Original Article

Journal of Pacific Rim Psychology

Volume 15: 1–13
Student and Teacher Characteristics on ! The Author(s) 2021
Article reuse guidelines:
Student Math Achievement sagepub.com/journals-permissions
DOI: 10.1177/1834490921991428
journals.sagepub.com/home/pac

Sukkyung You1 , Eui Kyung Kim2, Sun Ah Lim3 and

Myley Dang4

Abstract
Using data from the Programme for International Student Assessment (PISA), this study implements two statistical
analyses to investigate the effects of student and teacher characteristics on students’ mathematical achievement. First,
the authors conduct an exploratory factor analysis to explore the factor structure for the various student and teacher
variables of interest in this study. Second, they perform hierarchical linear modeling to analyze students’ and teachers’
multilevel structure in a school. The results suggest that student characteristics such as mathematics interest, instrument
motivation, mathematics self-efficacy, mathematics anxiety, mathematics self-concept, and out-of-school study time
predicted 39.9% of mathematical achievement variance. The results also suggest that mathematics self-efficacy had
the largest effect on mathematical achievement. Teacher characteristics such as teacher-directed instruction, cognitive
activation, teacher support, classroom management, and student–teacher relations predicted 34.9% of mathematical
achievement variance. This study’s results have implications for educators in fostering a positive learning environment to
increase students’ mathematics interest and self-efficacy, and focus on specific teacher characteristics to increase
students’ mathematical achievement.

Keywords
mathematical achievement, multilevel, PISA data, student characteristics, teacher characteristics
Received 25 August 2018; accepted 20 November 2020

Increasing students’ mathematical achievement is of There has been a plethora of empirical research
national concern. Results from international compar- focusing on the student characteristics and teacher
ison studies such as the Programme for International characteristics that influence students’ mathematical
Student Assessment (PISA) suggest that 15-year-old achievement scores. In terms of student characteristics,
U.S. students continually score lower on standard- the extant literature suggests a number of factors,
ized mathematical achievement tests than 15-year- including students’ mathematics interests, motivation,
old students in other countries (OECD, 2012). mathematics self-efficacy, mathematics anxiety, mathe-
Moreover, within the United States, the nation’s matics self-concept, and out-of-school study time, influ-
report card revealed that 42% of fourth-graders, ence academic achievement (Beaton et al., 1996;
35% of eighth-graders, and 26% of twelfth-graders
were placed at or above the proficient levels for their
respective grades, where proficient indicates that stu- 1
College of Education, Hankuk University of Foreign Studies
dents demonstrated competency with challenging 2
Graduate School of Education, University of California, Riverside
subject matter (Kena et al., 2015). Given these 3
Department of Education, Chonbuk National University
4
statistics, there is a critical need to understand the Department of Education, University of California, Santa Barbara
factors influencing students’ mathematical
Corresponding author:
achievement and address this achievement gap Sukkyung You, College of Education, Hankuk University of Foreign
between U.S. students and their international Studies, Imun-dong, Dongdaemun-gu, Seoul 130-791, South Korea.
counterparts. Email: skyou@hufs.ac.kr
Creative Commons NonCommercial-NoDerivs CC BY-NC-ND: This article is distributed under the terms of the Creative Commons
Attribution-NonCommercial-NoDerivs 4.0 License (https://creativecommons.org/licenses/by-nc-nd/4.0/) which permits non-commercial
use, reproduction and distribution of the work as published without adaptation or alteration, without further permission provided the original work is
attributed as specified on the SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage).
2 Journal of Pacific Rim Psychology

Else-Quest et al., 2013; Fast et al., 2010; Lee, 2009; a self-evaluation of one’s own general ability in math-
Marsh & Martin, 2011; Pajares & Miller, 1994; ematics; Lee, 2009; Marsh & Martin, 2011; Pajares &
Stevens et al., 2004). In terms of teacher characteristics, Miller, 1994), mathematics anxiety (i.e., a feeling of
there has been research suggesting that teacher-directed worry or discomfort toward mathematics that impedes
instruction, cognitive action, teacher support, classroom performance; Bandalos et al., 1995; Lee, 2009; Ma,
management, and student–teacher relations influence 1999; Schulz, 2005), mathematics interest (i.e., interest
students’ mathematical achievement (Baumert et al., in learning and achievement in mathematics; Else-
2010; Dever & Karabenick, 2011; Hochweber et al., Quest et al., 2013), mathematics motivation (i.e., the
2014; Hughes et al., 2011; Levpuscek & Zupancic, extent to which individuals embrace challenges and
2009). These student- and school-level factors have are motivated to perform well in mathematics;
been demonstrated to increase students’ college and Gottfried et al., 2013; Stevens et al., 2004), and out-
career opportunities, as well as their future income pros- of-school study time (i.e., the amount of time spent
pects (National Mathematics Advisory Panel, 2008). studying outside school; Beaton et al., 1996). These
characteristics may act as primary determinants of stu-
Conceptual Framework dents’ mathematical achievement.
Although many existing studies have supported the
Students’ mathematical achievement scores have been positive impacts of selected student characteristics,
widely used as indicators of students’ academic perfor- some studies have provided mixed results. For instance,
mance in school. However, there is much debate as to Pajares and Miller (1994, 1995) found that mathemat-
which factors influence students’ mathematical ics self-efficacy and mathematics self-concept were pos-
achievement and how they vary across countries. itively related to mathematical achievement among U.
Previous studies have provided empirical evidence for S. undergraduate students. However, Lee’s (2009)
the effects of student- and school-level factors on U.S. study used the PISA 2003 data to study how mathe-
students’ mathematical achievement. However, most matics self-efficacy, mathematics self-concept, and
studies have examined only one or some of these char- mathematics anxiety related to students’ mathematical
acteristics as predictors of mathematical achievement, performance and found different results across coun-
but not yet explored various student and teacher char- tries. Specifically, students in Asian countries, such as
acteristics simultaneously. Thus, based on the existing Korea and Japan, tended to have low mathematics
theoretical and empirical research, this study utilizes a self-efficacy, low mathematics self-concept, and high
conceptual framework that student academic outcome mathematics anxiety, yet had high mathematical
(i.e., mathematical achievement in this study) is achievement. Conversely, students in western countries
explained by various student- and school-level factors.
such as Finland, the Netherlands, and Liechtenstein,
The conceptual model is depicted in Figure 1.
tended to have high mathematics self-efficacy and
mathematics self-concept and low mathematics anxi-
Student-Level Variables ety, yet had high mathematical performance.
The previous literature has identified several student- Mathematics anxiety has often been negatively associ-
level variables that explain mathematical achievement. ated with mathematical achievement (e.g., Tomasetto,
Specifically, these student characteristics include math- 2020), but some studies have found a positive associa-
ematics self-efficacy (i.e., perceived confidence in math- tion between them (e.g., Hunt et al., 2017). These
ematics; Fast et al., 2010; Lee, 2009; Pajares & results suggest that mathematics self-efficacy and self-
Kranzler, 1995; Pajares & Miller, 1994, 1995; concept may be differentially related to mathematical
Stevens et al., 2004), mathematics self-concept (i.e., achievement across countries.

