Thinking Skills and Creativity 40 (2021) 100831

Contents lists available at ScienceDirect

Thinking Skills and Creativity

journal homepage: www.elsevier.com/locate/tsc

The effect of creative thinking on academic performance:

Mechanisms, heterogeneity, and implication
Juan Yang, Xinhui Zhao *
Business School, Beijing Normal University, No.19 Xinjiekou Wai St., Beijing, 100875, PR China

A R T I C L E I N F O A B S T R A C T

Keywords: Based on data collected from 2,355 students from a high school in China, we find a positive
Creative thinking relation between creative thinking and academic performance. As two vital dimensions of crea­
Divergent thinking tive thinking, convergent thinking plays a more important role in students’ academic perfor­
Convergent thinking
mance than divergent thinking. There is evidence that creative thinking influences academic
Academic performance
Self-esteem
performance through students’ self-esteem and internal locus of control. Further research notes
Internal locus of control that the impact of creative thinking on academic performance is dissimilar for different groups. It
plays a more positive role in the academic performance of boys, low-grade students, and those
from disadvantaged backgrounds. Education policymakers should focus on cultivating students’
creative thinking, and intervention policies aimed at creative thinking for disadvantaged students
will produce higher returns. The main contributions of this study lie in unravelling the effect of
creative thinking on academic performance, exploring the mechanisms through which this rela­
tionship develops, and discussing the heterogeneity effects among different groups.

1. Introduction

During the past decades, creative thinking has been found to have a crucial influence on different areas, such as the sciences,
technology, economics and education (OECD, 2019). For education, Tan (2000) and Greenstein (2012) pointed out that creative
thinking was one of the higher-order thinking skills that was vital to help individuals succeed in school and in later career (Torrance,
1972). Therefore, most scholars have shown increasing enthusiasm for this issue, which has added to the popularity of the topic of
creative thinking and moved it from the fringe of education to the core of education (Craft, 2006). In many countries, fostering creative
thinking is emphasized in studies and policy documents (Tabach & Friedlander, 2017). A main motivation for this phenomenon is the
wide acceptance of the thought that building students’ creative thinking is important to their academic achievement (Akpur, 2020;
Anwar, 2012; Gralewski & Karwowski, 2012; Gregory et al., 2013; Huang et al., 2017; Sebastian & Huang, 2016).
Since the reform of the China National College Entrance Examination in 2017, the Ministry of Education has published policy
documents on the recruitment of colleges and universities to “increase the examination of creativity and innovation.” For Chinese
middle school students there is a requirement to “include experimental operations in the junior high school entrance examination
before 2023” to foster students’ creative thinking ability. This means that creative thinking will play a role in screening throughout the
Chinese education system.
In the context of the intense interest in this topic from Chinese policy makers, it is of profound policy significance to investigate the

* Corresponding author.
E-mail address: [email protected] (X. Zhao).

https://doi.org/10.1016/j.tsc.2021.100831
Received 16 February 2021; Received in revised form 5 April 2021; Accepted 15 April 2021
Available online 20 April 2021
1871-1871/© 2021 Elsevier Ltd. All rights reserved.
J. Yang and X. Zhao Thinking Skills and Creativity 40 (2021) 100831

association between creative thinking and academic performance for at least two reasons. First, creative thinking may be regarded as a
key indicator of educational evaluation in China (CPC Central Committee and State Council, 2020). Educators and parents can thus
promote children’s creative thinking to improve their future academic achievement. Second, schools can select students who might
have higher creative thinking based on examination results for admission or advancement to new education levels and institutions. Put
simply, creative thinking might contribute to the prediction of academic achievement (Mourgues et al., 2016). The selection of top
creative talent can also be performed according to academic performance (Wo et al., 2007).
Unfortunately, the few empirical studies that have focused on this topic have notable limitations. Researchers have either reported
the correlation between creative thinking and academic performance (Anwar, 2012; Gralewski & Karwowski, 2012; Niaz et al., 2000;
Powers & Kaufman, 2004) and rarely controlled for other important factors (e.g., demographic variables and family background) or
overlooked the differentiation among different subgroups. Additionally, most scholars have limited their research to the direct in­
fluence of creative thinking on academic performance (Ai, 1999; Cai & Zhu, 2007; Deng & Zhang, 2011; Zheng & Xiao, 1983; Zhou
et al., 2006) and neglected the indirect influence. To the best of our knowledge, no research has explored the mechanism of the as­
sociation between creative thinking and academic performance in China. This is important because creative thinking may affect ac­
ademic performance by affecting students’ other outcomes.
The main purpose of this paper is to explore the association between creative thinking and high school students’ academic per­
formance when the demographic variables (e.g., gender, siblings, grade level) and family background are controlled. An additional
goal is to investigate the mechanism through which this association develops. The third goal is to examine the possible differences in
this association among different groups. In general, we examine whether the postulated relationship and mechanism between creative
thinking and academic performance is consistent among different genders, grade levels, and family backgrounds. These demographic
variables are used to determine whether they have heterogeneity effects on creative thinking and academic performance and to control
their effects on the association between creative thinking and academic performance.
Given the aims of this paper, the specific research questions we solve are as follows:
Research Question 1: What is the relationship between creative thinking and academic performance?
Research Question 2: What is the mechanism through which creative thinking affects academic performance?
Research Question 3: Is the association between creative thinking and academic performance dissimilar for different groups?
The paper proceeds as follows. Section 2 gives a brief review of the previous literature. Section 3 describes the data and outlines the
empirical strategy used in this paper. Section 4 shows the main results. Section 5 provides a detailed discussion. Section 6 presents the
limitations and further research. Section 7 draws final conclusions.

2. Literature review

2.1. Creative thinking and students’ academic performance

Creative thinking is traditionally described as the ability to detect previously unidentified relationships and produce novel and
original experiences as a new pattern (Davut, 2008). A recent update to this definition considers creative thinking the skill to generate,
evaluate and improve thought that can bring about novel solutions, promotion of knowledge and influential expressions of imagination
(OECD, 2019).
The association between creative thinking and academic performance is a classic topic. However, there are ambiguous or con­
tradictory results in the existing literature. Previous research studies can be divided into three groups on the basis of their results with
respect to the way creative thinking relates to academic performance. Some research has reported a positive association between
creative thinking and academic performance (Akpur, 2020; Anwar, 2012; Gralewski & Karwowski, 2012; Huang et al., 2017; Niaz
et al., 2000; Powers & Kaufman, 2004; Sebastian & Huang, 2016). For example, Gajda et al. (2017) reported a meta-analysis over one
hundred articles carried out since the 1960s estimating the association between creativity thinking and academic performance and
concluded that the average correlation between creative thinking and academic performance was 0.22. This correlation was not
changed over time but higher when creativity tests were employed to measure creative thinking and when standardized tests rather
than GPA were used to measure academic performance.
Other studies have shown little or no relation (Ai, 1999; Grigorenko et al., 2009; Naderi et al., 2009; Olatoye et al., 2010; Zheng &
Xiao, 1983). By testing thinking exercises for 812 high school students, Zheng and Xiao (1983) discovered that students’ creativity was
moderately related to their academic achievement scores. Based on a survey of a private preparatory school, Grigorenko et al. (2009)
assessed creativity through a creative writing task and a scientific reasoning task. They found that the indicators of creativity did not
demonstrate predictive validity with regard to grade point average, suggesting that there is no correlation between creativity and
academic performance. Furthermore, creative thinking (divergent thinking) was not found to be associated with intelligence (Wallach,
1985).
However, other studies have reported negative associations. Anderson et al. (1969) randomly selected approximately 50 students
from each grade in grades nine through twelve to examine the association between students’ creativity, intelligence and academic
performance. They found that half of the significant correlations between the different creativity factors and intelligence and academic
performance were negative. Ademola et al. (2010) administered the Nicolas Holt Creativity Test to 235 fin. l-year higher education
students and revealed a negative, non-significant association between creativity and students’ academic achievement.
Some scholars have identified all three associations within the same research. Gralewski and Karwowski (2012) studied almost 600
participants from 34 high schools all over Poland and noted different relationships between academic performance and creativity in
different schools. In some schools, the association was significant positive, while in other schools, there was no relation or negative. In

2
J. Yang and X. Zhao Thinking Skills and Creativity 40 (2021) 100831

schools where the role of intelligence is relatively small, creativity is most important for academic performance, while the influence of
creativity on academic performance is lower when intelligence is a critical factor. This is because some teachers focus more on teaching
by building students’ creative thinking, while others are more analytical and traditional.
In terms of different subjects, Sebastian and Huang (2016) applied the Program for International Student Assessment (PISA) 2012
data to examine the association between math-creative thinking competency and math performance. Their results showed that
math-creative thinking was positive associated with students’ academic performance. According to 187 primary school sixth graders,
Huang et al. (2017) found a significant positive relationship between scientific creativity and scientific achievement as well as between
mathematic creativity and mathematic achievement. Tong and Wo (2002) found that the association between senior students’ creative
thinking and writing achievement was positive correlated, but creative thinking was not associated with math achievement. When
estimating the association between creative thinking and different subjects’ performance, Wang (2007) found that creative thinking
was significantly and positively related to English writing and reading test scores. By conducting the Torrance Test for 196 university
students, Wang (2012) further shown that creative thinking was significantly related to attitudes towards writing or reading and the
time spent on writing or reading.
These contradictory findings might be attributed to the employ of different data and/or measures of creative thinking and academic
performance. Based on the perspectives of most of the aforementioned research, the following hypothesis is proposed:
Hypothesis 1. Creative thinking is associated with academic performance.
Creative thinking generally includes divergent thinking and convergent thinking (Hadar & Tirosh, 2019). In divergent thinking,
students generate numerous answers to a question that does not have a standard solution (Volle, 2017). In convergent thinking,
students determine how segments are integrated by identifying the critical elements of a problem (Sternberg & Lubart, 1999). Taking
mathematics as an example, Hadar and Tirosh (2019) note that divergent thinking requires thinking about various methods to deal
with problems, suggesting new solutions, and using mathematical thought in different situations, whereas convergent thinking re­
quires the integration of ideas to connect mathematical thought with other fields as a new basis for understanding mathematics and
even identifying broader mathematical thought. Colzato et al. (2013) showed that divergent thinking requires multiple novel ideas to
be developed in a situation where more than one answer is appropriate, such as a brainstorming session, which aims to generate as
many ideas on a specific problem as possible. Therefore, convergent thinking represents a process of providing one answer to a specific
question, stresses speed and relies on high logic and accuracy. Stamovlasis et al. (2015) applied structural equation modelling to assess
the direct and indirect effects of divergent and convergent thinking on middle school students’ academic performance in chemistry.
They found that these two types of thinking sufficiently explained students’ ability to understand chemical phenomena. Wo et al.
(2006) explored the development of creative thinking in teenagers with 3311 students from six middle schools and found that the
divergent thinking and convergent thinking of the students in the key class were all higher than those in the ordinary class. In addition,
there is a stable ascending developmental trend of convergent thinking among adolescents; specifically, convergent thinking advances
more rapidly in the junior school stage, with a milder development trend in the senior school stage.
On the basis of the existing studies, we introduce the following hypotheses:
Hypothesis 1a. The dimension of creative thinking that involves convergent thinking is associated with academic performance.
Hypothesis 1b. The dimension of creative thinking that involves divergent thinking is associated with academic performance.

