Students’ Expectations and Students’ Satisfaction: The Mediating Role of Excellent Teacher Behaviors

Abstract

This study sought to investigate whether teacher behaviors, within the Teacher Behavior Checklist dimensions of caring and supportive and professional competency and communication skills, relate to students’ satisfaction. Additionally, it assessed the influence of the importance students set on teacher behaviors on students’ satisfaction. Cross-sectional survey data were analyzed using hierarchical linear regression and mediation analysis. Results suggested teacher behaviors within the caring and supportive dimension is the dominant predictor of students’ satisfaction. Behaviors according to professional competency and communication skills were also related to students’ satisfaction but to a lesser degree. Results showed that students’ importance for behaviors did mediate the relationship between teacher behaviors and students’ satisfaction with both caring and supportive and professional competency and communication skills. Theoretical and practical implications for excellent teacher behaviors are discussed including recommendations for future research.

Keywords

Teacher Behavior Checklist excellent teaching student satisfaction

Teacher behaviors are known to influence student learning including cognitive and affective learning, participation, involvement, motivation, academic performance, and satisfaction (Frisby & Marin, 2010; Granitz et al., 2009; M. Hill & Epps, 2010). The behavior exhibited by a teacher can influence the student’s learning environment which in turn influences the student’s satisfaction and learning gains (Micari & Pazos, 2012; Ryan & Wilson, 2014). For example, Frisby and Myers (2008) found that rapport-related behaviors are significantly associated with student satisfaction. Student satisfaction is an expression of the quality of their learning experience (Piccoli et al., 2001). Student satisfaction is widely used as the evidence of successful learning. For example, McFarland and Hamilton (2005) and Summers et al. (2005) used student satisfaction to compare the quality of online classes versus traditional classes taking place in a classroom. Furthermore, student satisfaction is also related to retention rates (Edwards & Waters, 1982; Schertzer & Schertzer, 2004). High dropout rates can be very costly for universities. Therefore, student satisfaction plays a vital role in general but even more in online/distance education which has higher dropout rates and lower levels of student satisfaction (Bolliger & Martindale, 2004; McFarland & Hamilton, 2005; Summers et al., 2005; Willging & Johnson, 2009).

Buskist et al. (2002) determined 28 qualities (i.e., behaviors) of a “master teacher” at the college and university level (for a detailed description of these, see Keeley et al., 2006). These 28 behavioral typologies were combined into the Teacher Behavior Checklist (TBC). Based on Keeley et al. (2006), the TBC may be used as a one-factor construct using all 28 items or as a two-factor construct with the dimensions caring and supportive (“caring”) and professional competency and communication skills (“professional”) utilizing 24 of the 28 items (13 and 11, respectively). They preferred the two-factor model based on earlier research by Buskist et al. (2002) and Schaeffer et al. (2003), as the construct is more inclusive of specific factors (i.e., behaviors) as reflected in the two subscales. Further, Keeley et al. (2006) stated that the two-factor structure is similar to earlier research by Lowman’s (1995) model based on interpersonal rapport and intellectual excitement. This study will utilize the two-factor model.

The TBC framework has been used in two ways in research. First, it is used to rate student and faculty importance of teacher behaviors concerning excellent teaching. Buskist et al. (2002) had students (N = 917) and faculty (N = 118) of a university rate the 28 behaviors based on their importance to determine the top 10 behaviors of a master teacher. Students and faculty agreed on six teacher behaviors: (1) realistic student expectations, (2) knowledgeable about the subject, (3) approachable and personable, (4) respectful, (5) creative and interesting, and (6) enthusiastic about teaching. Schaeffer et al. (2003) replicated the study by Buskist et al. (2002) in a community college. Other studies with the TBC found consensus on the importance of teacher behaviors including a Canadian sample (Vulcano, 2007), across the three academic disciplines of psychology, education, and chemical engineering (Liu et al., 2016), and award-winning teachers compared to other faculty (Keeley et al., 2016). Second, the TBC has been used to rate the actual behaviors of teachers. The TBC has been shown to be a reliable tool to rate teacher behaviors by having different students rate the same teacher utilizing video clips (Landrum & Stowell, 2013), having students rate different teacher experiences (i.e., rate the behaviors of their best, worst, and most recent teacher; Keeley et al., 2010), and cultural comparison of excellent teacher behaviors, as perceived and rated by students, of a Chinese sample to an existing sample from Japan and the United States of America (Liu et al., 2015). The studies showed that the TBC is capable of detecting differences among teachers (Keeley et al., 2010; Liu et al., 2015) and consistency when rating the same teacher (Landrum & Stowell, 2013).

Research with the TBC has focused on the importance of teacher behaviors (e.g., Buskist et al., 2002; Keeley et al., 2016; Liu et al., 2016; Schaeffer et al., 2003; Vulcano, 2007) or the rating of perceived teacher behaviors (e.g., Keeley et al., 2010; Landrum & Stowell, 2013; Liu et al., 2015). Other studies investigated student attributes affecting TBC ratings such as academic self-concept and academic motivation to predict TBC behavior importance ratings (Komarraju, 2013) and the influence of gender, course level, and required student effort on TBC behavior ratings of teachers (Stigall & Blincoe, 2015). However, there has been little research to predict class outcomes. Lizzio et al. (2002) define class outcomes as academic achievement (i.e., grade point average [GPA]), development of skills, and students’ satisfaction. Specifically, there has been no study, to my knowledge, to predict students’ satisfaction based on teacher behaviors, as rated by the students using the TBC. Furthermore, for a relationship of perceived teacher behaviors and student satisfaction, it is unknown what role the students’ importance of teacher behaviors plays, as rated with the TBC. Therefore, the purpose of this study is twofold. First, it will explore the relationship of the student’s perception of teacher behavior, based on the two main dimensions of the TBC (i.e., caring and professional), and student’s satisfaction. Second, it will aim to understand the influence of students’ importance of teacher behaviors, within the two main dimensions of the TBC, on the relationship of perceived teacher behavior and students’ satisfaction.

