Abstract
Evaluating teachers has always been controversial, in part because inferences about their behaviors based on student performance are drawn in situations where they do not have full control of who they teach, students’ prior knowledge, and amounts of time needed in order to adapt teaching to individual differences (E. L. Baker, 2014). Over the years, more process-oriented approaches to teacher evaluation have used judgments of the quality of classroom artifacts, observation of teachers’ classroom behavior using high inference judgments, and even the neatness of the room and blinds. Needless to say, a more scientific and defensible approach is desirable. Moreover, poor teacher evaluation systems may hurt not only teachers but, inadvertently, students as well (Popham, 2013). But teacher evaluation is a necessity. In addition to their competency in assisting students to learn the content and skills expected in schools, teachers serve many other roles. They can be attentive to individual students’ needs, recognize and respond to their cultural differences, and provide support and model desirable behaviors. Many of these behaviors can be combined in the category of noncognitive factors or summarized by the term social psychological perspectives and behaviors.
The focus of this chapter is on the description and assessment of teachers’ social psychological factors, using the scientific literature as a base. Research on teachers’ social psychological domains has an ultimate goal of populating classrooms with competent people who can model and incite behaviors that assist students in their own learning. Social psychological based activities may include teachers’ showing that they care for students’ well-being; managing classrooms to keep interest, fairness, and effort at the fore; intervening in situations that may hurt students, such as bullying; and all the while focusing on helping students achieve learning goals. Clearly there are cognitive components of these behaviors. However, this chapter will examine, not exhaustively, key topics in the teacher social psychological literature; methods used to determine teachers’ status on variables, in this domain; and their association and impact in a variety of settings.
One useful word at the outset about student outcomes is appropriate here, as they are increasingly proposed as the ultimate dependent measures used to judge teacher effectiveness. Even in the relatively well-studied areas of teaching content and skills, there is considerable contention about the utility of student test performance as a major component of teacher evaluation. For instance, there is great and vocal disagreement about the design, utility, and fairness of cognitive or subject matter examinations (E. L. Baker et al., 2010; Gordon Commission on the Future of Assessment in Education, 2013). Periodic efforts to redesign and refocus these assessments have occurred (see, e.g., MET Project, 2012). Nonetheless, whatever their form, there is an abiding belief that teachers should be responsible for student performance. Without debating the details and merits of the belief, the claim has been made that teachers’ social psychological behaviors are important and should be considered in their evaluation. That said, the question is how?
Since the early 1990s, with the passage of the Improving America’s Schools Act (1994), social psychological concerns have broken through the policy barrier, although weakly and dependent upon archival data. With regard to students, there are three major threads: (a) noncognitive behaviors that signal their overall attitudes toward school, such as absences and tardiness, and (b) socioemotional perspectives that can be partially developed through educators’ efforts, that is, a sense of balance, tolerance for other viewpoints and people, teamwork, and exploratory behaviors. Educators and parents have legitimate concerns about the boundaries and appropriateness of schools’ prerogatives with regard to social psychological goals and outcomes, especially in connection to the roles and responsibilities of the family. The third major thread includes social psychological processes that support instrumentally cognitive learning by students. Since the 1990s, school goals have been operationalized as standards to be shared within and, most recently, among states. Although all three threads combine into a strong argument for research attention, our emphasis will be on the factors that contribute to student cognitive learning. For example, extensive lines of research have considered cultural interactions, including conscious and unconscious stereotyping (Dijksterhuis & Nordgren, 2006), experienced by all of us. Our review includes this range for context but will concentrate on teacher factors that are potentially influential on a student’s own behaviors and accomplishments. We include empirical studies from the past 10 years, drawn from peer related journals, with the exception of older conceptual or seminal works included in books, edited books, or other study reports. 1
Classification of Teacher Social Psychological Perspectives
Although not mutually exclusive and often intercorrelated, teacher social psychological factors are conceptualized here on a dimension moving from traits or tendencies that are relatively impervious to change, such as global temperament, specific personality variables, and long-held predispositions on one hand, and those attitudes and behaviors that are susceptible to change, even if taking a long period. Although many of these factors could have relevancy for all teachers, such as “responsibility,” other variables may be especially applicable in classroom settings, for instance, “ability to motivate others.” We will consider the range between persistent to malleable social and psychological factors.
Personality
Relatively intransigent factors, those that characterize the routine reactions and responses of the individual, are called personality variables. According to Olver and Mooradian (2003), the bulk of personality variables derive from the five-factor model (FFM). Personality traits are enduring dispositions that lead to consistent patterns of self-perception and behaviors with others in various environments. The FFM organizes individual personality traits into five broad categories. These five dimensions include Conscientiousness (i.e., responsible, dependable), Emotional Stability (i.e., calm under pressure, not neurotic), Extraversion (i.e., outgoing, assertive), Agreeableness (i.e., cooperative, loyal), and Openness to Experience (i.e., curious, imaginative). Multiple meta-analyses have shown consistent relationships between personality and performance in various domains (e.g., Barrick, Mount, & Judge, 2001; Hurtz & Donovan, 2000). The FFM is commonly measured using self-report Likert-type scales such as John, Donahue, and Kentle’s (1991) Big 5 Inventory, Goldberg’s (1999) International Personality Item Pool, and Costa and McCrae’s (1992) NEO Personality Inventory–Revised. The NEO Personality Inventory–Revised is one of the most widely used measures of the FFM, and it has 6-year test-retest reliability ranging from .63 to .83; strong consensual validity between self-, peer, and spouse reports; and good convergent validity with other personality and well-being measures (Costa & McCrae, 1992).
Two interesting perspectives may be considered in the light of teachers and their personality. First, personality attributes may count explicitly during the accession process, that is, in teachers’ interviews for school positions. The ability to display an appropriately extraverted (enthusiastic and outgoing) style, complemented by agreeableness and conscientiousness, is desirable for teachers in a variety of settings, in addition, of course, to mastery of subject matter, knowledge of human development, and pedagogical skills. In times of teacher shortage, these factors remain desirable but may have less weight in selection. Second, some of these variables may not be “traits,” which are exhibited in all or most situations, but may vary by the state in which teachers find themselves. For example, a teacher might be very responsible about planning instruction and giving assignments but less consistent about grading student responses promptly. In an elementary or other intact classroom, teachers may be responsible for a wide range of content. As a result, teachers’ exhibitions of behaviors may vary with their confidence in their own mastery of the subject matter. In any case, in the work we have undertaken, we are focusing on operational definitions of teacher attitudes, predispositions, and practices, as well as those that may imply developmental courses of action. Simply describing teachers is probably not the straightest path toward a quality educational system.
The target variables of this review also coincide with the deeper learning competencies identified as crucial for students to succeed in 21st-century jobs and civic life (Herman & Linn, 2013; O’Neil, Perez, & Baker, 2014; Webb, 2007). At the heart of the summary term deeper learning is a set of competencies that students must master in order to develop a keen understanding of academic content, achieve at high levels, and apply their knowledge to problems in the classroom and on the job. These include mastery of core academic content, ability to think critically and solve complex problems, ability to work collaboratively, communicating effectively, learning how to learn, and developing adaptive academic mindsets. These competencies also connect to the teacher variables discussed in this review. Before delving into a select review of our target constructs, we turn to one of the main factors that influence teachers’ ability to guide students and their learning as well as their own teaching process: motivation.