2.2. The mediating effects of self-esteem and locus of control

As mentioned above, some studies have estimated the association between creative thinking and academic performance, but the
conclusions are far from definite. These articles all advocate that more studies are needed to examine the association between creative
thinking and academic performance. On the one hand, creative thinking has an extremely important impact on noncognitive ability
(Mourgues, 2016), such as self-esteem and locus of control. On the other hand, most literature has demonstrated associations between
these noncognitive variables and academic performance (Arshad et al., 2015; Baumeister et al., 2003; Gifford et al., 2006; Kaiser,
1975; Prociuk & Breen, 1974; Pullmann & Allik, 2008; Zuffianò et al., 2013). Under these circumstances, possible mediation paths
should be explored to analyse the complicated associations between creative thinking and academic performance.

2.2.1. Creative thinking, self-esteem and academic performance

Since Coopersmith (1967) began investigating the association between creative thinking and self-esteem, empirical studies on the
relationship between the two have varied greatly, ranging from a moderate negative association among Chinese students (Lau et al.,
2004) to a powerful and positive association among US preschoolers (Kemple et al., 1996). However, most research has reported
positive associations within the 0.15− 0.30 range (Cantero et al., 2016; Goldsmith & Matherly, 1988; Rank et al., 2009; Wang & Wang,
2016). Based on a meta-analysis, Deng and Zhang (2011) found the correlation between creative thinking and self-esteem to be 0.289.
Barbot (2020) further explored the multivariate relationship between creative thinking in three domains and self-esteem in seven
domains among 170 adolescents and suggested that creativity-based interventions could support the development of positive
self-esteem in adolescence.
Self-esteem has also been related to academic performance (Baumeister et al., 2003), students may try to perform best at school to
keep a general perfect image of themselves (Zuffianò et al., 2013). Some research has investigated the association between students’
self-esteem and academic performance, and the results have shown that self-esteem is strongly associated with academic performance
(Arshad et al., 2015). Based on a large-scale study in Estonia, Pullmann & Allik (2008) also found that the association between

3
J. Yang and X. Zhao Thinking Skills and Creativity 40 (2021) 100831

self-esteem and academic performance was significantly positive. Zuffianò et al. (2013) further suggested that self-esteem was lightly
related to the performance of 11-year-old students but not 13-year-old students.
Earlier literature has recorded that self-esteem is associated with creative thinking and academic performance. However, no
previous research has examined the association between the three factors. It thus remains unknown whether creative thinking affects
academic performance through self-esteem. Therefore, we formulate the following hypothesis:
Hypothesis 2a. Self-esteem may be a mediating factor in the pathway from creative thinking to academic performance.

2.2.2. Creative thinking, locus of control and academic performance

Most studies have explored differences between the creative thinking of students with internal locus of control (LOC) and those
with external LOC. Torrance (1971) noted that creative students are likely to perform better when external strength is not as general; a
potential reason is that creative students are likely to be dependent on the internal LOC. Glover and Sautter (1976) showed that
students with internal LOC performed better on the flexibility scale and originality scale of the creative thinking test. A similar
conclusion was found among high school students in China (Li & Qu, 1998). Based on 171 university students, Pannells and Claxton
(2008) identified a significant association between creative thinking and internal LOC.
Students with internal LOC tend to have not only higher creative thinking but also better academic performance. Prociuk and
Breen’s (1974) research suggested that the internal LOC was positively related to academic performance, while the external LOC was
negatively related to academic performance. This may be because students with internal LOC obtain better academic achievement and
attribute their academic performance to internal reasons rather than destiny, fortune or other strong factors used by students with
external LOC (Kaiser, 1975). Based on a survey of 158 undergraduate students, Melissa and Jason (2016) found that academic per­
formance improves with higher internal LOC.
One of the purposes of this paper is to explore the possible relationships among creative thinking, the locus of control and academic
performance. In consideration of the high importance attached to internal LOC in creativity and academic success (e.g., Prociuk &
Breen, 1974; Torrance, 1971), it is conjectured that internal LOC may play a vital role in the association between creative thinking and
academic performance. Thus, we put forward the following hypothesis:
Hypothesis 2b. Internal locus of control may be a mediating factor in the pathway from creative thinking to academic performance.

2.3. Creative thinking and academic performance among different subgroups

According to the discussion above, we can observe that the association between creative thinking and academic performance is a
longstanding and largely unresolved problem (Gajda et al., 2017). What is the possible answer for the absence of an eventual
conclusion in existing literature on creative thinking, and how is creative thinking associated with academic performance? One po­
tential reason is that the association between creative thinking and academic performance may be various for different groups (e.g.,
gender, grade level, family background). However, most existing literature has concentrated only on whether there is an association
between creative thinking and academic performance without focusing on whether this association may be various among different
subgroups (Naderi et al., 2009).
Few studies have explored gender differences in the association between creative thinking and academic performance (Ai, 1999;
Nori, 2002; Naderi et al., 2010). Specifically, Ai (1999) investigated 2,264 students randomly selected from 68 schools in Spain and
found differences among different genders in creative thinking related to academic achievement. These differences might be defined as
the different gender roles of boys and girls. Another study of 306 high school students explored the association between creative
thinking and academic achievement (Nori, 2002) and concluded no significant association between the two, but the correlations were
different for boys and girls. Naderi et al. (2010) presented empirical evidence for the association between creative thinking and ac­
ademic performance; they also found that this association appears to differ by gender. Taken together, the hypothesis is proposed as
follows:
Hypothesis 3a. The association between creative thinking and academic performance is dissimilar for boys and girls.
For different grade levels, Cai and Zhu (2007) employed Williams’ Creativity Assessment Packet to test primary and middle school
students’ creative thinking and found that a creative inclination in students was related to Chinese and mathematics achievement.
Anwar et al. (2012) administered the Torrance Test of Creative Thinking to 256 middle school students to measure the creative potential
of participants and provide empirical evidence for the positive association between creative thinking and academic performance. But
this association may change when different levels of academic achievement are tested and different measures of creative thinking are
used. Based on a sample of 342 test takers of the Graduate Record Examinations (GRE), Powers and Kaufman (2004) concluded that
creativity is correlated with GRE scores. By investigating 453 college students in 4 key universities in China, Zhou et al. (2006) showed
that the scores of creative thinking among students in different majors were significantly different. The creative thinking scores of
students majoring in arts and engineering were greater than those of students majoring in medicine and science. Gajda et al. (2017)
also revealed higher correlations between creative thinking and academic performance in the students from lower grade levels
compared to those from higher grade levels. One potential interpretation is that education becomes more specialized at high school
and university. In light of the above arguments, we propose the hypothesis as follows:
Hypothesis 3b. The association between creative thinking and academic performance is dissimilar for students from different grade
levels.

4
 J. Yang and X. Zhao
5

Fig. 1. Kernel Density Distribution of Standardized Scores.

Note: The solid line is the normal distribution curve, and the dotted line is the kernel density estimation curve.

Thinking Skills and Creativity 40 (2021) 100831

J. Yang and X. Zhao Thinking Skills and Creativity 40 (2021) 100831

Additionally, the research has presented that differences in students’ family background (e.g., urban or rural, parents’ education)
are significant in interpreting their academic achievement (Bravo et al., 2015). Based on a meta-analysis of participants in 41
countries, Chiu and Xihua (2008) concluded that students with more family resources and fewer competing siblings performed better
in mathematics. Research on family background differences in relation to creative thinking is an important and complex issue in the
research of creativity. Mauricio et al. (2018) showed that as the rise of socioeconomic status, so did creative thinking competency, with
standardized test results following the same tendency as applied creative thinking tests. One of the aims of this paper is to investigate
whether creative thinking is associated with academic performance when controlling for family background variables and to explore
whether the conclusions present significant differences according to different family backgrounds. Therefore, we present the following
hypothesis:
Hypothesis 3c. The association between creative thinking and academic performance is dissimilar for students from different family
backgrounds.

2.4. The present study

To summarize, very few articles have estimated the association between creative thinking (including convergent thinking and
divergent thinking) and students’ academic achievements when controlling for other vital variables (e.g., gender, age or grade levels,
parents’ education). Additionally, the differences in students’ demographic variables and family backgrounds have not been
considered. To date, only a few articles (Ai, 1999; Nori, 2002; Naderi et al., 2010) have examined gender differences, and only one
study (Mauricio et al., 2018) has explored family background differences in the association between creative thinking and academic
achievement. To our knowledge, there is also no research exploring the possible mechanism, that is, the mediating role of self-esteem
or locus of control, in the association between creative thinking and academic performance.
Within this context, our present study aims to fill in these gaps. The main contributions of this paper are as follows. First, both
convergent and divergent thinking in creative thinking are discussed to attain a full picture of creative thinking and its associations to
academic performance. Second, we examine whether the relations between creative thinking and academic performance vary among
different subgroups to provide advice for intervention policies in the future. To be specific, we investigate heterogeneity effects across
three aspects: gender, grade level and family background. Our third contribution is identifying how creative thinking results in
different outcomes for students. By which paths can creative thinking improve students’ academic performance? Identifying the
mechanism is essential when introducing policies to improve academic achievement. This paper gains an insight into this issue by
using self-esteem and internal locus of control, to extend previous research and show the impact of these mediating factors on academic
performance as well as the impact of creative thinking on these mediating factors. Moreover, our conclusions might be helpful in
providing a scientific plan to foster middle school students’ creative thinking in China.

3. Method

3.1. Participants

The participants were all from a high school in a city in northern China in the 2019–2020 school year. A total of 2,611 individuals
from Grade 10 to Grade 12 originally completed our questionnaires. However, 256 students were excluded from the analyses because
of the poor reliability of their answers (i.e., their responses were incomplete or they offered contradictory responses to reversed items).
The final data consisted of 2,355 students (girls: 57.92 %, boys: 42.08 %), and the students’ ages ranged from 15 to 19 years
(M = 16.46, SD = 0.97).

3.2. Measures

3.2.1. Dependent variable

The dependent variable, academic performance, is measured by exam test scores provided by the school administration office.
China generally has 3 years of high school education (10th to 12th grades). At each grade level, all students should major in three core
subjects, that is, Chinese, mathematics, and English. They may also take biology, physics, chemistry, history, politics, or geography as
elective subjects. All students use the same test for each subject if they are at the same grade level, so the standardized academic
performance at the grade level are comparable. In this paper, we restrict our study to academic performance in the three core subjects.
Additionally, this paper normalizes academic performance within the subject and grade level to facilitate explanation (0 is the mean
value and 1 is the standard deviation). Fig. 1 presents the kernel density estimate of standardized academic performance, from which
the distribution of test scores can be observed clearly.