Theoretical Framework and Hypotheses

The TBC: Behavior and satisfaction

Student satisfaction can be defined as the perceived value or quality of the education received while attending an educational institution (Astin, 1993). Student satisfaction is based on the student’s perception of the educational experience (Piccoli et al., 2001). Various items are known to influence student satisfaction including teaching ability of the staff, advising, IT facilities, and learning management systems (DeShields et al., 2005; Douglas et al., 2006).

According to Bolliger and Martindale (2004), teacher behavior may be a significant contributor to students’ satisfaction. The following two paragraphs will categorize previous research regarding teacher behaviors and students’ satisfaction within the TBC dimensions of caring and professional behaviors.

The following caring behaviors were found to contribute to students’ satisfaction: accessibility (DeShields et al., 2005; Moore, 2005), flexibility (Moore, 2005), interaction and relationship with the students (Y. Hill et al., 2003; Moore, 2005), understanding (DeShields et al., 2005), helpful (DeShields et al., 2005), rapport (Frisby & Myers, 2008), responsiveness (Douglas et al., 2008), and feedback (DeShields et al., 2005; Y. Hill et al., 2003; Moore, 2005).

Behaviors that can be categorized as professional and contribute to students’ satisfaction are as follows: communication (Douglas et al., 2008; Moore, 2005), reliable (DeShields et al., 2005), subject expertise (Douglas et al., 2006), usefulness of the material (Douglas et al., 2008), and real-world relevance (DeShields et al., 2005).

The stated research represents support for the hypothesis that there is a relationship between both TBC caring and TBC professional scales and students’ satisfaction. However, which of the dimensions is the primary contributor? Reviewing previous research in more detail, it appears to be that more caring behaviors such as feedback contribute to students’ satisfaction than behaviors within the professional dimension. Furthermore, DeShields et al. (2005) showed the following associations with students’ satisfaction: categorized within the caring dimension, understanding (r = .16, p < .05), accessibility (r = .19, p < .05), and feedback (r = .20, p < .05); and for the professional dimension, reliable (r = .14, p < .05) and real-world relevance (r = .16, p < .05). The associations by DeShields et al. (2005) are all weak associations. Adhikary (2017) did a leadership study of teachers investigating how teachers’ leadership behaviors relate to students’ satisfaction. The leadership theory used was transformational leadership. Transformational leadership has four dimensions. The dimension individualized consideration can be classified as caring behavior, and the dimension intellectual stimulation can be classified as professional behavior. Individualized consideration was related to students’ satisfaction (β = .46, p < .001), and intellectual stimulation was not related to students’ satisfaction (β = .00, ns); the sample size was N = 113. Therefore, I hypothesize that the caring dimension of the TBC may be the dominant predictor of students’ satisfaction.
Hypothesis 1a: The behavior ratings of the TBC caring dimension, as perceived and rated by the students, will be positively related to students’ satisfaction.

Hypothesis 1b: The behavior ratings of the TBC professional dimension, as perceived and rated by the students, will be positively related to students’ satisfaction but to a lesser degree than the behavior ratings of the TBC caring dimension.

The TBC: Importance

The level of importance a student sets to the TBC behaviors can be seen as an expectation that the teacher should meet. In business, customer expectations are important when forming a level of satisfaction (Parasuraman et al., 1994). In educational services, the student is the “primary customer” (Douglas et al., 2006), and the students’ expectations have a high influence on the formation of satisfaction (Anderson & Sullivan, 1993; Yi, 1993). For instance, students’ satisfaction has been significantly related to the extent to which the overall course structure aligns with student expectations and preferences (Westerman et al., 2002). Elliot and Healy (2001) state that student’s satisfaction is a result of when actual teacher performance meets or exceeds the students’ expectations. The question is: How do students’ expectations (i.e., importance) and teacher behaviors interact when students’ satisfaction is formed? According to Moore (2005), effective teacher behavior strongly reflecting student expectations leads to high levels of student satisfaction. However, according to Bolliger and Martindale (2004), the teacher may be the primary predictor of student satisfaction. Therefore, the assumption is that there is a direct influence between teacher behavior and students’ satisfaction (see also Hypothesis 1). Furthermore, there may be an indirect relationship between importance and students’ satisfaction. From a student’s point of view, under the presence of importance, teacher behavior is experienced, and satisfaction is formed. The mentioned sequence of events leads to mediation (Stone-Romero & Rosopa, 2004). Mediation is a time-based model of events in which, in this case, importance (X) occurs before behavior (M) which in turn occurs before satisfaction (Y). Therefore, it can be assumed that teacher behavior mediates the relationship between students’ importance and students’ satisfaction. Based on the mentioned findings by Moore (2005) and Bolliger and Matindale (2004), I propose that there is no direct relationship between students’ importance and students’ satisfaction (i.e., mediation).
Hypothesis 2: For both dimension, caring and professional, the teacher behavior ratings will mediate the relationship of the students’ importance ratings, as perceived by the students, and students’ satisfaction.

Method

Sample

This was a cross-sectional study utilizing a convenience sample. The data were collected at a medium-sized, private university, located in the Southeastern Unites States during the spring semester 2018. Fifteen faculty members of the school of business were contacted for permission to administer the surveys in their classes, and eight faculty members gave permission (53.3%). Of the eight classes recruited, five were undergraduate courses (economics, management, global business, operations, and business applications) and three were graduate courses (marketing, finance, and quantitative decision making).

