Professional Documents
Culture Documents
To cite this article: Tzu-Yang Chao & Yao-Ting Sung (2019): An investigation of the reasons for
test anxiety, time spent studying, and achievement among adolescents in Taiwan, Asia Pacific
Journal of Education, DOI: 10.1080/02188791.2019.1671804
Article views: 2
Introduction
Test anxiety and examination stress remain prevalent among adolescents (Beidel, Turner, & Taylor-
Ferreira, 1999). More recently, Putwain and Daly (2014) demonstrated that 15%–40% of secondary
school students in the UK reported having high test anxiety. Thomas, Cassady, and Finch (2018) used
latent class analysis to reveal that 25% of US undergraduate students fell into the high test anxiety
class. Researchers are concerned about test anxiety because it may have an adverse impact on
student performance. Seipp’s (1991) meta-analysis reported a negative correlation between test
anxiety and achievement (r = −.23). On the basis of this linear negative correlation between test
anxiety and performance in adolescents, researchers might easily infer that students with the worst
grades would have the most severe test anxiety. However, do students with the lowest levels of
achievement really report the highest test anxiety? Sung, Chao, and Tseng (2016) indicated that
among Taiwanese students, the relationship between levels of test anxiety and achievement was
represented by an inverted-U shape, that is, students of middle achievement experienced the
highest level of test anxiety. Sung et al. (2016) mentioned that students with medium levels of
CONTACT Tzu-Yang Chao ziyang.ccu@gmail.com Research Center for Psychological and Educational Testing, National
Taiwan Normal University, No. 162, Sec. 1, Heping E. Rd., Da’an Dist., Taipei 106, Taiwan
© 2019 National Institute of Education, Singapore
2 T.-Y. CHAO AND Y.-T. SUNG
achievement experienced greater levels of test anxiety because they faced greater uncertainty about
qualifying for entry into state schools, which are more prestigious than private high schools in
Taiwan. The present study therefore attempted to corroborate the inverted-U-shaped relationship. In
addition, this study uses a survey to investigate the reasons for this inverted-U phenomenon; the
impact on test anxiety from factors, such as time spent studying and days spent in cram school (i.e.,
non-formal schools where students are trained to pass high-stakes tests) are also examined.
Literature review
Test anxiety and examination stress
Test anxiety is an anxious response evoked by test situations. However, its definition has changed
over time. When first coined, test anxiety was identified with various sorts of pretest physical arousal,
such as elevated heart rate. Subsequently, it was divided into two components: “emotionality”
(physiological arousal) and “worry” (negative thoughts about tests) (Spielberger, 1980). Later, test
anxiety was seen as a multidimensional phenomenon including cognitive, behavioural, physiologi-
cal, and emotional components (Wren & Benson, 2004; Zeidner, 1998) and was treated as being
synonymous with “examination stress.” Therefore, some researchers have argued that test anxiety is
difficult to distinguish from examination stress, that is, the two terms can be used interchangeably
(Schwarzer & Buchwald, 2003; Sung et al., 2016). Sung and Chao (2015) constructed an Examination
Stress Scale (ExamSS) for Taiwanese adolescents that divides examination stress into the physiolo-
gical anxiety response (PA), the cognitive–behavioural response (CB), and perceived social expecta-
tions and social comparisons (SS). The first two components represent test anxiety stress responses
when students face situational and societal examination stressors, such as an imminent regular test,
and the third represents the social atmosphere that continually reminds students to get higher
grades (Sung & Chao, 2015). In this study, we used the ExamSS to measure test anxiety among
Taiwanese junior high school students.