Figure 1. Conceptual Framework.

You et al. 3

Many studies have reported positive relations to maintain a positive classroom environment;
between mathematics interest and mathematical Hochweber et al., 2014), student–teacher relationships
achievement (e.g., Aunola et al., 2006; Fisher et al., (i.e., students’ perception of the closeness with teachers;
2012). Interest in a particular subject can be accompa- Hughes et al., 2011), cognitive activation (i.e., encour-
nied by attention and concentration (Hidi, 2006). aging students to think more deeply to find solutions
However, mixed results have been reported on the and reach the answer, rather than focusing on the
effect of mathematics interest on achievement. For answer itself; Baumert et al., 2010), and teacher-direct-
example, some researchers have found a significant ed instruction (i.e., instruction in which teachers are
positive relationship between mathematics interest primarily communicating mathematics to students;
and achievement (Liu, 2009; Singh et al., 2002), while National Mathematics Advisory Panel, 2008).
others have reported insignificant or inconsistent Perceived teacher support and positive student–
results (K€oller et al., 2001; Marsh et al., 2005). There teacher relationships facilitate student learning as
have also been some mixed reports on the effects of teachers are the primary organizers of classroom activ-
out-of-school study time on mathematical achieve- ities (Patrick et al., 2007; Simons-Morton & Chen,
ment. Cheema and Sheridan (2015) found that time 2009). They help students to develop academic capac-
spent on homework outside of school had a significant ities within their classroom (Bedeck, 2015) and to con-
impact on mathematical achievement. However, centrate on academic tasks better (Ryan & Patrick,
Beaton et al.’s (1996) study, which used Trends in 2001). The existing literature provided mixed empirical
International Mathematics and Science Study evidence of the impacts of teacher support and stu-
(TIMSS) data, suggested that the highest mathematical dent–teacher relationships on mathematical achieve-
achievement was associated with students who studied ment. For example, some studies have found positive
a moderate amount outside of school (i.e., 1 to 3 hours relations between perceived teacher support and stu-
per day). More study time outside of school did not dent–teacher relationships and mathematical achieve-
necessarily relate to increases in mathematical achieve- ment scores (e.g., Klem & Connell, 2004; Paulo et al.,
ment scores. Students who have academic difficulties 2007), but others have reported inconsistent or no sig-
may need additional time to study outside of school to nificant relations (e.g., Rueger et al., 2010; Valiente
catch up. et al., 2019).
These mixed results on the associations between stu- Teachers’ use of cognitive activation as a teaching
dent characteristics and mathematical achievement strategy has also been linked to higher mathematical
indicate that student characteristics may not play a achievement (e.g., Burge et al., 2015). Examples of cog-
consistent role in mathematical achievement for all stu- nitive activation tasks include having classroom discus-
dents, suggesting the need to examine their relations sions, encouraging students to explain and validate
among U.S. students specifically. Furthermore, math- their solutions, and prompting students to discover
ematics motivation has been consistently positively multiple solutions to a problem (Baumert et al.,
associated with mathematical achievement (e.g., 2010). Cognitive activation has been found to mediate
Areepattamannil et al., 2011; Zhu & Leung, 2011), teachers’ pedagogical content knowledge and
but there are limited studies that have examined math- students’ progress in mathematical performance, indi-
ematics motivation with other various student charac- cating the importance of cognitive activation on the
teristics simultaneously as achievement predictors. path to mathematical achievement (Baumert et al.,
Thus, this study will use selected student-level variables 2010). Additionally, classroom management, defined
to understand their effects on U.S. students’ mathemat- as teachers’ ability to create a functioning learning
ical achievement. The study also aims to include environment (Hochweber et al., 2014), helps students
teacher-level variables to understand how student- stay on task and enables teachers to better monitor
and teacher-level variables relate to students’ mathe- students’ progress (Blair & Dennis, 2010). Children in
matical achievement. well-managed classrooms have shown higher mathe-
matical achievement than those in ineffectively man-
aged classrooms (Stronge et al., 2011). However,
School-Level Variables Blazar (2015) found no significant classroom manage-
Based on an extensive literature review, several teacher ment impact on mathematical achievement when other
characteristics were identified as primary school-level classroom variables were examined simultaneously,
variables contributing to mathematical achievement. suggesting the need to further examine various class-
These teacher characteristics include teachers’ support room effects on mathematical achievement.
(i.e., perceived psychological and practical support Although there is no doubt that teachers play a sig-
from teachers; Levpuscek & Zupancic, 2009), classroom nificant role in students’ academic achievement, there
management (i.e., methods and strategies a teacher uses has been an ongoing debate on which instructional
4 Journal of Pacific Rim Psychology