Participants

The survey data were collected in two stages. First, participants completed the TBC importance survey at the beginning of the semester, then the TBC behavior survey at the end of the semester. The eight classes had a total of 312 students. Two-hundred and ninety-one TBC importance surveys were returned (93.3%) and 226 TBC behavior surveys (72.4%) leading to 212 matched records (i.e., paired the TBC importance survey data with the TBC behavior survey data per participant). The data included incomplete records (i.e., omissions). Records with missing survey data were removed from the data set, as recommended by Sekaran and Bougie (2011). A total of 18 records were removed leading to a final sample of N = 194. The overall response rate was 62.2%, which was determined using N = 194 and 312 potential participants. The student standings were 17.5% freshmen, 19.6% sophomore, 27.8% junior, 17.0% senior, and 18.0% graduate. The sample consisted of 51.5% males, and the mean age was 22.26 years (standard deviation [SD] = 6.19). The racial/ethnic backgrounds were as follows: 73.2% White, 9.3% Black or African American, 13.9% Hispanic or Latino, 2.1% Asian/Pacific Islander, and 1.5% Other.

Data Collection Procedure

The data were collected using paper surveys. As recommended by Hayes et al. (2016) for mediation analysis, the measurement was staggered in time. The TBC importance data were collected in Week 1 of the spring semester 2018. The survey included an informed consent, the TBC importance survey, and a demographics survey. The TBC behavior data were collected in Week 15 of the spring semester 2018 to ensure the students had sufficient exposure to the behaviors exhibited by the teacher. The survey included an informed consent, the TBC behavior survey, a student satisfaction survey, and a question about the expected class grade. The order in which the measures were presented remained the same across all participants.

Measures

TBC

The measure was used twice for the assessment of the teacher behavior importance and teacher behavior as perceived by the student. The TBC consists of 28 items that can be broken down into two dimensions, caring and professional (Keeley et al., 2006). The caring dimension consists of 11 items and a sample item is “accessible (Posts office hours, gives out phone number and email information).” The professional dimension consists of 13 items and a sample item is “confident (Speaks clearly, makes eye contact, and answers questions correctly).” Although four items are not used for hypothesis testing, all 28 items were presented in the survey. For the TBC importance assessment, participants were asked to rate how important each behavior is to them about what they define as a “master teacher” using a 4-point Likert-type scale from 1 (irrelevant) to 4 (very important; Schaeffer et al., 2003). For the TBC behavior assessment, participants were asked to rate how often their professor/instructor manifests each behavior/quality using a 5-point Likert-type scale from 1 (never exhibits this quality) to 5 (frequently exhibits this quality; Liu et al., 2015; Liu et al., 2016). Cronbach’s α for the TBC importance was as follows: caring = .78, professional = .72, and overall = .87. Cronbach’s α for the TBC behavior was as follows: caring = .89, professional = .87, and overall = .94.

Students’ satisfaction

The measure used to assess students’ satisfaction was developed by Hackman and Oldham (1974) and consists of three items. The survey was developed to measure job satisfaction; therefore, the text was slightly modified to fit the context. For example, the original item “The overall quality of the supervision I receive in my work” was adopted to “The overall quality of the teaching I receive in my class.” The other two items were “The degree of respect and fair treatment I receive from my instructor is satisfying” and “The amount of support and guidance I receive from my instructor are satisfying.” The survey used a 5-point Likert agreement scale from 1 (extremely dissatisfied) to 5 (extremely satisfied). The Cronbach’s α was .81.

Control variables

Stigall and Blincoe (2015) investigated which student and course qualities affect student ratings via the TBC for both the one-factor TBC and the two-factor TBC. Therefore, control variables were based on Stigall and Blincoe (2015) and included gender, race/ethnicity, age, and class standing. Further, previous publications mention an association of expected or actual course grades and student satisfaction (Hanus & Fox, 2015; Howard & Maxwell, 1980; Zabaleta, 2007). Gender and race/ethnicity were dummy coded. Class standings were 1 = first year, 2 = sophomore, 3 = junior, 4 = senior, and 5 = graduate, essentially representing the time spent in the university (i.e., tenure). Grades were on a plus and minus grading system: for example, A = 12, A− = 11, B+ = 10, to F = 1.

Analyses

All data analyses were performed in R (R Core Team, 2018) and IBM SPSS Statistics 24. Descriptive statistics were performed including means, SDs, correlations (Sproull, 1995), and reliability of the instruments (Sekaran & Bougie, 2011). The data distribution was checked using an Anderson–Darling (AD) test, which is one of the most potent empirical distribution function tests for a normality check of the data (Razali & Wah, 2011). The test revealed that the data were not normally distributed.

Before performing inferential statistics, a measurement model assessment was conducted in R to empirically justify using data collected for the TBC behaviors and students’ satisfaction. A confirmatory factor analysis (CFA) was performed to demonstrate convergence and discriminant validity (Rosseel, 2012). For the TBC behaviors, the CFA was conducted for the two-factor model (i.e., caring and professional). A multiindexes strategy was used to examine the goodness of fit (Hu & Bentler, 1999). The data indicated marginal but acceptable fit, χ²(316) = 818.23, p < .001, χ²/df = 2.6, root mean square error of approximation = .09, comparative fit index = .82, standardized root mean residual = .07. The model fit was comparable to Keeley et al. (2006). Item loadings were as proposed and significant (p < .001), providing evidence for convergence.

Based on the results of the AD test, inferential statistics employed nonparametric hierarchical linear regression (HLR) to test Hypotheses 1a and b using R. HLR involves the stepwise addition of predictor variables to the model while focusing on the significant change in the level of model significance (Petrocelli, 2003). Predictor variables are added into the analysis based on a specified order, typically control variables first. HLR is a framework for model comparison to determine which predictor variables explain the data best (i.e., determine the dominant predictor). HLR was performed according to Cohen et al. (2003) and Pertocelli (2003).