Micro-level perspective
The micro-level focuses on the short term, and test anxiety and performance are seen as cause and
effect, respectively. Three theories explain this causal relationship: the cognitive interference model,
the motivational enhancement model, and the Yerkes–Dodson law. In the cognitive interference
model, test anxiety interferes with students’ cognitive abilities during exams, resulting in reduced
performance. In past research on test anxiety, a significant correlation between test anxiety and low
academic achievement has been generally observed (Chapell et al., 2005; Hembree, 1988; McCarthy
& Goffin, 2005; Musch & Broeder, 1999; Seipp, 1991; Zeidner, 1998). Further empirical studies have
also revealed that students worry that their scores will fall short of expectations, and this worry may
lead to impaired working memory and thus reduced performance (Beilock, 2008; Beilock & Carr,
2005; Korhonen, Nyroos, Jonsson, & Eklöf, 2018).
The motivational enhancement model shows how test anxiety motivates students to study for or
focus more effectively on tests. While many studies have revealed a negative relationship between
test anxiety and performance, some have also concluded that a positive correlation exists: For
example, Cassady and Johnson (2002) found a weak and positive correlation between emotional
ASIA PACIFIC JOURNAL OF EDUCATION 3
anxiety and academic performance. Chapell et al. (2005) concluded that undergraduate girls’ test
anxiety and their grade point average were both significantly higher than those of boys. Struthers,
Perry, and Menec (2000) argued that students with test anxiety used more problem-focused coping
strategies, which led to higher scores.
The Yerkes–Dodson law describes a curvilinear relationship between anxiety and performance,
that is, slight anxiety could enhance motivation, encouraging students to expend more effort on
tasks; however, extreme anxiety could suppress performance. Srivastava and Krishna (1991) demon-
strated this phenomenon in industry: People who experienced moderate levels of stress performed
most efficiently.
Macro-level perspective
The macro-level, by contrast, focuses on the long term (e. g., the 3 years from Grades 7 to 9), and
performance is seen as the cause, while test anxiety is seen as the effect (Sung et al., 2016). In the
short term, the relationship between test anxiety and performance may relate to each stage of the
test experience, including preparation, the actual test, and the post-exam period (Stöber, 2004);
therefore, cognitive interference, motivation enhancement, and the effects of the Yerkes–Dodson
law may all come into play (Muse, Harris, & Feild, 2003). However, in the long term, students face
long-term high-stakes testing, and test anxiety can be seen as long-term stress. Carey, Hill, Devine,
and Szücs (2015) indicated that longitudinal studies have shown that students’ poor performance
could lead to test anxiety. Long-term stress among students could be due to their achievement
levels, which will determine their future careers. For example, Sung et al. (2016) reported that ninth
graders of middle achievement experienced the highest levels of test anxiety, with anxiety level
gradually tapering off towards the two extremes of the achievement spectrum, and this phenom-
enon may be explained by students of middle achievement facing greater uncertainty about
examination results. Uncertainty can be a major source of stress and is linked to anxiety (Gati
et al., 2011). According to the entropy model of uncertainty (EMU) put forward by Hirsh, Mar, and
Peterson (2012), uncertainty can be reflected in an individual’s perceptions and action: When a given
situation has more perceptual and action possibilities, there is greater uncertainty anxiety. As
described above, from a macro perspective which looks at students’ achievements as a cause and
their test anxiety as an effect, one may wonder why middling students feel and experience more test
anxiety and uncertainty.
No return on investment
The social environment created by the aforementioned examination culture means that in countries
where examinations are held in great esteem, students take exam results seriously and spend more
time preparing for them. Cram school attendance is widespread throughout Japan, Hong Kong,
South Korea, Singapore, Taiwan, Malaysia, and the Philippines (Bray, 1999; Johnson & Johnson, 1996;
Kim & Lee, 2010). Furthermore, studies have discovered that in 1998, the proportion of the average
South Korean family income spent on cram schooling (2.9%) was not far below that spent on formal
education (3.4%) (Kim, 2005). In addition, schools hold frequent mock exams to hone students’
examination skills (Lee & Larson, 2000). Such a social environment can affect students’ natural
rhythms of work and rest. For example, Lee and Larson (2000) revealed that South Korean students
spend more time studying and doing homework than American students, and this greatly reduces
the amount of free time spent on leisure activities. In addition, they found that spending more time
on study and homework correlates with higher rates of depression reported by students (between
.35 and .54). Thus, Korean students experience more negative emotions than American students.