style is effective for students’ mathematical achieve- student had the opportunity to attempt all mathematics
ment (National Mathematics Advisory Panel, 2008). items during the 2-hour assessment. The scaled achieve-
More specifically, there have been mixed results as to ment scores took the form of five plausible values for
whether a teacher-directed instructional approach each student, where each was generated as a random
effectively improves students’ mathematical perfor- draw from an estimated ability distribution of students
mance. Teacher-directed instruction is focused on the with similar item response patterns and backgrounds
teacher and ranges from highly scripted direct instruction (OECD, 2012). For all measures, we used Cronbach’s
to interactive lecture styles. Similar to other teacher var- a to assess the reliability of a set of scale or test items.
iables, there have been mixed results about the impacts The reliability coefficient for the mathematical assess-
of teacher-directed instruction on academic achievement. ment scores (i.e., five plausible values) was .985.
Morgan et al. (2015) reported that teacher-directed
instruction was significantly associated with the mathe- Student-Level Variables
matical achievement of students with mathematics diffi- Mathematics Interest. Mathematics interest measures
culties. Other studies, however, have indicated that too the degree to which students think about their views on
much teacher-directed instruction might slow the devel- mathematics, such as enjoying reading about mathe-
opment of students’ conceptual understanding (Hiebert, matics, looking forward to mathematics lessons, enjoy-
1999; Woodward & Howard, 1994). ing mathematics, and being interested in learning
Since the legislation of No Child Left Behind, a large in mathematics (Organisation for Economic Co-
body of research has been conducted to identify factors operation and Development, 2013). The mathematics
related to students’ academic achievement. Previous interest variable included four items, where students
studies have attempted to divide variables into school responded to the items on a 4-point Likert scale that
and student levels to extract school effects and student ranged from 1 ¼ strongly agree to 4 ¼ strongly disagree.
effects (Kang et al., 2005; Lee & Chung, 2011; Mu~ noz These responses were reverse-coded so that higher
& Chang, 2007). We conducted a multilevel analysis to scores indicate a higher interest related to mathematics.
address the data’s nested structure, where students are The reliability coefficient for mathematics interest was
nested within schools. Statistically, mixing school-level .910.
variance with student-level variance can generate mis-
Instrumental Motivation. Instrumental motivation
leading results. Therefore, the research design needs to
measures the degree to which students are motivated
be multilayered, with student and teacher characteris-
to learn mathematics because they perceive mathemat-
tics in the same school. This will reveal whether stu-
ics as being useful to them and to their future studies
dents’ academic achievement varies depending on the
and careers (Organisation for Economic Co-operation
teachers or their teaching abilities. Thus, the present
and Development, 2013). The instrumental motivation
study aims to investigate the effects of student and
variable was measured by four items, including stu-
teacher characteristics on students’ academic achieve-
dents’ perception of mathematics as worthwhile for
ments through multilevel analysis.
work, worthwhile for their career chances, important
for future study, and helping them to get a job.
Method Students had the option to respond on a 4-point
Likert scale that ranged from 1 ¼ strongly agree to
Data and Sample 4 ¼ strongly disagree. These responses were reverse-
This study used the fifth PISA survey, which tested 15- coded so that higher scores indicate higher instrumen-
year-olds on reading, mathematics, and science in 2012. tal motivation. The reliability coefficient for instrumen-
Specifically, this study investigated the relation of these tal motivation was .910.
student-level variables and school-level variables to stu-
Mathematics Self-Efficacy. Mathematics self-efficacy
dents’ mathematical achievement using a U.S. sample
measures the degree to which a student believes in
drawn from the PISA 2012 data. The U.S. sample con-
their ability to successfully perform or accomplish a
tains 4,978 students and 162 schools.
particular mathematical task or problem (Hackett &
Betz, 1989). Mathematics self-efficacy was measured
Measures using eight items with response options on a 4-point
Outcome Variable. The outcome variable measuring Likert scale ranging from 1 ¼ very confident to 4 ¼ not
mathematical achievement was based on a total of 85 at all confident. These responses were reverse-coded so
mathematics test items distributed over 13 booklets. that higher scores indicate higher mathematics self-
The balanced incomplete block design was used to max- efficacy. The reliability coefficient for mathematics
imize subject-matter coverage, which meant that no self-efficacy was .852.
You et al. 5

Mathematics Anxiety. Mathematics anxiety measures encouraging students to explain and validate their
the degree to which a student feels helpless or emotion- solutions, and prompting students to discover multiple
al stress when dealing with mathematics (Schulz, 2005). solutions to a problem (Baumert et al., 2010).
The mathematics anxiety items were measured using Cognitive activation was measured using nine items,
five items with response options on a 4-point Likert including the extent to which a student perceives that
scale ranging from 1 ¼ strongly agree to 4 ¼ strongly their mathematics teacher encourages them to reflect
disagree. These responses were reverse-coded so that on a problem; gives problems that require students to
higher scores indicate higher anxiety about mathemat- think for an extended period of time; asks students to
ics. The reliability coefficient for mathematics anxiety use their own procedures; presents problems with no
was .877. obvious solutions; presents problems in different con-
texts; helps students learn from their mistakes; asks for
Mathematics Self-Concept. Mathematics self-concept explanations; presents problems that require students
measures the degree to which a student evaluates their to apply what they have learned; and gives problems
own self-worth related to mathematics (Pajares & with multiple solutions. Students had the option to
Schunk, 2001). Mathematics self-concept was mea- respond using a 4-point Likert scale ranging from
sured using five items on a 4-point Likert scale ranging 1 ¼ always or almost always to 4 ¼ never or rarely.
from 1 ¼ strongly agree to 4 ¼ strongly disagree. These These values were reverse-coded so that higher scores
responses were reverse-coded so that higher scores indi- indicate higher cognitive activation by mathematics
cate a higher self-concept of mathematics. The reliabil- teachers in the classroom. The reliability coefficient
ity coefficient for the mathematics self-concept variable for mathematics cognitive activation was .869.
was .898.
Teacher Support. Teacher support measures the
Out-of-School Study Time. The out-of-school study degree to which students perceive that their mathemat-
time variable indicates the number of hours per week ics teacher provides support. Four teacher support
that a student spends on out-of-school study time for items measured students’ perception of the extent to
all school subjects. Out-of-school study time was mea- which their mathematics teacher lets students know
sured using six items: homework, guided homework, they have to work hard; provides extra help when
personal tutor, classes organized by a commercial com- needed; helps students with learning; and gives students
pany, study with a parent or other family member, and opportunities to express opinions. Students responded
online lessons. Students had the option to freely indi- to these items on a 4-point Likert scale ranging from
cate how many hours per week they spent on each out- 1 ¼ strongly agree to 4 ¼ strongly disagree. The values
of-school study activity. The reliability coefficient for were reverse-coded so that higher scores indicate a
out-of-school study time was .618. more positive student perception of teacher support.
The reliability coefficient for the mathematics teacher
School-Level Variables support variable was .840.
Teacher-Directed Instruction. The teacher-directed
instruction variable measures students’ perceptions of Classroom Management. Classroom management
their teacher’s directed mathematics instruction. There measures the degree to which students perceive how
were five items used to measure students’ perceptions their mathematics teacher manages the classroom.
of teacher-directed instruction, including the extent to The classroom management variable was measured
which the teacher sets clear goals; encourages thinking using four items, including the extent to which students
and reasoning; checks understanding; summarizes pre- perceive their teacher gets students to listen; keeps the
vious lessons; and informs about learning goals. class orderly; starts lessons on time; and has to wait a
Students had the option to respond on a 4-point long time for students to quieten down. Students
Likert scale: 1 ¼ every lesson, 2 ¼ most lessons, 3 ¼ some responded on a 4-point Likert scale ranging from
lessons, and 4 ¼ never or hardly ever. The values were 1 ¼ strongly agree to 4 ¼ strongly disagree. All of the
reverse-coded so that higher scores indicate the higher items except for one (i.e., the teacher has to wait a
frequency of a student’s perceptions of teacher-directed long time for students to quieten down) were reverse-
instruction. The reliability coefficient for teacher- coded. The reliability coefficient for classroom man-
directed instruction was .763. agement was .746.