3.2.2. Independent variable

To evaluate creative thinking, the core independent variable in this study, we employ the Chinese adaptation of the Creative

6
J. Yang and X. Zhao Thinking Skills and Creativity 40 (2021) 100831

Thinking Questionnaire (Luo & Meng, 2005)1, which has high reliability and validity in the Chinese context. Cronbach’s alpha for this
questionnaire was 0.702 in the present study. Specifically, this questionnaire includes 8 indicators: understanding key concepts (3
items, e.g., “I have a keen awareness of the key to solving problems”), synthesis (3 items, e.g., “I often classify the questions according
to the type of questions when I study”), summarily explaining (3 items, e.g., “I can explain some complicated things in a few words”),
evaluation (3 items, e.g., “I can usually judge a piece of work or invention correctly”), association (3 items, e.g., “I often have pictures
or melodies in my mind when I read poetry or prose”), originality (3 items, e.g., “I can always think of various methods to solve the
same problem”), pointing towards the future (3 items, e.g., “I am very curious about life and the mystery of the universe”), and insight
(3 items, e.g., “I can often see how things might relate to each other”). Participants answer these 24 items on a 5-point scale (1=never,
5=always). Table A1 is the rotation component matrix which presents factor loadings of each item. Factor 1 represents the process of
integrating ideas to seek a solution (Cronbach’s α = 0.696), which has the biggest variance contribution rate (37.265 %), demon­
strating that it is a key factor for composing creative thinking. Factor 2 represents generating multiple ideas to find a solution
(Cronbach’s α = 0.714). According to the research of Hadar and Tirosh (2019), Sternberg and Lubart (1999) and Volle (2017), we
define the former as convergent thinking and the latter as divergent thinking. Finally, the total score of creative thinking is obtained by
factor analysis. A higher score represents higher creative thinking. We also analyse the impact of the two dimensions of creative
thinking on academic performance in this paper.

3.2.3. Mediator variables

One of the mediator variables in our paper, self-esteem, is measured using 10-item scales developed by Rosenberg (1965) (e.g., “In
general, I am satisfied with myself”). Students answer a 5-point Likert scale ranging from “strongly disagree” (1) to “strongly agree”
(5). Higher scores represent greater self-esteem. Internal locus of control is another mediator variable measured using 10 items adapted
from Burger’s Locus of Control Scale (Burger, 1986) (e.g., “I can learn anything if I want to”). Students answer a 7-point Likert scale.
Higher scores represent greater internal locus of control. These two scales have been presented to have adequate reliability and validity
in mainland Chinese high school students (Deng & Zhang, 2011; Li & Qu, 1998). In the current research, the Cronbach’s alpha was
0.835 for the self-esteem scale and 0.715 for the locus of control scale. The self-esteem and internal locus of control scores were
obtained by factor analysis.

3.2.4. Control variables

This paper includes the following control variables. Gender is recorded as a dummy variable (1 if a student is a boy and 0 is a girl).
Hukou is also measured as a dummy variable (1 if a student is registered as rural and 0 is registered as urban). Grade level is recorded as
“grade 10 (1),” “grade 11 (2),” or “grade 12 (3).” Only child is recorded as a dummy variable (1 if a student is the only child and 0 is not
the only child in the family). Father’s years of education is recorded as “illiterate (0 years),” “primary school (6 years),” “middle school
(9 years),” “high school (12 years),” “3-year college (15 years),” “4-year college (16 years),” “bachelor (19 years),” and “postgraduate
(22 years).” In addition, we controlled for class fixed effects to alleviate the heterogeneity between different classes.

3.3. Procedure

The students were required to answer a questionnaire that measured creative thinking, self-esteem and locus of control. We also
collected their demographic characteristics, including gender, hukou (urban or rural), grade level, siblings, and parents’ education. All
the surveys were distributed in the regular classrooms after the school principals and teachers approved the research, and all students
offered written informed consent. The research assistants interpreted the instructions for the questionnaires and emphasized that the
questionnaires were not tests and that there were no standard answers. During the assessment, if participants had any questions, they
could raise their hand and ask the research assistants. Finally, using the student number (or students’ names), we matched the data
obtained by both demographic characteristics and questionnaires with the final examination performance. Therefore, those surveys
were not employed anonymously, but we guaranteed confidentiality.

3.4. Identification strategy

All statistical analyses were conducted with STATA 15.1 and SPSS 17.0 software. First, as shown in Table 1, Pearson’s correlation
analyses and descriptive statistics were computed for all original variables. Second, ordinary least squares regression (OLS regression)
was used to investigate the impact of creative thinking on academic performance after controlling variables such as gender, hukou, only
child, grade level and father’s years of education. Third, the mediation effect model was employed to test the mechanism by which
creative thinking affects academic performance. In addition, heterogeneity analysis was used to study the relationships and mediating
effects between creative thinking and academic performance among different subgroups.

3.4.1. Ordinary least squares regression

One of the aims of our study is the association between creative thinking and academic performance. Therefore, ordinary least
squares regression (OLS regression) was applied to estimate the effect of creative thinking on the scores for Chinese, mathematics and

1
In fact, creative thinking competency is both domain general and domain specific (Hong & Milgram, 2010). In the present study, the ques­
tionnaire we employed focuses on measuring domain-general creative thinking.

7
 J. Yang and X. Zhao
Table 1
Means, standardized deviation, and correlations for all variables.
Variables 1 2 3 4 5 6 7 8 9 10 11 12 13

1. Chinese –
2. Mathematics 0.52*** –
3. English 0.56*** 0.51*** –
4. Creative thinking 0.15*** 0.15*** 0.11*** –
5. Convergent thinking 0.13*** 0.17*** 0.10*** 0.46*** –
6. Divergent thinking 0.29*** 0.27*** 0.14*** 0.05*** 0.09*** –
7. Self-esteem 0.14*** 0.14*** 0.19*** 0.15*** 0.11*** 0.01
8

–
8. Internal LOC 0.06*** 0.03 0.08*** 0.08*** 0.05** 0.03 − 0.01 –
9. Gender (boy = 1) − 0.05** − 0.04** − 0.07*** − 0.15*** − 0.10*** 0.02 0.04 − 0.31*** –
10. Hukou (rural = 1) − 0.16*** − 0.18*** − 0.14*** 0.86*** 0.47*** 0.05** 0.12*** − 0.07*** 0.12*** –
11. Only child 0.10*** 0.09*** 0.05** 0.88*** 0.33*** 0.04** 0.14*** − 0.07*** 0.13*** − 0.51*** –
12. Grade levels − 0.01 − 0.05*** − 0.01 − 0.08*** − 0.03 − 0.05*** − 0.17*** 0.21*** − 0.11*** − 0.08*** − 0.06*** –
13. Feduc 0.06*** 0.09*** 0.09*** 0.41*** 0.46*** − 0.01 − 0.00 0.04 0.04** 0.44*** 0.28*** 0.02 –
Mean 95.86 78.16 84.32 77.90 37.90 40.00 24.21 32.12 0.42 0.86 0.91 1.92 9.10
Std. Dev 10.90 25.70 23.21 10.55 5.83 6.29 5.50 4.55 0.49 0.35 0.29 0.76 2.44

Note: internal LOC = internal locus of control; feduc = father’s years of education. N = 2,355. All variables are original values. *** p < 0.001, ** p < 0.01, and * p < 0.05.

Thinking Skills and Creativity 40 (2021) 100831

J. Yang and X. Zhao Thinking Skills and Creativity 40 (2021) 100831

Table 2
The regression analysis of academic performance.
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Chinese Mathematics English

0.198*** 0.104*** 0.133***

Creative thinking
(0.031) (0.032) (0.031)
0.111*** 0.129*** 0.160***
Convergent thinking
(0.018) (0.018) (0.017)
0.065*** − 0.003 − 0.001
Divergent thinking
(0.018) (0.018) (0.018)
− 0.335*** − 0.329*** − 0.318*** 0.256*** 0.245*** 0.272*** − 0.417*** − 0.431*** − 0.398***
Gender (boy = 1)
(0.038) (0.038) (0.038) (0.037) (0.037) (0.037) (0.038) (0.038) (0.038)
− 0.116** − 0.121** − 0.120** 0.006 0.006 0.001 − 0.092* − 0.093* − 0.098*
Hukou (rural = 1)
(0.059) (0.059) (0.060) (0.055) (0.054) (0.055) (0.056) (0.055) (0.056)
Only child 0.042 0.046 0.035 0.040 0.050 0.035 0.029 0.041 0.023
(0.064) (0.064) (0.065) (0.061) (0.060) (0.061) (0.065) (0.063) (0.065)
Grade 0.006 0.004 0.010 − 0.009 − 0.015 − 0.007 − 0.009 − 0.016 − 0.006
(0.025) (0.025) (0.025) (0.025) (0.025) (0.025) (0.025) (0.024) (0.025)
0.010 0.011 0.013 0.002 0.000 0.005 0.021*** 0.019** 0.024***
Father’s years of education
(0.008) (0.008) (0.008) (0.008) (0.008) (0.008) (0.008) (0.008) (0.008)
Constant 0.140 0.131 0.114 − 0.131 − 0.113 − 0.156 0.153 0.174 0.123
(0.174) (0.171) (0.175) (0.176) (0.175) (0.175) (0.180) (0.175) (0.178)
Adjusted R2 0.213 0.213 0.205 0.266 0.280 0.263 0.274 0.293 0.268

Note: ***, **, and * indicate significance at the 1%, 5%, and levels respectively. The table above also controls class-fixed effects. Standard errors
shown in parentheses are clustered at the individual level.

English. The basic equation is as follows:

Yis = α1 + α2 creativityis + α’3 Xis’ + φs + εis , (1)

where Yis represents the performance of Chinese, mathematics and English; creativityis represents creative thinking, convergent
thinking or divergent thinking; and α2 refers to the effect of creative thinking or the two dimensions of creative thinking on students’
academic achievement. Xisn’
refers to a series of controlling variables that influence students’ academic performance, consisting of
gender, hukou (rural = 1), grade levels, siblings, and father’s years of education. φs is the fixed effect at the class level, and εis is the
error term.

3.4.2. Recentred influence function

To examine the coefficients for the impact of creative thinking on different quantiles of academic performance, this paper uses the
regression of the recentred influence function (RIF) proposed by Firpo et al. (2009). We add the influence function and the distri­
butional statistic to obtain the recentred influence function:
RIF(Yisb ; v) = v(Fi ) + IF(Yisb ; v), (2)

where IF(Yisbi ; v) is the influence function of Yisb . The RIF regression is an unconditional quantile regression when v represents the
quantile. Specifically, RIF(Yisb ; v) can be generalized to the linear model:
RIF(Yisb ; v) = Xi θvi , (3)

where Yisb represents standardized academic achievement, v represents the quantile, Xi is the series of independent variables, and θ can
be obtained by the estimation of OLS regression. The above identification strategy can estimate the effect of variations in the distri­
bution of independent variables (e.g., creative thinking) on quantiles of the unconditional distribution of an outcome variable (e.g.,
academic performance).