Researchers also refer to this phenomenon of cram schooling and over studying as “examination
hell” (Lee, 2003; Lee & Larson, 2000).
In summary, the social environment of countries with an examination culture can increase the
amount of time students spend studying and in cram school; furthermore, students who spend more
time engaging in these activities may experience more negative emotions. Thus, one reason why
medium-achieving students suffer greater levels of test anxiety may be that these students spend
more time studying and at cram schools; moreover, even if they do so, their grades may remain
middling. This situation may create uncertainty because students feel that they earn no “return” on
their “investment.” Therefore, the time spend on study and private tuition and its relationship with
their performance is worth investigating. The present study examined whether medium-achieving
students spend more time in these pursuits than high- or low-performing students.
Research objectives
The main objective of the present study is to answer two research questions: Does the test anxiety
experienced by medium-achieving students originate in their failure to enter a state school, and do
medium-achieving students spend the most time in study and private tuition? These two research
questions are based on the findings of previous research; namely, that medium-achieving students
experience greater levels of test anxiety than other students (Sung et al., 2016). However, the non-linear
inverted-U relationship between test anxiety and academic achievement is very rare observed in previous
research. Thus, this study first attempted to corroborate the inverted-U phenomenon and then investi-
gated medium-achieving students’ perception of uncertainty about obtaining a grade that will permit
entry into a state school and thereby encourage them to spend more time studying than others.
Method
Participants
Initial data were collected in 2012 from 5,292 students. Stratified random sampling was used to
select junior high schools from Northern, Central, Southern, and Eastern Taiwan at a ratio of 3:2:2:1.
A total of 120 schools participated in the study, with students selected at random – 40, 80, or 120
students from each school, according to the size of the student body.
The post data filtering phase filtered out incomplete student data, defined as follows:
(1) incomplete basic information such as sex or national ID card number (which enables access to
the participant’s Basic Competence Test (BCTEST) scores),
(2) failure to answer all ExamSS questions,
ASIA PACIFIC JOURNAL OF EDUCATION 5
(3) missing information on time spent studying and the number of days attending cram school,
(4) student not going on to sit the BCTEST.
After data filtering, we retained the data of 5,220 students (98.6% of the original sample), 2,448
(46.90%) boys and 2,772 (53.10%) girls. Sampling covered junior high schools in every region of
Taiwan and students of all levels of academic performance in Taiwan; therefore, the sample was
representative.
Research materials
Examination stress scale (ExamSS)
This study used the ExamSS developed by Sung and Chao (2015), which covers three dimensions
that reflect the levels of test anxiety and stress experienced by high school students: physiological
anxiety responses (PA, 10 items), cognitive and behavioural responses (CB, 8 items), and perceived
social expectations and social comparisons (SS, 9 items). PA are physical responses arising from
examinations, such as “I have a stomach ache before or during a test.” CB are thoughts and
behaviours triggered by test anxiety, such as “I force myself to study hard in preparation for an
exam.” SS are feelings of anxiety that students have because of their parents’ and teachers’
expectations or comparison with their peers, such as “Whenever I think of how my teacher expects
me to perform on the exam, I feel overwhelmed.” Responses to the ExamSS use a 5-point Likert scale,
with scores of 1, 2, 3, 4, and 5 denoting “totally disagree,” “slightly disagree,” “partially agree,”
“agree,” and “totally agree,” respectively. The Cronbach’s α value for the three subscales were .91 for
PA, .86 for CB, and .88 for SS; the Cronbach’s α value for the whole scale was .93. These reliability
coefficients are similar to those of Sung and Chao (2015), where the Cronbach’s α values were .89,
.85, .88, and .92 for PA, CB, SS, and the whole scale, respectively. In addition, confirmatory factor
analysis (CFA) of the three-dimensional model was conducted to examine the present study’s
construct of the ExamSS, with the comparative fit index (CFI) = .87, the normed fit index
(NFI) = .87, and the root mean square error of approximation = .07. Although the CFI and NFI of
the CFA model were below the .90 threshold, the decreased chi-square relative values could be
explained by an excessive sample size, a large number of items, or a stronger correlation between
variables (Kline, 2010). Thus, according to the reliability and CFA evidence, the present study’s
ExamSS is an acceptable measurement tool.
