Content uploaded by Xintong Li
Author content
All content in this area was uploaded by Xintong Li on Aug 16, 2023
Content may be subject to copyright.
POSITIVE TSR MAY LEAD TO BETTER TEACHING 1
Positive Teacher-student Relationships May Lead to Better Teaching
Xintong Li1, Christi Bergin1, Amanda A. Olsen2
1 University of Missouri, USA
2 The University of Texas at Arlington, USA
Author Note
This is a preprint, please go to https://doi.org/10.1016/j.learninstruc.2022.101581 for the
published article.
Correspondence concerning this preprint should be addressed to
Xintong Li
Assessment Resource Center
College of Education and Human Development
University of Missouri.
2800 Maguire blvd, Columbia, MO, 65201, USA.
lixint@missouri.edu
POSITIVE TSR MAY LEAD TO BETTER TEACHING 2
Xintong Li, Christi Bergin & Amanda A. Olsen (2022)
1. Introduction
Substantial research literature indicates that positive teacher-student relationships (TSRs)
promote students’ academic achievement. One explanation is that students are more motivated to
learn when they have positive relationships with teachers (Urdan & Schoenfelder, 2006).
However, another plausible explanation is that teachers engage in higher quality teaching
practices when they have positive relationships with students. This is important because research
on school effectiveness consistently identifies high-quality instruction as one of the largest
school-based contributors to students’ academic achievement (Scheerens, 2001; Thoonen et al.,
2011). In the current study, we explored this alternative explanation. Figure 1 depicts the
relationship among TSRs, teaching practices, and student outcomes. There is extant research
evidence to support the bold arrows, which we briefly review next. Our study investigated the
dashed arrow representing the bidirectionality between positive TSRs and high-quality teaching
practices which, to our knowledge, no study has investigated. We used archival data from an
authentic teacher evaluation system in the U.S.
Figure 1
Relationships Among Teacher-Student Relationships, Student Outcomes, and Teaching Quality
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 3
1.1 Positive Teacher-Student Relationships (TSRs) and Student Outcomes
TSRs can be characterized on multiple dimensions such as conflict, closeness,
dependency, circularity, communication, or involvement (Pianta, 2001; Roorda et al., 2011;
Wubbels et al., 2006). In this study we focused on positive affect, specifically caring and mutual
enjoyment. When students feel teachers care about them, they work harder, engage in more
challenging academic activities, behave more appropriately for the school environment, are
genuinely happy to see their teacher, and meet or exceed their teacher’s expectations (Allen et
al., 2011; Bergin & Bergin, 2009; Hughes et al., 2008; Prewett et al., 2018; Quinn, 2017;
Wentzel, 2009). As positive Teacher-Student Relationships motivate students to become more
engaged in the classroom, learning should improve. Indeed, research confirms that positive TSRs
are associated with higher grade point averages and test scores from kindergarten through high
school (e.g., Baker, 2006; Cornelius-White, 2007; Curby et al., 2009; Hughes, 2011; Jia et al.,
2009; McCormick et al., 2013; O’Connor & McCartney, 2007). A meta-analysis found that
TSRs characterized by positive affect were strongly linked to student engagement and modestly
linked to achievement (Roorda et al., 2011). Importantly, the effect size may be larger for at-risk
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 4
students, suggesting that positive TSRs may help narrow the achievement gap (Gehlbach et al.,
2016).
Self-determination Theory (SDT) has been used to explain a potential mechanism for the
effect of positive TSRs on students’ engagement and achievement (Kincade et al., 2020; Roorda
et al., 2011). According to SDT, caring relationships meet a basic, innate need for relatedness or
feeling connected to others (Jang et al., 2010; Ryan & Deci, 2017). When relatedness needs are
met in a specific context, such as the classroom, individuals are more motivated behave in
adaptive ways, engage with tasks, persist in the face of failure, and respond creatively to
challenges in that context (Deci & Ryan, 2000).
1.2 Positive Teacher-Student Relationships (TSRs) and Teacher Outcomes
Just as positive TSRs may increase students’ motivation, positive TSRs may also increase
teachers’ motivation, effort, engagement, happiness, and confidence, which in turn, may result in
greater use of complex, high-impact teaching practices (van der Lans et al., 2020). Despite
teachers’ presumed basic need for relatedness with students in their classroom as predicted by
SDT, there is a paucity of research on the effect of TSRs on teachers. Spilt and colleagues state,
“there is little recognition of the internal needs that teachers themselves may have for positive,
personal relationships with individual students” (Spilt et al., 2011, p. 458). Studies have found
that positive TSRs may promote teachers’ self-efficacy (Mashburn et al., 2006), emotional well-
being (Milatz et al., 2015), and job satisfaction (Admiraal et al., 2019; Veldman et al., 2016), but
they do not address how positive TSRs may be associated with quality of teaching practices.
Although research is limited in the field of education, the field of organizational
psychology has addressed how positive relationships may drive adult performance. The Leader-
Member Exchange (LMX) theory that is commonly applied to adult workplaces is related to
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 5
SDT (Graves & Luciano, 2013). LMX theory asserts that the quality of a relationship between a
leader and a follower (akin to the relationship between a teacher and a student) predicts
performance (Gerstner & Day, 1997). A meta-analysis found that in adults, the link between
relationship quality and performance is likely mediated by motivation (Martin et al., 2016). The
development of deep connections at work results in higher quality work, increased effort, and
decreased attrition among adults (Manning, 2016). It is plausible that a similar effect occurs in
schools such that positive TSRs results in teachers’ increased teaching effort and performance.
1.3 Teaching Practices and Student Outcomes
Our study uses archival data from an authentic teacher growth and evaluation system
called the Network for Educator Effectiveness (NEE). In this evaluation system, teachers are
rated on four effortful, complex, high-impact teaching practices that are known to promote
student achievement (Hattie, 2009; van de Grift et al., 2014; van der Lans et al., 2017). The
association between positive TSRs and quality of teaching practices may depend on the specific
teaching practices under consideration. Other related lines of research have found that teachers’
self-efficacy (Burić & Kim, 2020; Holzberger et al., 2013), job satisfaction, and motivation
(Klusmann et al., 2008) were associated differently with specific teaching practices. Therefore,
our study examined how positive TSRs are associated with the quality of each of the four
different teaching practices. We use the definition of each teaching practice as operationalized by
NEE which are described next.
1.3.1 Cognitive Engagement (CE) in the Content
In the NEE system, Cognitive Engagement refers to active mental involvement by
students in learning activities, such as meaningful processing, strategy use, concentration, and
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 6
metacognition (Fredricks et al., 2004; Wang & Degol, 2014; Wang et al., 2014). Teachers
engage students cognitively in the content when they use strategies such as advanced organizers,
K-W-L charts, share-out, and shoulder-partner to connect instruction/activities with students’
lives. This allows teachers to show relevance, use authentic examples, present a puzzling
problem, and invite responses from all students.
1.3.2 Problem-Solving and Critical Thinking (PCT)
In the NEE system, Problem-Solving and Critical Thinking refers to skillfully applying,
analyzing, synthesizing, and evaluating information to reach a conclusion or solve a problem
(McCormick et al., 2015). Teachers promote their students’ critical thinking by requiring
students to explain or justify their thinking, evaluate others’ thinking, formulate challenging
questions, predict, determine what makes an argument valid, assess possible solutions, categorize
problems, or create map concepts (Wirkala & Kuhn, 2011). Teachers can also give students
challenging tasks that require persistent effort and various cognitive or metacognitive strategies.
In typical classrooms, the promotion of critical thinking is not always appropriate because there
are times when students should be practicing and over-learning skills that are foundational to
higher-level processing. Yet, this complex teaching practice may be too rare in typical
classrooms (Willingham, 2008).
1.3.3 Affective Engagement (AE) in the Content
In the NEE system, Affective Engagement in the Content refers to experiencing positive
emotion during the lesson. Teachers engage students affectively in the content with strategies
such as using materials and activities that students find interesting, using authentic examples,
pointing out progress, communicating enthusiasm for the content, helping students set achievable
but challenging goals, promoting self-efficacy, and giving students choices for classwork
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 7
(Bergin, 1999; Hidi & Renninger, 2006). As they use such strategies, teachers make lessons
more enjoyable, fun, and interesting (Archambault et al., 2017). AE refers to positive emotions
toward the content or lesson activities, not toward the teacher.
1.3.4 Instructional Monitoring (IM) During the Flow of the Lesson
In the NEE system, Instructional Monitoring refers to the teacher engaging in formative
assessment for the whole class and/or individuals and taking corrective action when needed. This
particular kind of formative assessment involves quick checks for understanding as the lesson is
progressing (Reddy et al., 2017). The purpose is to inform the modification of teaching and
learning activities in real-time to guide instruction. Strategies teachers may use include
questioning, asking students to solve problems on a whiteboard, or answering spot quizzes with
fist-to-five, thumbs up, or clicker techniques (Chien et al., 2016). IM has been associated with
student motivation, engagement, and achievement (Simons & Klein, 2007; Zhang & Hyland,
2018).
1.4 Research Questions
Teaching is a relentless and demanding job. It is particularly demanding when teachers
are expected to consistently implement effortful, complex, high-impact teaching practices. Given
that both SDT and LMX theory predict that positive TSRs should contribute to increased
motivation and performance among teachers, we hypothesize that variation in the quality of
complex teaching practices will result from variation in TSRs. Thus, our main research question
is “Do positive teacher-student relationships promote high-quality teaching practices?” In
addition, we have two ancillary research questions: “Does the effect vary by teaching practice?”
and “Does the effect vary by grade level?” We are interested in grade level variation because
studies find that close, positive teacher-student relationships often decrease steadily from 1st
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 8
grade to the end of primary school, and across middle school, before possibly stabilizing in high
school (Gillen-O'Neel & Fuligni, 2013; Hughes et al., 2012). The change in TSRs across grades
may alter the relationship between TSRs and the quality of instruction.