3.4.3. Mediating effect model

We employ the mediation effect model to identify the pathway by which creative thinking affects academic performance. The
regression equations are as follows:

mis = β1 + β2 creativityis + β’3 Xis’ + φs + εis (4)

Yis = γ1 + γ 2 creativityis + γ3 mis + γ ’4 Xis’ + φs + εis, (5)

where mis refers to the mediating variable (i.e., self-esteem and internal locus of control). α2 refers to the total effect of creative
thinking on academic performance. β2 refers to the effect of creative thinking on self-esteem or internal locus of control. γ 3 refers to the
effect of self-esteem or internal locus of control on academic performance. β2 ∙γ3 is regarded as the indirect effect or called mediating
effect, and γ2 refers to the direct effect of creative thinking on academic performance. By calculating β2 γ 3 /α2 , the proportion of the

9
 J. Yang and X. Zhao
10

Thinking Skills and Creativity 40 (2021) 100831

Fig. 2. The estimation coefficient of the recentred influence function regression.
Note: The figure above controls students’ gender, hukou, siblings, grade, father’s years of education and class-fixed effect.
 J. Yang and X. Zhao
Table 3
Mechanism of creative thinking affecting academic performance.
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
Mediating variable Chinese Mathematics English

Panel A: Mediating variable = self-esteem

0.697*** 0.153*** 0.037 0.074**
Creative thinking
(0.036) (0.034) (0.034) (0.034)
0.401*** 0.086*** 0.104*** 0.142***
Convergent thinking
(0.021) (0.020) (0.019) (0.018)
0.226*** 0.046** − 0.028 − 0.025
Divergent thinking
(0.022) (0.018) (0.018) (0.018)
0.064*** 0.063*** 0.087*** 0.096*** 0.062*** 0.110*** 0.085*** 0.044** 0.106***
Self-esteem
(0.019) (0.019) (0.018) (0.018) (0.018) (0.018) (0.019) (0.019) (0.018)
0.044*** 0.025*** 0.020*** 0.067*** 0.025*** 0.025*** 0.059*** 0.018*** 0.024***
Mediating effect (Sobel test)
(0.014) (0.008) (0.005) (0.014) (0.008) (0.005) (0.014) (0.008) (0.005)
Adjusted R2 0.192 0.200 0.096 0.217 0.217 0.212 0.274 0.283 0.275 0.280 0.295 0.279
11

Panel B: Mediating variable = internal locus of control

0.511*** 0.201*** 0.083** 0.104***
Creative thinking
(0.038) (0.032) (0.033) (0.033)
0.387*** 0.119*** 0.126*** 0.155***
Convergent thinking
(0.021) (0.020) (0.019) (0.018)
0.116*** 0.063*** − 0.009 − 0.010
Divergent thinking
(0.022) (0.018) (0.018) (0.018)
− 0.006 − 0.020 0.018 0.041** 0.006 0.055*** 0.056*** 0.014 0.073***
Internal locus of control
(0.018) (0.019) (0.018) (0.018) (0.018) (0.018) (0.018) (0.019) (0.018)
− 0.003 − 0.008 0.002 0.021** 0.003 0.006*** 0.029** 0.005*** 0.008***
Mediating effect (Sobel test)
(0.009) (0.007) (0.002) (0.009) (0.007) (0.002) (0.010) (0.007) (0.003)

Thinking Skills and Creativity 40 (2021) 100831

Adjusted R2 0.103 0.168 0.038 0.213 0.213 0.205 0.268 0.280 0.266 0.276 0.294 0.273

Note: ***, **, and * indicate significance at the 1%, 5%, and levels respectively. The table above also controls students’ gender, hukou, siblings, grade, father’s years of education and class-fixed effects.
Standard errors shown in parentheses are clustered at the individual level.
J. Yang and X. Zhao Thinking Skills and Creativity 40 (2021) 100831

mediating effect accounting for the total effect can be obtained.

4. Main results

4.1. Effect of creative thinking on academic performance

The influence of creative thinking on students’ academic performance is the focus of our research. According to Eq. (1), Table 2
presents the OLS regression results controlling for gender, hukou, only child, grade level, and father’s years of education to test the
relationships between creative thinking and its two dimensions and academic performance, described in Hypotheses 1, 1a, 1b.
Consistent with the predictions, in Columns (1), (4) and (7), creative thinking has a significant positive impact on academic perfor­
mance in all subjects. It has a greater effect on the Chinese test score with an estimated coefficient of 0.198, which is significant at the
1% level, while it has a smaller effect on the mathematics test score, with an estimated coefficient of 0.104 at the 1% level of sig­
nificance. This result is in favour of Hypothesis 1. To test Hypothesis 1a, Columns (2), (5) and (8) present the impact of convergent
thinking on academic performance. Convergent thinking is also positively related to achievement in all subjects, with a particularly
greater effect on English test scores (estimated coefficient = 0.160, p < 0.01). In addition, Columns (3), (6) and (9) show the effect of
divergent thinking on academic performance. We can observe that divergent thinking is positively associated only with students’
Chinese test scores (estimated coefficient = 0.065, p < 0.01). This finding partially supports Hypothesis 1b.
Furthermore, because the association between creative thinking and academic performance may change when academic
achievements are at different levels (Anwar et al., 2012), we employ the recentred influence function (RIF) to research the effect of
creative thinking on different quantiles of academic performance. As presented in Fig. 2, creative thinking has a positive effect on
different quantiles of test scores. To be specific, there are two findings. First, the effect of creative thinking on the lowest and highest
quantiles of every subject’s score is small. Second, creative thinking has a greater impact on the mid-low quantile for Chinese and
mathematics test scores, while there is a greater impact on the mid-high quantile for English achievement. In other words, creative
thinking has a greater impact on students with middle academic performance. Similar findings are identified for the impact of
convergent thinking on academic performance in different quantiles. Convergent thinking also has a greater effect on students with
mid-low Chinese and mathematics achievement or mid-high English achievement. With regard to divergent thinking for the Chinese
test score, it has a greater effect on students with mid-high scores and the smallest effect on the highest quantiles. Since divergent
thinking has no significant effect on mathematics and English test scores at any quantile in the regression results, we will not discuss it
in detail. To some extent, this finding also verifies the above conclusion, that is, divergent thinking has a significant positive effect only
on the Chinese performance.

4.2. Mechanism of creative thinking affecting academic performance

To examine the mediation effect of creative thinking, convergent thinking and divergent thinking on academic performance
through self-esteem and internal locus of control, we performed a regression analysis based on a mediating effect model. The results are
shown in Table 32 . As the first step, we regress on self-esteem and internal locus of control based on Eq. (4). As expected, creative
thinking, convergent thinking and divergent thinking all have a significant positive influence on these mediating variables. As the
second step, we introduce self-esteem and internal locus of control to regress on academic performance based on Eq. (5).
In panel A of Table 3, taking the Chinese test score as an example first, the original effect of creative thinking on the Chinese test
score is positive in Table 2 (estimated coefficient = 0.198, p < 0.01). However, the effect of creative thinking decreases in Table 3
(estimated coefficient = 0.153, p < 0.01), which reports that creative thinking significantly increases self-esteem (the coefficient is
0.697, significant at the 1% level) by improving self-esteem to improve the Chinese test score. These findings indicate that self-esteem
might mediate the positive relationship between creative thinking and Chinese achievement. The mediating effect is 0.044, and the
proportion of the mediating effect in the total effect is 22.65 %. Similar findings are observed for the effect of convergent thinking and
divergent thinking on Chinese achievement through self-esteem. The mediating effect of the former is 0.025, accounting for 22.83 % of
the total effect, while the mediating effect of the latter is 0.020, accounting for 30.10 % of the total effect.
In terms of mathematics, the effect of creative thinking is no longer significant after including self-esteem in the model, while self-
esteem has a significant positive impact on mathematics achievement (estimated coefficient = 0.096, p < 0.01). This result shows
complete mediation of the association between creative thinking and mathematics achievement and suggests that self-esteem mediates
this relationship (Baron & Kenny, 1986). Additionally, the mediating effect of convergent thinking on mathematics achievement is
0.025, and the proportion of the mediating effect in the total effect is 19.38 %. For English achievement, self-esteem might mediate the
positive relationship between creative thinking and academic performance with a mediating effect of 0.059, accounting for 44.50 % of
the total effect. The mediating effect of self-esteem between convergent thinking and the English test score is 0.018, and the proportion
of the mediating effect in the total effect is 10.99 %.

2
Considering one reviewer’s suggestion, appendix Table A2 also shows the influence and mechanisms of creative thinking on the total score. We
can see that creative thinking and convergent thinking all have a significant positive impact on the total score, while divergent thinking has no
significant impact on the total score. Secondly, self-esteem influences the association between creative thinking (including convergent thinking and
divergent thinking) and the total score, while internal locus of control presents a nearly similar function but does not mediate the effect of
convergent thinking on the total score.

12
J. Yang and X. Zhao Thinking Skills and Creativity 40 (2021) 100831

Table 4
The association between creative thinking and academic performance among different genders.
(1) (2) (3) (4) (5) (6) (7) (8)
Self-esteem Chinese Mathematics English

Groups = gender

girl boy girl boy girl boy girl boy

0.673*** 0.711*** 0.149*** 0.162*** 0.000 0.075* 0.017 0.146***

Creative thinking
(0.047) (0.053) (0.043) (0.054) (0.044) (0.042) (0.041) (0.056)
0.107*** 0.012 0.106*** 0.097*** 0.123*** 0.054*
Self-esteem
(0.026) (0.029) (0.025) (0.028) (0.023) (0.032)
0.072*** 0.008 0.071*** 0.069*** 0.083*** 0.038*
Mediating effect (Sobel test)
(0.018) (0.022) (0.018) (0.022) (0.017) (0.023)
Adjusted R2 0.187 0.224 0.236 0.197 0.289 0.279 0.286 0.281

Note: ***, **, and * indicate significance at the 1%, 5%, and levels respectively. The table above also controls students’ gender, hukou, siblings, grade,
father’s years of education and class-fixed effects. Standard errors shown in parentheses are clustered at the individual level.

Although there is no significant association between divergent thinking and mathematics and English achievement in Table 2,
Sobel tests confirm the proposed mediating effects of self-esteem between divergent thinking and mathematics achievement (medi­
ating effect = 0.025, p < 0.01) and English achievement (mediating effect = 0.024, p < 0.01). Thus, our findings show support for
Hypothesis 2a.
In panel B of Table 3, inconsistent with the predictions, internal locus of control does not significantly mediate students’ creative
thinking and Chinese achievement, failing to provide full support for Hypothesis 2b. Specifically, the mediating effects are not sig­
nificant between Chinese achievement and creative thinking, convergent thinking, or divergent thinking.
Note that the significance of the effect of creative thinking on academic performance in other subjects decreased in the existence of
internal locus of control, providing evidence of a partial mediation effect. Indeed, Column (7) presents that internal locus of control
might mediate the positive association between creative thinking and mathematics achievement with a mediating effect of 0.021,
accounting for 20.31 % of the total effect. Column (10) also shows that internal locus of control might mediate the positive association
between creative thinking and English achievement. The mediating effect is 0.029, and the proportion of the mediating effect in the
total effect is 21.45 %. In addition, Sobel tests confirm the proposed mediating effects of internal locus of control between divergent
thinking and mathematics achievement (mediating effect = 0.006, p < 0.01) and English achievement (mediating effect = 0.008,
p < 0.01). When the internal locus of control is introduced in Columns (8) and (11), it does not mediate the association between
convergent thinking and mathematics achievement but mediates the association between convergent thinking and English achieve­
ment (mediating effect = 0.005, p < 0.01). Taken together, the findings do not fully in favour of the proposed mediating effect of
internal locus of control, thus only partially supporting Hypothesis 2b.