The BCTEST
The BCTEST is the most widely administered official standardized test and was developed in 1998 by
the Research Centre for Psychological and Educational Testing (RCPET) at National Taiwan Normal
University. It is one of the largest academic achievement tests in Taiwan, and its results are used as
6 T.-Y. CHAO AND Y.-T. SUNG
admission criteria for junior high school students seeking to enter senior high school or vocational
school (Sung et al., 2014; Wang, Sung, Sung, Lu, & Li, 2008; Yeh, Chang, Hu, Yeh, & Lin, 2009). The test
covers six subjects: Mandarin, English, mathematics, social sciences, sciences, and writing. The first
five subjects are available exclusively in the multiple choice format, with scoring scales ranging from
1 to 80, while the writing section has 12 points, allocating a total of 412 points. Test items come from
an item bank, and all test items are rigorously reviewed by domain-subject experts and testing
experts and undergo pretest item analysis to ensure they meet standards of difficulty and discrimi-
nation, such as having discrimination coefficients over .4. As a result, the BCTEST is currently the
standardized test that best reflects the levels of academic achievement of junior high school
students in Taiwan. The present study accessed the test scores for the BCTEST held in May 2012
and used the scores as indices of students’ academic performance. The total number of items
(Cronbach’s α) for each of the five subjects were as follows: Mandarin, 48 items (.92); English, 45
items (.96); mathematics, 34 items (.92); social sciences, 63 items (.95); and sciences, 58 items (.94).
Results
Achievement and test anxiety
Table 1 indicates that girls had significantly higher test anxiety than boys on the ExamSS total
score or three subscale scores. Furthermore, we sought to determine whether the two genders
had the same correlation between the BCTEST and ExamSS scores. A Fisher’s z transformation
conducted for the correlation coefficients revealed that the difference between the correlation
coefficients for girls and boys for the BCTEST and ExamSS was not significant (z = 1.15, p = .25),
and non-significant differences were observed between the BCTEST scores and PA (z = −0.34,
p = .73), CB (z = 1.60, p = .11), and SS (z = 1.31, p = .19). On the basis of these results, gender
does not appear to moderate the correlation between academic performance and test anxiety;
thus, we combined the data for boys and girls for all subsequent analysis. First, we analysed the
descriptive statistics for the BCTEST and ExamSS scores for each PR group (Table 2) and found
that the students of middle achievement (PR40–49, PR50–59, and PR60–69 groups) reported the
highest levels of test anxiety, while anxiety levels tapered off towards the two extremes of the
ASIA PACIFIC JOURNAL OF EDUCATION 7
Table 1. Descriptive statistics and coefficients for correlations between the BCTEST, total ExamSS, and ExamSS subscales.
Boys
1 2 3 4 5
1. BCTEST –
2. PA .14* –
3. CB .24* .50* –
4. SS -.03 .53* .53* –
5. ExamSS .13* .83* .80* .85* –
Mean 267.36 20.79 24.14 27.19 72.13
SD 96.20 8.59 7.44 8.99 20.67
Girls
1. BCTEST –
2. PA .10* –
3. CB .24* .51* –
4. SS -.08* .50* .50* –
5. ExamSS .09* .84* .79* .83* –
Mean 266.94 23.95 25.85 29.90 79.70
SD 92.18 8.54 6.62 8.28 19.19
All students
1. BCTEST –
2. PA .11* –
3. CB .24* .52* –
4. SS -.06* .53* .52* –
5. ExamSS .11* .84* .80* .84* –
Mean 267.13 22.47 25.05 28.63 76.15
SD 94.08 8.71 7.07 8.72 20.25
BCTEST = Basic Competence Test; PA = physiological anxiety; CB = cognitive and behavioural; SS = social expectations and social
comparisons; ExamSS = Examination Stress Scale; SD = standard deviation
*p < .001.