Cognitive Activation. Cognitive activation measures Student–Teacher Relationship. Student–teacher rela-

the degree to which students perceive how often their tionship measures the degree to which a student per-
mathematics teacher incorporates cognitive activation ceives their relations with teachers at their school. Five
tasks in class, such as having classroom discussions, student–teacher relationship items measured the extent
6 Journal of Pacific Rim Psychology

to which students perceive their teacher gets along with Results

most students; is interested in students’ well-being; lis-
tens to students; helps students; and treats students Descriptive Statistics
fairly. Students responded to these items on a 4-point We implemented a correlation analysis to assess the
Likert scale ranging from 1 ¼ strongly agree to relations between student characteristics, teacher char-
4 ¼ strongly disagree. The items were reverse-coded so acteristics, and students’ mathematical achievement.
that higher scores indicate students’ perception of more As shown in Table 1, there are significant correlations
positive relationships with teachers. The reliability between student and school characteristics with varying
coefficient for student–teacher relationship was .833.
levels of significance, except for the relations of out-of-
school study time to mathematical anxiety, mathemat-
Data Analysis ics self-concept, and classroom management. The
Two statistical analyses were carried out: exploratory mathematical achievement score was significantly asso-
factor analysis and multilevel analysis. For the data han- ciated with all student and teacher characteristics.
dling and descriptive analysis, SPSS 16.0 was used.
Correlations among the variables were examined to Exploratory Factor Analysis
verify the descriptive statistics and model. Multilevel
A robust weighted least squares estimation was used
analysis was conducted using hierarchical linear model-
with a promax rotation. Empirical approaches, such as
ing. We conducted hierarchical linear modeling to
examining a scree test and patterns of the factor loading,
address the nested structure of the data, where students
were considered within a theoretical framework and the
are nested within schools. Statistically, mixing school-
extant literature to confirm that the final factor selection
level variance with student-level variance can generate
was interpretable and substantively plausible. Through
misleading results. Student and teacher characteristics
this process, an 11-factor model emerged as the most
from the same school are not independent of one another
meaningful and parsimonious model. The 5-factor solu-
and therefore their responses do not meet the indepen-
tion yielded a root mean square error of approximation
dence assumption for regression analysis. The estimates
of .06 and standardized root mean square residual of
thus reduce the statistical variance and yield more liberal
.04, which is considered acceptable (Hu & Bentler,
results in significant tests (Bryk & Raudenbush, 1992;
1999). Table 2 displays the factor loadings of the 11
Salvucci & Weng, 1995).
different factors obtained from the exploratory factor
We built three models, starting with a one-way anal-
analysis, and the Cronbach’s a and intercorrelations
ysis of variance model, which allowed for partitioning
for these factors.
of the total variance in overall mathematical achieve-
ment into within-school and between-school variances.
The second model—regression with the means-as- Multilevel Analysis
outcomes model—incorporated the student-level We conducted hierarchical linear modeling to identify
characteristics. The third model incorporated the mathematical achievement predictors by examining
school-level characteristics. Our results are presented student and teacher characteristics in their respective
in the following section. schools (see Table 3). First, we implemented a null

Table 1. Intercorrelation Matrix for Study Variables.

1 2 3 4 5 6 7 8 9 10 11

1. Mathematics score
2. Mathematics interest .140***
3. Instrumental motivation .153*** .607***
4. Mathematics self-efficacy .569*** .407*** .361***
5. Mathematics anxiety .418*** .440*** .305*** .466***
6. Mathematics self-concept .414*** .636*** .452*** .548*** .776***
7. Out-of-school study time .085*** .132*** .110*** .139*** .001 .040
8. Teacher-directed instruction .041* .332*** .309*** .270*** .187*** .233*** .091**
9. Cognitive activation .053** .279*** .275*** .276*** .137*** .217*** .109*** .617***
10. Teacher support .057** .278*** .295*** .219*** .182*** .208*** .065* .548*** .593***
11. Classroom management .214*** .231*** .252*** .287*** .229*** .233*** .035 .412*** .414*** .569***
12. Student–teacher relationship .150*** .283*** .316*** .274*** .213*** .243*** .129*** .483*** .411*** .460*** .391***
*p < .05. **p < .01. ***p < .001.
You et al. 7

Table 2. Factor Loadings of Items Used for This Study. Table 2. Continued.

Factor Factor
Items loadinga Items loadinga

Mathematics interest (Cronbach’s a ¼ .910) 3. Asks to use own procedures .663

1. Enjoy reading .840 4. Presents problems with no obvious solutions .575
2. Look forward to lessons .899 5. Presents problems in different contexts .752
3. Enjoy mathematics .913 6. Helps learn from mistakes .732
4. Interested .894 7. Asks for explanations .691
Percentage of variance explained 78.677 8. Apply what we learned .733
Instrumental motivation (Cronbach’s a ¼ .906) 9. Problems with multiple solutions .697
1. Worthwhile for work .871 Percentage of variance explained 49.337
2. Worthwhile for career chances .886 Teacher support (Cronbach’s a ¼ .840)
3. Important for future study .893 1. Lets us know we have to work hard .772
4. Helps to get a job .884 2. Provides extra help when needed .862
Percentage of variance explained 78.051 3. Helps students with learning .877
Mathematics self-efficacy (Cronbach’s a ¼ .852) 4. Gives opportunity to express opinions .789
1. Using a <train timetable> .682 Percentage of variance explained 68.255
2. Calculating TV discount .750 Classroom management (Cronbach’s a ¼ .746)
3. Calculating square meters of tiles .775 1. Students listen .850
4. Understanding graphs in newspapers .730 2. Teacher keeps class orderly .886
5. Solving Equation 1 .609 3. Teacher starts on time .777
6. Distance to scale .717 4. Wait long to <quiet down>b .524
7. Solving Equation 2 .645 Percentage of variance explained 59.627
8. Calculate petrol consumption rate .700 Student–teacher relationship (Cronbach’s a ¼ .833)
Percentage of variance explained 49.419 1. Get along with teachers .686
Mathematics anxiety (Cronbach’s a ¼ .877) 2. Teachers are interested .797
1. Worry that it will be difficult .826 3. Teachers listen to students .817
2. Get very tense .841 4. Teachers help students .782
3. Get very nervous .831 5. Teachers treat students fairly .787
4. Feel helpless .803 Percentage of variance explained 60.113
5. Worry about getting poor <grades> .805 a
Standardized factor loadings from exploratory factor analysis (factor
Percentage of variance explained 67.456
loading > .30). bA reversed item.
Mathematics self-concept (Cronbach’s a ¼ .898)
1. Not good at mathematicsb .836
2. Get good <grades> .805
model (Model 1), which provided information regard-
3. Learn quickly .885
4. One of best subjects .863 ing the partitioned total variance in overall mathemat-
5. Understand difficult work .837 ical achievement scores into within-school and
Percentage of variance explained 71.483 between-school variances. The intra-class correlation
Out-of-school study time (Cronbach’s a ¼ .618) was .24, indicating that 24% of the variance in overall
1. Homework .437 mathematical achievement scores is between school-
2. Guided homework .729 level characteristics.
3. Personal tutor .685 Second, Model 2 included student characteristics to
4. Commercial company .598 examine their effects on mathematical achievement.
5. With parent .727
The results from Model 2 show that the test language
6. Computer .644
conducted in English (b ¼ 11.926, p < .001) positively
Percentage of variance explained 41.538
Teacher-directed instruction (Cronbach’s a ¼ .763) affected mathematical achievement scores. Students’
1. Sets clear goals .768 gender, parental education, and economic wealth did
2. Encourages thinking and reasoning .714 not significantly affect mathematical achievement
3. Checks understanding .754 scores. Mathematics interest (b ¼ 6.351, p < .001) and
4. Summarizes previous lessons .664 mathematics self-efficacy (b ¼ 10.118, p < .001) had a
5. Informs about learning goals .698 statistically significant positive effect on mathematical
Percentage of variance explained 51.928 achievement, whereas mathematics anxiety
Cognitive activation (Cronbach’s a ¼ .869) (b ¼ 7.180, p < .001) had a significant negative effect
1. Teacher encourages to reflect on problems .747
on mathematical achievement.
2. Gives problems that required to think .714
Third, Model 3 added teacher characteristics to
(continued) Model 2 to assess the school-level effects on student
8 Journal of Pacific Rim Psychology