2. Methods
We addressed our research questions using archival data from a state-wide teacher
growth and evaluation system, the Network for Educator Effectiveness (NEE). This system is
used by 285 preK-12 school districts across the state of Missouri in the United States. NEE’s
teaching effectiveness measures are aligned with the Interstate Teacher Assessment and Support
Consortium (InTASC) teaching standards (Council of Chief State School Officers, 2011), which
apply to all subject areas and grade levels.
2.1 Participants
NEE member school districts select three to six teaching practices to focus on from a
pool of 26 teaching practices based on district priorities. Thus, data for this study comes from a
subset of member districts that decided to focus on the four target teaching practices of interest.
As a result, our analysis of each teaching practice involves overlapping but different samples
(see Table 1). The data used in the current study are based on the means of student reports
aggregated at the teacher level and are nested in schools. These data include two waves of
student surveys in the 2017-2018 and 2018-2019 academic years. Only teachers of regular
education classrooms from 4th to 10th grade were included (i.e., no students in our sample
indicated teacher subject areas as either English language learning or special education).
The 2019 Missouri average proficiency rate on the state test was 49% for English
language arts and 42% for mathematics. In comparison, the averages of our total sample were
also 49% for English language arts, but 31% for mathematics. In Missouri, 73% of k-12 students
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 9
were White, non-Hispanic and 50% were eligible for free or reduced-price meals. Our total
sample was 80% White and 26% eligible for free or reduced-price meals.
Table 1
Number of Teachers in Each Sample with Sample Overlap
CE sample
PCT sample
IM sample
AE sample
CE sample
566
448
405
140
PCT sample
118
672
564
181
IM sample
110
383
733
180
AE sample
-
-
144
217
Total: 844
AE sample
CE sample
PCT sample
Overlap of four
samples: 88
Note. The main diagonal shows the sample sizes of the four samples for separate analyses. The upper-right panel
above the main diagonal shows the number of cases in the overlap of two teaching-practice samples. The lower
left panel below the main diagonal shows the number of cases in the overlap of three teaching-practice samples.
CE means cognitive engagement, PCT means problem solving and critical thinking, IM means instructional
monitoring, and AE means affective engagement in the content.
2.2 Measures
2.2.1 Teaching Effectiveness & Teacher-Student Relationships
Data are based on the student survey of teaching effectiveness in the NEE database, called
the Teacher Effectiveness Student Survey (TESS; see https://neeadvantage.com). TESS is
modular, meaning member districts can survey students only on the teaching practices of interest
to the district. When a district selects a specific teaching practice, all items associated with that
teaching practice (3 to 5 items) automatically populates an online student survey.
1. Cognitive Engagement (CE). Four items addressed students’ perception of the degree to
which a teacher used strategies that cognitively engaged students in the content (α = .89).
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 10
2. Problem-Solving and Critical Thinking (PCT). Four items addressed students’ perception
of the degree to which a teacher used strategies that promote PCT (α = .91).
3. Affective Engagement in the Content (AE). Five items addressed students’ perception of
the degree to which a teacher used strategies that affectively engaged students in the
content (α = .92).
4. Instructional Monitoring (IM). Four items addressed students’ perception of the degree to
which a teacher monitored learning at the individual and whole-class level as the lesson
was progressing, and adjusted teaching as needed (α = .96).
On the same survey, students responded to five items that asked about their perspective on their
teacher-student relationship (α = .96). Asking for student perceptions is noteworthy because a
recent meta-analysis found that most studies measured TSRs from the teachers’ perspective
(Kincade et al., 2020). See Table 2 for a list of the items.
Table 2
Items Comprising Teaching Practices and Teacher-student Relationship Variables
Variable
Items
Cognitive
Engagement
(CE)
• This teacher expects us to think a lot and concentrate in this class.
• This teacher’s lessons make us think deeply.
• This teacher's lessons make us think the whole class time.
• This teacher wants us to ask questions during lessons.
Problem-Solving
and Critical
Thinking
(PCT)
• This teacher asks “how?” and “why?” questions to make us think more.
• This teacher waits a while before letting us answer questions, so we have time to
think.
• This teacher makes us compare different ideas or things.
• This teacher makes us use what we learn to come up with ways to solve problems.
Affective
Engagement in
the Content
(AE)
• This teacher makes lessons interesting.
• This teacher points out how this topic is important to our lives.
• This teacher gives us choices in our classwork.
• This teacher tells us that we can all be successful if we try hard.
• This teacher gets us excited about the subject.
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 11
Instructional
Monitoring
(IM)
• This teacher checks often to make sure we understand the lesson as we go along.
• This teacher explains the lesson in different ways if we don’t get it at first.
• This teacher knows when we understand the lesson.
• This teacher has a way to make sure everyone is learning.
Teacher-student
Relationships
(TSRs)
• This teacher knows me and cares about me.
• Students enjoy being with this teacher.
• This teacher enjoys working with students.
• This teacher is friendly.
• Students can talk to this teacher if they have a problem.
TESS uses a 4-point Likert scale (0 = Not true; 1 = Sort of true; 2 = True; 3 = Very true).
TESS has shown good factor structure and the subscales comprising each teaching practice have
a Cronbach’s Alpha ranging from .77 to .90 (Grajeda et al., 2017). Latent factor scores were
generated for the five measures based on the mean item scores from student ratings of their
teachers, and teachers were rated by at least 5 students (very few exceptions were rated by 4). In
addition, multidimensionality of teaching was distinguishable by 4th – 12th grade students. That
is, students might rate the same teacher as better at one teaching practice than another (Grajeda et
al., 2017).
2.2.2 Other Variables
We included the following variables as covariates because they may affect teaching
practices and teacher-student relationships.
Years of Experience. Teachers reported their years of teaching experience. More years
of experience predicts better student behavior and achievement (Ladd & Sorensen, 2017; Nye et
al., 2004). Furthermore, one meta-analysis found that the effect of TSRs on achievement was
larger in studies with teachers who had more years of teaching experience (Roorda et al., 2011).
Core Subject Area. Students reported whether or not a teacher primarily taught a core
subject (e.g., mathematics, English language arts, science or social studies). We tested whether
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 12
results were similar between core and other courses. Cohen et al. (2018) found that teaching
quality is not a uniform construct across varying subjects. Compared with teachers in core
subjects, teachers in non-tested subjects may use different instructional practices and are often
evaluated differently (Goe & Holdheide, 2011).
Grade Level. Teacher grade level was computed as a continuous variable based on the
mean grade level identified by student responses. Previous research has demonstrated that
teachers at varying grade levels had different levels of self-efficacy for implementing different
teaching practices (Klassen & Chiu, 2010). Student engagement, teacher behavior, and teacher-
student interactions are often different across subject areas and grade levels (den Brok et al.,
2004; Marks, 2000).
Demographics. Student responses on the TESS were strictly anonymous. Thus, student-
level demographic data were not available. School-level demographic data were obtained from
the Missouri Department of Elementary & Secondary Education (DESE). This included the
percentage of students receiving free or reduced-price lunch, total student enrollment, and the
percentage of White students. School level demographics predict teacher job satisfaction, stress,
and burnout, which may be associated with the quality of teaching practices used (Hamre et al.,
2008; Johnson et al., 2012; Kyriacou, 2001; Olsen & Huang, 2019; Perry & McConney, 2010).
Achievement. Due to students’ anonymity, student-level achievement data was not
available. School-level data on the state proficiency test –Missouri Assessment Program (MAP)
– was obtained from the state department of education. We retrieved the percentage of students
at or above the proficiency level for mathematics and English language arts and computed the
mean percentage as a school-level achievement variable. See http://dese.mo.gov for more details.
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 13
2.3 Procedures
TESS was delivered online during a window of time specified by the school. An access
code unique to each teacher needed to be entered to ensure students evaluated the correct teacher
and each access to the survey was authorized. A proctor other than the evaluated teacher
administered the survey using standard administration scripts provided by NEE. The proctor read
the instructions to the students, informed them of the purpose of the survey, ensured the
anonymity of their responses, emphasized the voluntary nature of the survey, and its importance
for school improvement. Students were encouraged to ask questions, but the proctors were
instructed not to interpret any survey items to avoid influencing student responses.
Three screening items were distributed across the survey (e.g., “I am being totally honest
on this survey”), which help improve survey validity and identify inattentive responses (Cornell
et al., 2012). Surveys that failed two of the three screening items were flagged for manual
review. Students who indicated in a fourth screening item that they were new to the class (less
than a month) were also excluded. Data use was approved by the university’s Institutional
Review Board. All member districts consented to using their data for research purposes.
2.4 Statistical Analysis
We hypothesized that positive teacher-student relationships (TSRs) would promote
higher-quality teaching practices (TPs). To test this, we conducted four regression models with
TPs as the outcomes, and TSRs, grade level, and other variables as predictors, including an
interaction term between TSRs and grade level. Teachers’ TPs and TSRs scores in the models
and following analyses all have controlled for prior scores (see the Appendix for more method
details). However, it is also possible that the direction of the effect is reversed. That is, TPs may
predict TSRs, or that both are due to an unspecified confounding variable. Since an experimental
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 14
design is not an option for the current inquiry due to ethical concerns, we have to rely on archival
data to examine the hypothesized direction of effect, and at the same time address the consequent
methodological challenges, including an alternative direction of effect (i.e., TP → TSR),
potential confounding variables, an illusory halo effect, and common method variance.
To examine the direction of the effect, we employed direction dependence analysis
(DDA; Li & Wiedermann, 2020b; Wiedermann et al., 2020; Wiedermann & Li, 2018), an
innovative causal model selection method, as a supplementary procedure to regression analysis.