PR scale (Figure 1). To examine the inverted-U shape relationship between achievement and test
anxiety, we also conducted a quadratic regression. As Table 3 displays, the quadratic term in the
regression was negative and significant (β = −.86, p < .01), which indicates that the pattern of
test anxiety in relation to achievement is an inverted-U shape.
We also calculated the correlation between the BCTEST and ExamSS scores among top-achieving
(PR50–99) students and those at the bottom of the academic scale (PR1–49). As Table 4 shows,
a positive relationship was observed between the BCTEST and ExamSS scores of students with the
lowest academic performance, r = .24. However, among the top-achieving students a slight negative
correlation was observed, r = −.10. These results are consistent with those presented by Sung et al.
(2016), indicating that the inverted-U shape relationship can be corroborated.
8 T.-Y. CHAO AND Y.-T. SUNG
95.00
90.00
85.00
80.00
ExamSS score
75.00
70.00
65.00
60.00
55.00
50.00
BCTEST PR gourps
Figure 1. Examination Stress Scale (ExamSS) scores of students with different levels of learning achievement. BCTEST = Basic
Competence Test; PR = percentile rank.
“However hard I work, my grades do not improve” and “I don’t understand what the teachers teach.”
For the PR1–9 group, these were the top two reasons for 18.51% and 21.87% of the students. For
those in the PR10–19 group, they were the top reasons for 24.21% and 19.55% of the students. In
groups with higher academic achievement, reasons such as “Not being able to get into a state
school” and “Not meeting my own standards” start to became more important. For example, in the
PR40–49 group, these two reasons were listed as the top two reasons for test anxiety by 19.32% and
20.31% of the participants. As the level of academic achievement increased, the reason “Not being
able to maintain a certain performance standard” started taking precedence, and the percentage of
students listing the top reasons as “Not meeting my own standards” also increased. For example, in
the PR70–79 group, these two reasons were ranked top by 16.71% and 33.87% of the students;
however, in the PR90–99 group, the percentage of students selecting “Not meeting my own
standards” as the top reason increased to 56.19%.
3.50
3.00
Days per week
2.50
2.00
1.50
1.00
0.50
0.00
BCTSET PR groupings
Figure 2. Average number of days in cram school per week by levels of learning achievement. BCTEST = Basic Competence Test;
PR = percentile rank.
ASIA PACIFIC JOURNAL OF EDUCATION 11
Table 6. ANOVA and Scheffé Post-Hoc tests of days in Cram School and time spent studying.
Source SS DF MS F p Scheffé’s post-hoc
Days in cram school PR1–9, PR10–19 < PR20–29~PR90–99
PR group 1793.77 9 199.31 52.38 < .01 PR20–29 < PR30–39~PR80–89
Within 19824.29 5210 3.81 PR30–39 < PR50–59 ~ PR70–79
PR50–59~PR70–79 > PR90–99
Hr studying (schooldays) PR1–9, PR10–19 < PR20–29~PR90–99
PR group 1232.13 9 136.90 94.97 < .01 PR20–29 < PR30–39~PR90–99
Within 7510.11 5210 1.44 PR30–39 < PR50–59~PR90–99
PR40–49 < PR60–69~PR90–99
Hr studying (weekend) PR1–9, PR10–19 < PR20–29~PR90–99
PR group 2857.85 9 317.54 157.32 < .01 PR20–29 < PR30–39~PR90–99
Within 10515.74 5210 2.02 PR30–39 < PR40–49~PR90–99
PR40–49 < PR60–69~PR90–99
PR50–59 < PR60–69, PR80–89, PR90–99
PR60–69, PR70–79 < PR90–99
ExamSS = Examination Stress Scale; SS = sum of squares; DF = degrees of freedom; MS = mean square
5.00
4.50
4.00
3.50
Time studying per day
3.00
2.50
2.00
1.50
Schooldays
1.00
0.50 Weekend
0.00
BCTEST PR gourps
Figure 3. Average number of hours spent studying per day by levels of learning achievement. BCTEST = Basic Competence Test;
PR = percentile rank.