4.3. Heterogeneous effects

To examine whether the above effects of creative thinking on academic performance differ according to students’ characteristics,
we present the heterogeneity analysis in this section. Tables 4–6 show the regression analyses for academic performance as well as
analyses for self-esteem as a mediating factor separately for different genders, grade levels and family backgrounds3 . Due to limi­
tations of space, this section does not discuss the impact of convergent thinking and divergent thinking on academic performance or
another mediating variable, i.e., the mediating role of internal locus of control in the association between creative thinking and ac­
ademic performance. Appendix Tables A4–A6 also show the effects of convergent thinking on academic performance among different
groups, and the results are similar to the above; therefore, we will not discuss them again.
As shown in Table 4, creative thinking is more beneficial to boys’ Chinese achievement (estimated coefficient = 0.162, p < 0.01)
and has a lower positive effect for girls (estimated coefficient = 0.149, p < 0.01). The impact of creative thinking on mathematics
achievement is very slight for boys (estimated coefficient = 0.075, p < 0.1) and is still not significant for girls. There is a positive
association between creative thinking and English achievement, but only for boys (estimated coefficient = 0.146, p < 0.01).
When self-esteem is added as a mediating factor, the mediating effect differs between girls and boys. Specifically, girls show a
significant positive mediating effect between creative thinking and Chinese achievement. This result is not found for boys. In terms of
mathematics and English achievement, self-esteem mediates creative thinking and academic performance for both girls and boys, but
with different coefficient sizes and/or significance. For girls, the mediating effect is 0.071 with 1% significance for mathematics
achievement, while the mediating effect is 0.069 with 1% significance for boys. Additionally, for girls, the mediating effect is 0.083
with 1% significance for English achievement, while it is 0.038 with only 10 % significance for boys.
As shown in Table 5, Hypothesis 1 is again partially supported, and creative thinking still has a significant positive impact on the
Chinese achievement of 10th- and 11th-grade students. The effect of creative thinking is 0.141 for the former and 0.116 for the latter,

3
Appendix Table A3 shows the heterogeneity of creative thinking on the total score. It also plays a more positive role in the total score of boys and
those from disadvantaged backgrounds.

13
 J. Yang and X. Zhao
Table 5
The association between creative thinking and academic performance among different grade levels.
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
Self-esteem Chinese Mathematics English

Groups = grade levels

10th Grade 11th Grade 12th Grade 10th Grade 11th Grade 12th Grade 10th Grade 11th Grade 12th Grade 10th Grade 11th Grade 12th Grade

0.667*** 0.642*** 0.717*** 0.141** 0.116** 0.066 0.028 − 0.016 − 0.121** − 0.003 0.016 0.032
14

Creative thinking
(0.068) (0.055) (0.070) (0.058) (0.050) (0.061) (0.047) (0.041) (0.059) (0.047) (0.045) (0.064)
0.034 0.045 0.033 0.005 0.058*** 0.137*** 0.026 0.077*** 0.071*
Self-esteem
(0.030) (0.027) (0.035) (0.026) (0.022) (0.035) (0.025) (0.025) (0.036)
0.023 0.029 0.023 0.003 0.037** 0.098*** 0.017 0.050** 0.051**
Mediating effect (Sobel test)
(0.020) (0.018) (0.027) (0.016) (0.016) (0.027) (0.017) (0.017) (0.026)
Adjusted R2 0.193 0.195 0.245 0.366 0.416 0.353 0.588 0.564 0.488 0.555 0.557 0.444

Note: ***, **, and * indicate significance at the 1%, 5%, and levels respectively. The table above also controls students’ gender, hukou, siblings, grade, father’s years of education and class-fixed effects.
Standard errors shown in parentheses are clustered at the individual level.

Thinking Skills and Creativity 40 (2021) 100831

J. Yang and X. Zhao Thinking Skills and Creativity 40 (2021) 100831

Table 6
The association between creative thinking and academic performance among different family backgrounds (hukou, only child, father’s years of
education).
(1) (2) (3) (4) (5) (6) (7) (8)
Self-esteem Chinese Mathematics English

Panel A: Groups=hukou
urban rural urban rural urban rural urban rural
0.784*** 0.685*** 0.145*** 0.204** 0.029 0.087 0.053 0.176**
Creative thinking
(0.101) (0.039) (0.037) (0.091) (0.037) (0.090) (0.037) (0.085)
0.053** 0.145*** 0.101*** 0.062 0.094*** 0.048
Self-esteem
(0.021) (0.042) (0.020) (0.048) (0.021) (0.051)
Mediating effect 0.036** 0.113*** 0.069*** 0.049 0.064*** 0.038
(Sobel test) (0.015) (0.041) (0.015) (0.042) (0.015) (0.040)
Adjusted R2 0.279 0.189 0.195 0.419 0.275 0.323 0.264 0.432

Panel B: Groups = only child

non-only non-only non-only non-only
only child only child only child only child
child child child child
0.729*** 0.694*** 0.139 0.159*** − 0.007 0.043 0.170 0.060*
Creative thinking
(0.120) (0.038) (0.126) (0.036) (0.113) (0.036) (0.120) (0.036)
0.146** 0.050** 0.091 0.094*** 0.083 0.084***
Self-esteem
(0.072) (0.020) (0.062) (0.020) (0.070) (0.020)
Mediating effect 0.107** 0.035** 0.066 0.066*** 0.060 0.058***
(Sobel test) (0.052) (0.014) (0.046) (0.015) (0.052) (0.014)
Adjusted R2 0.354 0.188 0.396 0.212 0.472 0.269 0.462 0.272

Panel C: Groups = father’s years of education

high- high- high- high-
low-educated low-educated low-educated low-educated
educated educated educated educated
father father father father
father father father father
0.669*** 0.702*** 0.145*** 0.193*** − 0.093 0.054 0.031 0.077**
Creative thinking
(0.077) (0.041) (0.039) (0.074) (0.071) (0.039) (0.073) (0.039)
0.053** 0.103** 0.109*** 0.092*** 0.055 0.093***
Self-esteem
(0.022) (0.042) (0.039) (0.021) (0.043) (0.021)
Mediating effect 0.037** 0.069** 0.073*** 0.064*** 0.037 0.065***
(Sobel test) (0.016) (0.030) (0.029) (0.016) (0.029) (0.016)
Adjusted R2 0.242 0.188 0.204 0.285 0.345 0.272 0.356 0.261

Note: ***, **, and * indicate significance at the 1%, 5%, and levels respectively. The table above also controls students’ gender, hukou, siblings, grade,
father’s years of education and class-fixed effects. Standard errors shown in parentheses are clustered at the individual level.

which is significant at the 1% level and the 5% level, respectively. However, this finding is not true for 12th-grade students, who
instead show a nonsignificant relation between creative thinking and Chinese achievement and even reveal a negative relation be­
tween creative thinking and mathematics achievement (estimated coefficient= − 0.121, p < 0.05). Columns (7) - (12) show that
creative thinking does not have any significant positive effects on mathematics and English achievement among different grade levels.
Columns (1) - (3) show that creative thinking is associated with self-esteem among different grade levels significantly at the 1%
level. This finding suggests that self-esteem might mediate the relationship between creative thinking and academic performance
among different grade levels. The Sobel tests partially confirm this proposed mediating effect. The mediating effect of self-esteem is
0.037 and 0.098 for 11th-grade and 12th-grade students’ mathematical achievement, respectively, and 0.050 and 0.051 for 11th-grade
and 12th-grade students’ English achievement, respectively.
To test the heterogeneous effects of creative thinking on academic performance among students from different family backgrounds,
Table 6 presents the association between creative thinking and academic performance for students with different socioeconomic
statuses (e.g., hukou, only child, father’s years of education4). Academic performance is also regressed on self-esteem in the model to
examine the mediation effect. Columns (1), (3), (5), and (7) describe the analysis for students from advantaged family backgrounds (i.
e., urban, only child, highly educated father), and Columns (2), (4), (6), and (8) describe the analysis for students from disadvantaged
family backgrounds (i.e., rural, non-only child, low-educated father). Note that mathematics and creative thinking do not predict
mathematics test scores. The findings indicate that there is no significant positive association between creative thinking and math­
ematics achievement among any subgroup when the sample is grouped by family background.
For the other two subjects, in panel A, the effect of creative thinking on Chinese achievement is 0.145 for urban students and 0.204
for rural students, significant at the 1% level and 5% level, respectively. Additionally, the effect of creative thinking on English
achievement for rural students is 0.176, which is significant at the 5% level. However, for urban students, there is no significant
association between creative thinking and English achievement.

4
Father’s years of education are grouped according to the mean value in our sample. Since the mean value of father’s years of education is
approximately 9 years, we define the group below the mean value as students with low-educated fathers.

15
J. Yang and X. Zhao Thinking Skills and Creativity 40 (2021) 100831

In panel B, students with siblings (i.e., students from non-only child families) show a positive association between creative thinking
and Chinese achievement (estimated coefficient = 0.159, p < 0.01) and English achievement (estimated coefficient = 0.060, p < 0.1).
However, these relationships are not found for students from only child families.
Panel C shows similar conclusions to panels A and B. The effect of creative thinking on the Chinese achievement of students with
highly educated fathers is smaller than that of students with less educated fathers. Specifically, the estimated coefficients are 0.145 for
the former and 0.193 for the latter, which are all significant at the 1% level. For English, the effect of creative thinking on English
achievement for students with low-educated fathers is 0.176 at the 5% level of significance, while there is no significant association
between creative thinking and English achievement for students with highly educated fathers.
Additionally, the mediating effects of creative thinking on academic performance present differences for different subgroups and
different subjects. Regardless of the group, self-esteem mediates the positive association between creative thinking and Chinese
achievement. In terms of students with different hukou or those with different numbers of siblings, self-esteem mediates the effect of
creative thinking on mathematics and English achievement only for urban students and those from non-only child families. Moreover,
the mediating effect is not significant between creative thinking, self-esteem and English achievement for students with highly
educated fathers.
In general, the above results indicate that the association between creative thinking and academic performance is dissimilar for
boys and girls, which supports Hypothesis 3a. Creative thinking has a greater effect on boys’ academic performance. The findings
indicate that the association between creative thinking and academic performance is dissimilar for students from different grade levels,
which supports Hypothesis 3b. Additionally, the association between creative thinking and academic performance is dissimilar for
students from different family backgrounds, which supports Hypothesis 3c. Those conclusions are specifically discussed below in
terms of their contributions for previous research, implications for policymakers, and further study.