DDA is a framework that consists of various statistical tests that can generate different result
patterns to differentiate competing models with different directions of effect (i.e., TP → TSR or
TSR → TP) and to detect potential confounding (i.e., TP confounder → TSR). When none of
the models are selected, the results are inconclusive, which may be due to the lack of sample
size. This framework was proposed to bridge the recent advancements in data science and
common modeling practices among educational researchers. Compared to other methods widely
used by educational researchers to make causal claims with non-experimental data, DDA has
minimal theoretical assumptions; instead, it primarily holds testable distributional assumptions
and requires a large sample size to achieve desirable statistical power (Li & Wiedermann, 2020b;
Wiedermann & Li, 2018). Specifically, for the current study, we applied Conditional Direction
Dependence Analysis (CDDA; Li & Wiedermann, 2020b) as a supplementary procedure to
examine the hypothesized direction of the effect. CDDA is an extension of standard DDA and is
capable of examining the direction of effect when there is an interaction between the effect of
interest. This extension can also achieve more statistical power and avoid some bias in
moderation models. The DDA framework has been used effectively in prevention science (Musci
& Stuart, 2019; Wiedermann et al., 2020), psychology (Wendt & Bartoli, 2019), public health
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 15
(Chew et al., 2020; Pérez-Mengual et al., 2021), and education (Sebastian et al., 2018;
Wiedermann et al., 2020). We conducted direction dependence analysis using the SPSS add-ons
that are publicly available at https://www.ddaproject.com and followed the guidelines for the
analysis and interpretation (Li et al., 2022; Li & Wiedermann, 2020a, 2020b). To control for
time-invariant confounding, teachers’ TP and TSR ratings in the models and following analyses
all have controlled for prior scores in advance, meaning the student ratings in the following
analyses were residualized variables, and therefore, prior scores are not displayed in the results.
Other potential confounding can be detected by DDA procedures (Wiedermann & Li, 2019;
Wiedermann & Sebastian, 2019). See the Appendix for more details about the method and its
implementation.
Any halo effect should be minimal as the aggregation of student ratings were used for
teacher scores (Feeley, 2006; Feldman, 1986). Moreover, various types of common method
variance should also be minimal as a result of strict student anonymity, the use of mean
aggregation, and districts’ changing of prioritized teaching practices across the years of
implementation. See the Appendix for more details about how challenges regarding the halo
effect and common method variance were addressed.
Though multilevel modeling (MLM) is a preferred method as teachers were nested within
schools, we used ordinary least squares (OLS) regression because DDA does not apply to MLM.
When a clustering effect is present, standard errors may be biased and erroneous conclusions
may be made when using OLS instead of MLM (Guo & Zhao, 2000; Hox, 1998). Therefore, we
used cluster robust standard errors for parameter testing, which is considered to be a valid
alternative to MLM (Berger et al., 2017; Cameron et al., 2011; Huang, 2016; Huang & Li, in
press; Kauermann & Carroll, 2001).
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 16
3. Results
Table 3 shows the directional conclusions across teaching practices and grade levels. The
hypothesized TSR → TP models were supported across all grade levels in the Cognitive
Engagement in the Content (CE), Problem-Solving and Critical Thinking (PCT), and
Instructional Monitoring (IM) samples, except that a confounder effect was detected between
TSR and IM scores in middle schools. However, contrary to our hypothesis, TSRs appeared to be
an outcome, rather than antecedent, of teachers’ Affective Engagement (AE) in the Content
teaching practice in secondary schools.
Table 4 shows the regression results of the supported models. TSRs and grade level were
statistically significant predictors of CE, PCT, and IM across all grade levels after controlling for
the covariates. The local effect sizes (Cohen's ; Cohen, 2013) were small (
,
) to medium (
,
) in the four supported
models. There was a grade-level main effect on teachers’ CE, PCT, and IM scores. After
controlling for TSRs and other covariates, teachers in higher grades tended to receive better
ratings in the three teaching practices from students. Grade level did not affect TSRs after
controlling for teachers’ Affective Engagement in the Content practices and other covariates.
Furthermore, the interaction term was statistically significant for Cognitive Engagement and
Instructional Monitoring. That is, for students in higher grades, TSRs had a greater effect on
student ratings of teachers’ CE and IM practices. There was no statistically significant interaction
for Problem-Solving and Critical Thinking and Affective Engagement in the Content. Figure 2
provides the interaction plots illustrating the difference in effects among 4th, 7th, and 10th grades.
Due to the smaller AE sample size, statistical power may be inadequate to detect the interaction
(Shieh, 2008), but this should not be the case for PCT.
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 17
Table 3
Data Supported Models across Samples and Grade Levels
Grade Levels
Direction of Effect
CE Sample
PCT Sample
IM Sample
AE Sample
4th
TSR→CE
TSR→PCT
TSR→IM
Inconclusive
5th
TSR→CE
TSR→PCT
TSR→IM
Inconclusive
6th
TSR→CE
TSR→PCT
TSR→IM
Inconclusive
7th
TSR→CE
TSR→PCT
Confounder
AE→TSR
8th
TSR→CE
TSR→PCT
Confounder
AE→TSR
9th
TSR→CE
TSR→PCT
TSR→IM
AE→TSR
10th
TSR→CE
TSR→PCT
TSR→IM
AE→TSR
Note. The grade levels represent the mean of student reported grade levels. Conditional Directional
Dependence Analysis (CDDA; Li & Wiedermann, 2020b) was used to examine the direction of effect
between TSRs and TPs at particular mean grade levels, and for the current analysis, Hilbert Schmidt
Independence Criterion (HSIC; based on 200 bootstrap resamples) and the cube-based Higher Order
Correlation Difference test (ΔHOCc; based on 1000 bootstrap resamples and 95% bootstrap confidence
interval) are used. See the Appendix for more details.
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 18
Table 4
Regression Results of the Supported Models
Outcomes
Predictors
B(SE)
Cognitive Engagement
(R2 = .384,
n = 566)
Problem-solving and
Critical Thinking
(R2 = .476,
n = 672)
Instructional Monitoring
(R2 = .629,
n = 733)
Teacher Student
Relationships
(R2 = .625,
n = 217)
Intercept
-.091(.367)
.530(.285)
.414(.267)
.436(.258)
Affective Engagement
−
−
−
.871(.071)***
TSR
.425(.086)***
.639(.082)***
.662(.058)***
−
Grade level
.069(.030)***
.092(.019)***
.043(.016)**
-.023(.021)
Teacher years of experience
.002(.003)
-.001(.002)
-.001(.003)
-.004(.004)
Free/Reduced lunch (%)
-.001(.002)
-.001(.002)
.002(.001)
.003(.003)
Enrollment (100ct)
-.002(.007)
-.010(.006)
-.008(.005)
.024(.009)*
White enrollment (%)
-.002(.004)
-.007(.002)***
-.005(.002)*
-.007(.003)*
MAP proficiency rate (%)
-.002(.006)
-.005(.004)
-.002(.004)
.001(.004)
Core subject
.203(.051)***
.160(.040)***
.033(.030)
.117(.047)*
TSR × Mean Grade levels
.058(.020)*
.022(.022)
.047(.016)**
−
AE × Mean Grade levels
−
−
−
-.005(.034)
Note. Both TSRs and TPs are residualized scores for 2018-2019 school year controlling both prior TSR and TP scores. Therefore, prior scores are not listed here
as predictors. See Table A3 in the Appendix for the results of prior-score-only models. The parameter tests are based on adjusted cluster-robust standard errors
(Cameron et al., 2011).
*** p < .001, ** p < .01, * p < .05
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 19
Figure 2
Predicted Outcome Scores Across Selected Mean Grade Levels
Note. Shading represents the 95% confidence intervals.
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 20
Core subject teachers, compared to teachers in other subject areas (e.g., physical and
health education, fine arts, foreign languages), were rated higher in Cognitive Engagement and
Problem-Solving and Critical Thinking, but similar in Instructional Monitoring. In the AE
sample, core subject teachers tended to have better relationships with students, when controlling
for other variables. Other control variables – teacher years of experience, school free or reduced
lunch percentage, or school-level MAP proficiency rate – did not show any significant influence
on student ratings of either TSR or teaching practices. Thus, these variables do not have a
detectable effect within the scope of one school year, after controlling for the previous year’s
survey ratings, in this sample.
4. Discussion
Substantial research indicates that positive teacher-student relationships (TSRs) are
linked to increased motivation and achievement in students. We hypothesize the same effect for
teachers, although this has not been a focus of the research literature. Our hypothesis is
supported by both Self-Determination Theory and Leader-Member Exchange Theory which
predicts that adults increase effort and performance in work contexts where they have positive
relationships. That is, positive TSRs should lead to teacher’s greater use of high-impact teaching
practices that are complex and effortful to implement. We explored our hypothesis using archival
data from a state-wide authentic teacher growth and evaluation system, the Network for Educator
Effectiveness (NEE), in Missouri, USA. Our analysis revealed three key results: (1) positive
TSRs predict higher-quality instruction and this effect varies by (2) teaching practice and (3)
grade level.
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 21
4.1 Positive Teacher-Student Relationships Lead to Higher-Quality Teaching
Our results, using student ratings, confirm our hypothesis that positive TSRs lead primary
and secondary teachers to more effectively enact three complex teaching practices examined in
this study – Cognitive Engagement (CE) in the Content, Problem-Solving and Critical Thinking
(PCT), and Instructional Monitoring (IM). Our study supports previous studies that found
classrooms with more positive TSRs, have teachers who are more likely to check-in, monitor,
scaffold, and/or provide constructive feedback to students (Reddy & Dudek, 2014; van de Pol et
al., 2010), have greater confidence in their students’ abilities (Summers et al., 2017), and use
better scaffolding strategies for critical thinking (Daws, 2005). We extended these studies by
testing the direction of effect between positive TSRs and high-quality teaching practices.
Furthermore, our results extend research using LMX theory (Gerstner & Day, 1997) from the
field of organizational psychology to the field of education by demonstrating that positive
relationships may lead to improved performance among adult workers.