Teacher Motivation
Motivation is considered a noncognitive factor that significantly influences performance. It is often referred to as the inner drive or force that sustains and directs our goal-directed behavior. Thus, motivation is a precursor for choosing and performing many other processes effectively, including, but not limited to, mastering content, thinking critically, problem solving, collaborating, and communicating. The indicators of motivated behavior include active choice, persistence, and mental effort or performance. Essentially, when individuals choose to do one thing as opposed to another, they are exhibiting active choice. Persistence involves individuals continuing to pursue a goal or task in the face of challenges and difficulty. Mental effort involves the amount of energy, thought, and involvement a person puts into a task.
In this section, we will address various variables and models that have been posited to underlie motivation. Rueda et al. (2010) describe the roots of motivation as comprising four major components: Interest, Self-Efficacy, Attribution, and Achievement Goals. As one reads the chapter, it will become obvious that even as we attempt to define, describe, and illustrate these aspects, inevitably one of these factors implicates the others. They are intertwined, and though defined as separate constructs, they most always co-occur. As a result, discussions focused on one factor will typically involve other perspectives, either of the self or of the self in various contexts. Let us consider some of the factors described by Rueda et al. Interest is defined as present when individuals seek and value what they learn. Self-efficacy broadly construed means that individuals perceive themselves as capable of success at particular tasks. Attribution is in play when appropriate inferences are drawn about success; students believe that their effort will lead to success and do not attribute failure to their own innate capability, luck, or other uncontrollable factors. Students who possess achievement goals with a mastery orientation in contrast to a performance orientation will be motivated to work because their goal is to understand a concept, context, or process rather than to compete with others.
One long-standing perspective on motivation and how an individual’s choice, persistence, and performance can be explained is the expectancy-value model (Wigfield & Eccles, 2000; Appendix A). This model is well researched, and in a general sense it explains the beliefs individuals have about how well they will do on an activity (i.e., expectancy) and the extent to which they value the activity. From this perspective, individuals construct motivational beliefs through social cognitive processes, grounded in larger social and cultural contexts. This model is similar to Bandura’s (1986) social cognitive theory, which posits the model of triadic reciprocity or reciprocal causation. The model of reciprocal causation postulates that individual factors as part of the person (e.g., self-efficacy), environmental factors (e.g., classroom environment, teacher expectations, etc.), and behavioral factors (e.g., behavioral indicators or manifestations) mutually influence each other to produce motivated behavior and performance.
Essentially, a student might perceive a peer successfully engaging in a certain task (i.e., behavioral factors). This observation is likely to influence positively the students’ sense of efficacy (i.e., individual factors) with respect to completing the same task. The student, in turn, choses to persist and to exert mental effort with respect to completing the task (i.e., behavioral factors). A teacher observes the same student performing the task (i.e., behavioral factors) and changes his/her expectations (i.e., environmental factors) with respect to how well the student can perform. In considering teachers, we are therefore interested in the extent to which they themselves have high values or measures in these areas as well as their ability to communicate and transform students so that they are able to exhibit similar components in their behaviors.
Recent research has suggested that teachers suffer from motivational problems more than other professional groups (de Jesus & Lens, 2005). More important, a teacher’s level of motivation influences his or her own behaviors as well as the behavior and motivational beliefs of students (Bandura, 1986). Thus, it is important to understand the internal factors that most readily affect teachers’ levels of motivation, their perceptions of students, and their capacity to influence student behaviors. As mentioned, the most salient factors that influence these relationships include teacher interest, teacher expectations, teacher efficacy, teacher attributions, and teacher goal orientations. These factors are considered various types of teacher beliefs or orientations.
Teacher Beliefs
A substantial body of research indicates that the beliefs that teachers hold about teaching and learning influence their instructional decisions and thereby student performance (Bohlmann & Weinstein, 2013). Teacher beliefs are broadly defined as “tacit, often unconsciously held assumptions about students, classrooms, and the academic material taught” (Kagan, 1992, p. 65). The general constructs of educational and teacher beliefs are expansive and include the study of more specific beliefs, such as epistemological beliefs (teachers’ view about the nature of acquisition of knowledge), teacher efficacy (beliefs about teachers’ ability to influence student performance), and attributions (beliefs about the causes of teachers’ and students’ performance; Pajares, 1992). These various teacher beliefs are critical to understand because they affect more specific pedagogical beliefs and instructional practices, and ultimately student outcomes. Beliefs and practices share a complex relationship in that beliefs and practices mutually affect one another. For example, if teachers have a more traditional view of education, they might believe that their role as instructor is to present information that students should store and remember. These teachers may engage in direct instructional practices and incorporate drill and practice activities. In contrast, if teachers have a more constructivist view of education, they might believe that their role is to guide discovery and model active learning. These teachers may develop a more student-centered curriculum that involves students working through authentic problems (Judson, 2006). It should be noted that both types of beliefs and practices may have positive results.
Teacher Beliefs About Intelligence
Teacher beliefs about the nature of students’ intelligence have also been found to predict student performance as well as to foster certain types of teaching practices. In essence, teacher beliefs about intelligence fall into two main categories: entity (i.e., fixed) mind-set and incremental (i.e., malleable) mind-set (Dweck, 1999; Hong, Chiu, Dweck, Lin, & Wan, 1999). Teachers who tend to have an entity mind-set believe that intelligence is a fixed human attribute, that one’s level of intelligence is predetermined by genetics, and that functional ability stays the same. In contrast, teachers who have an incremental mind-set believe that intelligence is a malleable human attribute, that individuals have some levels of control over their own intelligence or mental abilities, and that intelligence can be amplified through study and learning (Dweck, 1999; Hong et al., 1999).
From a student perspective, implicit theories of ability have been found to affect student learning, motivation, and achievement outcomes. When students hold an entity (or fixed) belief about intelligence, they are more likely to give up easily when faced with difficult or challenging situations (i.e., less persistent), and more likely to draw conclusions about their own ability from setbacks (Dweck, 1999, 2007). On the other hand, students with an incremental belief about intelligence are more likely to devote more effort to a learning task, to persist, and to try again in the face of failure (Hong et al., 1999).
Much of the work related to implicit theories of intelligence has focused on understanding the impact of students’ implicit beliefs, whereas less research has examined teachers’ mind-sets (Jones, Bryant, Snyder, & Malone, 2012). However, because teachers work closely with students during their academic learning, they have the potential to influence students’ implicit theories of intelligence (Jones et al., 2012). In fact, there is empirical evidence that teachers’ conceptions of intelligence affect their students’ beliefs about intelligence (Pretzlik, Olsson, Nabuco, & Cruz, 2003; Watanabe, 2006), which can in turn influence students’ motivation and achievement (Dweck, 1999). Teachers’ mind-sets have also been found to influence their teaching practices and their self-perceptions. Teachers’ endorsement of an incremental mind-set (as compared to a fixed mind-set) has been associated with higher levels of self-efficacy (i.e., teachers feeling like they can help students overcome difficulties in school) and greater likelihood of creating a classroom environment that supports students’ needs for autonomy, competence, and empathy and promotion of students’ intrinsic motivation (Leroy, Bressoux, Sarrazin, & Trouilloud, 2007).