Finally, to test Hypotheses 2 for the two dimensions of the TBC, a test for simple mediation was performed following the procedure outlined by Preacher and Hayes (2004) and Hayes (2013) using IBM SPSS Statistics 24 with PROCESS Version 3.1. The method by Hayes (2013) is preferred over Baron and Kenny (1986) as it solves various limitations such as lower ability to detect the mediation effect and the inability to explicitly quantify the magnitude of the mediation effect (Hayes, 2013). Further, the bootstrapping approach is a more valid and powerful method for testing mediation effects and, for this reason, it should be the method of choice in mediation analysis.

Results

Descriptive Statistics, Correlations, and Reliabilities

Table 1 presents a descriptive analysis including correlations for the study variables. Correlations include the control variables. The control variable race and expected grades showed significant correlations (r = −.15 and r = .28, respectively, both ps < .05) with student satisfaction. Therefore, both race and expected grades were used in hypothesis testing. Both TBC behavior caring and professional showed significant correlations (r = .75 and r = .82, respectively, both ps < .001) with student satisfaction. Further, the TBC importance caring dimension showed a significant correlation with student satisfaction (r = .22, p < .01). All measures produced adequate estimates of reliability (α = .72–.89).

Table 1.
Means, Standard Deviations (SDs), α Coefficients, and Intercorrelations.

Variable M SD 1 2 3 4 5 6 7 8 9 10

1. SSt 2.98 1.34 —

2. Age 22.23 6.19 .64*** —

3. Gender — — .11 −.03 —

4. Race — — −.02 .00 −.03 —

5. EG 10.03 1.85 −.04 .01 −.02 .02 —

6. TB_C 4.11 0.61 .10 .18* −.02 −.07 .18* (.89)

7. TB_P 4.40 0.53 .09 .19** −.07 −.01 .18* .85 (.87)

8. SS 4.27 0.67 .03 .13 .00 −.15* .28* .82* .75*** (.81)

9. TI_C 3.36 0.37 −.09 −.11 .08 .10 −.11 .27*** .16* .22** (.78)

10. TI_P 3.46 0.35 −.04 −.04 .12 .10 −.10 .18* .15* .08 .69 (.72)

Note. N = 194. The values in italics represent Cronbach’s α for the measures used. Gender and race are dummy variables. SSt = student standing; EG = expected grade; TB_C = TBC behavior caring; TB_P = TBC behavior professional; SS = student satisfaction; TI_C = TBC importance caring; TI_P = TBC importance professional.

p < .05. p < .01. p < .001.

Hypotheses 1 Tests: Teacher Behavior and Students’ Satisfaction

HLR was performed according to Cohen et al. (2003) and Pertocelli (2003) to test Hypotheses 1a and b. The predictor variables were added into the analysis consecutively in the specified order. The control variables (i.e., race and expected grade) were first, followed by the theorized order of caring TBC behavior, and finally, professional TBC behavior. Further, to thoroughly test the hypothesized relationship, caring TBC behavior and professional TBC behavior were also entered in reverse order to assess the significant model change of the independent variables. Specifically, if caring adds significantly after accounting for professional, the argument that caring is more important has additional support. The parameters evaluated at each step to determine the best model were as follows: (a) adjusted R ², (b) the change in adjusted R ² (▵R ²), and (c) predicted residual sums of squares (PRESS). Better model fit means a larger adjusted R ², significant change in ▵R ², and a smaller PRESS. The standardized β coefficient indicates how strongly each independent variable influences the dependent variable and can, therefore, be used to determine the dominant predictor (Hauke & Kossowski, 2011). The outcomes of the HLR are displayed in Table 2.

Table 2.
Parameters of the Hierarchical Linear Regression Analysis for Hypothesis 1.

Dependent Variable: Students’ Satisfaction (SS)

Model 1 Model 2 Model 3 Model 4

Control variables

Expected grade .27* .13 .13 .14

Race −.12 −.08 −.09* −.13

Independent variables

TB_C .79* .62*

TB_P .20 .73***

F value 9.7* 145.7* 114.8* 96.3*

Adjusted R ² .08 .69 .70 .60

▵ Adjusted R ² .61* .01 .52*

Predicted residual sums of square 182.3 60.8 59.5 80.2

Note. Relationships are standardized β coefficients. Dependent variable is SS. TB_C = TBC behavior caring; TB_P = TBC behavior professional; TI_C = TBC importance caring; TI_P = TBC importance professional.

p < .05. p < .01. p < .001.

Caring TBC behavior and professional TBC behavior were entered for Model 2 and Model 3, respectively, causing a significant change in ▵R ² providing evidence that both caring and professional were relevant variables in the overall model. Model 3 had the highest adjusted R ² (.70, p < .001) and also the lowest PRESS. Thus, Model 3 was the best model and used to address Hypothesis 1a. Race (β = −.09, p < .05) and expected grade (β = .13, p < .01) explained 8% of the variability. Caring TBC behavior was the dominant predictor with β = .62, p < .001, therefore supporting Hypothesis 1a.

Professional TBC behavior was a significant predictor with β = .22, p < .01 but to a lesser degree than caring TBC behavior. Finally, caring TBC behavior and professional TBC behavior were entered in reverse order. Professional TBC behavior was a significant predictor with β = .73, p < .001 when entered first (Model 4). Adding caring TBC behavior last caused a significant change in ▵R ² (p < .01) providing additional support that TBC behavior caring is the dominant predictor or, in other words, that TBC behavior professional predicts student satisfaction to a lesser degree, therefore supporting Hypothesis 1b. Note that no additional model is shown in Table 2 for adding caring TBC behavior last as it is statistically the same as Model 3.

Supplemental Analysis

A supplemental analysis was performed based on Hypothesis 1. Hypothesis 1 was based on the two TBC behavior dimensions of caring and professionalism and their relationship with students’ satisfaction. This supplemental analysis will investigate the relationship of the TBC behavior with students’ satisfaction on an item level (i.e., for all 28 behavior items of the TBC).