Table 3. Multilevel Analysis Results for Mathematics Achievement.

Model 1: Null model Model 2: Student level Model 3: School level

b (SE) b (SE) b (SE)

Control variables
Intercept 481.012*** (3.746) 460.217*** (7.057) 461.703*** (6.842)
Gender (Female ¼ 1) 1.533 (2.134) 1.431 (2.131)
Mother’s school 0.205 (1.538) 0.236 (1.537)
Father’s school 2.200 (1.396) 2.054 (1.393)
Wealth 1.388 (1.061) 1.206 (1.071)
Language (English ¼ 1) 11.926*** (3.347) 11.859** (3.350)
Effect of student characteristics
Mathematics interest 6.351*** (0.617) 6.368*** (0.616)
Instrument motivation 0.341 (0.554) 0.339 (0.554)
Mathematics self-efficacy 10.118*** (0.438) 10.128*** (0.438)
Mathematics anxiety 7.180*** (0.590) 7.187*** (0.590)
Mathematics self-concept 0.735 (0.541) 0.736 (0.541)
Out-of-school study time 0.087 (0.140) 0.091 (0.141)
Effect of teacher characteristics
Teacher-directed instruction 12.931* (5.820)
Cognitive activation 8.540* (4.230)
Teacher support 36.615** (10.612)
Classroom management 28.447** (7.913)
Student–teacher relationship 39.319*** (6.310)
Variance
School level 1938.130 1894.976 1233.035
Student level 6087.792 3656.817 3658.189
Total 8025.922 5551.793 4891.224
Intra-class correlation 0.241 0.341 0.252
Accumulated variance explained (R2)
School level 0.022 0.349
Student level 0.399 0.000
Total 0.308 0.119
*p < .05. **p < .01. ***p < .001.

mathematical achievement. The results from Model 3 predicting variables included in this analysis was 12%
show that among all of the school-level variables, of the variance.
teacher-directed instruction (b ¼ 12.931, p < .05) and
teacher support (b ¼ 36.615, p < .01) negatively pre-
dicted mathematical achievement scores. On the other Discussion
hand, cognitive activation, classroom management, and
student–teacher relationship were significant positive The current study aimed to investigate factors that pre-
predictors of student mathematical achievement scores. dict student mathematical achievement by evaluating
Student–teacher relationship (b ¼ 39.319, p < .001) was student- and school-level factors. The hierarchical
the strongest predictor of the achievement outcome, fol- linear modeling analysis found that student variables
lowed by classroom management (b ¼ 28.447, p < .01) predicted 39.9% of students’ mathematics achievement
and cognitive activation (b ¼ 8.540, p < .05). and the teacher variables predicted 34.9%. Specifically,
The effect of the conditional model is shown in the among the student variables, mathematics interest and
accumulated explained variance (R2). The student-level mathematics self-efficacy had positive impacts on
model (Model 2) explained 39.9% of the variance in mathematical achievement, whereas mathematics anx-
mathematical achievement among classrooms and the iety had a negative effect. Among the teacher variables,
school-level model (Model 3) explained 34.9% of the teacher-directed instruction and teacher academic sup-
variance in mathematical achievement among schools. port had negative effects on students’ mathematical
The accumulated variance explained by all the achievement, while cognitive activation, classroom
You et al. 9

management, and student–teacher relationship had mathematics anxiety may interfere with manipulating
positive results. numbers and solving mathematical problems
(Richardson & Suinn, 1972), mathematics interventions
Student Characteristics and Mathematical could incorporate the means to reduce students’ mathe-
Achievement matics anxiety to improve their mathematical
performance.
First, mathematics interest significantly influenced
mathematical achievement, which is consistent with
Teacher Characteristics and Mathematical
previous studies (e.g., Fisher et al., 2012). For example,
Singh et al. (2002) provided empirical evidence for the Achievement
impact of mathematics interest on mathematical The student–teacher relationship was the strongest pre-
achievement among 24,599 middle school and high dictor of students’ mathematical achievement at the
school students collected by the National Center for school level, followed by teacher support, classroom
Education Statistics. The current study also reports management, teacher-directed instruction, and cognitive
that mathematics interest positively predicted academic activation. Specifically, the student–teacher relationship,
achievement. These results suggest that students who classroom management, and cognitive activation posi-
are interested in mathematics are more immersed and tively predicted mathematical achievement, whereas
invest more time in studying mathematics, leading to teacher support and teacher-directed instruction nega-
higher achievement. tively predicted mathematical achievement.
Second, although mixed results exist, many previous First, our results provide additional empirical sup-
studies have reported the positive effect of mathematics port to the existing literature for student–teacher rela-
self-efficacy on mathematical achievement. For exam- tionships’ positive influence on students’ mathematical
ple, self-efficacy has been shown to be more strongly achievement (e.g., Croninger & Lee, 2001). The quality
associated with mathematical performance than of the student–teacher relationship plays a primary role
English and writing (e.g., Pajares, 1996). Another in students’ cognition and academic achievement (Kim
study, conducted by Shores and Shannon (2007), also & Lee, 2015). The current study suggests that the stu-
found that self-efficacy was a significant variable that dent–teacher relationship has the most considerable
increased mathematical achievement among 761 fifth- effect on mathematical achievement—more than
and sixth-graders. Similarly, Yum and Park (2011) other student or teacher characteristics—emphasizing
studied the relation between mathematics self-efficacy the critical role of a positive student–teacher relation-
and mathematical achievement among seventh- to ship in improving students’ mathematical achievement.
ninth-graders through a 3-year latent growth model. Second, teacher support had a negative relation to
They reported a continuous positive effect of mathe- mathematical achievement. It is important to note that
matics self-efficacy on mathematical achievement. The this does not necessarily imply that more academic sup-
current study supports these previous results by show- port from teachers results in decreased mathematical
ing that mathematics self-efficacy had the largest achievement scores. Instead, we interpret this finding
impact on mathematical achievement among all the as students who had lower mathematical achievement
student variables, suggesting the importance of improv- needed more support from teachers. Given that previ-
ing students’ perceived confidence in mathematics to ous studies have reported mixed results on teacher sup-
improve their mathematical performance. port and mathematical achievement (e.g., Klem &
Third, this study’s findings support mathematics Connell, 2004; Rueger et al., 2010), more research is
anxiety’s adverse impact on mathematical achievement. needed to understand this relationship.
Previous studies have reported inconsistent results on Third, this study shows that teachers’ classroom
the relation between mathematics anxiety and mathe- management positively predicted students’ mathemati-
matical achievement. For instance, Ma’s (1999) meta- cal achievement. Our finding provides additional sup-
analysis of 26 studies indicated a significant negative port for the positive impacts of classroom management
correlation between anxiety and academic achievement. on mathematical achievement (Kim, 2000; Stronge
Similarly, Shim’s (2000) study of 219 high school stu- et al., 2011), suggesting that better organized and man-
dents showed that mathematics anxiety was negatively aged classrooms enable students to perform to the best
correlated with mathematical achievement. Shim (2000) of their ability and promote academic achievement.
argued that negative emotions, such as anxiety, interfere Fourth, this study shows the negative effect of
with learning, and negatively affect achievement. The teacher-centered classes on students’ mathematical
present study reveals comparable findings, where math- achievement. This result contributes to the existing lit-
ematics anxiety had a statistically significant negative erature by supporting the findings of previous studies
effect on mathematical achievement. Given that (e.g., Hiebert, 1999; Woodward & Howard, 1994) that
10 Journal of Pacific Rim Psychology

