[−0.18, 0.10]), math
(indirect effect: 0.01, p =.86, 90% CI =[−0.16, 0.13]), and science
(indirect effect: = − 0.01, p =.86, 90% CI =[−0.17, 0.14]), thereby
failing to support H4.
In terms of the covariates, SES negatively predicted test anxiety (β =
−0.04, p <.001) but positively predicted achievement in reading (β =
0.08, p <.001), math (β =0.08, p <.001), and science (β =0.08, p <
.001). This meant that students from more advantaged backgrounds had
lower test anxiety but higher levels of achievement. Males had lower test
anxiety (β = − 0.39, p <.001) and reading achievement (β = − 0.26, p <
.001), but higher math (β =0.07, p <.001) and science (β =0.03, p <
.001) achievement.
At the country level, country afuence was not a signicant predictor
of test anxiety (β =0.10, p =.39), but was a positive predictor of
achievement in reading (β =0.58, p <.001), math (β =0.44, p <.001),
and science (β =0.51, p <.001). This meant that richer countries had
students with higher levels of achievement in reading, math, and
science.
3.2.4. Model 4: income inequality as a moderator
To test whether inequality moderates the association between test
anxiety and achievement (H5), we added a random slope component to
Model 3 above. The random slope component freed the parameter es-
timate between test anxiety and achievement to vary across countries.
Income inequality and country afuence were designated as predictors
of the random slope component. A signicant effect of inequality on the
random slope could be taken as evidence of a cross-level interaction.
Results indicated that, after controlling for the effect of country
afuence, income inequality did not moderate the association between
test anxiety and achievement, failing to support H5. Inequality was not a
signicant predictor of the test anxiety to reading achievement slope (b
=0.03, p =.74), the test anxiety to math achievement slope (b =0.04, p
=.64), nor the test anxiety to science achievement slope (b =0.22, p =
.27).
3
These results indicated that the association between test anxiety
and achievement did not vary as a function of income inequality.
3.3. Supplementary analyses
Our supplementary analyses involved testing the curvilinear rela-
tionship between test anxiety and academic achievement through the
addition of a quadratic term. Results of the supplementary analyses
showed that there was a small curvilinear effect such that at relatively
low levels, test anxiety was associated with slightly higher levels of
reading, math, and science achievement. More details can be found in
the Supplementary Materials.
4. Discussion
The aim of this study was to examine how income inequality is
associated with test anxiety and achievement in adolescents, and
whether inequality moderates the relationship between income
Table 1
Correlations among the variables at the country and student levels.
Income Inequality Country Afuence Test Anxiety Reading Achievement Math Achievement Science Achievement
Income Inequality – −.581** .612** −.529** −.592** −.519**
Country Afuence – – −.291* .684** .631** .636**
Test anxiety – – – −.332* −.353** −.304*
Reading achievement – – −.094** – .925** .966**
Math achievement – – −.165** .808** – .960**
Science achievement – – −.151** .882** .893** –
SES – – −.114** .389** .404** .399**
Gender – – −.208** −.110** .059** .037**
Note: *p <.05; **p <.01. Correlations above the diagonal are at the country-level; those below the diagonal are at the student-level.
Fig. 1. Relationship between income inequality and test anxiety at the country-
level. Note. ZGINI =income inequality standardized; ZTanx =test anxiety
standardized. Numbers represent countries.
3
Only unstandardized estimates are provided when testing random slopes in
Mplus.
R.B. King et al.
Author's personal copy - R.B. King

Learning and Instruction 89 (2024) 101825
6
inequality and achievement. Our study yielded three key ndings: First,
income inequality was associated with greater test anxiety, supporting
H1. Second, test anxiety was associated with lower achievement, sup-
porting H2. Furthermore, income inequality negatively predicted
achievement in math but not in reading and science, providing only
partial support to H3. However, the effect of income inequality on ac-
ademic achievement was not mediated by test anxiety (H4), and income
inequality did not moderate the association between test anxiety and
achievement (H5). Hence, H4 and H5 were not supported. We turn to
the specic ndings below.
Our rst hypothesis — that income inequality is positively associated
with test anxiety —was supported. Our ndings are novel and make an
important contribution to the test anxiety literature. Although research
on test anxiety has a long history, much of this work has focused on the
effects of test anxiety and the psychological factors that predict it. More
recently, the role of the environment has been highlighted, including the
testing environment, parents, teachers, and peers (von der Embse et al.,
2018). Relatively little work has been done on socio-ecological pre-
dictors of test anxiety. Our study extends the existing literature by
highlighting the role of the broader socio-ecological environment, spe-
cically national income inequality, in predicting test anxiety.