A multiple linear regression was performed. The sample size was N = 194, and adjusted R ² was .76 (p < .001). Control variables were expected grade and race. There were four TBC behavior items that showed a significant relationship with students’ satisfaction: two from the TBC caring behavior and two from TBC professional behavior. For TBC caring behavior, significant predictor items were provides positive feedback (β = .16, p < .01) and understanding (β = .16, p < .01); for TBC professional behavior, significant predictor items were confident (β = .21, p < .001) and knowledgeable about the subject (β = −.12, p < .05). The control variable expected grades also showed a significant relationship with students’ satisfaction (β = .11, p < .05).

Hypotheses 2 Tests: TBC Importance and Students’ Satisfaction and the Mediation of TBC Behavior

IBM SPSS Statistics 24 with PROCESS Version 3.1 was used to estimate confidence intervals (CIs) for the indirect effect (Hayes, 2013). Control variables were race and expected grade. The results of the mediation are displayed in Table 3.

Table 3.
Parameters of the Mediation Analysis for Hypothesis 2.

95% Confidence Intervals a Path (X → M) b Path (M|X → Y) c′ Path (X|M → Y)

Lower Upper β SE β SE β SE

TBC caring .23 .64 .50* .11 .86* .05 .06 .08

Expected grade .07 .02 .05 .02 .05** .02

Race −.18 .14 −.19* .09 −.19* .09

TBC professional .05 .44 .26* .11 .93* .06 −.01 .09

Expected grade .06 .02 .05 .02 .05 .02

Race −.03 .13 −.29 .11 −.29 .11

Note. a path denotes the relationship between the independent variable (X) and mediator variable (M). b path denotes the relationship of the mediator variable (M) and the dependent variable (Y) in the presence of the independent variable (X). c′ path denotes no relationship between the independent variable (X) and the dependent variable (Y) in the presence of the mediator (M). For TBC caring, the independent variable (X) was TBC importance caring, the moderator (M) was TBC behavior professional, and the dependent variable (Y) was students’ satisfaction (SS). For TBC professional, the independent variable (X) was TBC importance professional, the moderator (M) was TBC behavior caring, and the dependent variable was SS. Control variables were expected grade and race.

p < .05. p < .01. p < .001.

For both, TBC caring and professional, I examined whether TBC behavior mediated the relationship between TBC importance and students’ satisfaction. In both cases, with TBC behavior as the mediator, the a path (the relationship between the independent variable [X] and mediator variable [M]; TBC importance to TBC behavior) and the b path (the relationship of the mediator variable [M] and the dependent variable [Y] in the presence of the independent variable [X]; TBC behavior to students’ satisfaction in the presence of TBC importance) were both significant and positive. The c′ path (no relationship between the independent variable [X] and the dependent variable [Y] in the presence of the mediator [M]; TBC importance to students’ satisfaction in the presence of TBC behavior) was not significant. The 95% CI for the indirect effect excluded zero indicating that TBC behavior mediates the relationship between TBC importance and students’ satisfaction, therefore providing support for Hypothesis 2.

Discussion

This study found that there is a positive relationship between teacher behaviors, as measured with the TBC, and students’ satisfaction. Specifically, the caring dimension was the dominant predictor explaining most of the variability. According to Keeley et al. (2006), the TBC caring dimension includes the following behaviors: providing constructive feedback, rapport, realistic expectations of students, sensitive and persistent, flexible/open-minded, humble, promoting class discussion, understanding, encouraging and caring for students, accessible, promoting critical thinking/intellectually stimulating, strives to be a better teacher, and enthusiastic about teaching and about the topic. The TBC caring dimension includes behaviors that were supported by previous research to relate to student satisfaction such as understanding (DeShields et al., 2005) and feedback (Bolliger & Martindale, 2004; DeShields et al., 2005; Y. Hill et al., 2003).

A supplemental analysis showed that four teacher behaviors contributed to students’ satisfaction: two from the caring dimension (i.e., understanding and provides positive feedback) and two from the professional dimension (i.e., confident and knowledgeable about the subject). Interestingly, knowledgeable about the subject showed a negative relationship with students’ satisfaction (all others were positive), which will be further discussed under practical implications.

Furthermore, the students’ importance expectations have a role in contributing to the formation of satisfaction, as indicated by previous research (e.g., Anderson & Sullivan, 1993; Yi, 1993). The mediation analysis revealed that there is a positive relationship between the expectations students have toward teacher behaviors and the perceived teacher behaviors that ultimately form students’ satisfaction.

Implications

The study, in part, provides insight into teacher behaviors that relate to class outcomes (i.e., student satisfaction) using the TBC framework. The study provides, to my knowledge, one of the first quantitative perspectives on how the TBC behavioral dimensions relate to class outcome. Further, the study indicates how student expectations, as rated with the TBC, and the teacher behaviors that the students experience interact toward the formation of students’ satisfaction.

Theoretical implications

This study provides multiple contributions. First, within the theoretical framework of the TBC, it provides further evidence of specific behaviors that are related to student satisfaction. Findings extend previous research findings using a different theoretical framework (e.g., Bolliger & Martindale, 2004; DeShields et al., 2005; Douglas et al., 2006, 2008; Y. Hill et al., 2003). Second, it was one of the first studies to utilize TBC behaviors and importance ratings to predict a class outcome statistically. Third, based on the indication that teacher behaviors, as rated by the TBC, mediate TBC importance and students’ satisfaction, the study incorporates student expectations expanding knowledge toward a more comprehensive view of how student satisfaction is formed. Finally, the study provides further evidence of the validity of the TBC. Model fit of the two dimensions of caring and professional was confirmed yielding similar results to Keeley et al. (2006).