4.2 The Effect Varies by Teaching Practice
An exception to our results discussed above was use of teaching strategies to promote
Affective Engagement in the Content. Contrary to our hypothesis, our results indicate that AE is
more likely to lead to, rather than result from, positive TSRs in secondary schools. That is, when
teachers put effort into making lessons interesting and enjoyable for students, and provide
opportunities for autonomy in choosing learning tasks, secondary students view their teacher as
more caring. This study supports previous literature indicating that making lessons interesting is
perceived by students as an act of caring (Jeffrey et al., 2013; Wentzel, 1997, 2009). We extend
these studies by using quantitative data to test the direction of the effect.
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 22
Our results support the logic model in Figure 3 for secondary grades. That is, teachers’
use of strategies that affectively engage students in the content leads to positive TSRs. Recall
that Affective Engagement in the Content refers to positive emotions toward the content or lesson
activities, not toward the teacher. Other factors may be involved in promoting positive TSRs as
well, such as being sensitive and warm, using non-coercive discipline, and behaving prosocially
toward students (Bergin & Bergin, 2009; Prewett et al., 2019). When positive TSRs are
established, they may motivate teachers to engage in more complex teaching practices (CE, IM,
PCT) with greater frequency and higher quality.
Figure 3
Association between Teaching Practices and Teacher-Student Relationships
Given that our hypothesized TSR → AE model was rejected, we are more confident that
a halo effect was not strongly present. That is, the mean aggregated student ratings of the other
teaching practices were not a biased result of the halo effect from student relationships with their
teachers, after prior teachers’ scores were controlled. Our reasoning is that neither a consistent
confounding effect nor a consistent driving force of all four teaching practices were identified.
That is, neither the general impression halo model nor the salient dimension halo model was
supported. See the Appendix for more details.
4.3 The Effect Varies by Grade Level
4.3.1. Main Effect
Affective
Engagement in
the Content
+ Other Factors
Positive
Teacher-
Student
Relationships
Increased Use of
Complex
Teaching Practices
(Cognitive engagement, Critical
Thinking, Instruction Monitoring)
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 23
There was a statistically significant positive main effect for grade level on teaching
practices while controlling for other covariates. That is, older students reported teachers more
frequently used the effortful, complex, high-impact teaching practices we studied. One possible
explanation is that it is easier for teachers to use Cognitive Engagement, Problem-Solving and
Critical Thinking, and Instruction Monitoring strategies with older students because of their
increased cognitive abilities. In adolescence, information processing speed, complex working
memory, executive functioning (Coyle et al., 2011; Kail & Ferrer, 2007; Luciana et al., 2005;
Steinberg, 2009), use of memory strategies, and vivid recall (Brainerd et al., 2004) increase.
These cognitive developments allow adolescents to reason abstractly, incorporate new
information, and monitor learning progress faster and easier. In addition, as course content
becomes more challenging in higher grades, teachers may tend to use these complex strategies
more often. Furthermore, teacher certification requires teachers in secondary grades to have
deeper content expertise. As content specialists, rather than generalists, secondary teachers may
engage in more complex teaching practices regarding their subject of expertise. However, there
is mixed evidence on whether teacher certifications are actually associated with student
outcomes (Goldhaber & Brewer, 2000; Kusumawardhani, 2017). These explanations are
speculative and need to be confirmed by further research.
4.3.2 Interaction Effect
There was also a statistically significant interaction effect between TSRs and grade level
for CE and IM (see Figure 2). As we mentioned above, positive teacher-student relationships
tended to decrease steadily from 1st grade to the end of middle school, before stabilizing in high
school (Gillen-O'Neel & Fuligni, 2012; Hughes et al., 2012). Because TSRs are lowest for
teachers in higher grades, the effect of positive relationships on teaching practices may be
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 24
stronger for teachers who manage to have positive relationships despite the developmental trend.
Supporting this logic, a meta-analysis found a stronger effect between TSRs and achievement in
secondary schools compared to primary schools, with student engagement as a mediator (Roorda
et al., 2017).
In contrast, there was not a statistically significant interaction effect between TSRs and
grade level for PCT and AE. This suggests that TSRs have an equal effect on Problem-Solving
and Critical Thinking across all grade levels. A possible explanation is that PCT is a challenging
and rare teaching practice (Willingham, 2008), therefore, teachers may not use it frequently
during class even though they have good relationships with the students. As for Affective
Engagement in the Content, the effect may simply be undetectable due to the small sample size.
4.4 Implications for Practice
Our results suggest that when teacher-student relationships (TSRs) are positive, teachers
are more likely to engage in effortful, complex, high-impact teaching practices that are
associated with student learning. Our results suggest that when schools aim to improve quality of
teaching practices, they would do well to attend to improving the quality of TSRs first. One way
to improve TSRs may be to use teaching strategies that affectively engage students in the
content. Effects are likely to occur for all students but may be strongest for secondary students.
4.5 Limitations
A major limitation of the current study is the small size of the AE sample for primary
grades, which reduced the interpretability of results. We cannot rule out a confounding variable
in the relationship between Affective Engagement in the Content and TSRs in lower grade levels.
Another limitation is that, since our archival data were collected from an authentic teacher
evaluation system where student-level information was strictly anonymous, student-level
HIGH QUALITY TSR LEADS TO BETTER TEACHING PRACTICES 25
measurement error cannot be fully taken into consideration, although it was attenuated by using
mean aggregation (Dunn et al., 2015; Richter & Brorsen, 2006). Further, schools in the samples
have somewhat lower math proficiency rates than the state average, which may be a result of
selection bias into the teacher evaluation system. Therefore, the samples of the study are
representative of schools with somewhat lower student performance in math
26
Appendix
Supplementary Statistical Analysis and Results
We hypothesized that positive teacher-student relationships would predict higher-quality
teaching practices. To test this, we conducted four regression models. The outcome variables
were high-quality teacher practices (Cognitive Engagement, Problem-Solving and Critical
Thinking, Instruction Monitoring, and Affective Engagement in the Content). The predictor
variables included TSR, teacher-level variables (taught a core subject, grade level, years of
experience), and school-level variables (total enrollment, White enrollment, free lunch, MAP
proficiency rate). In addition, an interaction between TSRs and grade level was included.
Contrary to our hypothesis, it is also possible that the direction of the effect is reversed.
That is, Teaching practices (TPs) may predict TSRs, or that both are due to an unspecified
confounding variable. Since an experimental design is not an option for the current inquiry due
to ethical concerns, we have to rely on observational data to examine the hypothesized direction
of effect, and at the same time address the consequential methodological challenges, including an
alternative direction of effect (i.e., TPs → TSR), potential confounder, illusory halo, and
common method variance.
Alternative Direction of Effect
Establishing cause-effect relationships between variables is one of the focal interests of
educational researchers. Though conventionally causal direction or the direction of effect is
examined with experimental or quasi-experimental studies in the educational context, we used an
innovative method, Direction Dependence Analysis (DDA; Li & Wiedermann, 2020b;
Wiedermann & Li, 2018; Wiedermann & von Eye, 2015) to validate the hypothesized models
(TSR TP) against models with the alternative direction of effect (TP TSR).
27
In the current study, we investigated whether our hypothesized direction of effect – that
positive TSRs cause higher-quality teaching practices – was supported across grade levels. That
is, CDDA was used to validate the hypothesized explanatory model (i.e., TSR TP) against the
plausible alternative model (i.e., TP TSR), with a reversed flow of causality between the two
variables across grade levels.
Potential Confounder
Although we have included variables that may affect both TSRs and TP, to minimize
potential confounding, we also adjusted both the TP and TSR scores for the 2018-2019 school
year using the prior scores. Specifically, as CDDA requires the same set of covariates in the
competing models (Li et al., 2020; Li & Wiedermann, 2020b), we use both prior scores to obtain
the adjusted scores (i.e. and ) shown in Eq. (A.1) and Eq. (A.2),
where the subscripts of TP and TSR indicate the school year; ,, and are the
coefficients ( are not statistically significant); and and are the residuals
representing the teacher-level variance unexplained by the prior scores, and are the focal
variables we used for the analysis. Using the residualized variables, the models we build are to
examine the synchronous effect between TSR and TP controlling their prior scores and cross-
lagged effects (Li et al., 2020). That is to say, the scope of the analysis will be narrowed within
the 2018-2019 school year and potential time-invariant confounders can be greatly reduced
(VanderWeele, 2019). Besides the efforts to minimize potential confounding, we also used the
DDA framework to test confounding effects (Wiedermann & Li, 2019; Wiedermann &
Sebastian, 2019).
28
Assumption Check and Implementation
Table A1
The List of Outliers Excluded from Conditional Direction Dependence Analysis
Outliers
Mahalanobis' Distance
DFBETA
Mean Grade Level | Years of
Experience | Core subject (T/F)
CE
Sample
PCT
Sample
IM
Sample
CE
Sample
PCT
Sample
IM
Sample
CE
Sample
PCT
Sample
IM
Sample
#1
36.87
61.84
48.20
-0.000
.000
.000
8 |14| F
5| 22| T
5 |22| T
#2
33.17
61.15
47.53
-0.002
-.002
.001
5 | 0 | F
6 | 2 | T
6 | 2 | T
#3
29.59
60.74
46.57
0.002
.000
.000
6 | 9 | F
4 | 1 | F
4 | 1 | F
#4
27.68
58.46
44.45
0.000
.000
.001
6 |14| F
6 | 7 | F
6 | 3 | T
#5
58.30
44.24
.000
.000
4 | 4 | T
5 | 0 | T
#6
58.23
44.20
.001
.000
6 | 3 | T
4 | 4 | T
#7
57.97
44.09
.000
.000
4 | 9 | T
6 | 7 | F
#8
57.06
44.05
.000
.000
5 | 5 | T
4 | 9 | T
#9
28.90
43.06
-.001
.000
5 | 0 | F
5 | 5 | T
#10
26.76
30.28
.001
-.001
4 | 3 | F
4 | 0 | F
#11
26.71
.001
4 |15| F
Note. Mahalanobis' Distance is used to detect multivariate outliers ( ), and no outliers
were identified in the AE sample. DFBETA indicates the influence of individual cases on the main effect of Teacher-
student relationships on respective teaching practices.