5. Discussion

Understanding the association between students’ creative thinking and academic performance is of considerable interest and
importance. The present study focuses on the effects of creative thinking (including convergent thinking and divergent thinking) on
academic performance after controlling for demographic variables such as gender and grade level. Moreover, we investigate the extent
to which self-esteem and internal locus of control mediate those associations as well as the possible differentiation among different
groups. Some potentially important findings are generated.
First, there is a positive association between creative thinking and academic performance, which is similar to existing conclusions
(Huang et al., 2017; Niaz et al., 2000; Powers & Kaufman, 2004; Sebastian & Huang, 2016). Creative thinking is related to higher
academic achievement. Moreover, we examine the effect of creative thinking on different quantiles of academic performance and find
that creative thinking has a greater impact on students with middle academic achievement. This finding demonstrates that the as­
sociation between creative thinking and academic performance may change when academic achievements are at different levels
(Anwar et al., 2012).
Second, our study extends creative thinking from a traditionally focused single dimension to two dimensions, convergent thinking
and divergent thinking, to try to fill this gap in the previous research (Ai, 1999; Anderson et al., 1969; Gralewski & Karwowski, 2012;
Naderi et al., 2009; Olatoye et al., 2010; Zheng & Xiao, 1983). The results suggest that the effect of convergent thinking on academic
performance is greater than that of divergent thinking. Specifically, the former has a significant positive impact on academic per­
formance in all subjects, while the latter has a significant positive impact only on Chinese performance. The results are similar to the
study of Wallach (1985). That is, there is a weak correlation between divergent thinking and academic performance. The fact that
convergent thinking plays a more important role in students’ academic performance than divergent thinking may contribute to the
types of tasks they are asked to perform while learning. Convergent thinking is also integral to critical thinking (making connections,
integrating ideas), an aspect that is common through schooling. Curricula around the world place much more emphasis on convergent
thinking than on divergent thinking.
Third, this paper contributes to previous research by analysing the possible mediating mechanisms that might explain the impacts
of creative thinking (including convergent thinking and divergent thinking) on academic performance. Drawing on the ideas of the
mediating effect model, we find that the inclusion of mediating factors into the regression equation sheds light on the reasons of the
positive effect of creative thinking on academic performance. Specifically, this paper identifies two interesting potential mechanisms,
which further research could examine in other populations. Our results present that self-esteem influences the association between
creative thinking (including convergent thinking and divergent thinking) and academic performance for each subject, while internal
locus of control presents a nearly similar function but does not mediate the effect of creative thinking on Chinese achievement. For
example, the results indicate that complete mediation of the relation between creative thinking and mathematics achievement and self-
esteem mediates this relationship. That is, when we bring self-esteem into the equation, the direct effect of creative thinking on
mathematics is no longer significant since it is reflected by the indirect effect, which is consistent with previous research conclusions
(Tong & Wo, 2002). One possible reason is that the mathematics examination in our survey does not require students to implement
creative thinking abilities while solving the tasks; this might explain the strong mediating effect found in the above result. In other
words, when examinations ask students to implement creating thinking, that specific ability influences their scores; when they do not,
creative thinking loses its effect, while the effects of self-esteem and internal locus of control increase.
Fourth, we contribute to the study by testing underlying gender differences in the association between creative thinking and ac­
ademic performance. Consistent with our hypothesis, the association between creative thinking and academic performance is dis­
similar for boys and girls, and creative thinking has a higher impact on boys’ academic performance. The popular media, teachers, and

16
J. Yang and X. Zhao Thinking Skills and Creativity 40 (2021) 100831

education policymakers have pointed out the issue that many countries are faced with a “boy crisis” issue in which boys are falling
behind girls in all aspects of education, such as average performance gaps in reading, the presence of learning disabilities, rate of grade
repetition, and rate of college enrolment (Husain & Millimet, 2009). The effect of creative thinking identified in this paper hints that
the gender gap in learning will narrow when educators focus on fostering children’s creative thinking. This key result may be applied to
solve the dilemma of “boy crisis”.
Fifth, this paper examines the association between creative thinking and academic performance among students from different
grade levels and finds that creative thinking has a significant positive effect only on Chinese achievement for 10th- and 11th-grade
students, while it has no significant positive effect on mathematics and English achievement at all grade levels. The reason may be
that the effects of creative thinking on mathematics and English achievement are converted to indirect effects through two mediating
paths of self-esteem and internal locus of control (Tong & Wo, 2002). With the increase in grade level (or age), the coefficient of
creative thinking on academic achievement decreases, and the effect on mathematics achievement in 12th grade is even significant
negative. Tong and Wo (2002) and Wo et al. (2006) find that higher-grade students show a low point in the development of creativity,
and there is a weak association between creativity and academic performance. This is because higher-grade students’ learning in­
terests, methods, strategies and personalities are basically fixed. Most students at this age have stable cognitive and noncognitive
abilities, and the development of creative thinking enters a flat stage. Another possible reason is that education becomes more pro­
fessional at higher grade levels (Gajda et al., 2017). More broadly, creative thinking may no longer be the critical factor that affects
academic performance for higher-grade students. This hints that compensatory investments in creative thinking are likely to have
lower effect in later stages in the life cycle. This finding is in line with a series of results showing that remediation in later stages is
usually less effective than cultivation in early stages for skills such as creative thinking competency (Heckman & Mosso, 2014).
Sixth, as predicted, the association between creative thinking and academic performance is dissimilar for students from different
family backgrounds. Perhaps the most surprising and meaningful result of this paper is that creative thinking has a greater effect on the
academic performance of students from disadvantaged family backgrounds (i.e., rural, non-only child, or low-educated father).
Specifically, the Chinese and English achievements of rural students and those with siblings or less-educated fathers benefit more from
creative thinking. That is, creative thinking had a powerful positive impact on academic performance among students with disad­
vantaged backgrounds (e.g., those who share educational resources with siblings). The possible explanation is that children with better
family environments may have more resources for education. Their parents may not only build their creativity but also spend more
money and time to foster other noncognitive skills to promote their academic success (Fang et al., 2018). The finding is in line with the
assumption that disadvantaged children gain less family investment for education, so creative thinking is a more important factor that
affects disadvantaged students’ academic development. Our results suggest that targeted interventions of creative thinking for
disadvantaged students can produce higher returns and compensate for this gap in academic performance due to differences in family
background. Therefore, policymakers can help disadvantaged children build creative thinking by introducing creativity exercises into
curriculums at present to effectively improve their future academic performance (Karpova et al., 2011).
The subject of the paper is important as the Programme for International Student Assessment (PISA) 2012 proposed a survey-based
measure of problem-solving competency targeting both creative thinking and critical thinking. Items are developed to measure how
well students perform when there are two different types of problem situations across a range of contexts. Specifically, problems are set
in a technology context (e.g., mobile phones, ticket vending machines) or a nontechnology context (e.g., route planning, task
scheduling). Subsequently, the PISA 2021 will assess creative thinking by using more direct and scientific measurement tools, such as
focusing on two extensive thematic content fields, “creative expression” and “knowledge creation and creative problem solving.” The
former represents written expression and visual expression, while the latter includes scientific problem solving and social problem
solving (OECD, 2019). This may encourage policymakers to make positive changes in education, and the assessment will provide a
wider discussion on supporting this important ability through education. Therefore, our topic is timely.
One of our most important contributions is the demonstration that creative thinking (including convergent thinking and divergent
thinking) improves academic performance and the testing of two mechanisms between creative thinking and academic performance.
Similar to other skills, creative thinking can also be fostered by targeted training (Fleith et al., 2002). Schools, teachers and parents all
play an important role in nurturing students’ creative thinking. In fact, the importance of engaging students with creative thinking is
broadly realized by educators around the world. Gregory et al. (2013) claim that it cannot afford to postpone the teaching and
cultivating of creative thinking in schools and education institutions. Many countries also regard promoting creative thinking as an
educational goal to develop innovation in the next generation. For example, Singapore considers “inventive thinking” a national
educational goal in the new curriculum, and the new curriculum of Canada includes creative thinking as one of the three key skills
(Hadar & Tirosh, 2019).

6. Limitations and further research

Although this paper makes valuable contributions, some limitations should be noted for further research.
First, this paper concentrates on the effect of creative thinking on students’ academic performance. Although many important
results are revealed, it does not discuss the long-run impact of creative thinking on students. It would be worth investigating the impact
of creative thinking on students’ other important outcomes. For example, future research could explore how creative thinking in­
fluences performance in the labour market by affecting students’ academic achievement.
Second, we only investigated creative thinking and academic performance in high school due to range restrictions, and this study
therefore might be subject to measurement error. Thus, this paper should be repeated using a larger population. While our paper
provides insights into the impact of creative thinking, further study could solve other important problems. Considering that there is a

17
J. Yang and X. Zhao Thinking Skills and Creativity 40 (2021) 100831

stable ascending developmental trend of creative thinking that advances more rapidly in the junior school stage and then changes to a
mild development trend in the senior school stage (Wo et al., 2006), it would be valuable to replicate with a younger population.
Third, creative thinking competency is both domain general and domain specific (Hong & Milgram, 2010). In the present study, the
questionnaire we employed measured general creative thinking rather than subject-specific measures of creative thinking; this might
have lower validity when interpreting more professional academic performance (Baer, 2014). Indeed, many researchers note that
creative thinking and academic performance are domain specific (Baer, 2014; Gajda et al., 2017). Further study should thus apply
measures of domain-specific to explore whether these measures change the association between creative thinking and academic
performance as well as whether there are potentially crucial differences across other domains.
Fourth, this paper does not discuss how to promote academic achievement by improving creative thinking. Future studies can focus
on intervention through the family environment or teachers on students’ creative thinking. On the one hand, socioeconomic status may
affect the development of creative thinking. For example, students from advantaged classes are more likely to be encouraged to think
creatively (Fan & Zhang, 2014). It would be interesting to research whether students from advantaged backgrounds are more creative.
On the other hand, the most critical part of improving creative thinking is related to teachers’ attitudes and behaviours towards
students. Compared to parents, teachers play a more important role in fostering creative thinking in students. Therefore, teachers need
to consider how they can effectively instruct children to promote creative thinking in future teaching, such as introducing a great body
of short creativity exercises throughout the curriculums they teach (Karpova et al., 2011; Ozkal, 2014).

7. Conclusion

As a necessary capacity, creative thinking allows students to perform better in the classroom, on educational evaluation, and well
beyond as they advance in the labour market (Gregory et al., 2013). More broadly, it can help students adapt to a future fierce
competitive society that needs individuals equipped with “21st-century” competences that exceed core literacy and numeracy (OECD,
2019). However, there are still ambiguous or contradictory results regarding the association between creative thinking and academic
performance. Additionally, few scholars have explored the mechanism between creative thinking and academic performance. This
paper examines the effect and mechanism of creative thinking on academic performance and further explores the heterogeneous effect
of creative thinking among different groups.
In summary, our findings suggest that there is a positive association between creative thinking and academic performance. Creative
thinking has a greater effect on students with middle academic achievement. We also find that as two important dimensions of creative
thinking, convergent thinking plays a more important role in students’ academic performance than divergent thinking. In addition, we
employ a mediating effect model and find that creative thinking influences academic performance through students’ self-esteem and
that internal locus of control promotes academic development. Further research shows that the impact of creative thinking on aca­
demic performance varies among different groups. Creative thinking has a stronger effect on boys, low-grade students and students
from disadvantaged backgrounds (e.g., those who share educational resources with siblings). Education policymakers need to pay
more attention to disadvantaged students, especially to allocate more resources for them, thereby helping them enhance creative
thinking and further improve their academic performance.

Author statement

Juan Yang: Conceptualization, Supervision, Project administration, Funding acquisition, Writing- Original draft, Writing-
Reviewing and Editing.
Xinhui Zhao: Software, Validation, Formal analysis, Methodology, Data Curation, Writing - Original draft, Writing- Reviewing and
Editing.

Funding acknowledgement

We gratefully acknowledge the support from the National Natural Science Foundation of China (71973012) and Philosophy and
Social Science Research Base Project of Beijing (CN) (19JDJYA003).