teacher-centered classes, or teacher-directed instruction, achievement. In future studies, researchers need to

might hinder students’ voluntary learning and thus neg- investigate other factors that might influence mathe-
atively influence mathematical achievement. However, matical achievement, such as cognition, metacognition,
Morgan et al. (2015) noted that students who are strug- and motivation.
gling with mathematics benefit from teacher-directed Future studies could also expand on the current
instruction more than student-centered instruction. findings by using various achievement scores (i.e., read-
Thus, more research is needed to understand which ing and science) and including family and peer charac-
type of instruction works better for students depending teristics. Additionally, the present study relied on
on their mathematical performance level. Lastly, cogni- students’ self-reported data. Although it is arguably a
tive activation positively influenced mathematical unique source of information regarding student and
achievement, supporting previous findings (Burge teacher characteristics, it is possible that variables relat-
et al., 2015; Lampert, 2001; Sizmur et al., 2015) that ed to student perceptions of teacher characteristics may
classes emphasizing students’ cognitive activities could be underreported or overreported by the students. The
help students construct their knowledge and show present study was also limited to only teacher-related
improved mathematical performance. factors at the school level and did not examine other
This study’s results contribute to the existing litera- contextual variables such as peer factors. Despite these
ture by using a large nationally representative data set limitations, however, this study’s results contribute to
of 15-year-old students across the United States to our understanding of various student and teacher char-
examine student and teacher factors related to student acteristics and their impacts on student mathematical
mathematical achievement. The study emphasizes the achievement.
importance of student mathematics interest, student
mathematics self-efficacy, the student–teacher relation- Declaration of Conflicting Interests
ship, teacher cognitive activation, and classroom The authors declared no potential conflicts of interest with
management, providing implications for educators. respect to the research, authorship, and/or publication of this
Specifically, interventions designed to improve stu- article.
dents’ mathematics interest and mathematics self-
efficacy may promote students’ mathematical achieve- Funding
ment. We have found that teacher characteristics had The author(s) disclosed receipt of the following financial sup-
the most effect on students’ mathematical achievement, port for the research, authorship, and/or publication of this
especially the student–teacher relationship. It is widely article: This work was partially supported by the Hankuk
acknowledged that teachers are an essential component University of Foreign Studies research fund granted to
Sukkyung You.
in education (Driel et al., 2001) and determine the class
(Borich, 2000). Building positive relationships with stu-
ORCID iD
dents, using cognitive activation, and more student-
centered instruction are suggested to improve mathe- Sukkyung You https://orcid.org/0000-0002-7779-7965
matical performance. This can come from inquiry-
based activities that are student-centered and challenge References
students’ mathematical thinking while listening to and Areepattamannil, S., Freeman, J. G., & Klinger, D. A.
encouraging students (Savery, 2006). Fostering these (2011). Intrinsic motivation, extrinsic motivation, and aca-
student and teacher characteristics can be the first demic achievement among Indian adolescents in Canada
step in improving U.S. students’ mathematical achieve- and India. Social Psychology of Education, 14, 427–439.
ment and, ultimately, rankings in international mathe- Aunola, K., Leskinen, E., & Nurmi, J.-E. (2006).
matical achievement comparisons. Developmental dynamics between mathematical perfor-
mance, task motivation, and teachers’ goals during the
transition to primary school. British Journal of
Limitations and Future Directions Educational Psychology, 76(1), 21–40. https://doi.org/10.
The findings of this study should be interpreted with 1348/000709905X51608
consideration of the study’s limitations. The study Bandalos, D. L., Yates, K., & Thorndike-Christ, T. (1995).
Effects of math self-concept, perceived self-efficacy, and
examined the influence of teacher and student charac-
attributions for failure and success on test anxiety. Journal
teristics on student mathematical achievement.
of Educational Psychology, 87(4), 611–623. https://doi.org/
Although the student characteristics mostly focused 10.1037//0022-0663.87.4.611
on mathematics-related factors, out-of-school study Baumert, J., Kunter, M., Blum, W., Brunner, M., Voss, T.,
time was the amount of time spent studying any sub- Jordan, A., Klusmann, U., Krauss, S., Neubrand, M., &
ject. Out-of-school study time specific to mathematics Tsai, Y.-M. (2010). Teachers’ mathematical knowledge,
may have more substantial impacts on mathematical cognitive activation in the classroom, and student
You et al. 11