Furthermore, teachers’ implicit theories of intelligence are related to the types of attributions teachers make about low student achievement, expectations about students, and the type of feedback and praise they deliver to students (Mueller & Dweck, 1998; Rattan, Good, & Dweck, 2012). This train of perceptions and practices can affect student motivation and achievement. For example, holding an entity theory of math intelligence is associated with making attributions to low ability (vs. effort) and not expecting appreciable future improvement in students (Rattan et al., 2012). Teachers’ holding a fixed mind-set are also more likely to praise students’ intelligence (vs. effort), a practice associated with negative student behaviors, such as lower levels of task persistence, and attitudes in which “being challenged” and “learning a lot” are rejected in favor of “seeming smart” (Mueller & Dweck, 1998). In experimental studies in which undergraduate students imagined themselves in the role of a teacher, those who held an entity theory were more likely to comfort students for their supposed low ability by engaging in pedagogical practices that could reduce engagement (e.g., consoling a student for a poor grade in math by telling him/her that plenty of people have trouble in this field but go on to be very successful in other fields), as compared to those who held more of an incremental theory where expectations were maintained (Rattan et al., 2012). Rattan et al. also found that students who were exposed to these types of comfort-oriented feedback practices were more likely to feel less motivated and had lower expectations about doing well in the class.
Teachers’ implicit theories of intelligence have been most commonly measured using self-report questionnaires (e.g., Dweck, 1999; Nature of Ability Beliefs Questionnaire: Leroy et al., 2007). One of the most popularly used measures is Dweck’s (1999) three-item scale in which individuals rate their level of agreement (1 = strongly disagree to 6 = strongly agree) with the following statements: You have a certain amount of intelligence, and you really can’t do much to change it; Your intelligence is something about you that you can’t change very much; and You can learn new things, but you can’t really change your basic intelligence. Evaluation of the psychometric properties of this scale has shown this scale to have strong factorial validity and reliability (e.g., Dweck, Chiu, & Hong, 1995; Levy, Stroessner, & Dweck, 1998). An alternative to this teacher report questionnaire is to use a student report measure such as the four-item Perceptions of an Environmental Entity Theory scale (e.g., My professor believes that I have a certain amount of math intelligence, and I can’t really do much to change it; C. Good, Rattan, & Dweck, 2012) to assess students’ perceptions of their teacher’s mind-set. Self-report measures have well-known and long-standing limitations, such as social desirability (see, e.g., Podsakoff & Organ, 1986; Constantine & Ladany, 2000). They also suffer from scoring approaches that average or total responses; however, improvements in their psychometric properties have been achieved using item response theory approaches (Fraley, Waller, & Brennan, 2000).
Teacher Efficacy
Teacher efficacy encompasses a teacher’s individual beliefs about his or her ability to carry out a course of action and influence student performance, engagement, and learning across various student groups, including students with learning challenges and motivational difficulties (Hoy & Spero, 2005; Tschannen-Moran & Hoy, 2001). Teacher efficacy is considered a key motivational belief influencing teachers’ professional behaviors and style, and student motivation and learning (Bandura, 1977, 1986; Hoy & Spero, 2005; Klassen, Tze, Betts, & Gordon, 2011; Tschannen-Moran & Hoy, 2001; Tschannen-Moran, Hoy, & Hoy, 1998).
With regard to teacher outcomes, teacher efficacy has been shown to positively affect teacher beliefs about teaching and their instructional behaviors (Skaalvik & Skaalvik, 2007). In particular, teachers with low self-efficacy experience greater difficulties in teaching, lower levels of job satisfaction, and higher levels of job-related stress (e.g., Betoret, 2006). However, teachers with high efficacy are willing to take on more challenges and risks, and tend to be more enthusiastic and committed to their teaching practice (Tschannen-Moran & Hoy, 2001). Teachers with a high sense of efficacy also experience more job-related satisfaction, expend more effort, set goals for their practice, and display more creativity in their instruction. Together these behaviors positively influence student achievement, motivation, and sense of self-efficacy (Klassen et al., 2011; Tschannen-Moran & Hoy, 2001).
Several measures of teacher efficacy have been developed and implemented over the past 50 years. These have roots in either Rotter’s social learning theory or Bandura’s social cognitive theory. Among the most prominent measures based on Rotter’s tradition include the 2-item Rand Measure of the 1970s, Guskey’s (1984) 30-item Responsibility for Student Achievement, Rose and Medway’s (1981) 28-item Teacher Locus of Control, and the Webb Efficacy Scale (Ashton, Oljenik, Crocker, & McAuliffe, 1982). Among the most prominent measures based on Bandura’s tradition and his construct of self-efficacy include the Ashton vignettes (Ashton, Buhr, & Crocker, 1984), Gibson and Dembo’s (1984) 30-item Teacher Efficacy Scale, Midgley, Feldlaufer, and Eccles’s (1989) 5-item personal teaching efficacy measure, and Bandura’s (1997) Teacher Self-Efficacy Scale. Measures have also been developed for specific domains and for multiple populations. Some of these include the Science Teaching Efficacy Belief Instrument (Enoch & Riggs, 1990), and Coladarci and Breton’s (1997) 30-item instrument for measuring efficacy in the context of special education. The commonality among most of these measures is that they are either forced-choice or self-report Likert-type scales based on scenarios, statements, or vignettes.
Despite this targeted effort, teacher efficacy research is troubled by divergent conceptualizations, measurement inconsistencies, and validity and reliability concerns and has an extensive research history to address these concerns (Tschannen-Moran et al., 1998; Tschannen-Moran & Hoy, 2001; Klassen et al., 2011). The main research trajectories related to validating measures of teacher efficacy have involved improving measurement of the construct and clarifying the two-factor structure that is revealed through factor analysis of most teacher efficacy measures.
Among the reasons offered to explain measurement difficulties are a lack of conceptual clarity in that a majority of measures focus on teachers’ beliefs about their control of student outcomes as opposed to their perceptions about a mediating construct, their capability to teach students. In fact, Klassen et al. (2011) reviewed 218 studies on teacher efficacy and found that approximately 50% included measures that were not consistent with the Bandurian concept of self- and collective efficacy and did not assess teachers’ confidence of their own capability to carry out a course of action.
Moreover, several measures are inconsistent with the tenets of self-efficacy as they focus on capability based on past performance or outcomes that are expected to result from their performance, instead of the capability to carry out a specific task (Klassen & Chiu, 2010; Tschannen-Moran et al., 1998). Bandura (1997, 2006) clearly distinguished between self-efficacy beliefs and outcome expectancies, with the former assessing judgment of capabilities for various types of performances and the latter measuring outcomes that are expected to result from these performances. Many of the teacher efficacy measures also reveal a two-factor structure, and there are confusion and debate about the meaning of and differentiation between these two factors (Tschannen-Moran & Hoy, 2001).
To address some of the issues in the measurement of teacher efficacy, Tschannen-Moran and Hoy (2001) developed a new scale, the Ohio State Teacher Efficacy Scale (OSTES) based on Bandura’s (1997) teacher self-efficacy scale. The instrument was tested in three separate studies and was reduced from 52 original items in the first study to a long form with 24 items and a short form with 12 items based on factor and reliability analyses. For Study 1, items loading from .62 to .78 were selected for further testing. For Study 2, alpha reliabilities for the three subscales (i.e., engagement, instruction, and management) ranged from .72 to .82. With respect to construct validity, total scores on the OSTES had small positive relationships with both Rand items (r = .30 and .28, p < .01), as well as the personal teaching efficacy factor of the Gibson and Dembo (1984) measure (r = .48, p < .01), and the general teacher efficacy factor (r = .30, p < .001). For the third study, the intercorrelations between the short and long forms for the total scale and the three subscales were high, ranging from .95 to .98. Similar to Study 2, the OSTES items were positively related to both the Rand (r = .18 and .53, p < .01) and Gibson and Dembo’s personal teaching efficacy (r = .64, p < .01) and general teacher efficacy (r = .16, p < .01) factors. Perhaps if more modern scoring approaches were used, findings would have been stronger.