Students are embedded within larger economic structures and in-
come inequality may shape their school experiences even before they
participate in the labor market. In unequal societies, it is possible that
the stakes associated with educational success may be much higher and
there might be a larger number of students competing for a few privi-
leged spots, where most of the rewards are concentrated. Hence, stu-
dents might experience higher levels of test anxiety in unequal societies,
as performing poorly might mean having fewer job opportunities and
lower social status. However, we acknowledge that income inequality as
a macro-environmental factor is relatively more distal. There are likely
different intermediary mechanisms that could link distal inequality to
anxiety and achievement. Hence, future studies may need to test the
potential mediating mechanisms.
Our research also makes an important contribution to the income
inequality literature by extending it to the adolescent population. Past
studies on income inequality and anxiety have primarily focused on
adult populations. Furthermore, these past studies examined other forms
of anxiety such as status anxiety. Our study demonstrated that income
inequality is also a relevant predictor of test anxiety among adolescent
students. This nding indicates that income inequality not only affects
adults who are participating in the labor market, but also adolescents
who are still in school (Elgar et al., 2017).
Our second hypothesis was that test anxiety is negatively associated
with reading, math, and science achievement. This hypothesis was also
supported. The negative association between test anxiety and achieve-
ment held after we included SES and gender as covariates. This nding is
in line with previous studies that have shown that anxiety interferes
with cognitive functioning leading to lower levels of performance
(Eysenck & Calvo, 1992; Eysenck et al., 2007; Gass & Curiel, 2011; Ng &
Lee, 2015; Schillinger et al., 2021). However, it is important to note that
the PISA test is relatively low stakes and that higher levels of test anxiety
occur when the examinations are perceived as important and high stakes
(Putwain & Best, 2011). Studies that explore test anxiety in the context
of high-stakes examinations (e.g., college entrance exams) might yield
more ecologically valid ndings.
Our supplementary analyses further indicated a small curvilinear
effect, such that at relatively low levels test anxiety was associated with
slightly higher achievement. This nding seems to corroborate past
studies noting the need to explore non-linear relationships in terms of
test anxiety, beyond just linear associations (Cassady & Finch, 2020).
However, we note that these curvilinear effects are relatively small in
magnitude and overall the relationship between test anxiety and
achievement was negative.
Our third hypothesis was that income inequality is associated with
lower achievement. This hypothesis was only partially supported. In
Model 2, where income inequality was the only country-level variable,
income inequality negatively predicted reading, math, and science
achievement. However, when country afuence was added as a covar-
iate to Model 3, income inequality was only a negative predictor of math
achievement. The effects on reading and science achievement, albeit in
the predicted negative direction, became non-signicant. Hence, the
relation between income inequality and math achievement appears to be
particularly robust; evidence for the role of income inequality on
Fig. 2. Relationship between income inequality and achievement in reading,
math, and science. Note. ZGINI =income inequality standardized; Zread =
reading achievement standardized; Zmath =math achievement standardized;
Zscie =science achievement standardized. Numbers represent countries. Please
refer to Table 2 for the country codes.
R.B. King et al.
Author's personal copy - R.B. King

Learning and Instruction 89 (2024) 101825
7
reading and science achievement is much less robust.
We found a strong positive association between income inequality
and country afuence (r = − 0.58), which corresponded to ndings from
past studies (King et al., 2022; Zagorski et al., 2014). In general, poorer
countries have higher levels of income inequality (Solt, 2016). It is
possible that the effects of income inequality were masked given the
overlap between country afuence and income inequality. Future
studies that examine a wider range of countries are needed, given that
we only included 51 countries in the current study. These countries do
not represent the full range of inequality across the globe, and our sta-
tistical analyses might have suffered from a restricted range.
It is also possible that income inequality might have stronger dele-
terious effects on performance in some subjects than others. A study by
Pickett and Wilkinson (2007) found that there was a negative correla-
tion between income inequality and math scores, but not with reading
and science scores. Perhaps math, given its cumulative nature, is more
sensitive to inequality’s effects. The math curriculum is more tightly
interlinked across different developmental stages, and failure to master
key mathematical concepts presented earlier in one’s school career (e.g.,
multiplication), might make it harder to understand more advanced
concepts presented later (e.g., algebra). However, for language and
science, the link between concepts presented earlier and those presented
later may not be so tight.
Our fourth hypothesis was that test anxiety partially mediates the
negative association between income inequality and achievement. The
results of the mediation analyses were not signicant, suggesting that
other mechanisms might be able to better account for the negative as-
sociation between income inequality and achievement than test anxiety.
Past studies have shown that economic mechanisms might explain how
income inequality is associated with lower achievement (Chiu, 2015;
Chiu & Khoo, 2005). In highly unequal societies, there are fewer
learning resources for everyone and there is a lack of proper school
Table 2
Income inequality, test anxiety, and achievement.