Practical implications

Student satisfaction is commonly used as an indicator of academic success (McFarland & Hamilton, 2005; Summers et al., 2005). Another class outcome that measures success is students’ final grades. However, student satisfaction is interesting as it is known to contribute to student retention (Edwards & Waters, 1982; Schertzer & Schertzer, 2004). Compared to a traditional classroom setting, students in an online class typically have lower levels of satisfaction (Summer et al., 2005) and, therefore, higher dropout rates (Bolliger & Martindale, 2004; Willging & Johnson, 2009). That may be due to the difficulty of exhibiting caring behaviors in an online environment.

The findings of this study may contribute toward teacher development with the goal of achieving excellent teaching. A teacher should demonstrate behaviors of both the caring and the professional dimensions to achieve high levels of student satisfaction. Teacher behaviors listed in the caring dimension should receive special attention as this study showed that it may be the dominant predictor of students’ satisfaction.

More specifically, behaviors associated with being understanding, providing positive feedback, being confident, and being knowledgeable about the subject should be exhibited. Surprisingly, being knowledgeable about the subject had an inverse relationship with students’ satisfaction indicating that being too knowledgeable decreases students’ satisfaction. It needs to be noted that this behavior was found to be important by students and faculty in the study by Buskist et al. (2002). Demonstrating too much knowledge may be intimidating. According to Scott et al. (2015), intimidation in teaching can have an adverse effect on learning outcomes and even create the intent to change majors. Therefore, the key might be to be very knowledgeable about the subject, but instead of showing off this knowledge, the teacher should explain the subject in simple words, provide examples, and answer questions about the subject.

Finally, while there is no direct relationship between students’ expectation/importance and students’ satisfaction, however, they do matter toward the formation of students’ satisfaction. Thus, it is important for a teacher to understand their “customers.” For example, using the TBC importance survey at the beginning of the semester can provide insight into the expectations students have when it comes to teacher behavior.

Limitations and Future Research

The correlational study had a cross-sectional design which requires caution for an assessment of cause and effect. Additionally, the survey order was the same for all participants. Method bias may be minimized by utilizing counterbalancing to control for repeated measures design elements. Further, the teacher behaviors were limited to the behavioral typologies within the TBC model and its two dimensions providing a general guideline of behaviors, but no specific behaviors. Lastly, the study was limited to a university business school and therefore may have limitations in its generalizability.

There has been interesting research on the TBC. Studies explored TBC importance ratings (e.g., Buskist et al., 2002; Keeley et al., 2016; Liu et al., 2016; Schaeffer et al., 2003; Vulcano, 2007). Interestingly, how the level of importance affects class outcomes, such as academic achievement (i.e., GPA) and development of skills (Lizzio et al., 2002), is widely unknown. Similar, there has been limited research on the relationship between TBC behaviors and class outcomes. Studies investigating excellent teacher behaviors (e.g., Keeley et al., 2010; Liu et al., 2015) were mainly focused on detecting differences in behaviors of different teachers, which was an essential contribution. More research is required to explore how TBC behavior and TBC importance ratings influence class outcomes.

An interesting finding in this study was the negative relationship between being knowledgeable about the subject and students’ satisfaction. The adverse effect indicates that too much of a good thing may be a bad thing, which may be the case more likely in the professional dimension of the TBC such as being punctual/manage time, being confident, and being authoritative. Caring behaviors, on the other hand, may counter the negative effect of certain professional behaviors. It could be explored if other behaviors have negative effects on class outcomes and if there is an interaction between the TBC dimensions of caring and professional.

Variable	M	SD	1	2	3	4	5	6	7	8	9	10
1. SSt	2.98	1.34	—
2. Age	22.23	6.19	.64***	—
3. Gender	—	—	.11	−.03	—
4. Race	—	—	−.02	.00	−.03	—
5. EG	10.03	1.85	−.04	.01	−.02	.02	—
6. TB_C	4.11	0.61	.10	.18*	−.02	−.07	.18*	(.89)
7. TB_P	4.40	0.53	.09	.19**	−.07	−.01	.18*	.85	(.87)
8. SS	4.27	0.67	.03	.13	.00	−.15*	.28***	.82***	.75***	(.81)
9. TI_C	3.36	0.37	−.09	−.11	.08	.10	−.11	.27***	.16*	.22**	(.78)
10. TI_P	3.46	0.35	−.04	−.04	.12	.10	−.10	.18*	.15*	.08	.69	(.72)

	Dependent Variable: Students’ Satisfaction (SS)
Control variables
Expected grade	.27***	.13**	.13**	.14**
Race	−.12	−.08	−.09*	−.13**
Independent variables
TB_C		.79***	.62***
TB_P			.20**	.73***
F value	9.7*	145.7***	114.8***	96.3***
Adjusted R ²	.08	.69	.70	.60
▵ Adjusted R ²		.61***	.01**	.52***
Predicted residual sums of square	182.3	60.8	59.5	80.2

	95% Confidence Intervals	a Path (X → M)	b Path (M\|X → Y)	c′ Path (X\|M → Y)
TBC caring	.23	.64	.50***	.11	.86***	.05	.06	.08
Expected grade			.07**	.02	.05**	.02	.05**	.02
Race			−.18	.14	−.19*	.09	−.19*	.09
TBC professional	.05	.44	.26*	.11	.93***	.06	−.01	.09
Expected grade			.06**	.02	.05**	.02	.05**	.02
Race			−.03	.13	−.29**	.11	−.29**	.11

Footnotes

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

References

Adhikary

(2017). Leadership style and student satisfaction: Mediation of teacher effectiveness. Asian Journal of Business and Management, 5(2), 62–69. https://doi.org/10.24203/ajbm.v5i2.4632

Anderson

Sullivan

(1993). The antecedents and consequences of customer satisfaction for firms. Marketing Science, 12(2), 125–143.

Astin

A. W.

(1993). What matters in college? Four critical years revisited. Jossey-Bass.