Teacher Expectations
Following the classic Rosenthal and Jacobson (1968) study, a large body of experimental and correlational research has documented the effects of teachers’ expectations on student outcomes. Teacher expectations are referred to as perceptions teachers have about their students’ potential and what they are able to accomplish academically. There has been debate in the field about whether teachers’ expectations influence student outcomes, or whether differential outcomes indicate that teachers’ perceptions are accurate. Teachers typically draw these types of conclusions about their students early in the school year, forming opinions about their students’ strengths, weaknesses, and their potential for academic success (T. L. Good & Nichols, 2001; Ormrod, 2011). It is hypothesized that changes in student performance as a result of teachers’ expectations occur due to direct and indirect effects. Direct effects involve differential interactions with teachers that provide different opportunities to learn. Indirect effects involve social cues that communicate differential ability (Bohlmann & Weinstein, 2013).
As mentioned, teachers’ implicit beliefs about intelligence influence their expectations of students in that when they have an entity belief about intelligence, teachers make attributions to low ability, which makes them expect less future improvement from students (Rattan et al., 2012). Previous research has demonstrated that many teachers do in fact have entity views of intelligence (Dweck & Molden, 2005; Reyna, 2000). This entity attribution leads teachers to form fairly stable expectations for students’ performance, thereby influencing them to treat students differently (Ormrod, 2011). For instance, when teachers have high expectations for students, they tend to present more challenging tasks, interact with students more frequently, and give more positive and specific feedback. In contrast, when teachers have low expectations for students, they tend to provide easy tasks, offer few opportunities for speaking in class, and give minimal feedback (T. L. Good & Brophy, 1994; Graham, 1990; Rosenthal, 1991). These direct and indirect influences ultimately affect students’ self-expectations, motivation, and learning (Bohlmann & Weinstein, 2013).
Previous research has demonstrated that some teachers also underestimate the abilities of students who come from certain ethnic minority groups or low-income families (Tenenbaum & Ruck, 2007; Woolfolk-Hoy, Davis, & Pape, 2006). Specifically, McKown and Weinstein (2008) found that teacher expectations explained more of the year-end achievement gap between stereotyped and nonstereotyped groups in high-bias classrooms than in low-bias classrooms after controlling for prior achievement. Some recent research has also investigated teachers’ expectations on a class level, examining the differences between teachers who held high versus low expectations for all of their students, and found large effect size differences in expectancy outcomes between teachers who held high versus low expectations, and positive associations between high-expectancy teachers and students’ personal attributes (i.e., attitudes to schoolwork, relationships with others, and home support for school), and contrasting relationships with low-expectancy teachers and students’ characteristics (Rubie-Davis, 2006, 2010).
Contextual analyses of expectancy effects have largely focused on the degree of differential treatment by teachers present in classrooms. Most of these studies have measured students’ perceptions of their teacher’s interactions with high- versus low-achieving students in the 30-item Teacher Treatment Inventory (Weinstein, Marshall, Sharp, & Botkin, 1987), which has demonstrated adequate internal consistency (Weinstein et al., 1987). Some studies have also used adapted versions of St. George’s (1983) scale for teachers to rate students’ attributes and characteristics. Observational tools to determine differential treatment by teachers have also been used. These require extensive analysis of narrative records, and ultimately yield themes as opposed to more low-inference ratings (Bohlmann & Weinstein, 2013). An example of this approach includes the Classroom Ability-Based Practices scale, which has demonstrated interrater reliability with 86% agreement and a mean kappa of 0.71.
Much past research “prompted” teachers by presenting information and certain expectations for particular students. Students were randomly identified as low or high achievers, and their classification determined teachers’ subsequent behaviors and practices, as well as students’ perceptions thereof. More recent research has required teachers to rank students in order of their expected achievement and also asked teachers to rate them from 1 (poor) to 5 (outstanding) for another indication of the level of expectations they held for each students’ achievement (Bohlmann & Weinstein, 2013). Although research on teacher expectation has been performed in a variety of ways, a generally consistent finding across research endeavors is the predictive validity of teacher expectations as they adapt to different students (Bohlmann & Weinstein, 2013; Kuklinski & Weinstein, 2001).
Teacher Attributions
In general, an attribution is considered an explanation for a past event (i.e., an event that has already occurred). As related to the educational context, both students and teachers make attributions about the causes for academic successes and failures. Previous research has demonstrated that the attributions teachers make about their own and students’ performances influence their own as well as students’ emotional reactions and expectations for future performances, as well as their perceptions of self-efficacy. As such, attributions can serve as a mediating variable between teachers’ expectations and student motivation and performance (Zhou & Urhahne, 2013).
Teachers typically explain students’ academic success and failure by reference to three dimensions: locus (i.e., attributing the causes of an event to factors within themselves or factors outside themselves), stability (i.e., belief that the event was due to things that probably cannot change or things that can change from one time to the next), and controllability (i.e., attributing to events that they or someone else can influence and change, or things over which they have no influence). These three dimensions map onto five different causes that teachers and students use to explain student performance: ability/lack of ability, effort/lack of effort, task easiness/difficulty, and help/lack of help from teachers or parents (Brady & Woolfson, 2008; Graham & Williams, 2009; Ormrod, 2011; Weiner, 1985, 1986, 2004, 2005, 2010; Zhou & Urhahne, 2013). For instance, if a teacher attributes a students’ failure to stable factors such as the student lacking the ability to complete a task successfully, the teacher is also likely to have low expectations for the student and, therefore, may not encourage the student to expend more effort or experiment with academic strategies to succeed. If a teacher, on the other hand, attributes a student’s failure to lack of effort, he or she is likely to encourage the student to try by expending effort and implementing new strategies.
Current research has confirmed the mediating role of attributions in the relationship between teacher expectations and student performance, as well as the negative effects that teacher negative perceptions have on students’ own attributions. Zhou and Urhahne (2013) examined the impact of teacher judgment on students’ motivational patterns, and the mediating effect of attributions, and found that students whose teachers had underestimated their performance had maladaptive attribution patterns (i.e., attributed their performance to lack of ability—something internal to them that they cannot change and over which they have no influence).
Recent studies investigating teacher and student attributions and performance have measured teacher attributions using a 5-point Likert-type scale consistent with Weiner’s attributions of stability and controllability, with 5 signifying low stability (i.e., amenable to change; and low controllability), and adaptations of Clark’s (1997) vignettes and Woolfson, Grant, and Campbell’s (2007) Teacher Attribution Scale. Internal reliabilities for the Clark (1997) instruments have ranged from .84 to .86 across scales (Woolfson et al., 2007). Previous research has also noted that vignette-based measures, such as adaptations of Clark’s (1997) instruments, pose threats to external validity given the effort to provide simple scenarios that might not adequately depict the complexity of classroom dynamics (Brady & Woolfson, 2008). The Teacher Attribution Scale (Woolfson et al., 2007) measures teacher’s attributions of locus of causality, controllability, and stability regarding learners who are experiencing difficulties using vignettes. Reliabilities for the three scales (causality, controllability, and stability) have ranged from .86 to .91.
Teacher Goal Orientation
Researchers in recent years have explored the relationship between teachers’ goal orientations and associated perceptions, practices, and impacts on student perceptions and behavior. Butler (2007) proposed that achievement goal theory could provide a useful perspective for conceptualizing qualitative differences in teachers’ motives for teaching, as had been previously investigated with students’ motives for learning. Achievement goal theory assumes that students’ perceptions, strategies, and outcomes depend on their constructed goals for their schoolwork, and thereby on what they want to achieve (Butler, 2007; Butler & Shibaz, 2008; Retelsdorf, Butler, Streblow, & Schiefele, 2010).