Correlation between anxiety and achievement
Country Income Inequality Log GDP Per Capita Test Anxiety Reading Mathematics Science
1. Australia 33.5 4.80 0.21 −.061** −.129** −.120**
2. Austria 27.7 4.71 −0.11 −.184** −.241** −.246**
3. Belgium 25.7 4.68 −0.18 −.072** −.136** −.121**
4. Brazil 44.8 4.08 0.61 .027** −.089** −.049**
5. Bulgaria 34.1 3.90 −0.09 0 −0.017 −.041**
6. Canada 30.9 4.71 0.17 −.075** −.184** −.164**
7. Chile 44.7 4.17 0.06 −.172** −.219** −.232**
8. Chinese Taipei 29.8 4.36 0.38 −.025* −.035** −.027*
9. Colombia 48.4 3.91 0.54 .022* −.046** −.062**
10. Costa Rica 45.8 4.04 0.61 −.085** −.079** −.115**
11. Croatia 28.4 4.14 0.01 −0.013 −.088** −.082**
12. Czech Republic 25.4 4.30 −0.22 −.049** −.109** −.131**
13. Denmark 25.9 4.80 0.10 −.126** −.220** −.186**
14. Dominican Republic 44.5 3.82 0.42 0.019 0.001 −0.006
15. Estonia 34.0 4.31 −0.20 −.139** −.195** −.205**
16. Finland 25.3 4.70 −0.41 −.146** −.239** −.241**
17. France 29.5 4.63 −0.09 −.064** −.097** −.099**
18. Germany 29.2 4.68 −0.34 −.167** −.207** −.214**
19. Greece 33.7 4.33 −0.09 −.067** −.130** −.134**
20. Hong Kong 41.0 4.61 0.33 −.046** −.093** −.074**
21. Hungary 27.9 4.16 −0.10 −.072** −.119** −.124**
22. Iceland 24.1 4.74 −0.11 −.129** −.274** −.260**
23. Ireland 30.1 4.74 0.15 −.128** −.198** −.188**
24. Israel 36.2 4.58 −0.24 −.033** −.117** −.091**
25. Italy 33.3 4.55 0.29 −.061** −.116** −.150**
26. Japan 30.4 4.59 0.26 .036** −0.003 0.024
27. Korea 30.3 4.47 0.11 .075** 0.02 .037**
28. Latvia 35.3 4.20 −0.13 −.112** −.176** −.159**
29. Lithuania 34.9 4.22 −0.07 0.002 −.095** −.075**
30. Luxembourg 28.4 5.09 −0.16 −.149** −.211** −.211**
31. Mexico 44.9 4.04 0.26 −.097** −.181** −.160**
32. Montenegro 31.3 3.87 0.09 −.045** −.096** −.088**
33. Netherlands 26.9 4.72 −0.54 0.002 −.034* −0.026
34. New Zealand 33.2 4.65 0.27 −.084** −.210** −.191**
35. Norway 25.2 4.99 0.07 −.038** −.131** −.160**
36. Peru 44.8 3.82 0.13 −.060** −.087** −.081**
37. Poland 30.7 4.15 −0.11 −.082** −.179** −.160**
38. Portugal 34.0 4.34 0.47 −.059** −.096** −.117**
39. Russian Federation 39.2 4.15 −0.05 −.088** −.116** −.144**
40. Singapore 38.9 4.76 0.59 −.119** −.157** −.154**
41. Slovak Republic 24.8 4.27 −0.17 −0.013 −.092** −.086**
42. Slovenia 25.3 4.38 0.04 −0.007 −.141** −.108**
43. Spain 34.4 4.47 0.40 −.091** −.178** −.153**
44. Sweden 26.3 4.78 0.05 −.095** −.192** −.170**
45. Switzerland 29.2 4.95 −0.40 −.091** −.161** −.163**
46. Turkey 40.4 4.08 0.32 −0.002 −.066** −.065**
47. United Kingdom 33.1 4.68 0.25 −.097** −.167** −.137**
48. United States 38.0 4.74 0.15 −.087** −.203** −.157**
49. Uruguay 36.4 4.23 0.46 −.125** −.181** −.184**
50. B-S-J-G (China) 40.2 3.88 0.24 −.100** −.122** −.131**
51. Spain (Regions) 34.4 4.47 0.40 −.061** −.145** −.131**
Average correlation −.094** −.165** −.151**
Note. ***p <.001, **p <.01, *p <.05.
R.B. King et al.
Author's personal copy - R.B. King

Learning and Instruction 89 (2024) 101825
8
infrastructure. Perhaps these economic mechanisms might be more
relevant in explaining why income inequality is associated with lower
achievement compared to test anxiety. Furthermore, the relationship
between test anxiety and academic achievement was much smaller at
the country-level than at the student-level, which could also explain the
non-signicant mediation.