Appendix A

Table A1
Factor loadings of creative thinking variables.
1 2

Understanding key ideas 0.741

Synthesis 0.664
Summarily explaining 0.660
Evaluation 0.668
Association 0.759
Originality 0.528
Pointing towards the future 0.697
Insight 0.611
Eigenvalue 2.981 1.005
Variance contribution rate (%) 37.265 12.562
Cumulative variance contribution rate (%) 37.265 49.827

18
 J. Yang and X. Zhao
Table A2
The influence and mechanisms of creative thinking on the total score.
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Total score Mediating variable = self-esteem Mediating variable = internal locus of control

0.157*** 0.086*** 0.133***

Creative thinking
(0.030) (0.032) (0.031)
0.164*** 0.138*** 0.162***
Convergent thinking
(0.017) (0.018) (0.018)
0.012* − 0.016 0.004
Divergent thinking
(0.007) (0.017) (0.017)
− 0.134*** − 0.145*** − 0.113*** − 0.146*** − 0.154*** − 0.135*** − 0.127*** − 0.144*** − 0.107***
Gender (boy = 1)
(0.036) (0.036) (0.036) (0.036) (0.036) (0.036) (0.036) (0.036) (0.036)
− 0.068 − 0.069 − 0.074 − 0.065 − 0.066 − 0.068 − 0.074 − 0.070 − 0.081
Hukou (rural = 1)
(0.053) (0.052) (0.053) (0.053) (0.052) (0.053) (0.053) (0.052) (0.054)
19

Only child 0.043 0.054 0.036 0.039 0.051 0.034 0.040 0.054 0.034
(0.061) (0.059) (0.061) (0.060) (0.058) (0.060) (0.060) (0.059) (0.061)
Grade − 0.011 − 0.017 − 0.007 − 0.021 − 0.023 − 0.021 − 0.009 − 0.017 − 0.005
(0.024) (0.024) (0.024) (0.024) (0.024) (0.024) (0.024) (0.024) (0.024)
0.014* 0.011 0.017** 0.012 0.010 0.013* 0.013* 0.011 0.016**
Father’s years of education
(0.008) (0.008) (0.008) (0.008) (0.008) (0.008) (0.008) (0.008) (0.008)
Constant 0.034 0.051 0.002 0.077 0.081 0.066 0.041 0.052 0.017
(0.175) (0.171) (0.174) (0.173) (0.171) (0.172) (0.175) (0.171) (0.174)
0.102*** 0.065*** 0.124*** 0.046*** 0.006 0.066***
Mediating variable
(0.018) (0.018) (0.017) (0.017) (0.018) (0.017)
0.071*** 0.026*** 0.028*** 0.024*** 0.002 0.007***
Mediating effect (Sobel test)
(0.013) (0.007) (0.005) (0.009) (0.007) (0.002)

Thinking Skills and Creativity 40 (2021) 100831

Adjusted R2 0.315 0.335 0.308 0.324 0.339 0.323 0.317 0.335 0.312

Note: ***, **, and * indicate significance at the 1%, 5%, and levels respectively. The table above also controls class-fixed effects. Standard errors shown in parentheses are clustered at the individual level.
J. Yang and X. Zhao Thinking Skills and Creativity 40 (2021) 100831

Table A3
The heterogeneity of creative thinking on the total score.
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)
Groups = gender Groups=hukou Groups = only child Groups = feduc Groups = grade levels

girl boy urban rural only non-only high- low- 10th 11th 12th
child child feduc feduc Grade Grade Grade

Creative 0.040 0.140*** 0.071** 0.166** 0.109 0.083** 0.005 0.095*** 0.044 0.025 − 0.038
thinking
(0.041) (0.051) (0.035) (0.082) (0.115) (0.034) (0.069) (0.037) (0.040) (0.037) (0.056)
Self-esteem 0.134*** 0.076*** 0.106*** 0.088* 0.114* 0.098*** 0.105*** 0.100*** 0.022 0.074*** 0.111***
(0.024) (0.027) (0.020) (0.048) (0.063) (0.019) (0.039) (0.020) (0.021) (0.021) (0.031)
Mediating 0.090*** 0.054*** 0.073*** 0.069*** 0.083* 0.068*** 0.070** 0.071*** 0.015 0.047*** 0.080***
effect
(0.017) (0.021) (0.014) (0.040) (0.048) (0.014) (0.028) (0.015) (0.014) (0.015) (0.024)
(Sobel test)
Adjusted R2 0.356 0.345 0.310 0.446 0.523 0.316 0.398 0.311 0.701 0.648 0.567

Note: feduc = father’s years of education. ***, **, and * indicate significance at the 1%, 5%, and levels respectively. The table above also controls
students’ gender, hukou, siblings, grade, father’s years of education and class-fixed effects. Standard errors shown in parentheses are clustered at the
individual level.

Table A4
The association between convergent thinking and academic performance among different genders.
(1) (2) (3) (4) (5) (6) (7) (8)
Self-esteem Chinese Mathematics English

Groups = gender

girl boy girl boy girl boy girl boy

0.406*** 0.392*** 0.084*** 0.091*** 0.117*** 0.086*** 0.130*** 0.158***

Convergent thinking
(0.026) (0.032) (0.026) (0.031) (0.025) (0.029) (0.024) (0.029)
0.105*** 0.010 0.059** 0.078*** 0.076*** 0.021
Self-esteem
(0.025) (0.029) (0.025) (0.028) (0.023) (0.031)
0.043*** 0.004 0.024** 0.031*** 0.031*** 0.008
Mediating effect (Sobel test)
(0.011) (0.012) (0.010) (0.012) (0.010) (0.013)
Adjusted R2 0.202 0.227 0.236 0.197 0.301 0.284 0.302 0.296

Note: ***, **, and * indicate significance at the 1%, 5%, and levels respectively. The table above also controls students’ gender, hukou, siblings, grade,
father’s years of education and class-fixed effects. Standard errors shown in parentheses are clustered at the individual level.

Table A5
The association between convergent thinking and academic performance among different grade levels.
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
Self-esteem Chinese Mathematics English

Groups = grade levels

10th 11th 12th 10th 11th 12th 10th 11th 12th 10th 11th 12th
Grade Grade Grade Grade Grade Grade Grade Grade Grade Grade Grade Grade

0.667*** 0.642*** 0.717*** 0.141** 0.116** 0.066 0.028 − 0.016 − 0.121** − 0.003 0.016 0.032
Convergent thinking
(0.068) (0.055) (0.070) (0.058) (0.050) (0.061) (0.047) (0.041) (0.059) (0.047) (0.045) (0.064)
0.034 0.045 0.033 0.005 0.058*** 0.137*** 0.026 0.077*** 0.071*
Self-esteem
(0.030) (0.027) (0.035) (0.026) (0.022) (0.035) (0.025) (0.025) (0.036)
Mediating effect 0.023 0.029 0.023 0.003 0.037** 0.098*** 0.017 0.050** 0.051**
(Sobel test) (0.020) (0.018) (0.027) (0.016) (0.016) (0.027) (0.017) (0.017) (0.026)
Adjusted R2 0.193 0.195 0.245 0.366 0.416 0.353 0.588 0.564 0.488 0.555 0.557 0.444

Note: ***, **, and * indicate significance at the 1%, 5%, and levels respectively. The table above also controls students’ gender, hukou, siblings, grade,
father’s years of education and class-fixed effects. Standard errors shown in parentheses are clustered at the individual level.

20
J. Yang and X. Zhao Thinking Skills and Creativity 40 (2021) 100831

Table A6
The association between convergent thinking and academic performance among different family backgrounds (hukou, only child, father’s years of
education).
(1) (2) (3) (4) (5) (6) (7) (8)
Self-esteem Chinese Mathematics English

Panel A: Groups=hukou
urban rural urban rural urban rural urban rural
Convergent 0.398*** 0.449*** 0.076*** 0.124** 0.099*** 0.149*** 0.130*** 0.205***
thinking (0.022) (0.070) (0.022) (0.054) (0.021) (0.051) (0.020) (0.048)
0.054** 0.140*** 0.068*** 0.020 0.053*** 0.003
Self-esteem
(0.021) (0.042) (0.020) (0.050) (0.020) (0.051)
Mediating effect 0.021*** 0.063*** 0.027*** 0.009 0.021*** 0.001
(Sobel test) (0.009) (0.023) (0.008) (0.023) (0.008) (0.022)
Adjusted R2 0.200 0.285 0.194 0.420 0.283 0.339 0.277 0.455

Panel B: Groups = only child

non-only non-only non-only non-only
only child only child only child only child
child child child child
Convergent 0.396*** 0.430*** 0.077*** 0.184*** 0.100*** 0.174*** 0.126*** 0.295***
thinking (0.022) (0.070) (0.021) (0.067) (0.020) (0.058) (0.019) (0.069)
0.054*** 0.107 0.064*** 0.021 0.046** 0.008
Self-esteem
(0.020) (0.068) (0.020) (0.056) (0.020) (0.064)
Mediating effect 0.021*** 0.046 0.025*** 0.009 0.018** 0.003
(Sobel test) (0.008) (0.030) (0.008) (0.026) (0.008) (0.029)
Adjusted R2 0.197 0.350 0.210 0.413 0.277 0.492 0.284 0.504

Panel C: Groups = father’s years of education

high- high- high- high-
low-educated low-educated low-educated low-educated
educated educated educated educated
father father father father
father father father father
Convergent 0.351*** 0.413*** 0.079*** 0.127*** 0.074* 0.108*** 0.111*** 0.150***
thinking (0.050) (0.023) (0.022) (0.045) (0.040) (0.022) (0.038) (0.021)
0.052** 0.099** 0.062 0.059*** 0.021 0.047**
Self-esteem
(0.022) (0.040) (0.039) (0.021) (0.042) (0.021)
Mediating effect 0.021** 0.035** 0.022*** 0.024*** 0.007 0.019**
(Sobel test) (0.009) (0.016) (0.014) (0.009) (0.015) (0.009)
Adjusted R2 0.224 0.206 0.204 0.289 0.347 0.281 0.365 0.279

Note: ***, **, and * indicate significance at the 1 %, 5 %, and levels respectively. The table above also controls students’ gender, hukou, siblings,
grade, father’s years of education and class-fixed effects. Standard errors shown in parentheses are clustered at the individual level.

References

Ai, X. (1999). Creativity and academic achievement: An investigation of gender differences. Creativity Research Journal, 12(4), 329–337.
Akpur, U. (2020). Critical, reflective, creative thinking and their reflections on academic achievement. Thinking Skills and Creativity, 37.
Anderson, H. E., White, W. F., & Stevens, J. C. (1969). Student creativity, intelligence, achievement, and teacher classroom behavior. The Journal of Social Psychology,
78(1), 99.
Anwar, M. N., & Aness, M. (2012). An examination of the relationship between creative thinking and academic achievements of secondary school students.
International Interdisciplinary Journal of Education, 1(3), 1–4.
Arshad, M., Zaidi, S., & Mahmood, K. (2015). Self-esteem and academic performance among university students. Journal of Education and Practice, 6(1), 156–162.
Baer, J. (2014). Creativity and divergent thinking: A task-specific approach. London: Psychology Press.
Barbot, B. (2020). Creativity and self-esteem in adolescence: A study of their domain-specific, multivariate relationships. The Journal of Creative Behavior, 54(2).
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations.
Journal of Personality and Social Psychology, 51(6), 1173–1182.
Baumeister, R. F., Campbell, J. D., Kruegger, J. I., & Vohs, K. D. (2003). Does high self-esteem cause better performance, interpersonal success, happiness, or healthier
lifestyles? Psychological Science in the Public Interest, 4, 1–44.
Bravo, M., Salvo, S., & Muñoz, C. (2015). Profiles of Chilean students according to academic performance in mathematics: An exploratory study using classification
trees and random forests. Studies in Educational Evaluation, 44, 50–59.
Burger, J. M. (1986). Personality: Theory and research. Belmont, Calif: Wadsworth Publishing Company.
Cai, X., & Zhu, Y. (2007). The correlation of adolescent’s creative inclination, intelligence and academic achievement. Psychological Development and Education, (02),
36–41 (in Chinese).
Cantero, M.-J., Alfonso-Benlliure, V., & Melero, R. (2016). Creativity in middle childhood: Influence of perceived maternal sensitivity, self-esteem, and shyness.
Creativity Research Journal, 28, 105–113.
Chiu, M. M., & Xihua, Z. (2008). Family and motivation effects on mathematics achievement: Analyses of students in 41 countries. Learning and Instruction, 18(4),
321–336.
Colzato, L. S., Szapora, A., Pannekoek, J. N., & Hommel, B. (2013). The impact of physical exercise on convergent and divergent thinking. Frontiers in Human
Neuroscience, 7.
Coopersmith, S. (1967). The antecedents of self-esteem. San Fransisco: Freeman.
CPC Central Committee and State Council. (2020). The overall plan for deepening educational evaluation reform in the new era.
Craft, A. (2006). Fostering creativity with wisdom. Cambridge Journal of Education, 36(3).