A preponderance of research supports the fact that when students have a mastery goal orientation, they tend to have higher efficacy, be intrinsically motivated, attribute outcomes to effort, view setbacks as opportunities for improvement, and implement strategies that facilitate mastery of learning materials. In contrast, when students are performance oriented they tend to be focused toward outperforming other students, be oriented toward grades (i.e., extrinsically motivated), favor ability attributions, view difficulty as diagnostic of low ability, experience more academically related anxiety, and engage in strategies that facilitate rote memorization as opposed to meaningful learning (Butler, 2000; Elliot, 2005; Molden & Dweck, 2000).
As teacher goal orientation is a relatively new addition to the field, there are very few established measures for the construct. Also, a vast body of research regarding the impact on teacher and student outcomes does not exist. However, similar to research examining student goal orientations, research examining teachers’ goal orientation has demonstrated that teacher mastery goals are associated with higher levels of perceived teacher support and positive perceptions of instructional practices, whereas ability avoidance goals are associated with student reports of negative instructional practices and student cheating (Butler & Shibaz, 2008). With respect to teacher outcomes, Retelsdorf et al. (2010) found that teacher mastery orientation and work avoidance emerged as positive and negative predictors, respectively, of adaptive patterns of instruction and high interest in teaching and low burnout.
With respect to measurement, Midgley et al. (2000) developed a self-report measure for teachers’ mastery- and performance-oriented practices. This measure has been questioned for not including all of the facets of the goal orientation construct. Butler (2007) developed the Goal Orientation for Teaching Scale where teachers rate their agreement with statements such as “I would feel that I had a successful day in school if something that happened in class made me want to learn more about teaching.” Internal consistencies for this measure have ranged from .71 to .82 (Butler, 2007; Butler & Shibaz, 2008; Retelsdorf et al., 2010). Recent research has also examined and affirmed the structural validity of teacher goal orientation into mastery and performance goal orientations (Butler, 2007; Malmberg, Wanner, Nordmyr, & Little, 2004).
Teacher Intrapersonal Skills
Teachers’ intrapersonal skills are the personal behaviors and internal thought processes that can be systematically acquired or enhanced by instruction or learning experiences and where evidence of change can be directly or indirectly inferred (E. L. Baker, 2014). Teachers’ intrapersonal competencies fall within two categories: those related to emotional capacities and those related to metacognitive capacities and self-regulation.
Emotional Capacities
There is a substantial and growing body of research on the importance of fostering students’ social and emotional competencies because of its consequences for academic success and impact on lifelong learning (Zins, Bloodworth, Weissberg, & Walberg, 2007). However, less attention has been paid to understanding and supporting teachers’ own social and emotional skills. This is surprising given that teachers’ emotional processes (e.g., emotion regulation, ability to take the perspective and empathize with others from diverse backgrounds) and social/interpersonal skills (e.g., interacting positively with students, parents, and colleagues) are likely to influence their students’ social, emotional, and academic well-being, as well as students’ development of interpersonal competencies, which are a key component of 21st-century skills and the deeper learning competencies (Jennings & Greenberg, 2009).
Jennings and Greenberg (2009) proposed a prosocial classroom mediational model, which essentially highlights the importance of teachers’ social and emotional competence and well-being in supporting teacher-student relationships, effective classroom management, successful social and emotional learning program implementation, and preventing teacher burnout (Appendix B). Essentially, the model explains how deficits in teacher social and emotional competence and well-being provoke what the authors refer to as a “burnout cascade,” which contributes to a deteriorating classroom climate with increased troublesome student behaviors. Teachers drain emotionally in attempting to manage these types of classrooms, resorting to reactive and punitive responses that in turn do not teach self-regulation and may contribute to a self-sustaining cycle of classroom disruption (Jennings & Greenberg, 2009). In contrast, teachers who recognize students’ emotions and the underlying causes of students’ behaviors may also show greater concern and empathy for students and be better able to help students learn to self-regulate, thereby contributing to a positive classroom climate.
One aspect of teachers’ emotional processes that has received some attention is the construct of emotional intelligence (EI). EI refers to one’s ability to identify, process, and regulate emotions (Mayer, Salovey, Caruso, & Sitarenios, 2001), and is positively associated with teacher efficacy (Penrose, Perry, & Ball, 2007) and lower rates of teacher burnout (Chan, 2006).
There has been some debate over appropriate measures for assessing EI. One of the main issues is the difference between ability (i.e., performance) versus self-report measures. Brackett and Mayer (2003) examined the convergent, discriminant, and incremental validities of one widely used EI ability test measure (i.e., Mayer-Salovey-Caruso Emotional Intelligence Test [MSCEIT]) and two self-report EI measures (i.e., Emotional Quotient Inventory [EQ-I] and Self-Report EI Test [SREIT]).
The MSCEIT (Mayer, Salovey, & Caruso, 2002) has right or wrong answers based on consensus or expert scoring (Brackett & Mayer, 2003). For example, emotion perception is assessed by having individuals rate how much of a particular emotion is being expressed in faces, and emotion management is measured by asking people to select effective ways to manage emotions in hypothetical situations. In contrast, the EQ-I (Bar-On, 1997) and SREIT (Schutte et al., 1998) combine mental abilities (e.g., ability to perceive emotion) and self-reported characteristics such as optimism, and are thus considered to be “mixed models” (Brackett & Mayer, 2003, p. 1147).
In light of these differences, Brackett and Mayer (2003) found that whereas the EQ-I and SREIT self-report measures were moderately related, the MSCEIT ability measure was weakly related to the self-report measures. Furthermore, the MSCEIT proved to be distinct from the Big Five Personality Inventory. There was, however, considerable shared variance between the self-report measures and the Big Five Personality Inventory, indicating that the EI ability and self-report measures appear to assess different skills and qualities, producing different measurements for the same individual. Recent work has focused on teaching teachers and students to recognize emotion in video games and other computer environments (Griffin & Madni, 2013; Madni, Griffin, & Delacruz, 2013).
Metacognition and Self-Regulation
One prevalent model that explains metacognitive and self-regulatory processes is Zimmerman’s (2008) model of self-regulation (Appendix C). This model posits that the self-regulation process contains three separate phases with corresponding cognitive processes, actions, and motivational beliefs. The first phase is the forethought phase where individuals set goals and create strategic plans for goal attainment. This phase is influenced by motivational factors such as self-efficacy, outcome expectations, task interest/value, and goal orientation. For instance, an individual is likely to set more challenging goals and be more creative with the strategies for goal attainment when he or she has higher efficacy associated with the goal being set. The second phase of Zimmerman’s (2008) self-regulation process is the performance phase.
The performance phase involves individuals exhibiting self-control and self-observation as they are working toward attaining their goal. Self-control involves cognitive processes and actions such as self-instruction, engaging in learning strategies such as imagery, attention focusing, and selecting and implementing appropriate task strategies to complete target tasks. The last phase of Zimmerman’s (2008) model is the self-reflection phase, which involves self-judgment and self-reaction. Self-judgment involves individuals self-evaluating their process toward goal attainment and the outcome, as well as making attributions (i.e., coming up with explanations) for goal attainment or lack thereof. For instance, if an individual attributes his or her lack of goal attainment to insufficient task strategies, then he or she is likely to adjust the strategic plan for goal attainment. Self-reaction involves how individuals feel about their outcome and process and whether they have an adaptive as opposed to a defensive reaction.