Following Wiedermann and Li (2018), the distribution assumptions of the adjusted focal
variables were examined, and multivariate outliers were excluded from CDDA but not from the
multiple regression analysis, as these cases may bias the directional conclusion but are not
influential on the main effect (see Table A1 for a list of the excluded cases). Two CDDA tests
were used, including the Hilbert Schmidt Independence Criterion (HSIC; Gretton et al., 2005)
and the cube-based Higher-Order Correlation Difference test (ΔHOCc; Wiedermann & Li, 2018)
after we built the conventional multiple regression models and their alternative models with
reversed causal directions. That is, our focal variables, adjusted TSRs and TP, served in turns as
the dependent and the independent variables in pairs of competing models across each sample.
29
HSIC was used to examine residual-predictor independence in the regression models. A HSIC
value not significantly different from zero indicates the model is more likely to have the “true”
direction of effect when its directionally competing model shows a non-zero HSIC value. When
HSIC values in both competing models are significantly larger than zero (i.e., a predictor-
residual dependence was observed in both models) an unconsidered confounder may exist
(Wiedermann & Li, 2020). The ΔHOCc test was used to examine the distributional properties of
our hypothesized dependent and independent variables after controlling other covariates. A
positive value indicates a model is more likely to be directionally “true” when its competing
model is not shown positive. Both tests are based on bootstrap procedures.
The Halo Effect
Halo effect (Thorndike, 1920), or illusory halo (Cooper, 1981), is a pervasive challenge
when teachers are evaluated by students (Wolfe & Song, 2015). It represents the added constant
(positive or negative) to all of a given rater’s true scores (Feeley, 2006; Feldman, 1986). The
halo effect is likely to be minimal in the current study because it can be reduced by a factor equal
to the number of raters per teacher when mean aggregation is used and when raters have
adequate experience with the ratees (Feeley, 2006; Feldman, 1986). However, since CDDA is
capable of evaluating causal models, to probe the potential role of any remaining halo, we
examined how our data may fit the three causal models (Fisicaro & Lance, 1990; Solomonson &
Lance, 1997) that are often used to explain halo effect: 1) the general impression model, which
suggests that a separate general impression affects the observed scores; 2) the salient dimension
model, which indicates that one salient trait influences the evaluations of other traits; and 3) the
inadequate discrimination model, which attributes the halo to the raters’ failure to discriminate
among conceptually distinct aspects of a ratee’s behavior. Previous analysis of TESS indicated
30
that both upper primary and secondary students do distinguish the dimensionality of teacher
effectiveness (Grajeda et al., 2017), so the first and the second models are our models of interest.
A confounding effect is detectible with CDDA when the general impression model better
explains the halo, and the halo effect is big enough to bias our model estimates. Otherwise, the
salient dimension model may be a better model to explain the data when TSRs is shown to be a
driving force for all four teaching practices, though it is equally possible that the TSR→TPs
hypothesis may reflect the true direction of effects.
Common Method Variance (CMV)
As both the dependence and independent variables were from the same survey by the
same group of students through the same platform, common method variance (CMV) may exist
and bias the variance between the constructs of TSRs and TP. CMV is the spurious variance
between dependent and independent variables due to shared method and/or source (Podsakoff et
al., 2003). Since we have controlled teachers’ prior scores measured with the same items through
the same platform, CMV due to item characteristic effects and measurement context effects were
controlled. In addition, item context effects were also addressed as different districts prioritized
different teaching practices over years, and therefore, the context of the item was not the same
across districts and years. Moreover, mean aggregation of student ratings and strict student
anonymity also helped minimize common rater effects and other measurement errors (Feeley,
2002; Podsakoff et al., 2003; Richter & Brorsen, 2006).
Results
Before addressing our research questions, we investigated the association between
variables without regard to causal direction by examining the eight regression models. That is,
four multiple regression models as our hypothesized models with the interaction terms between
31
grade level and TSRs (i.e., TSR | Grade Level TP) and four corresponding alternative models
with the interaction terms between grade level and TPs(i.e., TSR | Grade Level TP). There
was a statistically significant relationship between TSRs and teaching practices in all of the eight
models. Therefore, before interpreting the results of the multiple regression analyses, it is
necessary to select the models with the direction of effect better supported by the data. The
CDDA results are displayed in Table A2.
Cognitive Engagement (CE)
To interpret Table 4 and Table 5, look at the column for CE first. For all grades
and
. This means that the hypothesized model (TSRCE) is
more likely to be the “true” model. Moreover supports the same conclusion for all grade
levels as
(the zero point is below the 95% bootstrap confidence interval
(BCI) in the target model), and
(the zero point is within or above the 95%
BCI).
Problem-Solving and Critical Thinking (PCT)
Similarly, the hypothesized TSR PCT model was also supported by both HSIC and
tests across grade levels, except for 5th and 6th grades. As HSIC remains powerful across
all the grades for the PCT sample, it is less likely that it consists of cases with different directions
of effect (Li & Wiedermann, 2020b). That is to say, Problem-Solving and Critical Thinking is
32
Table A2
Conditional Direction Dependence Analysis Results across Grades
Hypothesized Models
Grade
Levels
TSR → CE
TSR → PCT
TSR → IM
TSR → AE
HSIC
ΔHOCc
HSIC
ΔHOCc
HSIC
ΔHOCc
HSIC
ΔHOCc
4th
.033
.51~17.24
.003
3.87~14.72
.004
4.70~16.68
.002
2.61~23.49
5th
.004
3.23~28.15
.005
7.54~23.49
.008
8.10~22.87
.005
3.78~34.61
6th
.007
7.73~31.70
.009
11.26~27.65
.022
7.39~21.90
.012
-7.11~7.85
7th
.016
2.22~ 9.18
.016
1.77~7.91
.075**
-.88~3.10
.073***
-.82~ 3.77
8th
.050
1.16~7.76
.037
1.71~7.25
.046*
-.00~7.96
.027*
.74~9.28
9th
.033
1.68~12.72
.034
2.81~11.13
.017
1.63~16.77
.009*
-1.3~5.84
10th
.015
.92~ 6.58
.014
1.48~6.54
.008
.83~13.00
.004*
-.99~ 2.71
Alternative Models
CE→ TSR
PCT→ TSR
IM→ TSR
AE→ TSR
HSIC
ΔHOCc
HSIC
ΔHOCc
HSIC
ΔHOCc
HSIC
ΔHOCc
4th
.074**
-.83~ 7.24
.029***
-.52~13.40
.012*
5.89~38.08
.001
3.69~38.45
5th
.027***
-.50~9.21
.049***
.90~18.05
.025***
8.55~51.97
.002
7.32~55.53
6th
.058**
-.18~1.36
.084***
2.43~17.77
.057**
7.54~39.35
.004
3.14~26.73
7th
.149***
-3.06~ 1.88
.170***
-2.28~1.88
.121***
-.88~3.10
.011
-2.04~ 1.98
8th
.174***
-2.38~1.18
.137***
-2.24~.87
.112***
-.11~5.24
.006
-.76~5.15
9th
.074**
-1.31~5.75
.049*
-1.71~1.65
.045***
3.00~13.83
.002
-1.28~3.89
10th
.032***
-1.63~ 3.89
.021*
-2.29~1.25
.015*
2.78~12.25
.001
-1.05~ 2.48
Note. Both focal variables in each model, are for 2018-2019 school year while adjusting the prior year scores; Hilbert Schmidt Independence Criterion
(HSIC; based on 200 bootstrap resamples) and the cube-based Higher Order Correlation Difference test (ΔHOC; based on 1000 bootstrap resamples and
95% bootstrap confidence interval) are used in CDDA (Li & Wiedermann, 2020b)
*** p < .001, ** p < .01, * p < .05
33
less likely to be a mixture sample with some grades supporting the TSR→PCT model and the
others supporting the PCT→TSR model. In addition, the interaction term in neither the
hypothesized model nor the alternative model is significant. As neither scenario of moderation
was the case, we can exclude the possibility of grade-level being a moderator. Therefore, as
ΔHOCc seemed not powerful enough to identify the direction of effect for 5th and 6th grades, i.e.,
the test returned positive ranges for both competing models, it makes more sense to use standard
Direction Dependence Analysis instead. The results support our hypothesized model,
.
Instructional Monitoring (IM)
The hypothesized TSR IM model was supported by HSIC across grades except for the
7th, and 8th grades, where both
and
were smaller than .05. Therefore, an
unconsidered confounder may exist between the two variables in middle schools. does
not differentiate the competing models for the IM sample, as shown in Table 4 and Table 5,
because results would be biased when a confounder exists (Li & Wiedermann, 2020b), as
indicated by HSIC results here.
Affective Engagement in the Content (AE).
Finally, the hypothesized TSR AE model was not supported as
and
for 7th to 10th grades but the p-value of both models was greater than .05 for
lower grades. This means that for teachers at or above 7th grade, the alternative model was
supported instead. That is, teachers’ use of high-quality practices to affectively engage students
may result in better TSRs in secondary schools but not in upper primary schools. The inclusive
results were probably due to the smaller sample size. It is less likely that the hypothesized
34
TSR AE model would be supported for teachers in lower grade levels (41.5% of cases)
because HSIC is capable of identifying mixture samples consisting of two subsamples of similar
sizes but reversed causal direction such that some cases in a sample support the X→Y model
while the others support the Y→X model (Li & Wiedermann, 2020b). As the AE sample is
smaller than suggested by Wiedermann and Li (2020) to detect a potential confounder, we
cannot rule out the possibility that there was an unconsidered variable associated with both TSRs
and AE for lower grade teachers. In addition, the smaller sample size contributes to unreliable
test results for CDDA, because it does not have consistent performance when the sample
size is smaller than 400 (Li & Wiedermann, 2020b). Similar to the PCT sample, as grade levels
may not serve as a moderator, we also used standard DDA as a supplementary analysis for the
AE sample. The alternative model was also favored by both tests,
. Therefore, we can conclude with caution that the alternative
hypothesis is more likely to be “true” for Affective Engagement across secondary grade levels but
not necessarily for lower grade levels.