We found that income inequality’s effects on achievement were not
mediated by test anxiety. However, it is also possible that a different test
of this hypothesis might yield supportive data; for example, the use of a
more local (e.g., neighborhood) indicator of income inequality may
provide a more powerful test that yields signicant results. For example,
individuals are more sensitive to local income inequality, such as the
inequality in their local neighborhood or their district, rather than
inequality of their country (Newman et al., 2018).
Our fth hypothesis was that income inequality moderates the effect
of anxiety on achievement. This hypothesis was not supported. We
found that inequality did not account for variation in how anxiety was
associated with achievement across different societies. It seems that the
intrapsychic experience of test anxiety is such a powerful inuence on
achievement that it exerts its deleterious effects over and above the
inuence of the macro-environment. Existing test anxiety research
supports the contention that anxiety harms learning and disrupts
cognitive processes across many different socio-cultural contexts (Cav-
iola et al., 2022; von der Embse et al., 2018).
We also comment on the effect sizes in our study. In terms of the
country-level correlations, the relationship between income inequality
and test anxiety was r =.60, while the correlation between income
inequality and achievement ranged from r = − 0.52 to −0.59. These
correlations closely match other ecological correlations in the existing
literature between country factors and educational outcomes. For
example, He and colleagues (2017) found that the relationship between
country afuence and PISA achievement ranged from .48 to .50, while
Pickett and Wilkinson (2007) found that the country-level correlation
between income inequality and academic achievement was r = − 0.41.
In terms of the student-level variables, the association between test
anxiety and achievement ranged from r = − 0.09 to 0.19. Updated
guidelines in effect size research consider 0.10, 0.20, and 0.30 as rela-
tively small, typical, and large (Gignac & Szodorai, 2016). Hence, the
effect sizes of test anxiety in this study could be considered as ranging
from small to typical. These effect sizes also match the typical effect sizes
found for other emotional and motivational factors in terms of pre-
dicting academic achievement (e.g., Camacho-Morles et al., 2021).
When interpreting these effect sizes, it is worth noting two key
points. First, our measure of test anxiety was domain-general, and it is
possible that larger effect sizes would be obtained if we had domain-
specic measures of test anxiety (e.g., mathematics test anxiety). Sec-
ond, the achievement data in this study was operationalized as students’
scores in the PISA tests. However, the PISA test is relatively low stakes.
Hence, future studies that examine test anxiety data during high-stakes
situations (e.g., students taking college entrance exams or taking their
nal exams) might likely yield larger effect sizes (e.g., Segool et al.,
2013).
We note the strong yet distinct effects of country afuence from in-
come inequality. Country afuence was a more robust predictor of ac-
ademic achievement compared to income inequality, with effect sizes
nearly twice as large as that associated with income inequality. How-
ever, only income inequality signicantly predicted test anxiety and
country afuence was not a signicant predictor. This shows that these
country-level variables might be associated with distinct sets of out-
comes. These ndings converge with the existing literature showing that
researchers need to attend to the roles of both afuence and inequality,
as both are important aspects of the social ecology (King, 2022; Oishi,
2014).
4.1. Practical implications
One of the practical implications of our research is the importance of
psycho-educational programs to reduce test anxiety. These programs
might even be more necessary in highly unequal societies where test
anxiety is likely to be more prevalent. Given the high number of students
who have been shown to have test anxiety across the globe, test anxiety
interventions could potentially benet large numbers of students.
4.2. Limitations, and directions for future research
In addition to its strengths, our study has limitations as well. First,
PISA, as with most other large-scale international assessments, is cross-
sectional in nature. Longitudinal data are needed to afford stronger
conclusions about temporal precedence.
Second, our study was conned to 15-year-old students and the PISA
Fig. 3. 2-1-1 model depicting the role of income inequality on test anxiety and achievement.
R.B. King et al.
Author's personal copy - R.B. King

Learning and Instruction 89 (2024) 101825
9
test was low stakes. The impact of test anxiety on achievement outcomes
might vary across age groups and across test situations. It is possible that
test anxiety might have more deleterious consequences for older stu-
dents such as senior secondary students who are taking their university
admissions exams and higher education students whose grades will have
direct implications for their future employment opportunities. Hence,
we encourage future studies to include a wider range of age groups and
different types of tests.
Third, we only examined the linkages among income inequality, test
anxiety, and academic achievement. However, there are other medi-
ating variables that might be important to account for. Income
inequality is a relatively distal environmental factor and its inuence on
key outcomes might be shaped by more proximal processes. For
example, past inequality studies have focused on the role of both psycho-
social variables (e.g., impaired social relationships) and material re-
sources (e.g., less investment in educational resources) in mediating the
effects of income inequality on key outcomes (e.g., Du et al., 2022; King
et al., 2022; Oishi, 2014). These variables might also be relevant in
understanding how national income inequality is associated with test
anxiety and achievement.