Social and Interpersonal Skills
Social and interpersonal skills require cognition and metacognition; however, they are applied in interpersonal situations, such as when teachers are collaborating with other teachers, when teachers are working and communicating with students, and last, when teachers are communicating and working with parents. These skills and processes do not require an extraverted personality but instead require concrete abilities to clarify goals, modify behavior, and adapt to situations and people (E. L. Baker, 2014).
Collaboration and Teamwork
Teachers’ interpersonal competencies associated with student academic achievement depend in part on teachers’ ability to collaborate effectively and successfully with each other. Teacher collaboration is thought to provide opportunities for teacher learning in which individuals engage in professional discourse and further develop their content, pedagogical, and experiential knowledge to improve instruction (Y. L. Goddard, Goddard, & Tschannen-Moran, 2007), thereby enhancing student outcomes.
Teacher collaboration has been studied and assessed using qualitative case study methods (e.g., Levine & Marcus, 2007), survey methods (Y. L. Goddard et al., 2007), and social network analysis (Moolenaar, 2012). Y. L. Goddard et al. (2007) conducted a large-scale naturalistic study using survey methods to test the relationship between teacher collaboration for school improvement and student achievement. They used a 6-point Likert-type scale that asked teachers to rate on a scale of 1 (not at all) to 6 (very much) “To what extent do teachers work collectively to influence these types of decisions?—planning school improvement, selecting instructional methods and activities, evaluating curriculum and programs, determining professional development needs and goals, and planning professional development activities.” All items loaded onto a single factor using principal axis factor analysis and showed strong internal consistency. The authors conceptualized the level of teacher collaboration as an important aspect of a school’s normative and behavioral environment. Therefore, individual teacher collaboration data were then aggregated to the school level. Results showed that even after controlling for children’s academic and social background characteristics and school context, fourth-grade students in a large urban school district had higher achievement in mathematics and reading when they attended schools characterized by higher levels of teacher collaboration for school improvement. In evaluation of schools, models involving social capital have been used, and a prominent part of these is teachers’ perception of their collaborative effort (Huang et al., 2007; Lee, 2010).
A social network analysis approach focuses on the pattern of social relationships among teachers, captures the multilevel nature of teacher collaboration (e.g., teachers in schools), and provides a visual representation of teacher interactions (Moolenaar, 2012). Social network analysts typically examine issues of centrality (e.g., total number of relationships a teacher maintains) and ego reciprocity (to capture the two-way nature of relationships; Moolenaar, 2012). Examples of the types of questions asked of respondents when conducting social network analysis have included the following: “During this school year, to whom in your school have you turned for advice on strategies to assist low-performing students ?” (Cole & Weiss, 2009), “To whom have you turned to for advice or information about math teaching strategies and content?” (Pitts & Spillane, 2009), “Whom do you go to for work-related advice?” and “Whom do you go to for guidance on more personal matters?” (Moolenaar, Sleegers, & Daly, 2012). Moolenaar et al. (2012) found that well-connected teacher networks were associated with strong teacher collective efficacy, which in turn supported student achievement. Thus, there is some evidence that the construct of collective efficacy may help explain the relationship between teacher collaboration and student achievement outcomes.
Collective Efficacy
Collective efficacy refers to teachers’ beliefs at a school that the entire faculty as a whole is capable of organizing and executing the necessary actions to have a positive impact on its students (R. D. Goddard, Hoy, & Hoy, 2004). Collective efficacy depends on the interaction of perceived competence to perform a given task and the context in which the task will take place. With regard to teacher collective efficacy, these two components have been conceptualized in terms of group competence and task analysis (R. Goddard, 2002). Group competence refers to the capabilities the faculty brings to a teaching situation (e.g., teaching methods, skills, training, expertise), whereas task analysis refers to perceptions of constraints and opportunities due to the nature of the task (e.g., level of support provided by students’ home and the community; R. Goddard, 2002). A strong sense of collective teacher efficacy enhances individual teacher’s sense of efficacy and has been linked to student academic achievement (R. D. Goddard et al., 2004; Tschannen-Moran & Barr, 2004). Teachers’ collective efficacy beliefs are an important construct of interest because it can be affected by educational leadership efforts to foster strong teacher relationships (Moolenaar, Daly, & Sleegers, 2011).
There are various approaches to measuring collective efficacy. One approach is to aggregate individual teacher’s self-efficacy beliefs (e.g., “I have what it takes to get my students to learn”) to develop a group mean of self-efficacy perceptions at a school. Another approach is to ask teachers to discuss questions as a group and arrive at a group consensus of how they view their collective efficacy. However, this method may suffer from social desirability issues affecting the validity of the assessment and may also hide the variability that exists within a group (Bandura, 1997).
A method that is more widely endorsed by researchers (R. D. Goddard et al., 2004) is to aggregate measures of individual teacher’s perceptions of the group’s capability by responding to statements such as “Teachers in this school have what it takes to educate students here” (p. 6). This approach aligns more closely with Bandura’s (1997) stance that the collective efficacy is an “emergent organizational property” (R. D. Goddard et al., 2004, p. 7) that reflects the perceptions of teachers about a school’s conjoint capability to successfully influence students (R. Goddard, 2002). Thus, the assessment of collective efficacy necessitates the combination of individual-level perceptual measures, which can be achieved through group-level (i.e., school-level) aggregates. Furthermore, R. D. Goddard (2001) showed that within-school variabilities among teachers’ reports of collective efficacy were not predictive of student achievement differences between schools but rather that the use of a central tendency measure (i.e., school-level aggregates of teacher collective efficacy) was a better predictor. R. D. Goddard, Hoy, and Hoy (2000) developed a 21-item Collective Efficacy Scale, which R. Goddard (2002) later modified to develop a short form using only 12 of the original items. This abbreviated measure is more theoretically pure and parsimonious and is highly correlated with the original longer scale (r = .983). Furthermore, the short form has high internal consistency based on Cronbach’s alpha and demonstrated predictive validity using multilevel modeling (R. Goddard, 2002). Modern practical models of school and other institutional evaluation use collective effort, efficacy, and transparency and trust as key attributes to be studied (Cai, Baker, Choi, & Buschang, 2014; Huang et al., 2007).
Collaborating With Students
High quality teacher-child relationships provide students with a supportive and emotionally secure environment. Students are better able to regulate their emotions, interact with others, and focus on academics (Pianta, 1999). Furthermore, teacher-student relationships are associated with school satisfaction, engagement, and academic and behavioral outcomes (J. A. Baker, 2006, Elias & Haynes, 2008; Hamre & Pianta, 2001; Maldonado-Carreño & Votruba-Drzal, 2011; Murray, 2009).
Methods to assess teacher-student relationships are often tied to a particular perspective (e.g., attachment, motivation, sociocultural; Davis, 2003). For example, researchers from a sociocultural perspective view relationships as “dynamic, changing, and culturally bound” (p. 225) and tend to employ qualitative methodologies such as case studies and ethnography, conducting in-depth interviews and holistic observations (Davis, 2003). In contrast, those who work within attachment theory use observational methods, checklists, and teacher-rating scales to evaluate the quality of relationships according to theoretically defined dimensions and social developmental outcomes. Clearly choice of method depends substantially on the age of the student.
Furthermore, motivation researchers are inclined to use student report for older students (e.g., Network of Relationships Inventory; Furman & Buhrmester, 1985) and teacher self-report instruments (e.g., Student-Teacher Relationship Scale; Pianta & Steinberg, 1992) to evaluate relationships based on their own set of dimensions believed to shape cognitive and academic outcomes because they view relationships as embedded in complex classroom contexts (Davis, 2003).