Since the prior scores were partitioned out before we build models for direction
dependence analysis and cluster-robust regression analysis, we examined the prior-score only
models that are supported by direction dependence analysis. The results are shown in Table A3.
All prior scores are significant predictors
35
Table A3
Results of Prior-Score Models Supported by CDDA
Outcomes
Predictors
B(SE)
Cognitive Engagement
(R2 = .433,
n = 566)
Problem-solving and
Critical Thinking
(R2 = .399,
n = 672)
Instructional Monitoring
(R2 = .461,
n = 733)
Teacher Student
Relationships
(R2 = .610,
n = 217)
Intercept
.014(.029)
.065(.027)*
.049(.024)*
.019 (.039)
Prior Scores
.664(.032)***
.606(.029)***
.633 (.025)***
.702 (.038)***
Note. Outcome variables in the models are the scores for the 2018-2019 school year. Prior Scores in the models are those for the 2017-2018 school year.
*** p < .001, ** p < .01, * p < .05
36
References
Admiraal, W., Veldman, I., Mainhard, T., & van Tartwijk, J. (2019). A typology of veteran
teachers' job satisfaction: Their relationships with their students and the nature of their
work. Social Psychology of Education, 22, 337-355. https://doi.org/10.1007/s11218-018-
09477-z
Allen, J. P., Pianta, R. C., Gregory, A., Mikami, A. Y., & Lun, J. (2011). An interaction-based
approach to enhancing secondary school instruction and student achievement. Science,
333(6045), 1034-1037.
Archambault, I., Vandenbossche-Makombo, J., & Fraser, S. L. (2017). Students’ oppositional
behaviors and engagement in school: The differential role of the student-teacher
relationship. Journal of Child and Family Studies, 26, 1702-1712.
Baker, J. A. (2006). Contributions of teacher-child relationships to positive school adjustment
during elementary school. Journal of School Psychology, 44(3), 211-229.
https://doi.org/10.1016/j.jsp.2006.02.002
Berger, S., Graham, N., & Zeileis, A. (2017). Various versatile variances: An object-oriented
implementation of clustered covariances in R. Retrieved from https://cran.r-
project.org/web/packages/sandwich/vignettes/sandwich-CL.pdf
Bergin, C., & Bergin, D. (2009). Attachment in the classroom. Educational Psychology Review,
21, 141-170. https://doi.org/10.1007/s10648-009-9104-0
Bergin, D. A. (1999). Influences on classroom interest. Educational Psychologist, 34(2), 87-98.
doi:10.1207/s15326985ep3402_2
37
Brainerd, C. J., & Holliday, R. E. (2004). Behavioral measurement of remembering
phenomenologies: So simple a child can do it. Child Development, 75(2), 505-522.
https://doi.org/10.1111/j.1467-8624.2004.00689.x
Burić, I., & Kim, L. E. (2020). Teacher self-efficacy, instructional quality, and student
motivational beliefs: An analysis using multilevel structural equation modeling. Learning
and Instruction, 66, 101302.
https://doi.org/https://doi.org/10.1016/j.learninstruc.2019.101302
Cameron, A. C., Gelbach, J. B., & Miller, D. L. (2011). Robust inference with multiway
clustering. Journal of Business & Economic Statistics, 29(2), 238-249.
https://doi.org/10.1198/jbes.2010.07136
Chew, Q. H., Chia, F. L.-A., Ng, W. K., Lee, W. C. I., Tan, P. L. L., Wong, C. S., Puah, S. H.,
Shelat, V. G., Seah, E.-J. D., Huey, C. W. T., Phua, E. J., & Sim, K. (2020). Perceived
stress, stigma, traumatic stress levels and coping responses amongst residents in training
across multiple specialties during covid-19 pandemic-a longitudinal study. International
Journal of Environmental Research and Public Health, 17(18), 6572.
https://doi.org/10.3390/ijerph17186572
Chien, Y.-T., Chang, Y.-H., & Chang, C.-Y. (2016). Do we click in the right way? A meta-
analytic review of clicker-integrated instruction. Educational Research Review, 17, 1-18.
https://doi.org/10.1016/j.edurev.2015.10.003
Cohen, J. (2013). Statistical power analysis for the behavioral sciences (2nd ed.). Academic
Press.
38
Cohen, J., Ruzek, E., & Sandilos, L. (2018). Does teaching quality cross subjects? Exploring
consistency in elementary teacher practice across subjects. AERA Open, 4(3), 1-16.
https://doi.org/10.1177%2F2332858418794492
Cooper, W. H. (1981). Ubiquitous halo. Psychological Bulletin, 90(2), 218.
Cornelius-White, J. (2007). Learner-centered teacher-student relationships are effective: A meta-
analysis. Review of Educational Research, 77(1), 113-143.
https://doi.org/10.3102%2F003465430298563
Cornell, D., Klein, J., Konold, T., & Huang, F. (2012). Effects of validity screening items on
adolescent survey data. Psychological Assessment, 24(1), 21.
Council of Chief State School Officers. (2011). InTASC model core teaching standards: A
resource for state dialogue. Washington, DC: Council of Chief State School Officers.
Coyle, T. R., Pillow, D. R., Snyder, A. C., & Kochunov, P. (2011). Processing speed mediates
the development of general intelligence (g) in adolescence. Psychological Science,
22(10), 1265-1269. https://doi.org/10.1177%2F0956797611418243
Curby, T. W., Rimm-Kaufman, S. E., & Ponitz, C. C. (2009). Teacher-child interactions and
children’s achievement trajectories across kindergarten and first grade. Journal of
Educational Psychology, 101(4), 912-925. https://psycnet.apa.org/doi/10.1037/a0016647
Daws, J. E. (2005). Teachers and students as co-learners: Possibilities and problems. Journal of
Educational Enquiry, 6(1), 110-125.
Deci, E. L., & Ryan, R. M. (2010). Self-determination. John Wiley & Sons.
den Brok, P., Brekelmans, M., & Wubbles, T. (2004). Interpersonal teacher behaviour and
student outcomes. School Effectiveness and School Improvement, 15(3-4), 407-442.
https://doi.org/10.1080/09243450512331383262
39
Dunn, E. C., Masyn, K. E., Jones, S. M., Subramanian, S. V., & Koenen, K. C. (2015).
Measuring psychosocial environments using individual responses: An application of
multilevel factor analysis to examining students in schools. Prevention Science, 16(5),
718-733. https://doi.org/10.1007/s11121-014-0523-x
Feeley, T. H. (2002). Comment on halo effects in rating and evaluation research. Human
Communication Research, 28(4), 578-586. https://psycnet.apa.org/doi/10.1111/j.1468-
2958.2002.tb00825.x
Feldman, J. M. (1986). A note on the statistical correction of halo error. Journal of Applied
Psychology, 71(1), 173.
Fisicaro, S. A., & Lance, C. E. (1990). Implications of three causal models for the measurement
of halo error. Applied Psychological Measurement, 14(4), 419-429.
Fredricks, J. A., Blumenfeld, P. C., & Paris, A. H. (2004). School engagement: Potential of the
concept, state of the evidence. Review of Educational Research, 74(1), 59-109.
https://doi.org/10.3102%2F00346543074001059
Gehlbach, H., Brinkworth, M. E., King, A. M., Hsu, L. M., McIntyre, J., & Rogers, T. (2016).
Creating birds of similar feathers: Leveraging similarity to improve teacher-student
relationships and academic achievement. Journal of Educational Psychology, 108(3),
342-352. https://psycnet.apa.org/doi/10.1037/edu0000042
Gerstner, C. R., & Day, D. V. (1997). Meta-analytic review of leader-member exchange theory:
Correlates and construct issues. Journal of Applied Psychology, 82(6), 827-844.
https://psycnet.apa.org/doi/10.1037/0021-9010.82.6.827
40
Gillen-O’Neel, C., & Fuligni, A. (2012). A longitudinal study of school belonging and academic
motivation across high school. Child Development, 84(2), 678-692.
https://doi.org/10.1111/j.1467-8624.2012.01862.x
Goe, L., & Holdheide, L. (2011). Measuring teachers' contributions to student learning growth
for nontested grades and subjects. Research & Policy Brief. National Comprehensive
Center for Teacher Quality.
Goldhaber, D. D., & Brewer, D. J. (2000). Does teacher certification matter? High school teacher
certification status and student achievement. Educational Evaluation and Policy Analysis,
22(2), 129-145. http://www.jstor.com/stable/1164392
Grajeda, S., Bergin, C., Jones, E., & Tsai, C. (2017). Reliability and factor structure of a student
survey of teacher effectiveness. Paper presented at the American Educational Research
Association, San Antonio.
Graves, L. M., & Luciano, M. M. (2013). Self-determination at work: Understanding the role of
leader-member exchange. Motivation and Emotion, 37, 518-536.
https://doi.org/10.1007/s11031-012-9336-z
Gretton, A., Bousquet, O., Smola, A., & Schölkopf, B. (2005, October). Measuring statistical
dependence with Hilbert-Schmidt norms. In International conference on algorithmic
learning theory (pp. 63-77). Springer.
Guo, G., & Zhao, H. (2000). Multilevel modeling for binary data. Annual Review of Sociology,
26(1), 441-462. https://doi.org/10.1146/annurev.soc.26.1.441
Hamre, B. K., Pianta, R. C., Downer, J. T., & Mashburn, A. J. (2008). Teachers' perceptions of
conflict with young students: Looking beyond problem behaviors. Social Development,
17(1), 115-136. doi:10.1111/j.1467-9507.2007.00418.x
41
Hattie, J. (2009). The black box of tertiary assessment: An impending revolution. In L. H.