Fourth, our measure of test anxiety was domain general. It might be
useful in future research to also examine domain-specic measures of
test anxiety and map out how they would be associated with academic
achievement across different domains. Using domain-specic measures
might be associated with stronger effect sizes.
Fifth, countries that participate in PISA are relatively wealthier;
extremely poor countries do not participate in PISA. However, the
poorest countries often have the highest levels of income inequality.
Hence, the countries included in this study do not cover the full range of
national differences in country afuence and income inequality. This
may lead to a restricted range that could reduce the power of our sta-
tistical analyses; it additionally limits the generalizability of our results.
Last, we focused on national income inequality but there are other
types of inequality such as wealth inequality, inequality among peers, or
subjective inequality. Future studies that include different types of
inequality are needed to explore how these types of inequality might
impact test anxiety and academic achievement.
4.3. Conclusion
Income inequality is becoming an increasingly prevalent feature of
contemporary society. The current study demonstrated that students
experience heightened test anxiety and lower academic achievement in
more unequal societies. This is the rst study to empirically link expe-
riences of test anxiety with income inequality. We believe that our un-
derstanding of test anxiety can be enriched by broadening our
theoretical purview beyond an exclusive focus on psychological factors
and the proximal environment. Taking the broader socio-ecological
context into account promises to yield a fuller understanding of the
factors that underpin students’ affective experiences and achievement in
academic settings.
Data availability statement
The data that support the ndings of this study are openly available
in OECD PISA 2015 dataset at: https://www.oecd.org/pisa/data/2015
database/
Consent to publication
All authors consent to publish this article in Learning and Instruction.
CRediT author statement
Ronnel B. KING: Conceptualization, Methodology, Writing- Original
draft preparation. Yuyang CAI: Conceptualization, Methodology,
Formal analysis, Writing- Original draft preparation, Funding acquisi-
tion. Andrew J. ELLIOT: Writing- Reviewing and Editing.
Acknowledgement
This work was supported by The Program for Professor of Special
Appointment (Eastern Scholar) at Shanghai Institutions of Higher
Learning (Code: TP2018068) given to the second author.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.learninstruc.2023.101825.
References
Andrews, D., & Leigh, A. (2009). More inequality, less social mobility. Applied Economics
Letters, 16(15), 1489–1492. https://doi.org/10.1080/13504850701720197
Bartram, D. (2022). Does inequality exacerbate status anxiety among higher earners? A
longitudinal evaluation. International Journal of Comparative Sociology, 63(4),
184–200. https://doi.org/10.1177/00207152221094815
Bond, T. G., & Fox, C. M. (2015). Applying the Rasch model: Fundamental measurement in
the human sciences (3rd ed.). Routledge.
Browman, A. S., Destin, M., Kearney, M. S., & Levine, P. B. (2019). How economic
inequality shapes mobility expectations and behaviour in disadvantaged youth.
Nature Human Behaviour, 3, 214–220. https://doi.org/10.1038/s41562-018-0523-0
Camacho-Morles, J., Slemp, G. R., Pekrun, R., Loderer, K., Hou, H., & Oades, L. G.
(2021). Activity achievement emotions and academic performance: A meta-analysis.
Educational Psychology Review, 33(3), 1051–1095. https://doi.org/10.1007/s10648-
020-09585-3
Cassady, J. C., & Finch, W. H. (2020). Revealing nuanced relationships among cognitive
test anxiety, motivation, and self-regulation through curvilinear analyses. Frontiers in
Psychology, 11, 1141. https://doi.org/10.3389/fpsyg.2020.01141
Cassady, J. C., & Johnson, R. E. (2002). Cognitive test anxiety and academic
performance. Contemporary Educational Psychology, 27(2), 270–295. https://doi.org/
10.1006/ceps.2001.1094
Cassady, J. C., Pierson, E. E., & Starling, J. M. (2019). Predicting student depression with
measures of general and academic anxieties. Frontiers in Education, 4, 11. https://doi.
org/10.3389/feduc.2019.00011
Chiu, M. M. (2015). Family inequality, school inequalities, and mathematics
achievement in 65 countries: Microeconomic mechanisms of rent seeking and
diminishing marginal returns. Teachers College Record, 117(1), 1–32. https://doi.