Collaborating With Parents
Although there is a considerable body of work on teacher-student relationships, less is known about teachers’ relationships with parents. Researchers have studied the benefits of parental involvement and engagement for student educational outcomes (Fan & Chen, 2001; Minke, Sheridan, Kim, Ryoo, & Koziol, 2014; Murray, 2009), yet less is known about affective quality of parent-teacher interactions. Some qualitative studies (Angelides, Theophanous, & Leigh, 2006; Bruckman & Blanton, 2003; Miretzky, 2004) and case studies (Billman, Geddes, & Hedges, 2005) assess teachers’ relationships with parents through observations, interviews, and focus groups. Such studies often highlight the challenges of developing positive relationships and true collaboration between teachers and parents, perhaps because they involve complex status relationships. Some research has implemented quantitative analyses of teacher self-report and parent self-report measures (Hughes & Kwok, 2007; Nzinga-Johnson, Baker, & Aupperlee, 2009). In studies by Herman and Baker (2003), structural models were used to document that synchronicity of teacher and parent goals influenced student performance on standardized and other content measures, although results were mediated by other factors such as attendance (Herman & Baker, 2003; Quigley, 2000).
Minke et al. (2014) findings substantiated that congruence between teachers and parents was an important factor in understanding reports of child behavior. With a more extensive set of measures, agreement on ways to support and influence positive student outcomes had effects (Minke et al., 2014). As part of this study, teachers and parents completed the Parent-Teacher Relationship Scale (Vickers & Minke, 1995), which contains two subscales and is rated on a 5-point Likert-type scale. The Joining subscale assesses the interpersonal connection between parents and teachers (e.g., “We understand each other”). The Communication-to-Other subscale evaluates the communication quality between parents and teachers (e.g., “I tell this parent/teacher when I am pleased”). The internal consistency reliabilities of this scale range from .93 to .95. These efforts to assess reliability are important because of the reputation of “softer” constructs such as collaboration that have a wide variety of definitions and hence interpretations.
Similarly, Murray (2009) examined the relationship between teacher-parent relationships and urban youth engagement and functioning. The researchers implemented the Research Assessment Package for Schools (RAPS; Connell & Wellborn, 1991), and adapted it for both parents and teachers. The RAPS is an 84-item Likert-type measure that can be used to assess student school engagement, student self-beliefs, and student perceptions of interpersonal support. Example items include “My parents are fair with me,” “My teachers think what I say is important,” and “My teachers don’t seem to have enough time for me.” The alpha reliabilities for the RAPS ranged from .65 to .84. No validity data were reported.
Taken together, the preceding findings on teacher-parent relationships suggest that these relationships predict parental involvement (Nzinga-Johnson et al., 2009) and children’s engagement, motivation, functioning, and performance in school (Hughes, Luo, Kwok, & Loyd, 2008; Minke et al., 2014; Murray, 2009).
Diversity and Acceptance
Underlying teachers’ ability to collaborate effectively with each other, parents, and students are their beliefs and perceptions related to diversity and acceptance. These beliefs are also of particular importance to English language learners, their teachers, and their parents. Teachers’ beliefs and ability to engage in perspective taking, especially when working with diverse students and/or students from different backgrounds, are imperative given their association with how teachers interact and respond to students (Pajares, 1992). Children who are at risk of school failure (e.g., low socioeconomic status) are most affected by the quality of their relationships with teachers (Hamre & Pianta, 2001); yet marginalized students and their parents may also experience less supportive relationships with teachers (e.g., Hughes & Kwok, 2007). Part of this relationship may depend on parents’ experiences with schools as students or their beliefs about institutional authority.
Teachers, however, bear strong responsibility to make these relationships work. Several existing measures in use to assess teachers’ beliefs about diversity lack technical information about reliability and validity (see Brown, 2004, for review; Pohan & Aguilar, 2001). However, Brown (2004) discusses two measures that are psychometrically sound. One is the Cultural and Educational Issues Survey (CEIS; Pettus & Allain, 1999), which assesses how individuals and groups feel about social and cultural issues that have consequences for how teachers make decisions about educational services in society. The CEIS contains four demographic items and 59 opinion statements graded on a 5-point Likert-type scale. Although the scale is considered a viable instrument, it has not been widely implemented or tested.
The CEIS contains both positively and negatively worded statements. Example positively worded items include “Teachers should draw on students’ experiences, cultures, and languages to make learning relevant and interesting for the students” and “A teacher should openly express dissatisfaction with a colleague who makes disparaging comments about a student’s sexual orientation.” Examples of negatively worded statements include “Compared to other problems, sexism is not a significant problem in the schools in the United States,” and “Generally, different racial groups have different abilities to learn different school subjects and activities.”
The other measure of presumed quality is the Personal and Professional Beliefs About Diversity Scale (Pohan & Aguilar, 2001), which addresses a spectrum of diversity issues including race and ethnicity, social class, gender, religion, language, and sexual orientation. This scale was specifically designed to assess varying levels of openness to a wide range of diversity topics and issues. It is a two-part open-ended survey and therefore provides the opportunity for assessment to vary widely (see Pohan & Aguilar, 2001). Example items include “There is nothing wrong with people from different racial backgrounds having/raising children,” “Making all public facilities accessible to the disabled is simply too costly,” “Teachers should not be expected to adjust their preferred mode of instruction to accommodate the needs of all students,” and “Students and teachers would benefit from having a basic understanding of different (diverse) religions.”
Psychometric Properties of Measures
In discussing measurement of teachers’ social psychological constructs, it is essential to also address the importance of technical quality as it pertains to validity. Validity has been reconceptualized over the years to focus not on the quantitative properties of the measure but on the quality of the inferences that can be drawn from results of the measures in light of their purposes. (Cai et al., 2014; Messick, 1995; O’Neil et al., 2014). This definition broadens the application of validity to measures that are not strictly psychometric in nature. The use of construct validity and quality of evidence and inference related to the purposes for which the results will be used permits the integration of alternative methods into the validity paradigm, heretofore seen as the province of quantitatively oriented analyses. Even so, there remain a number of challenges for the measurement of social, psychological, and emotional factors. For research purposes, there is always a need to vet new or replicate prior measures to determine their consistency and validity. These approaches can include refining and establishing validity of scoring approaches used for observation; the transformations of raw data into item response theory or other scales, for more precise interpretation; and clarification of contextual variables likely to influence process and outcome measures. For example, teachers’ motivation or attribution behavior may be observed in a short period of time, whereas the student outcome measure, particularly of an end-of-year standardized test, is unlikely to be sensitive to relatively short or variable interventions. Second, triangulating on various instruments (current validity in the 20th century) requires careful vetting of each of the measures assumed to corroborate validity. As in practical use, research studies need to relate the inferences they draw about constructs to the purposes they explicitly describe in their hypotheses. Is the interest in the change of self-perception over time? Is the interest in the ability to translate their attributions into positive support for students? These two different questions would depend on different data to obtain valid inferences, even if the same measures were used. In reviewing the literature it seems clear that the emphasis is on reliability, such as alpha coefficients and between-rater kappa, rather than extending inferences to the purposes for which the constructs are intended to be used—in other words, their validity.