Meyer, S. Davidson, H. Anderson, R. Fletcher, P.M. Johnston, & M. Rees (Eds.).
Tertiary assessment & higher education student outcomes: Policy, practice & research
(pp. 259-275). Ako Aotearoa.
Hidi, S., & Renninger, K. A. (2006). The four-phase model of interest development. Educational
Psychologist, 41(2), 111-127. doi:10.1207/s15326985ep4102_4
Holzberger, D., Philipp, A., & Kunter, M. (2013). How teachers’ self-efficacy is related to
instrucitonal quality: A longitudinal analysis. Journal of Educational Psychology, 105(3),
774-786. http://dx.doi.org/10.1037/a0032198
Hox, J. (1998). Multilevel modeling: When and why. Paper presented at the Classification, Data
Analysis, and Data Highways, Berlin, Heidelberg.
Huang, F. L. (2016). Alternatives to multilevel modeling for the analysis of clustered data. The
Journal of Experimental Education, 84(1), 175-196.
https://doi.org/10.1080/00220973.2014.952397
Huang, F. L., & Li, X. (in press). Using cluster robust standard errors when analyzing cluster
randomized trials with few clusters. Behavior Research Methods.
Hughes, J. N. (2011). Longitudinal effects of teacher and student perceptions of teacher-student
relationship qualities on academic adjustment. The Elementary School Journal, 112(1),
38-60. https://doi.org/10.1086/660686
Hughes, J. N., Luo, W., Kwok, O.-M., & Loyd, L. K. (2008). Teacher-student support, effortful
engagement, and achievement: A 3-year longitudinal study. Journal of Educational
Psychology, 100(1), 1-14. https://psycnet.apa.org/doi/10.1037/0022-0663.100.1.1
42
Hughes, J. N., Wu, J.-Y., Kwok, O.-m., Villarreal, V., & Johnson, A. Y. (2012). Indirect effects
of child reports of teacher-student relationship on achievement. Journal of Educational
Psychology, 104(2), 350-365. https://psycnet.apa.org/doi/10.1037/a0026339
Jang, H., Reeve, J., & Deci, E. L. (2010). Engaging students in learning activities: It is not
autonomy support or structure but autonomy support and structure. Journal of
Educational Psychology, 102(3), 588-600. https://psycnet.apa.org/doi/10.1037/a0019682
Jeffrey, A. J., Auger, R. W., & Pepperell, J. L. (2013). “If we're ever in trouble they're always
there”: A qualitative study of teacher-student caring. The Elementary School Journal,
114(1), 100-117. https://doi.org/10.1086/671062
Jennings, P. A., & Greenberg, M. T. (2009). The prosocial classroom: Teacher social and
emotional competence in relation to student and classroom outcomes. Review of
Educational Research, 79(1), 491-525. https://doi.org/10.3102%2F0034654308325693
Jia, Y., Way, N., Ling, G., Yoshikawa, H., Chen, X., Hughes, D., Ke, X., & Lu, Z. (2009). The
influence of student perceptions of school climate on socioemotional and academic
adjustment: A comparison of Chinese and American adolescents. Child Development,
80(5), 1514-1530. https://doi.org/10.1111/j.1467-8624.2009.01348.x
Johnson, S. M., Kraft, M. A., & Papay, J. P. (2012). How context matters in high-needs schools:
The effects of teachers’ working conditions on their professional satisfaction and their
students’ achievement. Teachers College Record, 114(10), 1-39.
Kail, R. V., & Ferrer, E. (2007). Processing speed in childhood and adolescence: Longitudinal
models for examining devlopmental change. Child Development, 78(6), 1760-1770.
https://doi.org/10.1111/j.1467-8624.2007.01088.x
43
Kauermann, G., & Carroll, R. J. (2001). A note on the efficiency of sandwich covariance matrix
estimation. Journal of the American Statistical Association, 96(456), 1387-1396.
https://doi.org/10.1198/016214501753382309
Kincade, L., Cook, C., & Goerdt, A. (2020). Meta-analysis and common practice elements of
universal approaches to improving student-teacher relationships. Review of Educational
Research, 90(5), 710-748. https://doi.org/10.3102%2F0034654320946836
Klassen, R. M., & Chiu, M. M. (2010). Effects on teachers’ self-efficacy and job satisfaction:
Teacher gender, years of experience, and job stress. Journal of Educational Psychology,
102(3), 741-756. https://psycnet.apa.org/doi/10.1037/a0019237
Klusmann, U., Kunter, M., Trautwein, U., Ludtke, O., & Baumert, J. (2008). Teachers’
occupational well-being and quality of instruction: The important role of self-regulatory
patterns. Journal of Educational Psychology, 100(3), 702-715.
https://psycnet.apa.org/doi/10.1037/0022-0663.100.3.702
Kusumawardhani, P. N. (2017). Does teacher certification program lead to better quality
teachers? Evidence from Indonesia. Education Economics, 25(6), 590-618.
https://doi.org/10.1080/09645292.2017.1329405
Kyriacou, C. (2001). Teacher stress: Directions for future research. Educational Review, 53, 27-
35. https://doi.org/10.1080/00131910120033628
Ladd, H. F., & Sorensen, L. C. (2017). Returns to teacher experience: Student achievement and
motivation in middle school. Education Finance and Policy, 12(2), 241-279.
Li, X., Martens, M. P., & Wiedermann, W. (2022). Conditional Direction of Dependence
Modeling: Application and Implementation in SPSS. Social Science Computer Review.
https://doi.org/10.1177/08944393211073168
44
Li, X., & Wiedermann, W. (2020a). Conditional Direction Dependence Analysis in Linear
Models with SPSS Macros and Custom Dialogue Implementation (Version 1.1)
[Computer software]. https://www.ddaproject.com
Li, X., & Wiedermann, W. (2020b). Conditional direction dependence analysis: Evaluating the
causal direction of effects in linear models with interaction terms. Multivariate
Behavioral Research, 1-25. https://doi.org/10.1080/00273171.2019.1687276
Li, X., Prewett, S., & Olsen, A. A. (2020). Predictor-residual independence tests for examining
causal assumptions of synchronous effects models. Paper presented at the American
Psychological Association Annual Convention, Virtual.
Luciana, M., Conklin, H. M., Hooper, C. J., & Yarger, R. S. (2005). The development of
nonverbal working memory and executive control processes in adolescents. Child
Development, 76(3), 697-712.
Manning, A. (2016). Unlocking the power of employee happiness: What top employees seek
from the workplace today. Strategic HR Review, 15(4), 191-192.
https://doi.org/10.1108/SHR-05-2016-0038
Marks, H. M. (2000). Student engagement in instructional activity: Patterns in elementary,
middle, and high school years. American Educational Research Journal, 37(1), 153-184.
https://doi.org/10.3102%2F00028312037001153
Martin, R., Guillaume, Y., Thomas, G., Lee, A., & Epitropaki, O. (2016). Leader-member
exchange (LMX) and performance: A meta-analytic review. Personnel Psychology,
69(1), 67-121. https://doi.org/10.1111/peps.12100
Mashburn, A. J., Hamre, B. K., Downer, J. T., & Pianta, R. C. (2006). Teacher and classroom
characteristics associated with teachers’ ratings of prekindergartners’ relationships and
45
behaviors. Journal of Psychoeducational Assessment, 24(4), 367-380.
https://psycnet.apa.org/doi/10.1177/0734282906290594
McCormick, M. P., O’Connor, E. E., Cappella, E., & McClowry, S. G. (2013). Teacher-child
relationships and academic achievement: A multilevel propensity score model approach.
Journal of School Psychology, 51(5), 611-624. https://doi.org/10.1016/j.jsp.2013.05.001
McCormick, N. J., Clark, L. M., & Raines, J. M. (2015). Engaging students in critical thinking
and problem solving: A brief review of the literature. Journal of Studies is Education,
5(4), 100-113.
Milatz, A., Lüftenegger, M., & Schober, B. (2015). Teachers' relationship closeness with
students as a resource for teacher wellbeing: A response surface analytical approach.
Frontiers in Psychology, 6, 1949. https://doi.org/10.3389/fpsyg.2015.01949
Musci, R. J., & Stuart, E. (2019). Ensuring causal, not casual, inference. Prevention Science,
20(3), 452-456. https://doi.org/10.1007/s11121-018-0971-9
Nye, B., Konstantopoulos, S., & Hedges, L. V. (2004). How large are teacher effects?
Educational Evaluation and Policy Analysis, 26(3), 237-257.
O’Connor, E., & McCartney, K. (2007). Examining teacher-child relationships and achievement
as part of an ecological model of development. American Educational Research Journal,
44(2), 340-369. https://doi.org/10.3102%2F0002831207302172
Olsen, A. A., & Huang, F. L. (2019). Teacher job satisfaction by principal support and teacher
cooperation: Results from the Schools and Staffing Survey. Education Policy Analysis
Archives, 27(11), 1-27. https://doi.org/10.14507/epaa.27.4174
46
Pérez-Mengual, N., Aragonés-Barbera, I., Moret-Tatay, C., & Moliner-Albero, A. R. (2021). The
relationship of fear of death between neuroticism and anxiety during the Covid-19
pandemic. Frontiers in Psychiatry, 12(501). https://doi.org/10.3389/fpsyt.2021.648498
Perry, L., & McConney, A. (2010). School socio-economic composition and student outcomes in
Australia: Implications for educational policy. Australian Journal of Education, 54(1),
72-85. https://doi.org/10.1177%2F000494411005400106
Pianta, R. C. (2001). Student-Teacher Relationship Scale: Professional manual. Odessa,
FL: Psychological Assessment Resources.
Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. P. (2003). Common method
biases in behavioral research: a critical review of the literature and recommended
remedies. Journal of Applied Psychology, 88(5), 879-903. https://doi.org/10.1037/0021-
9010.88.5.879
Prewett, S. L., Bergin, D. A., & Huang, F. L. (2018). Student and teacher perceptions on student-
teacher relationship quality: A middle school perspective. School Psychology
International, 40(1), 66-87. https://doi.org/10.1177%2F0143034318807743
Prewett, S., Bergin, D. A., & Huang, F. L. (2019). Student and teacher perceptions on student-
teacher relationship quality: A middle school perspective. School Psychology
International, 40(1), 66-87. https://doi.org/10.1177%2F0143034318807743
Quin, D. (2016). Longitudinal and contextual associations between teacher-student relationships
and student engagement: A systematic review. Review of Educational Research, 87(2),
345-387. https://doi.org/10.3102%2F0034654316669434
47
Reddy, L. A., & Dudek, C. M. (2014). Teacher progress monitoring of instructional and
behavioral management practices: An evidence-based approach to improving classroom
practices. International Journal of School & Educational Psychology, 2, 71-84.
Reddy, L. A., Dudek, C. M., & Lekwa, A. (2017). Classroom strategies coaching model:
Integration of formative assessment and instructional coaching. Theory into Practice,
56(1), 46-55. https://doi.org/10.1080/00405841.2016.1241944
Richter, F. G. C., & Brorsen, B. W. (2006). Aggregate versus disaggregate data in measuring
school quality. Journal of Productivity Analysis, 25(3), 279-289.
https://doi.org/10.1007/s11123-006-7644-6
Roorda, D. L., Jak, S., Zee, M., Oort, F. J., & Koomen, H. M. Y. (2017). Affective teacher-
student relationships and students’ engagement and achievement: A meta-analytic update
and test of the mediating role of engagement. School Psychology Review, 46(3), 239-261.
https://doi.org/10.17105/SPR-2017-0035.V46-3
Roorda, D. L., Koomen, H. M. Y., Spilt, J. L., & Oort, F. J. (2011). The influence of affective
teacher-student relationships on students’ school engagement and achievement: A meta-
analytic approach. Review of Educational Research, 81(4), 493-529.
https://doi.org/10.3102%2F0034654311421793
Ryan, R. M., & Deci, E. L. (2017). Self-determination theory: Basic psychological needs in
motivation, development, and wellness. The Guilford Press.
Scheerens, J. (2001). School effectiveness research. International Encyclopedia of the Social &
Behavioral Sciences, 13567-13572. https://doi.org/10.1016/B0-08-043076-7/02438-4
Sebastian, J., Allensworth, E., Wiedermann, W., Hochbein, C., & Cunningham, M. (2018).
Principal leadership and school performance: An examination of instructional leadership
48
and organizational management. Leadership and Policy in Schools, 1-23.
https://doi.org/10.1080/15700763.2018.1513151
Shieh, G. (2008). Detecting interaction effects in moderated multiple regression with continuous
variables power and sample size considerations. Organizational Research Methods,
12(3), 510-528. https://doi.org/10.1177%2F1094428108320370
Simons, K. D., & Klein, J. D. (2007). The impact of scaffolding and student achievement levels
in a problem-based learning environment. Instructional Science, 35, 41-72.
Solomonson, A. L., & Lance, C. E. (1997). Examination of the relationship between true halo
and halo error in performance ratings. Journal of Applied Psychology, 82(5), 665.
Spilt, J. L., Koomen, H. M. Y., & Thijs, J. T. (2011). Teacher wellbeing: The importance of
teacher-student relationships. Educational Psychology Review, 23, 457-477.
Steinberg, L. (2009). Adolescent development and juvenile justice. Annual Review of Clinical
Psychology, 5, 459-485. https://doi.org/10.1146/annurev.clinpsy.032408.153603
Summers, J. J., Davis, H. A., & Hoy, A. W. (2017). The effects of teachers’ efficacy beliefs on
students’ perceptions of teacher relationship quality. Learning and Individual
Differences, 53, 17-25.
Thoonen, E. E. J., Sleegers, P. J. C., Oort, F. J., Peetsma, T. T. D., & Geijsel, F. P. (2011). How
to improve teaching practices: The role of teacher motivation, organizational factors, and
leadership practices. Educational Administration Quarterly, 47(3), 496-536.
https://doi.org/10.1177%2F0013161X11400185
Thorndike, E. L. (1920). A constant error in psychological ratings. Journal of Applied
Psychology, 4(1), 25-29.
49
Urdan, T., & Schoenfelder, E. (2006). Classroom effects on student motivation: Goal structures,
social relationships, and competence beliefs. Journal of School Psychology, 44(5), 331-
349. https://doi.org/10.1016/j.jsp.2006.04.003
van de Grift, W., Helms-Lorenz, M., & Maulana, R. (2014). Teaching skills of student teachers:
Calibration of an evaluation instrument and its value in predicting student academic
engagement. Studies in Educational Evaluation, 43, 150-159.
https://doi.org/10.1016/j.stueduc.2014.09.003
van de Pol, J., Volman, M., & Beishuizen, J. (2010). Scaffolding in teacher-student interaction:
A decade of research. Educational Psychology Review, 22(3), 271-296.
van der Lans, R. M., Cremers, J., Klugkist, I., & Zwart, R. (2020) Teachers' interpersonal
relationships and instructional expertise: How are they related? Studies in Educational
Evaluation, 66, 100902. https://doi.org/10.1016/j.stueduc.2020.100902
van der Lans, R. M., van de Grift, W. J. C. M., & van Veen, K. (2017). Developing an
instrument for teacher feedback: Using the Rasch Model to explore teachers’
development of effective teaching strategies and behaviors. The Journal of Experimental
Education, 86(2), 247-264. https://doi.org/10.1080/00220973.2016.1268086
VanderWeele, T. J. (2019). Principles of confounder selection. European journal of
epidemiology, 34(3), 211-219. https://doi.org/10.1007/s10654-019-00494-6
Veldman, I., Admiraal, W., van Tartwijk, J., Mainhard, T., & Wubbels, T. (2016). Veteran
teachers' job satisfaction as a function of personal demands and resources in the
relationships with their students. Teachers and Teaching, 22(8), 913-926.
https://doi.org/10.1080/13540602.2016.1200546
50
Wang, M.T., & Degol, J. (2014). Staying engaged: Knowledge and research needs in student
engagement. Child Development Perspectives, 8(3), 137-143.
https://doi.org/10.1111/cdep.12073
Wang, Z., Bergin, C., & Bergin, D. A. (2014). Measuring engagement in fourth to twelfth grade
classrooms: The Classroom Engagement Inventory. School Psychology Quarterly, 29(4),
517-535. https://doi.org/10.1037/spq0000050
Wendt, G. W., & Bartoli, A. J. (2019). Understanding the psychopathy-stress association in
typical developing adults: The role of emotional deficits. Personality and Individual
Differences, 149, 296-301. https://doi.org/10.1016/j.paid.2019.06.016
Wentzel, K. R. (1997). Student motivation in middle school: The role of perceived pedagogical
caring. Journal of Educational Psychology, 89(3), 411.
https://psycnet.apa.org/doi/10.1037/0022-0663.89.3.411
Wentzel, K. R. (2009). Students' relationships with teachers as motivational contexts. In K. R.
Wenzel & A. Wigfield (Eds.), Educational psychology handbook series. Handbook of
motivation at school (p. 301–322). Routledge/Taylor & Francis Group
Wiedermann, W., & Li, X. (2020). Confounder detection in linear mediation models:
Performance of kernel-based tests of independence. Behavior Research Methods, 52(1),
342-359. https://doi.org/10.3758/s13428-019-01230-4
Wiedermann, W., & Li, X. (2018). Direction dependence analysis: A framework to test the
direction of effects in linear models with an implementation in SPSS. Behavior Research
Methods, 50(4), 1581-1601. https://doi.org/10.3758/s13428-018-1031-x
51
Wiedermann, W., Li, X., & von Eye, A. (2020). Direction dependence analysis: Statistical
foundations and applications. In Direction dependence in statistical modeling (pp. 9-46).
https://doi.org/https://doi.org/10.1002/9781119523024.ch2
Wiedermann, W., Reinke, W. M., & Herman, K. C. (2020). Prosocial skills causally mediate the
relation between effective classroom management and academic competence: An
application of direction dependence analysis. Developmental Psychology, 56(9), 1723–
1735.
Wiedermann, W., & Sebastian, J. (2019). Direction dependence analysis in the presence of
confounders: Applications to linear mediation models using observational data.
Multivariate Behavioral Research, 1-21. https://doi.org/10.1080/00273171.2018.1528542
Wiedermann, W., & von Eye, A. (2015). Direction-dependence analysis: A confirmatory
approach for testing directional theories. International Journal of Behavioral
Development, 39(6), 570-580. doi:10.1177/0165025415582056
Willingham, D. T. (2008). Critical thinking: Why is it so hard to teach? Arts Education Policy
Review, 109(4), 21-32. https://doi.org/10.3200/AEPR.109.4.21-32
Wirkala, C., & Kuhn, D. (2011). Problem-based learning in K-12 education: Is it effective and
how does it achieve its effects? American Educational Research Journal, 48(5), 1157-
1186. https://doi.org/10.3102%2F0002831211419491
Wolfe, E. W., & Song, T. (2015). Methods for monitoring and document rating quality. In H.
Jiao & R. W. Lissitz (Eds.), The next generation of testing: Common core standards,
smarter-balanced, PARCC, and the nationwide testing movement. Information Age
Publishing.
52
Wubbels, T., Brekelmans, M., den Brok, P., & van Tartwijk, J. (2006). An interpersonal
perspective on classroom management in secondary classrooms in the Netherlands. In C.
Evertson & C. Weinstein (Eds.), Handbook of classroom management: Research,
practice, and contemporary issues (pp. 1161-1191). Mahwah, NJ: Lawrence Erlbaum
Associates.
Zhang, Z., & Hyland, K. (2018). Student engagement with teacher and automated feedback on
L2 writing. Assessing Writing, 36, 90-102.