Baron

Kenny

(1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182.

Bolliger

Martindale

(2004). Key factors for determining student satisfaction in online courses. International Journal of E-Learning, 3(1), 61–67.

Buskist

Sikorski

Buckley

Saville

(2002). Elements of master teaching. In Davis

S. F.

Buskist

(Eds.), The teaching of psychology: Essays in honor of Wilbert J. McKeachie and Charles L. Brewer (pp. 27–39). Lawrence Erlbaum Associates.

Cohen

West

Aiken

(2003). Applied multiple regression/correlation analysis for the behavioral sciences (3rd ed.). Lawrence Erlbaum.

DeShields

Kara

Kaynak

(2005). Determinants of business student satisfaction and retention in higher education: Applying Herzberg’s two-factor theory. International Journal of Educational Management, 19(2), 128–139. https://doi.org/10.1108/09513540510582426

Douglas

Barnes

(2006). Measuring student satisfaction at a UK university. Quality Assurance in Education, 14(3), 251–267. https://doi.org/10.1108/09684880610678568

10.

Douglas

McClelland

Davies

(2008). The development of a conceptual model of student satisfaction with their experience in higher education. Quality Assurance in Education, 16(1), 19–35. https://doi.org/10.1108/09684880810848396

11.

Edwards

Waters

(1982). Involvement, ability, performance, and satisfaction as predictors of college attrition. Educational and Psychological Measurement, 42, 1149–1152. https://doi.org/10.1177/001316448204200421

12.

Elliot

Healy

(2001). Key factors influencing student satisfaction related to recruitment and retention. Journal of Marketing for Higher Education, 10(4), 1–33. https://doi.org/10.1300/J050v10n04_01

13.

Frisby

Martin

(2010). Instructor-student and student-instructor rapport in the classroom. Communication Education, 59, 146–164. https://doi.org/10.1080/03634520903564362

14.

Frisby

Myers

(2008). The relationship among perceived instructor rapport, student participation, and student learning outcomes. Texas Speech Communication Journal, 33, 27–34.

15.

Granitz

N. A

Koernig

S. K.

Harich

K. R.

(2009). Now it’s personal: Antecedents and outcomes of rapport between business faculty and their students. Journal of Marketing Education, 31, 52–65. https://doi.org/10.1177/0273475308326408

16.

Hackman

Oldham

(1974). The job diagnostics survey: An instrument of the diagnosis of jobs and the evaluation of job redesign projects (Tech. Rep. No. 4). Department of Administrative Sciences, Yale University.

17.

Hanus

Fox

(2015). Assessing the effects of gamification in the classroom: A longitudinal study on intrinsic motivation, social comparison, satisfaction, effort, and academic performance. Computers and Education, 80, 152–161. https://doi.org/10.1016/j.compedu.2014.08.019

18.

Hauke

Kossowski

(2011). Comparison of values of Pearson’s and Spearman’s correlation coefficients on the same sets of data. Quaestiones Geographicae, 30(2), 87–93. https://doi.org/10.2478/v10117-011-0021-1

19.

Hayes

(2013). Introduction to mediation, moderation, and conditional process analysis (2nd ed.). Guilford.

20.

Hayes

Montoya

Preacher

Selig

Page-Gould

Sharples

(2016, January). Advances in within-subject mediation analysis [Symposium]. Meeting of the Society for Personality and Social Psychology, San Diego, CA, United States.

21.

Hill

Epps

(2010). The impact of physical classroom environment on student satisfaction and student evaluation of teaching in the university environment. Academy of Educational Leadership Journal, 14(4), 65–79.

22.

Hill

Lomas

MacGregor

(2003). Students’ perceptions of quality in higher education. Quality Assurance in Education, 11(1), 15–20. https://doi.org/10.1108/09684880310462047

23.

Howard

G. S.

Maxwell

S. E.

(1980). Correlation between student satisfaction and grades: A case of mistaken causation. Journal of Educational Psychology, 72(6), 810–820. https://dx-doi-org.web.bisu.edu.cn/10.1037/0022-0663.72.6.810

24.

Bentler

P. M.

(1999). Cutoff criteria for fit indices in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55. https://doi.org/10.1080/10705519909540118

25.

Keeley

Furr

Buskist

(2010). Differentiating psychology students’ perceptions of teachers using the teacher behavior checklist. Teaching of Psychology, 37(1), 16–20. https://doi.org/10.1080/00986280903426282

26.

Keeley

Ismail

Buskist

(2016). Excellent teachers’ perspective on excellent teaching. Teaching of Psychology, 43(3), 175–179. https://doi.org/10.1177/0098628316649307

27.

Keeley

Smith

Buskist

(2006). The Teacher Behaviors Checklist: Factor analysis of its utility for evaluating teaching. Teaching of Psychology, 33(2), 84–91. https://doi.org/10.1207/s15328023top3302_1

28.

Komarraju

(2013). Ideal teacher behaviors: Student motivation and self-efficacy predict preferences. Teaching of Psychology, 40(2), 104–110. https://doi.org/10.1177/0098628312475029

29.

Landrum

Stowell

(2013). The reliability of student ratings of master teacher behaviors. Teaching of Psychology, 40(4), 300–303. https://doi.org/10.1177/0098628313501043

30.

Liu

Keeley

Buskist

(2015). Chinese college students’ perceptions of characteristics of excellent teachers. Teaching of Psychology, 42(1), 83–86. https://doi.org/10.1177/0098628314562684

31.

Liu

Keeley

Buskist

(2016). Chinese college students’ perceptions of excellent teachers across three disciplines: Psychology, chemical engineering, and education. Teaching of Psychology, 43(1), 70–74. https://doi.org/10.1177/0098628315620888

32.