org/10.1177/016146811511700110
Chiu, M. M., & Chow, B. W. Y. (2015). Classmate characteristics and student achievement
in 33 countries: Classmates’ past achievement, family socioeconomic status,
educational resources, and attitudes toward reading. Journal of Educational
Psychology, 107(1), 152–169. https://doi.org/10.1037/a0036897
Chiu, M. M., & Khoo, L. (2005). Effects of resources, inequality, and privilege bias on
achievement: Country, school, and student level analyses. American Educational
Research Journal, 42(4), 575–603. https://doi.org/10.3102/00028312042004575
Condron, D. J. (2011). Egalitarianism and educational excellence: Compatible goals for
afuent societies? Educational Researcher, 40(2), 47–55. https://doi.org/10.3102/
0013189X11401021
Du, H., Chi, P., & King, R. B. (2019a). Economic inequality is associated with long-term
harm on adolescent well-being in China. Child Development, 90(4), 1016–1026.
https://doi.org/10.1016/j.socscimed.2019.04.043.
Du, H., G¨
otz, F. M., King, R. B., & Rentfrow, P. J. (2022). The psychological imprint of
inequality: Economic inequality shapes achievement and power values in human
life. Journal of Personality. Advance online publication https://doi.org/10.1111/j
opy.12758.
Du, H., King, R. B., & Chi, P. (2019b). Income inequality is detrimental to long-term well-
being: A large-scale longitudinal investigation in China. Social Science & Medicine,
232, 120–128. https://doi.org/10.1016/j.socscimed.2019.04.043.
Elgar, F. J., Gari´
epy, G., Torsheim, T., & Currie, C. (2017). Early-life income inequality
and adolescent health and well-being. Social Science & Medicine, 174, 197–208.
https://doi.org/10.1016/j.socscimed.2016.10.014
von der Embse, N., Jester, D., Roy, D., & Post, J. (2018). Test anxiety effects, predictors,
and correlates: A 30-year meta-analytic review. Journal of Affective Disorders, 227,
483–493. https://doi.org/10.1016/j.jad.2017.11.048
Fr´
echette-Simard, C., Plante, I., Duchesne, S., & Chaffee, K. E. (2022). The mediating role
of test anxiety in the evolution of motivation and achievement of students
transitioning from elementary to high school. Contemporary Educational Psychology,
71, Article 102116. https://doi.org/10.1016/j.cedpsych.2022.102116
Gass, C. S., & Curiel, R. E. (2011). Test anxiety in relation to measures of cognitive and
intellectual functioning. Archives of Clinical Neuropsychology, 26(5), 396–404.
https://doi.org/10.1093/arclin/acr034
Gignac, G. E., & Szodorai, E. T. (2016). Effect size guidelines for individual differences
researchers. Personality and Individual Differences, 102, 74–78. https://doi.org/
10.1016/j.paid.2016.06.069
Heck, R. H., & Thomas, S. L. (2015). An introduction to multilevel modeling techniques: MLM
and SEM approaches using Mplus. Routledge.
R.B. King et al.
Author's personal copy - R.B. King

Learning and Instruction 89 (2024) 101825
10
Heisig, J. P., & Schaeffer, M. (2019). Why you should always include a random slope for
the lower-level variable involved in a cross-level interaction. European Sociological
Review, 35(2), 258–279. https://doi.org/10.1093/ESR/JCY053
Hembree, R. (1988). Correlates, causes, effects, and treatment of test anxiety. Review of
Educational Research, 58(1), 47–77. https://doi.org/10.3102/00346543058001047
Hill, F., Mammarella, I. C., Devine, A., Caviola, S., Passolunghi, M. C., & Szucs, D. (2016).
Maths anxiety in primary and secondary school students: Gender differences,
developmental changes and anxiety specicity. Learning and Individual Differences,
48, 45–53. https://doi.org/10.1016/j.lindif.2016.02.006
Hoferichter, F., Raufelder, D., & Eid, M. (2014). The mediating role of socio-motivational
relationships in the interplay of perceived stress, neuroticism, and test anxiety
among adolescent students. Psychology in the Schools, 51(7), 736–752. https://doi.
org/10.1002/pits.21778
Jerrim, J., & Macmillan, L. (2015). Income inequality, intergenerational mobility, and
the Great Gatsby Curve: Is education the key? Social Forces, 94(2), 505–533. https://
doi.org/10.1093/sf/sov075
Jiang, L., & Probst, T. M. (2017). The rich get richer and the poor get poorer: Country-
and state-level income inequality moderates the job insecurity-burnout relationship.
Journal of Applied Psychology, 102, 672–681. https://doi.org/10.1037/apl0000179
King, R. B. (2022). Sociocultural and ecological perspectives on achievement motivation.
Asian Journal of Social Psychology, 25(3), 433–448. https://doi.org/10.1111/ajs
p.12507.
King, R. B., Chiu, M. M., & Du, H. (2022). Greater income inequality, lower school
belonging: Multilevel and cross-temporal analyses of 65 countries. Journal of
Educational Psychology, 114(5), 1101–1120. https://doi.org/10.1037/edu0000736.