ANOVA and Scheffé post-hoc tests revealed that the PR90–99 group reported significantly more time
spent studying at weekends than other groups (ps < .01); however, no significant difference was
observed among students in the PR60–69, PR70–79, and PR80–89 groups (ps > .06).
Discussion
Learning achievement and test anxiety
The present study used the analytical method of Sung et al. (2016) to divide students into 10 groups
according to their PR scores on the BCTEST and calculate the average value of their ExamSS scores.
The results indicated that middle achievement students in the PR40–49, PR50–59, and PR60–69
groups had the highest ExamSS scores, while the scores tapered off towards the two extremes of the
BCTEST performance spectrum. Regression analysis also revealed a significant negative quadratic
term in the regression, which supports the hypothesis of an inverted-U shape relationship between
test anxiety and achievement. The correlations between the ExamSS and BCTEST scores in the low-
12 T.-Y. CHAO AND Y.-T. SUNG
achieving group and the high-achieving group also pointed to the same conclusion (Table 4). For
students at the bottom of the academic achievement ladder, test anxiety was positively correlated
with the BCTEST scores, with a correlation coefficient of .24. However, among high-achieving
students, a negative relationship was observed between test anxiety and the BCTEST scores, with
a correlation coefficient of −.10. These results replicate the findings of Sung et al. (2016).
career counselling system, including aptitude tests, interest tests, visits to various departments, and
career exploration, such as O*net (https://www.onetonline.org/), SIGI3 (http://www.sigi3.org/), and
the Computerized Career Testing System in Taiwan (http://career.ntnu.edu.tw/), could help students
with academic achievement in the PR30–69 range.
In addition to career advice, students of middle to high academic performance (PR70–99) may
need counselling to alleviate their test anxiety. Our findings revealed that the main reason for anxiety
among such students was “I don’t meet my own standards” – over half of the students with test
scores above PR90 cited this as their top reason. Although students in the PR70–99 group are more
likely to be admitted to a state school, they may be expected – by wider society, parents, teachers, or
even themselves – to win admission to an elite school. Thus, in order to alleviate test anxiety among
high-achieving students, in addition to career path counselling, such students may benefit from
counselling solutions that help them revise their goals and stop being too self-demanding.
Disclosure statement
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of
this article.
Funding
This work was supported by the Ministry of Science and Technology, Taiwan [101-2511-S-003-058-MY3,102-2511-S-003-
001-MY3,102-2911-I-003-301]; Higher Education Sprout Project of the Ministry of Education, Taiwan [HESP_2019].
ASIA PACIFIC JOURNAL OF EDUCATION 15
Notes on contributors
Tzu-Yang Chao, Ph. D., is a Senior Research Faculty in the Research Center for Psychological and Educational Testing,
National Taiwan Normal University. His research interests include measurement theory, test anxiety of adolescents, and
teacher education.
Yao-Ting Sung, Ph. D., is a Chair Professor in the Department of Educational Psychology and Counseling, National
Taiwan Normal University. His research interests include computer assisted testing, psychological and educational
testing, and career information analysis and application.
ORCID
Tzu-Yang Chao http://orcid.org/0000-0002-5049-5836
References
Beidel, D.C., Turner, S.M., & Taylor-Ferreira, J.C. (1999). Teaching study skills and test-taking strategies to elementary
school students. Behavior Modification, 123, 630–692.
Beilock, S.L. (2008). Math performance in stressful situations. Current Directions in Psychological Science, 17(5), 339–343.
Beilock, S.L., & Carr, T.H. (2005). When high-powered people fail: Working memory and “choking under pressure” in
math. Psychological Science, 16, 101–105.