21
J. Yang and X. Zhao Thinking Skills and Creativity 40 (2021) 100831

Davut, H. (2008). The examination of the basic skill levels of the students’ in accordance with the perceptions of teachers, parents and students. International Journal of
Instruction, 1.
Deng, X., & Zhang, X. (2011). Understanding the relationship between self-esteem and creativity: A meta-analysis. Advances in Psychological Science, 19(05), 645–651
(in Chinese).
Fan, J. Q., & Zhang, L. F. (2014). The role of perceived parenting styles in thinking styles. Learning and Individual Differences, 32, 204–211.
Fang, C. C., Hu, Y. M., & Zhang, P. P. (2018). Can pupils of primary schools benefit from the study time after class. Journal of Educational Science of Hunan Normal
University, (1), 69–77 (in Chinese).
Firpo, S., Fortin, N. M., & Lemieux, T. (2009). Unconditional quantile regressions. Econometrica, 77(3), 953–973.
Fleith, D. D., Renzulli, J. S., & Westberg, K. L. (2002). Effects of a creativity training program on divergent thinking abilities and self-concept in monolingual and
bilingual classrooms. Creativity Research Journal, 14(3–4), 373–386.
Gajda, A., Karwowski, M., & Beghetto, R. A. (2017). Creativity and academic achievement: A meta-analysis. Journal of Educational Psychology, 109(2), 269–299.
Gifford, D. D., Briceno-Perriott, J., & Mianzo, F. (2006). Locus of control: Academic achievement and retention in a sample of university first-year students. Journal of
College Admission, 191(191), 18–25.
Glover, J. A., & Sautter, F. (1976). An investigation of the relationship of four components of creativity to locus of control. Social Behavior and Personality, 4, 257–260.
Goldsmith, R. E., & Matherly, T. A. (1988). Creativity and self-esteem: A multiple operationalization validity study. The Journal of Psychology, 122, 47–56.
Gralewski, J., & Karwowski, M. (2012). Creativity and school grades: A case from Poland. Thinking Skills and Creativity, 7(3), 198–208.
Greenstein, L. (2012). Assesing skill 21st century. A guide to evaluating mastery and authentic learning. New York: Crowin A SAGE Company.
Gregory, E., Hardiman, M., Yarmolinskaya, J., Rinne, L., & Limb, C. (2013). Building creative thinking in the classroom: From research to practice. International
Journal of Educational Research, 62, 43–50.
Grigorenko, E. L., Diffley, R., Goodyear, J., Shanahan, E. J., Jarvin, L., & Sternberg, R. J. (2009). Are SSATs and GPA enough? A theory-based approach to predicting
academic success in secondary school. Journal of Educational Psychology, 101(4), 964–981.
Hadar, L. L., & Tirosh, M. (2019). Creative thinking in mathematics curriculum: An analytic framework. Thinking Skills and Creativity, 33.
Heckman, J., & Mosso, S. (2014). Web appendix for the economics of human development and social mobility James Heckman and Stefano Mosso February 4, 2014.
Annual Review of Economics, 6, 6.
Hong, E., & Milgram, R. M. (2010). Creative thinking ability: Domain generality and specificity. Creativity Research Journal, 22(3), 272–287.
Huang, P. S., Peng, S. L., Chen, H. C., Tseng, L. C., & Hsu, L. C. (2017). The relative influences of domain knowledge and domain-general divergent thinking on
scientific creativity and mathematical creativity. Thinking Skills and Creativity, 25, 1–9.
Husain, M., & Millimet, D. L. (2009). The mythical’ boy crisis’? Economics of Education Review, 28(1), 38–48.
Kaiser, D. L. (1975). Internal-external control and causal attributions of own and other’s performance. Pyschological Reports, 37, 423–426.
Karpova, E., Marcketti, S. B., & Barker, J. (2011). The efficacy of teaching creativity: Assessment of student creative thinking before and after exercises. Clothing and
Textiles Research Journal, 29(1), 52–66.
Kemple, K. M., David, G. M., & Wang, Y. (1996). Preschoolers’ creativity, shyness, and self-esteem. Creativity Research Journal, 9(4), 317–326.
Lau, S., Li, C. S., & Chu, D. H. (2004). Perceived creativity: Its relationship to social status and self-concept among Chinese high ability children. Creativity Research
Journal, 16, 59–67.
Li, L. H., & Qu, G. L. (1998). A study on the relationship between locus of control and creativity of high school students. Psychological Development and Education, (03),
38–41 (in Chinese).
Luo, F., & Meng, Q. (2005). Construction of self-assessment test on creative thinking abilities of high school students. Psychological Development and Education, (04),
94–98 (in Chinese).
Mauricio, C. V., Nicole, B. G., Cuello Laura, J., Alejandro, A. M., & Christian, A. O. (2018). Does socioeconomic status influence student creativity? Thinking Skills and
Creativity, 29, 142–152.
Melissa, A. A., & Jason, J. D. (2016). Learning goal orientation and locus of control interact to predict academic self-concept and academic performance in college
students. Personality and Individual Differences, 97, 245–248.
Mourgues, C., Tan, M., Hein, S., Elliott, J. G., & Grigorenko, E. L. (2016). Using creativity to predict future academic performance: An application of Aurora’s five
subtests for creativity. Learning and Individual Differences, 51, 378–386.
Naderi, H., Abdullah, R., Aizan, H. T., Sharir, J., & Kumar, V. (2010). Relationship between creativity and academic achievement: A study of gender differences. The
Journal of American Science, 6(1), 181–190.
Naderi, H., Abdullah, R., Hamid, T., Jamaluddin, S., & Kumar, V. (2009). Creativity, age and gender as predictors of academic achievement among undergraduate
students. Marsland Press Journal of American Science, 5, 101–112.
Niaz, M., De Nunez, G. S., & De Pineda, I. R. (2000). Academic performance of high school students as a function of mental capacity, cognitive style, mobility-fixity
dimension, and creativity. The Journal of Creative Behavior, 34(1), 18–29.
Nori, Z. (2002). Gender differences creativity, academic achievement (mathematics, sciences and language of literature) among high school in City of Shiraz, Iran. Shiraz:
University of Shiraz.
OECD. (2019). PISA 2021 creative thinking framework: Third draft. Paris: OECD Publishing.
Olatoye, R. A., Akintunde, S. O., & Yakasi, M. I. (2010). Emotional intelligence, creativity and academic achievement of business administration students. Electronic
Journal of Research in Educational Psychology, 8(2), 763–786.
Ozkal, N. (2014). Relationships between teachers creativity fostering behaviors and their self-efficacy beliefs. Educational Research and Reviews, 9, 724–733.
Pannells, T. C., & Claxton, A. F. (2008). Happiness, creative ideation, and locus of control. Creativity Research Journal, 20(1), 67–71.
Powers, D. E., & Kaufman, J. C. (2004). Do standardized tests penalize deep-thinking, creative, or conscientious students? Some personality correlates of Graduate
Record Examinations test scores. Intelligence, 32(2), 145–153.
Prociuk, T. J., & Breen, L. J. (1974). Locus of control, study habits and attitudes, and college academic performance. The Journal of Psychology, 88, 91–95.
Pullmann, H., & Allik, J. (2008). Relations of academic and general self-esteem to school achievement. Personality and Individual Differences, 45(6), 559–564.
Rank, J., Nelson, N. E., Allen, T. D., & Xu, X. (2009). Leadership predictors of innovation and task performance: Subordinates’ selfesteem and self-presentation as
moderators. Journal of Occupational and Organizational Psychology, 82, 465–489.
Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press.
Sebastian, J., & Huang, H. (2016). Examining the relationship of a survey based measure of math creativity with math achievement: Cross-national evidence from
PISA 2012. International Journal of Educational Research, 80, 74–92.
Stamovlasis, D., Kypraios, N., & Papageorgiou, G. (2015). A sem model in assessing the effect of convergent, divergent and logical thinking on students’ understanding
of chemical phenomena. Science Education International, 26(3), 284–306.
Sternberg, R. J., & Lubart, T. (1999). The concept of creativity: Prospects and paradigms. Handbook of Creativity, 1, 93–115.
Tabach, M., & Friedlander, A. (2017). Algebraic procedures and creative thinking. Zdm-Mathematics Education, 49(1), 53–63.
Tan, A. G. (2000). A review on the study of creativity in Singapore. The Journal of Creative Behavior, 34(4), 259–284.
Tong, X., & Wo, J. (2002). Research on the students′ development of creativity thinking in senior high school. Psychological Development and Education, (02), 22–26 (in
Chinese).
Torrance, E. (1972). Predictive validity of the Torrance tests of creative thinking. The Journal of Creative Behavior, 6.
Torrance, E. P. (1971). Encouraging creativity in the classroom. Dubuque, IA: Brown.
Volle, E. (2017). Associative and controlled cognition in divergent thinking: Theoretical, experimental, neuroimaging evidence, and new directions.
Wallach, M. A. (1985). Creativity testing and giftedness. In F. D. Horowitz, & M. O’Brien (Eds.), The gifted and talented: Developmental perspectives (pp. 99–123).
Washington, DC, US: American Psychological Association.
Wang, A. Y. (2007). Contexts of creative thinking: Teaching, learning, and creativity in Taiwan and the United States. Claremont Graduate University.

22
J. Yang and X. Zhao Thinking Skills and Creativity 40 (2021) 100831

Wang, A. Y. (2012). Exploring the relationship of creative thinking to reading and writing. Thinking Skills and Creativity, 7(1), 38–47.
Wang, Y., & Wang, L. (2016). Self-construal and creativity: The moderator effect of self-esteem. Personality and Individual Differences, 99, 184–189.
Wo, J., Wang, F., Lin, C., & Liu, C. (2007). A research on high and low academic achievement adolescents’ creativity. Psychological Development and Education, (02),
29–35 (in Chinese).
Wo, J., Yang, W., & Lin, C. (2006). A study on the development characteristics of convergent thinking of adolescents. Chinese Journal of Applied Psychology, (04),
297–304 (in Chinese).
Zheng, R., & Xiao, B. (1983). A study on the creativity of high school students. Acta Psychologica Sinica, (04), 445–452 (in Chinese).
Zhou, Y., Yang, W., & Zhao, X. (2006). Investigation an analysis on the students’ creative features. The Journal of Higher Education, (05), 78–82 (in Chinese).
Zuffianò, A., Alessandri, G., Gerbino, M., Luengo Kanacri, B. P., Di Giunta, L., Milioni, M., & Caprara, G. V. (2013). Academic achievement: The unique contribution of
self efficacy beliefs in self-regulated learning beyond intelligence, personality traits, and self-esteem. Learning and Individual Differences, 23(1), 158–162.

23