Kraus, M. W., Tan, J. J. X., & Tannenbaum, M. B. (2013). The social ladder: A rank-based
perspective on social class. Psychological Inquiry, 24(2), 81–96. https://doi.org/
10.1080/1047840X.2013.778803
Kuo, C. T., & Kawachi, I. (2023). County-level income inequality, social mobility, and
deaths of despair in the US, 2000-2019. JAMA Network Open, 6(7), Article e2323030-
e2323030. https://doi.org/10.1001/jamanetworkopen.2023.23030
Layte, R. (2012). The association between income inequality and mental health: Testing
status anxiety, social capital, and neo-materialist explanations. European Sociological
Review, 28(4), 498–511. https://doi.org/10.1093/esr/jcr012
Layte, R., & Whelan, C. T. (2014). Who feels inferior? A test of the status anxiety
hypothesis of social inequalities in health. European Sociological Review, 30(4),
525–535. https://doi.org/10.1093/esr/jcu057
Lowe, P. A., Grumbein, M. J., & Raad, J. M. (2011). Examination of the psychometric
properties of the test anxiety scale for elementary students (TAS-E) scores. Journal of
Psychoeducational Assessment, 29(6), 503–514.
Maloney, E. A., Sattizahn, J. R., & Beilock, S. L. (2014). Anxiety and cognition. Wiley
Interdisciplinary Reviews: Cognitive Science, 5(4), 403–411. https://doi.org/10.1002/
wcs.1299
Messias, E., Eaton, W. W., & Grooms, A. N. (2011). Economic grand rounds: Income
inequality and depression prevalence across the United States: An ecological study.
Psychiatric Services, 62(7), 710–712. https://doi.org/10.1176/ps.62.7.pss6207_0710
Ngamaba, K. H., Panagioti, M., & Armitage, C. J. (2018). Income inequality and
subjective well-being: A systematic review and meta-analysis. Quality of Life
Research, 27, 577–596. https://doi.org/10.1007/s11136-017-1719-x
Ng, E. L., & Lee, K. (2015). Effects of trait test anxiety and state anxiety on children’s
working memory task performance. Learning and Individual Differences, 40, 141–148.
https://doi.org/10.1016/j.lindif.2015.04.007
Nie, Y., Lau, S., & Liau, A. K. (2011). Role of academic self-efcacy in moderating the
relation between task importance and test anxiety. Learning and Individual
Differences, 21(6), 736–741. https://doi.org/10.1016/j.lindif.2011.09.005
OECD. (2015). How’s life? 2015: Measuring well-being. OECD Publishing.
OECD. (2016). PISA 2015 assessment and analytical framework: Science, reading,
mathematics, and Financial Literacy. OECD Publishing.
OECD. (2017). Pisa 2015: Technical report. OECD Publishing.
Oishi, S. (2014). Socioecological psychology. Annual Review of Psychology, 65(1),
581–609. https://doi.org/10.1146/annurev-psych-030413-152156
Owens, M., Stevenson, J., Hadwin, J. A., & Norgate, R. (2012). When does anxiety help
or hinder cognitive test performance? The role of working memory capacity. British
Journal of Psychology, 105, 92–101. https://doi.org/10.1111/bjop.12009
Owens, M., Stevenson, J., Norgate, R., & Hadwin, J. A. (2008). Processing efciency
theory in children: Working memory as a mediator between trait anxiety and
academic performance. Anxiety, Stress & Coping, 21, 417–430. https://doi.org/
10.1080/10615800701847823
Pekrun, R. (1992). The expectancy-value theory of anxiety: Overview and implications.
In D. G. Forgays, T. Sosnowski, & K. Wrzesniewski (Eds.), Anxiety: Recent
developments in self-appraisal, psychophysiological and health research (pp. 23–41).
Hemisphere.
Peleg-Popko, O. (2002). Children’s test anxiety and family interaction patterns. Anxiety,
Stress & Coping, 15(1), 45–59. https://doi.org/10.1080/10615800290007281
Pickett, K. E., & Wilkinson, R. G. (2007). Child wellbeing and income inequality in rich
societies: Ecological cross-sectional study. British Medical Journal, 335(7629),
1080–1085. https://doi.org/10.1136/bmj.39377.580162.55
Pickett, K. E., & Wilkinson, R. G. (2015). Income inequality and health: A causal review.