Bond, C.F., & Richardson, K. (2004). Seeing the Fisher Z-transformation. Psychometrika, 69(2), 291–303.
Bray, M. (1999). The shadow education system: Private tutoring and its implications for planners. Paris, France: International
Institute for Educational Planning.
Carey, E., Hill, F., Devine, A., & Szücs, D. (2015). The chicken or the egg? The direction of the relationship between
mathematics anxiety and mathematics performance. Frontiers in Psychology, 6, 1987.
Cassady, J.C., & Johnson, R.E. (2002). Cognitive test anxiety and academic performance. Contemporary Educational
Psychology, 27, 270–295.
Chao, T.Y., & Sung, Y.T. (2019). Examination stress and personal characteristics among taiwanese adolescents: A latent
class approach. Journal of Research in Education Sciences, 64(3), 253–285.
Chao, T.Y., Sung, Y.T., Kuo, H.N., & Cheng, Y.Y. (2018). Effectiveness of group counseling sessions aimed at alleviating
examination stress among junior high school students. Bulletin of Educational Psychology, 50(1), 31–52.
Chapell, M.S., Blanding, Z.B., Silverstein, M.E., Takahashi, M., Newman, B., Gubi, A., & McCann, N. (2005). Test anxiety and
academic performance in undergraduate and graduate students. Journal of Educational Psychology, 97, 268–274.
Gati, I., Gadassi, R., Saka, N., Hadadi, Y., Ansenberg, N., Friedmann, R., & Asulin-Peretz, L. (2011). Emotional and
personality-related aspects of career decision-making difficulties: Facets of career indecisiveness. Journal of Career
Assessment, 19(1), 3–20.
Hembree, R. (1988). Correlates, causes, effects, and treatment of test anxiety. Review of Educational Research, 58, 47–77.
Hirsh, J.B., Mar, R.A., & Peterson, J.B. (2012). Psychological entropy: A framework for understanding uncertainty-related
anxiety. Psychological Review, 119, 304–320.
Hui, E.K. (2000). Personal concerns and their causes: Perceptions of Hong Kong Chinese adolescent students. Journal of
Adolescence, 23(2), 189–203.
Hutchings, M. (2015). Exam factories? The impact of accountability measures on children and young people. London: The
National Union of Teachers.
Johnson, M.L., & Johnson, J.R. (1996). Daily life in Japanese high schools. ERIC Digest, 1–6. https://www. ericdigests.org/
1997-4/daily.htm
Kim, S., & Lee, J.H. (2010). Private tutoring and demand for education in South Korea. Economic Development and
Cultural Change, 58(2), 259–296.
Kim, T. (2005). Shadow education: School quality and demand for private tutoring in Korea. Kyoto: Kyoto University.
Kline. (2010). Principles and practice of structural equation modeling (2nd ed.). New York: Guilford.
Korhonen, J., Nyroos, M., Jonsson, B., & Eklöf, H. (2018). Additive and multiplicative effects of working memory and test
anxiety on mathematics performance in grade 3 students. Educational Psychology, 38(5), 572–595.
Lee, M. (2003). Korean adolescents’ “examination hell” and their use of free time. New Directions for Child and Adolescent
Development, 2003(99), 9–22.
Lee, M., & Larson, R. (2000). The Korean ‘examination hell’: Long hr of studying, distress, and depression. Journal of Youth
and Adolescence, 29, 249–272.
McCarthy, J.M., & Goffin, R.D. (2005). Selection test anxiety: Exploring tension and fear of failure across the sexes in
simulated selection scenarios. International Journal of Selection and Assessment, 13(4), 282–295.
Musch, J., & Broeder, A. (1999). Test anxiety versus academic skills: A comparison of two alternative models for
predicting performance in a statistics exam. British Journal of Educational Psychology, 69, 105–116.
16 T.-Y. CHAO AND Y.-T. SUNG
Muse, L.A., Harris, S.G., & Feild, H.S. (2003). Has the inverted-U theory of stress and job performance had a fair test?