progress. American Educational Research Journal, 47(1), accomplishments and educational attainment. Journal of
133–180. https://doi.org/10.3102/0002831209345157 Research on Educational Effectiveness, 6(1), 68–92. https://
Beaton, A. E., Mullis, I. V. S., Martin, M. O., Gonzales, E. J., doi.org/10.1080/19345747.2012.698376
Kelly, D. L., & Smith, T. A. (1996). Mathematics achieve- Hackett, G., & Betz, N. E. (1989). An exploration of the
ment in the middle years: IEA’s Third International mathematics self-efficacy/mathematics performance corre-
Mathematics and Science Study (TIMSS). Center for spondence. Journal for Research in Mathematics
the Study of Testing, Evaluation, and Educational Policy. Education, 20(3), 261–273.
Bedeck, K. (2015). Children’s bullying and victimization on Hidi, S. (2006). Interest: A unique motivational variable.
school engagement: The influence of teacher support Educational Psychology Review, 1(2), 69–82. https://doi.
[Unpublished honors thesis]. org/10.1016/j.edurev.2006.09.001
Blair, C., & Dennis, T. (2010). An optimal balance: The inte- Hiebert, J. (1999). Relationships between research and the
gration of emotion and cognition in context. In S. D. NCTM standards. Journal of Research in Mathematics
Calkins & M. A. Bell (Eds.), Child development at the inter- Education, 30(1), 3–19.
section of emotion and cognition (pp. 17–35). APA Books. Hochweber, J., Hosenfeld, I., & Klieme, E. (2014). Classroom
Blazar, D. (2015). Effective teaching in elementary mathemat- composition, classroom management, and the relationship
ics: Identifying classroom practices that support student between student attributes and grades. Journal of
achievement. Economics of Education Review, 48, 16–29. Educational Psychology, 106(1), 289–300. https://doi.org/
https://doi.org/10.1016/j.econedurev.2015.05.005 10.1037/a0033829
Borich, G. (2000). Effective teaching methods (4th ed.). Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes
Prentice-Hall. in covariance structure analysis: Conventional criteria
Bryk, A.S. and Raudenbush, S.W. (1992) Hierarchical linear versus new alternatives. Structural Equation Modeling,
models (applications and data analysis methods). Sage 6(1), 1–55. https://doi.org/10.1080/10705519909540118
Publications, New York. Hughes, J. N., Wu, J.-Y., Kwok, O.-M., Villarreal, V., &
Burge, B., Lenkeit, J., & Sizmur, J. (2015). PISA in practice: Johnson, A. Y. (2011). Indirect effects of child reports of
Cognitive activation in maths. National Foundation for teacher–student relationship on achievement. Journal of
Educational Research. Educational Psychology, 104(2), 350–365. https://doi.org/
Cheema, J. R. & Sheridan, K. (2015). Time spent on home- 10.1037/a0026339
work, mathematics anxiety and mathematics achievement: Hunt, T. E., Bhardwa, J., & Sheffield, D. (2017). Mental
Evidence from a U.S. sample. Issues in Educational arithmetic performance, physiological reactivity and
Research, 25(3), 246–259. mathematics anxiety amongst U.K. primary school chil-
Croninger, R., & Lee, V. (2001). Social capital and dropping out dren. Learning and Individual Differences, 57, 129–132.
of school: Benefits to at-risk students of teachers’ support https://doi.org/10.1016/j.lindif.2017.03.016
and guidance. Teachers College Record, 103(4), 548–581. Kena, G., Musu-Gillete, L., Robinson, J., Wang, X.,
Dever, B. V., & Karabenick, S. A. (2011). Is authoritative Rathburn, A., Zhang, J., Wilkinson-Flicker, S.,
teaching beneficial for all students? A multi-level model Barmer, A. & Dunlop Velez, E. (2015). The condition of
of the effects of teaching style on interest and achievement. education 2015 (NCES 2015-144). National Center for
School Psychology Quarterly, 26(2), 131–144. https://doi. Education Statistics, Institute of Education Sciences.
org/10.1037/a0022985 https://nces.ed.gov/pubs2015/2015144.pdf
Driel, J. H. van, Beijaard, D., & Verloop, N. (2001). Kim, K. J., & Lee, H. (2015). A panel data analysis on the
Professional development and reform in science educa- effects of the teacher–student relationship on academic
tion: The role of teachers’ practical knowledge. Journal achievement for Korean middle school students. Teacher
of Research in Science Teaching, 38(2), 137–158. Education Research, 54(1), 59–73.
Else-Quest, N. M., Mineo, C. C., & Higgins, A. (2013). Math Kim, Y. (2000). A study of effective teacher in the classroom.
and science attitudes and achievement at the intersection Journal of Social Science, 9, 1–29.
of gender and ethnicity. Psychology of Women Quarterly, Klem, A. M., & Connell, J. P. (2004). Relationships matter:
37(3), 293–309. https://doi.org/10.1177/0361684313480694 Linking teacher support to student engagement and
Fast, L. A., Lewis, J. L., Bryant, M. J., Bocian, K. A., achievement. Journal of School Health, 74(7), 262–273.
Cardullo, R. A., Rettig, M., & Hammond, K. A. (2010). https://doi.org/10.1111/j.1746-1561.2004.tb08283.x
Does mathematics self-efficacy mediate the effect of the per- K€
oller, O., Baumert, J., & Schnabel, K. (2001). Does interest
ceived classroom environment on standardized mathematics matter? The relationship between academic interest and
test performance? Journal of Educational Psychology, 102(3), achievement in mathematics. Journal for Research in
729–740. https://doi.org/10.1037/a0018863 Mathematics Education, 32(5), 448–470.
Fisher, P. H., Dobbs-Oates, J., Doctoroff, G. L., & Arnold, Lampert, M. (2001). Teaching problems and the problems of
D. H. (2012). Early math interest and the development of teaching. Yale University Press.
math skills. Journal of Educational Psychology, 104(3), Lee, H., & Chung, J. (2011). An analysis of the influence of
673–681. https://doi.org/10.1037/a0027756 teachers’ traits on student achievement: Focusing on
Gottfried, A. E., Marcoulides, G. A., Gottfried, A. W., & teachers’ characteristics to enhance professionality in
Oliver, P. H. (2013). Longitudinal pathways from math TIMSS 2007. Journal of Korean Teacher Education,
intrinsic motivation and achievement to math course 28(1), 243–266.
12 Journal of Pacific Rim Psychology