Social Science & Medicine, 128, 316–326. https://doi.org/10.1016/j.
socscimed.2014.12.031
Preacher, K. J., Zhang, Z., & Zyphur, M. J. (2016). Multilevel structural equation models
for assessing moderation within and across levels of analysis. Psychological Methods,
21(2), 189–205. https://doi.org/10.1037/met0000052
Putwain, D. W., & Best, N. (2011). Fear appeals in the primary classroom: Effects on test
anxiety and test grade. Learning and Individual Differences, 21(5), 580–584. https://
doi.org/10.1016/j.lindif.2011.07.007
Putwain, D. W., Daly, T., Chamberlain, S., & Saddredini, S. (2016). "Sink or swim":
Buoyancy and coping in the test anxiety and academic performance relationship.
Educational Psychology, 36(10), 1807–1825. https://doi.org/10.1080/
01443410.2015.1066493
Putwain, D. W., Gallard, D. G., Beaumont, J., Loderer, K., & von der Embse, N. (2021).
Does test anxiety predispose poor school-related wellbeing and enhanced risk of
emotional disorders? Cognitive Therapy and Research, 45(6), 1150–1163. https://doi.
org/10.1007/s10608-021-10211-x
Putwain, D. W., Nicholson, L. J., Connors, E., & Woods, K. A. (2013). More resilient
children are less test anxious and perform better in tests at the end of primary
schooling. Learning and Individual Differences, 28(1), 41–46. https://doi.org/
10.1016/j.lindif.2013.09.010
Putwain, D. W., Stockinger, K., von der Embse, N. P., Suldo, S. M., & Daumiller, M.
(2021). Test anxiety, anxiety disorders, and school-related wellbeing: Manifestations
of the same or different constructs? Journal of School Psychology, 88, 47–67. https://
doi.org/10.1016/j.jsp.2021.08.001
Putwain, D. W., & Symes, W. (2020). The four Ws of test anxiety: What is it, why is it
important, where does it come from, and what can be done about it. Psychologica, 63
(2), 31–52. https://doi.org/10.14195/1647-8606_63-2_2
Robson, D. A., Johnstone, S. J., Putwain, D. W., & Howard, S. (2023). Test anxiety in
primary school children: A 20-year systematic review and meta-analysis. Journal of
School Psychology, 98, 39–60. https://doi.org/10.1016/j.jsp.2023.02.003
Sarason, S. B., & Mandler, G. (1952). Some correlates of test anxiety. Journal of Abnormal
and Social Psychology, 47(4), 810–817. https://doi.org/10.1037/h0060009
Schillinger, F. L., Mosbacher, J. A., Bruner, C., Vogel, S. E., & Grabner, R. H. (2021).
Revisiting the role of worries in explaining the link between test anxiety and test
performance. Educational Psychology Review, 33, 1887–1906. https://doi.org/
10.1007/s10648-021-09601-0
Segool, N. K., Carlson, J. S., Goforth, A. N., Von der Embse, N., & Barterian, J. A. (2013).
Heightened test anxiety among young children: Elementary school students’ anxious
responses to high-stakes testing. Psychology in the Schools, 50(5), 489–499. https://
doi.org/10.1002/pits.21689
Seipp, B. (1991). Anxiety and academic performance: A meta-analysis of ndings.
Anxiety Research, 4(1), 27–41. https://doi.org/10.1080/08917779108248762
Solt, F. (2016). The standardized world income inequality database. Social Science
Quarterly, 97(5), 1267–1281. https://doi.org/10.1111/SSQU.12295
Sommet, N., Elliot, A. J., Jamieson, J., & Butera, F. (2019). Income inequality, perceived
competitiveness, and approach-avoidance motivation. Journal of Personality, 87,
767–784.
Sommet, N., Morselli, D., & Spini, D. (2018). Income inequality affects the psychological
health of only the people facing scarcity. Psychological Science, 29(12), 1911–1921.
https://doi.org/10.1177/0956797618798620
Sommet, N., Weissman, D. L., & Elliot, A. J. (2022). Income inequality predicts
competitiveness and cooperativeness at school. Journal of Educational Psychology.
Advance online publication. https://doi.org/10.1037/edu0000731
Spielberger, C. D. (1980). Preliminary professional manual for the test anxiety inventory
(TAI). Consulting Psychologists Press.
World Bank. (2018). World Bank open data. https://data.worldbank.org/.
Xie, F., Xin, Z., Chen, X., & Zhang, L. (2019). Gender difference of Chinese high school
students’ math anxiety: The effects of self-esteem, test anxiety and general anxiety.
Sex Roles, 81, 235–244. https://doi.org/10.1007/s11199-018-0982-9
Zagorski, K., Evans, M. D. R., Kelley, J., & Piotrowska, K. (2014). Does national income
inequality affect individuals’ quality of life in europe? Inequality, happiness,
nances, and health. Social Indicators Research, 117(3), 1089–1110. https://doi.org/
10.1007/s11205-013-0390-z
R.B. King et al.
Author's personal copy - R.B. King