The problems of reliability come into play in the use of such measures to make decisions about people. Minimally, one would want some consistency over a short time period, as measuring traits, or “types,” of personality or temperament would unlikely meet standards associated with employment law. When teacher processes are judged, such as their adaptive behavior based on student attributions, it would be important to have good measures of alternative evidence of such adaptation as well as evidence that the adaptations work as intended—that is, aligned with their purposes. Our review suggests that far more evidence is required for all elements in a validity argument, including measures of social or emotional perspective, the enacted teaching behaviors, students’ process responses, and the relevance and quality of outcome measures. To date, very few investigators have taken on this stream of evidence. Our suggestion rides not on a desire for comprehensiveness but on the imperative to marshal strong arguments if the measures are intended for high-stakes use. High stakes in this regard means not only typical accountability sanctions but the consequences to the respondent of the data and decision.
In the past 10 years, a new line of research has been undertaken by researchers intending to use these and other psychological constructs for support, advice, or therapeutic recommendations. Technology, as it has expanded, has an array of tools that can be combined to draw inferences about psychological states. These approaches may involve the use of cameras, gestural sensors, electroencephalography caps to read simple neurological states, and inferences from motion patterns during an extended period. In addition, there are voice analyses used to infer anxiety or disturbance from pitch, pauses, and repetitive word choice. Some of these methods are used with everyday activities, such as the use of smart phones, whereas others rely on structured interactions and where respondent answers are analyzed for voice and text features (Rizzo et al., 2013; Stripling, 2012; Stripling, Lee, & Cohn, 2014). Whether such approaches will become, or are already, routine by web providers raises the questions of informed consent, confidentiality, and privacy. All three of these issues are highly pertinent to affective behaviors, especially when inferences will be made about individuals. Nonetheless, the benefit of these approaches is that computers, well debugged and programmed, are reliable collectors of data. The focus then shifts to validity, in particular, the evidence as collected, its use, and the inferences drawn from it. This area will be a strong domain for research, particularly in the light of wearable computing, sending signals, whether desired or not, to developers and instigating web sites.
Conclusions and Implications
In this chapter, we reviewed research on teachers’ social and motivational behavior. Our interest was in drawing research attention to those factors that (1) can be shaped by appropriate interventions, (2) influence teacher instructional practices, and (3) predict student motivational and performance outcomes needed in the 21st century. Our research based on our review clearly suggests two overarching categories under which the research falls: teacher motivational constructs, and teacher intrapersonal and interpersonal competencies. We have also discussed limitations of measures throughout and in a final segment of the chapter.
Discussion Related to Measures of Teacher Social Psychological Constructs
Within the area of teacher social psychological constructs, it is apparent that a majority of the research data flow from variations of self-report questionnaires, including Likert-type scales. In some cases self-reports involve parents and student report measures as well. These may be based on stimulus statements, scenarios, or vignettes related to teacher attributions; teacher efficacy; collaboration; support; and teacher expectations. Self-report teacher measures of teacher expectations have also been measured through student reports about teacher treatment, through observational tools, by “prompting” teachers with preestablished expectations, or having teachers rank students in order of expected achievement. Further refinement of the construct of teacher expectations would be useful to create more replicable and consistent measurements of the construct.
Our chapter also revealed mixed findings across constructs with respect to factorial and construct validities, and reliability. The measures of teachers’ implicit theories of intelligence and teacher attributions have demonstrated appropriate psychometric properties. After a “troublesome” measurement history, it appears that a measure of teacher efficacy with stronger factorial and construct validities, and reliability has emerged (i.e., the OSTES; Tschannen-Moran & Hoy, 2001). The OSTES is nascent in the field and, therefore, needs further testing and refinement. Emerging from the field more recently is the construct of teacher goal orientation (Butler, 2007). This construct is demonstrating promise with respect to psychometric properties and influencing both teacher practice and student outcomes. However, the Midgley et al. (2000) measure has been critiqued for not including all facets of the construct. Further exploration and refinement of this construct would be a positive addition to the field of teacher motivation.
The teacher motivation constructs reviewed as part of this study also demonstrated predictive effects on salient and valued teacher perceptions and practices as well as student motivation and performance. This finding indicates that continued focus on exploring and enhancing measurement of these constructs, as part of early teacher education programs as well as later teacher training opportunities and program development, is warranted. Further refinement of these constructs would also aid in differentiating the constructs studied from each other, and in enhancing already-established models and producing new models of their interrelationships.
Technical Quality Summary and Recommendations
This chapter also determined that information related to reliability and validity information across teacher motivation constructs and measures is not consistently available in the literature. Thus, it is imperative that validation type studies are performed using measures in the anticipated structural model linking internal states to behaviors and then to desired student processes and outcomes
Measurement Related to Intrapersonal and Interpersonal Competencies
In contrast to the intrapersonal competencies, it appears that interpersonal teacher competencies have more variation with respect to different forms of measurement and research methodologies including self-report survey methods, case studies, ethnographies, in-depth interviews, holistic observations, and social network analysis. In addition, intrapersonal competencies demonstrate mixed findings with respect to the psychometric properties of certain measures. The measures of EI appear to assess different skills and qualities, producing inconsistent results. Measures for collective efficacy, on the other hand, have demonstrated high correlations across scales and high internal consistency and predictive validity. Collective efficacy may also serve as a moderating variable to explain the influence of teacher collaboration on student achievement outcomes. Teacher collaboration has been measured using survey methods and social network analysis with relevant measures producing internally consistent results, and has been found to influence higher math and reading achievement in students. Further using social network analysis to characterize what specific aspects of teacher collaboration result in positive student outcomes would provide a significant contribution to research within this area.
As evidenced by the current review, methods implemented to assess teacher-child relationships tend to be tied to a particular learning or psychological perspective and, therefore, span various qualitative and quantitative methods including ethnographies, in-depth interviews, observations, checklists, teacher rating scales, and student and teacher self-report measures. In contrast, there is minimal research and measurement history about teacher’s relationships with parents. However, findings in the area indicate that teacher-parent relationships are key to parent involvement. Given these sets of results, in future studies refining measures are needed to further determine the specific facets of teacher-parent relationships that contribute to positive student outcomes. In contrast to the research on teacher-parent relationships, research on teachers’ beliefs about diversity has used several measures, but some reflect only minimal reliability and validity information. As such, it would be pertinent for future scholars to perform studies evaluating and validating these measures. Such studies will be increasingly pertinent in the practical world as the demography of teachers, parents, and students changes.
Further recommendations for future studies include combining various forms of measurement to produce more robust results, thereby further solidifying constructs. Performance type assessments to determine teacher’s actual level and skill with respect to relevant constructs, such as collaboration and interpersonal engagement, could aid in refining the aspects of these processes that are salient in key predictive relationships. Moreover, focusing on defining and capturing behavioral indicators of instructional practices associated with internal non-cognitive factors, such as teacher efficacy, teacher attributions, and teacher expectations, can contribute to development of new measurements for robust triangulation, assuming concurrent measures also have strong evidence. Computational modeling of such behavioral indicators can also aid in composing and refining these types of constructs.
Limitations
In conclusion, we note that certain limitations in our study influence the findings presented. First, given the span of constructs, the chapter is not exhaustive relative to each construct. This is a practical constraint. Nevertheless, we strongly believe that our careful sampling of articles for this chapter adequately portrays the state of the field. Future reviewers are encouraged to consider searching non-English journals and graduate theses and dissertations to uncover novel patterns, themes, or new avenues within the field. Since our focus was on the measurement of key noncognitive teacher factors, a detailed exploration of the strengths of relationships among the noncognitive factors and a variety of teacher and student outcomes was beyond the scope of our chapter. Consequently, we chose to focus primarily on outcomes related to teacher practice and pedagogy, and student motivation and achievement.