Lee, J. (2009). Universals and specifics of math self-concept, Pajares, F., & Schunk, D. H. (2001). Self-beliefs and school
math self-efficacy, and math anxiety across 41 PISA 2003 success: Self-efficacy, self-concept, and school achieve-
participating countries. Learning and Individual ment. In: R. Riding & S. Rayner (Eds.), Perception (pp.
Differences, 19(3), 355–365. https://doi.org/10.1016/j. 239–266). Ablex.
lindif.2008.10.009 Patrick, H., Ryan, A. M., & Kaplan, A. (2007). Early ado-
Levpuscek, M. P., & Zupancic, M. (2009). Math achievement lescents’ perceptions of the classroom social environment,
in early adolescence: The role of parental involvement, motivational beliefs, and engagement. Journal of
teachers’ behavior, and students’ motivational beliefs Educational Psychology, 99(1), 83–98. https://doi.org/10.
about math. Journal of Early Adolescence, 29(4), 1037/0022- 0663.99.1.83
541–570. https://doi.org/10.1177/0272431608324189 Paulo, A. G., Rachael, D. R., Susan, P.K., Susan, D. C.
Liu, O. L. (2009). An investigation of factors affecting gender (2007). The role of emotion regulation in children’s early
differences in standardized math performance: Results academic success, Journal of School Psychology, 45(1),
from U.S. and Hong Kong 15 year olds. International 3–19. https://doi.org/10.1016/j.jsp.2006.09.002.
Journal of Testing, 9(3), 215–237. Richardson, F. C., & Suinn, R. M. (1972). The mathematics
Ma, X. (1999). A meta-analysis of the relationship between anxiety rating scale: Psychometric data. Journal of
anxiety toward mathematics and achievement in mathe- Counseling Psychology, 19(6), 551–554.
matics. Journal for Research in Mathematics Education, Rueger, S. Y., Malecki, C. K., & Demaray, M. K. (2010).
30(5), 520–540. Relationship between multiple sources of perceived social
Marsh, H. W., & Martin, A. J. (2011). Academic self-concept support and psychological and academic adjustment in
and academic achievement: Relations and causal ordering. early adolescence: Comparisons across gender. Journal of
British Journal of Educational Psychology, 81(1), 59–77. Youth and Adolescence, 39, 47–61. https://doi.org/10.1007/
https://doi.org/10.1348/000709910X503501 s10964-008-9368-6
Marsh, H. W., Trautwein, U., Ludtke, O., K€ oller, O., & Ryan, A. M., & Patrick, H. (2001). The classroom social
Baumert, J. (2005). Academic self-concept, interest, grades,
environment and changes in adolescents’ motivation and
and standardized test scores: Reciprocal effects models of
engagement during middle school. American Educational
causal ordering. Child Development, 76(2), 397–416.
Research Journal, 38(2), 437–460. https://doi.org/10.3102/
Morgan, P. L., Farkas, G., and Maczuga, S. (2015). Which
00028312038002437
instructional practices most help first-grade students with
Salvucci, S., & Weng, S. Design Effects and Generalized
and without mathematics difficulties? Educational
Variance Functions for the 1990-91 Schools and Staffing
Evaluation and Policy Analysis, 37(2), 184–205. https://
Survey (SASS).
doi.org/10.3102/0162373714536608
Savery, J. (2006). Overview of problem-based learning:
Mu~ noz, M. A., & Chang, F. C. (2007). The elusive relation-
Definitions and distinctions. Interdisciplinary Journal of
ship between teacher characteristics and student academic
Problem-Based Learning, 1(1), 9–20.
growth: A longitudinal multilevel model for change.
Schulz, W. (2005, April 11–15). Mathematics self-efficacy and
Journal of Personnel Evaluation in Education, 20(3–4),
expectations: Results from PISA 2003 [Paper presenta-
147–164.
National Mathematics Advisory Panel. (2008). Foundations tion]. Annual Meeting of the American Educational
for success: The final report of the National Mathematics Research Association, Montreal, Canada. http://files.
Advisory Panel (Report No. ED04CO0015/0006 & eric.ed.gov/fulltext/ED490044.pdf
ED04CO0082/0001). U.S. Department of Education. Shim, S. (2000). A study on the relation between mathematics
http://www2.ed.gov/about/bdscomm/list/mathpanel/ anxiety & aggressiveness and mathematics proficiency of
report/final-report.pdf high school students. Journal of the Korean School
Organisation for Economic Co-operation and Development Mathematics, 3(2), 99–109.
(OECD). (2012). Data Base. Programme for International Shores, M. L., & Shannon, D. M. (2007). The effects of self-
Student Assessment (PISA). http://hdl.voced.edu.au/ regulation, motivation, anxiety, and attributions on math-
10707/292200. ematics achievement for fifth and sixth grade students.
Pajares, F. (1996). Self-efficacy beliefs in academic settings. School Science and Mathematics, 107(6), 225–236.
Review of Educational Research, 66(4), 543–578. Simons-Morton, B., & Chen, R. (2009). Peer and parent
Pajares, F., & Kranzler, J. (1995). Self-efficacy beliefs and influences on school engagement among early adolescents.
general mental ability in mathematics problem-solving. Youth & Society, 41(1), 3–25. https://doi.org/10.1177/
Contemporary Educational Psychology, 20(4), 426–443. 0044118X09334861
Pajares, F., & Miller, M. D. (1994). Role of self-efficacy and Singh, K., Granvile, M., & Dika, S. (2002). Mathematics and
self-concept beliefs in mathematics problem solving: A science achievement: Effects of motivation, interest, and
path analysis. Journal of Educational Psychology, 86(2), academic engagement. Journal of Educational Research,
193–203. 95(6), 323–332.
Pajares, F., & Miller, M. D. (1995). Mathematics self-efficacy Stevens, T., Olivarez, A., Lan, W. Y., & Tallent-Runnels,
and mathematics performances: The need for specificity of M. K. (2004). Role of mathematics self-efficacy and moti-
assessment. Journal of Counseling Psychology, 42(2), vation in mathematics performance across ethnicity.
190–198. https://doi.org/10.1037//0022-0167.42.2.190 Journal of Educational Research, 97(4), 208–221.
You et al. 13

Stronge, J. H., Ward, T. J., & Grant, L. W. (2011). What achievement. Early Childhood Research Quarterly, 49,
makes good teachers good? A cross-case analysis of the 109–121. https://doi.org/10.1016/j.ecresq.2019.05.001
connection between teacher effectiveness and student Woodward, J., & Howard, L. (1994). The misconceptions of
achievement. Journal of Teacher Education, 62(4), 339–355. youth: Errors and their mathematical meaning.
Tomasetto, C., Kinga, M., Veronica, G., & Patrick, A. Exceptional Children, 61(2), 126–136.
(2020). Math Anxiety Interferes with Learning Novel Yum, S. C., & Park, C. Y. (2011). Mediating effect of learn-
Mathematics Contents in Early Elementary School. ing strategy in the relation of mathematics self-efficacy
Loughborough University. https://doi.org/10.1037/ and mathematics achievement: Latent growth model anal-
edu0000602 yses. Mathematics Education, 50(1), 103–118.
Valiente, C., Parker, J. H., Swanson, J., Bradley, R. H., & Zhu, Y., & Leung, F. K. (2011). Motivation and achievement:
Groh, B. M. (2019). Early elementary student–teacher rela- Is there an East Asian model? International Journal of
tionship trajectories predict girls’ math and boys’ reading Science and Mathematics Education, 9(5), 1189–1212.