Human Performance, 16(4), 349–364.
Nolen-Hoeksema, S., Girgus, J.S., & Seligman, M.E. (1992). Predictors and consequences of childhood depressive
symptoms: A 5-year longitudinal study. Journal of Abnormal Psychology, 101(3), 405.
Putwain, D., & Daly, A.L. (2014). Test anxiety prevalence and gender differences in a sample of English secondary school
students. Educational Studies, 40(5), 554–570.
Putwain, D.W. (2008). Test anxiety and GCSE performance: The effect of gender and socio-economic background.
Educational Psychology in Practice, 24(4), 319–334.
Qian, G., & Alvermann, D. (1995). Role of epistemological beliefs and learned helplessness in secondary school students’
learning science concepts from text. Journal of Educational Psychology, 87(2), 282.
Schwarzer, C., & Buchwald, P. (2003). Examination stress: Measurement and coping. Anxiety, Stress, and Coping, 16,
247–249.
Seipp, B. (1991). Anxiety and academic performance: A meta-analysis of findings. Anxiety Research, 4, 27–41.
Shek, D.T., & Chen, L.K. (1999). Hong Kong Chinese parents’ perceptions of the ideal child. The Journal of Psychology, 133,
291–303.
Spielberger, C.D. (1980). Preliminary professional manual for the test anxiety inventory. Palo Alto, CA: Consulting
Psychologists Press.
Srivastava, A.K., & Krishna, A. (1991). A test of inverted U hypothesis of stress performance relationship in the industrial
context. Psychological Studies, 34, 34–38.
Stöber, J. (2004). Dimensions of test anxiety: Relations to ways of coping with per-exam anxiety and uncertainty. Anxiety,
Stress, and Coping, 17, 213–226.
Struthers, C.W., Perry, R.P., & Menec, V.H. (2000). An examination of the relationship among academic stress, coping,
motivation, and performance in college. Research in Higher Education, 41, 581–592.
Sung, Y.T., & Chao, T.Y. (2015). Construction of the test anxiety scale for adolescent students. Measurement and
Evaluation in Counseling and Development, 48(1), 44–58.
Sung, Y.T., Chao, T.Y., & Tseng, F.L. (2016). Reexamining the relationship between test anxiety and learning achievement:
An individual- differences perspective. Contemporary Educational Psychology, 46, 241–252.
Sung, Y.T., Huang, L.Y., Tseng, F.L., & Chang, K.E. (2014). The aspects and ability groups in which little fish perform worse
than big fish: Examining the big-fish-little-pond effect in the context of school tracking. Contemporary Educational
Psychology, 39(3), 220–232.
Thomas, C.L., Cassady, J.C., & Finch, W.H. (2018). Identifying severity standards on the cognitive test anxiety scale: Cut score
determination using latent class and cluster analysis. Journal of Psychoeducational Assessment, 36(5), 492–508.
Wang, W.L., Sung, Y.T., Sung, F.C., Lu, T.H., & Li, C.Y. (2008). Low birth weight, prematurity, and paternal social status:
Impact on the basic competence test in Taiwanese adolescents. Journal of Pediatrics, 153, 333–338.
Wren, D.G., & Benson, J. (2004). Measuring test anxiety in children: Scale development and internal construct validation.
Anxiety, Stress, and Coping, 17, 227–240.
Yeh, T.K., Chang, C.Y., Hu, C.Y., Yeh, T.G., & Lin, M.Y. (2009). Association of catechol-O-methyltransferase (COMT)
polymorphism and academic achievement in a Chinese cohort. Brain and Cognition, 71(3), 300–305.
Yin, M., Sims, J., & Cothran, D. (2012). Scratching where they itch: Evaluation of feedback on a diagnostic English
grammar test for Taiwanese university students. Language Assessment Quarterly, 9(1), 78–104.
Zeidner, M. (1998). Test anxiety: The state of the art. New York: Plenum.