Lizzio

Wilson

Simons

(2002). University students’ perceptions of the learning environment and academic outcomes: Implications for theory and practice. Studies in Higher Education, 27(1), 27–52. https://doi.org/10.1080/03075070120099359

33.

Lowman

(1995). Mastering the techniques of teaching (2nd ed.). Jossey-Bass.

34.

McFarland

Hamilton

(2005). Factors affecting student performance and satisfaction: Online versus traditional course delivery. Journal of Computer Information Systems, 46(2), 25–32.

35.

Micari

Pazos

(2012). Connecting to the professor: Impact of the student-faculty relationship in a highly challenging course. College Teaching, 60, 41–47.

36.

Moore

(2005). The Sloan Consortium quality framework and the five pillars. The Sloan Consortium.

37.

Parasuraman

Zeithaml

Berry

L. L.

(1994). Reassessment of expectations as a comparison standard in measuring service quality: Implications and further research. Journal of Marketing, 1, 111–124. https://doi.org/10.2307/1252255

38.

Petrocelli

(2003). Hierarchical multiple regression: Regression in counseling research: Common problems and possible remedies. Journal of Measurement and Evaluation in Counseling and Development, 36, 9–21.

39.

Piccoli

Ahmad

Ives

(2001). Web-based virtual learning environments: A research framework and a preliminary assessment of effectiveness in basic IT skills training. MIS Quarterly, 25(4), 401–426. https://doi.org/10.2307/3250989

40.

Preacher

Hayes

(2004). SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behavior Research Methods, Instruments, and Computers, 36(4), 717–731. https://doi.org/10.3758/BF03206553

41.

Razali

Wah

(2011). Power comparisons of Shapiro-Wilk, Kolmogorov-Smirnov, Lilliefors and Anderson-Darling tests. Journal of Statistical Modeling and Analytics, 2(1), 21–31.

42.

R Core Team. (2018). R: A language and environment for statistical computing. R Foundation for Statistical Computing.

43.

Rosseel

(2012). Lavaan: An R package for structural equation modeling. Journal of Statistical Software, 48(2), 1–36. https://doi.org/10.18637/jss.v048.i02

44.

Ryan

Wilson

J. H.

(2014). Professor-Student Rapport Scale: Psychometric properties of the brief version. Journal of Scholarship of Teaching and Learning, 14, 64–74. https://doi.org/10.14434/josotl.v14i3.5162

45.

Schaeffer

Epting

Zinn

Buskist

(2003). Student and faculty perceptions of effective teaching: A successful replication. Teaching of Psychology, 30(2), 133–136.

46.

Schertzer

. (2004). Student satisfaction and retention: A conceptual model. Journal of Marketing for Higher Education, 14(1), 79–91. https://doi.org/10.1300/J050v14n01_05

47.

Scott

Caldwell

Barnes

Barrett

(2015). Teaching by humiliation and mistreatment of medical students in clinical rotations: A pilot study. The Medical Journal of Australia, 203(4), 185. https://doi.org/10.5694/mja15.00189

48.

Sekaran

Bougie

(2011). Research methods for business: A skill building approach. John Wiley and Sons.

49.

Sproull

(1995). Handbook of research methods (2nd ed.). Scarecrow.

50.

Stigall

Blincoe

(2015). Student and instructor use of the teacher behavior checklist. Teaching of Psychology, 42(4), 299–306. https://doi.org/10.1177/0098628315603061

51.

Stone-Romero

Rosopa

(2004). Inference problems with hierarchical multiple regression-based tests of mediating effects. In Martocchio

(Series Ed.), Research in personnel and human resources management (Vol. 23, pp. 249–290). Elsevier.

52.

Summers

Waigandt

Whittaker

(2005). A comparison of student achievement and satisfaction in an online versus a traditional face-to-face statistics class. Innovative Higher Education, 29(3), 233–250. https://doi.org/10.1007/s10755-005-1938-x

53.

Vulcano

(2007). Extending the generality of the qualities and behaviors constituting effective teaching. Teaching of Psychology, 34(2), 114–117. https://doi.org/10.1080/00986280701293198

54.

Westerman

Nowicki

Plante

(2002). Fit in the classroom: Predictors of student performance and satisfaction in management education. Journal of Management Education, 26(1), 5–18. https://doi.org/10.1177/105256290202600102

55.

Willging

Johnson

(2009). Factors that influence students’ decision to dropout of online courses. Journal of Asynchronous Learning Networks, 13(3), 115–127.

56.

(1993). The determinants of consumer satisfaction: The moderating role of ambiguity. Advances in Consumer Research, 20, 502–506.

57.

Zabaleta

(2007). The use and misuse of student evaluations of teaching. Teaching in Higher Education, 12(1), 55–76. https://doi.org/10.1080/13562510601102131

	Dependent Variable: Students’ Satisfaction (SS)
	Model 1	Model 2	Model 3	Model 4
Control variables
Expected grade	.27***	.13**	.13**	.14**
Race	−.12	−.08	−.09*	−.13**
Independent variables
TB_C		.79***	.62***
TB_P			.20**	.73***
F value	9.7*	145.7***	114.8***	96.3***
Adjusted R ²	.08	.69	.70	.60
▵ Adjusted R ²		.61***	.01**	.52***
Predicted residual sums of square	182.3	60.8	59.5	80.2

	95% Confidence Intervals		a Path (X → M)		b Path (M\|X → Y)		c′ Path (X\|M → Y)
	Lower	Upper	β	SE	β	SE	β	SE
TBC caring	.23	.64	.50***	.11	.86***	.05	.06	.08
Expected grade			.07**	.02	.05**	.02	.05**	.02
Race			−.18	.14	−.19*	.09	−.19*	.09
TBC professional	.05	.44	.26*	.11	.93***	.06	−.01	.09
Expected grade			.06**	.02	.05**	.02	.05**	.02
Race			−.03	.13	−.29**	.11	−.29**	.11