Who's in charge? The history and future of teacher learning orientation measures

Introduction

For over a century, researchers have attempted to capture a deceptively simple question at the heart of educational practice: who drives learning, the teacher or the student? These two instructional/learning orientations (or LOs) are referred to as teacher-oriented and student-oriented. The teacher-oriented perspective emphasizes the teacher's authority, structured curriculum, and efficient transmission of knowledge, whereas the student-oriented perspective prioritizes student autonomy, exploration, and the active construction of understanding (Bruner, 1961; Kirschner et al., 2006; Loveless, 2001; Rosenshine, 2012). These orientations continue to influence how teachers, researchers, and policymakers conceptualize effective teaching; the implied dichotomy (student or teacher-oriented) has not been resolved. Understanding how these two basic orientations have been measured and defined over time is essential as we continue to work towards a more optimal model for education (de Jong et al., 2023; Geary and Xu, 2022; Hirsh-Pasek et al., 2020).

In the current review, we examine more than 70 years of LO-related measures tapping into these two essential views, present how the domain of LO has been conceptualized over time, and group these changing conceptualizations into five theoretical perspectives. The theoretical perspectives roughly correspond to contemporary education culture as it existed when each measure was created: progressive, teacher control and student autonomy, developmental, constructivist, and teacher-/student-centered or -oriented. We focus on Likert scale measures due to their prevalence in this domain. The review's organization illustrates how conceptions of LO evolved and where domains overlap, highlighting both the persistence of the student-/teacher-oriented learning divide and efforts to bridge it. We conducted a narrative rather than a systematic review, consistent with our aims of conceptual synthesis and clarification of how the constructs have been defined and measured over time.

Some measures conceptualize these two essential LOs as opposite ends of a continuum (Deci et al., 1981; Willower et al., 1967), whereas others treat them as orthogonal, or independent, constructs that can coexist to varying degrees within an educator's belief system (Bunting, 1985; Kerlinger and Kaya, 1959a). Clarifying these distinctions is essential for interpreting findings across studies and for refining theoretical models of teacher cognition. Examining the perspectives reflected in LO measures elucidates how the LOs have been defined across educational movements and psychological theories, providing a theoretical foundation for future measure development and interpretation.

Three questions guide this review: (1) How have teachers’ learning orientations been conceptualized and operationalized in Likert scale measures over the past 75 years? (2) What theoretical perspectives organize these measures, and where do their construct domains overlap? (3) How have measures treated the relation between teacher- and student-oriented beliefs—as opposite ends of a single continuum, or as orthogonal, potentially coexisting constructs? In what follows, we first define the two LOs and review 15 existing LO measures organized by theoretical perspective. We end with a more refined LO definition and recommendations for future measure development.

Defining learning orientation

Since the first teacher belief measures of the 1950s, researchers have examined a key dichotomy in beliefs about teaching and learning, a dichotomy that essentially reflects teacher-oriented vs. student-oriented perspectives. This dichotomy has been described in various ways: traditional vs. progressive, developmentally inappropriate vs. appropriate, transmissive vs. constructivist, adult- vs. child-centered, and teacher- vs. student-centered beliefs. Given this plethora of terminology, we begin with clarification of the terms.

Aligning with the locus of control literature, we designate our main construct with the term orientation. Regarding attributions, Rotter (1966) described a continuum of external to internal orientations. People with an external orientation attribute outcomes to outside forces such as chance or other people; those with an internal orientation view outcomes as determined by their own effort and choices. We apply this distinction to the locus of agency within the learning process. In this interpretation, student-oriented beliefs locate that agency within the learner (internal orientation), where the student constructs their own understanding; teacher-oriented beliefs locate it outside the learner (external orientation), where the teacher transmits knowledge (Findley and Cooper, 1983). This theoretical parallel highlights how LOs can be conceptualized as reflecting teachers’ implicit assumptions about agency in the learning process. We use the term learning to emphasize teachers’ and students’ relative positions in the learning process. Accordingly, this narrative review examines measures assessing teachers’ dichotomous learning orientations (LOs) as teacher- vs. student-oriented.

Learning orientation measures

Search and selection procedure

Consistent with a narrative review, we used an iterative search procedure. We first identified peer-reviewed empirical articles assessing learning orientations among preK–12th-grade pre- and in-service teachers. Initial searches were conducted in Google Scholar using the terms “student-centered teacher beliefs”, “learner-centered teacher beliefs”, and “child-centered teacher beliefs”. We then examined the reference lists of retrieved articles to identify additional measures of the key constructs, and searched measure and author names in Google Scholar to locate the development and validity studies for each candidate measure. The procedure was iterative in that new orientation terms and measures surfaced during the review and were incorporated into subsequent searches. We reverified the search in June 2026 with Google Scholar, and on July 31, 2026, with ERIC, Scopus, and Web of Science. After addressing the inclusion criteria (described next), these searches recovered the same set of 15 instruments. Figure 1 summarizes the search and selection process.

Inclusion criteria

To be retained, a measure had to (1) use a Likert scale response format in which teachers evaluated statements or vignettes; (2) apply to teaching generally rather than to a specific subject; (3) yield scores for exactly two learning orientations; (4) have undergone some formal scale-development process for psychometric health; and (5) have been used in at least one peer-reviewed article by someone other than the scale's author. Although teacher-belief scales specific to particular subjects exist, we deliberately focused on general subject matter to remain inclusive of as many pedagogical methods as possible. In total, 15 measures met all five criteria. These measures were organized into domains based on LO construct labels and, in the case of the Minnesota Teacher Attitude Inventory, publication date (Cook et al., 1951).

In the following sections, we review each selected measure according to domain (see Table 1). Based on the earliest LO measures developed toward the end of the United States’ progressive education movement (i.e., late 1800s to mid-1900s; Cremin, 1961), the first domain assessed teachers’ progressive and traditional beliefs (Cook et al., 1951; Kerlinger and Kaya, 1959a; Schaefer and Edgerton, 1985). The second domain, based on three psychological frameworks that successively dominated psychological theory across the 1900s, assessed beliefs about teacher control and student autonomy: control ideology (Willower et al., 1967), locus of control (Wehling and Charters, 1969), and self-determination theory (Deci et al., 1981).

Table 1. Learning orientation measures by year published.

Item counts refer to the items composing the learning-orientation factors reported here, and may not sum to the instrument's full length.

The third domain used centered and oriented terms (e.g., teacher-centered, student-oriented) to describe LO (Bunting, 1985; de Vries et al., 2013; McCombs and Whisler, 1997; Meirink et al., 2009). This domain chronologically cuts across all other domains; centered and oriented terms have been used to describe education for nearly a century (Rugg and Shumaker, 1928) and the most recent LO measures fall within this domain. The fourth domain emerged as education organizations in the late 1980s pushed against the downward pressure to prepare young children early for rigorous academic learning; this led to a focus on beliefs about developmentally appropriate and inappropriate practices (Charlesworth et al., 1991, 1993; Hermans et al., 2008; Smith, 1993). Finally, the fifth domain assessed LO relative to constructivist theories of learning (Chan and Elliott, 2004; Woolley et al., 2004), inspired by the constructivism-based educational reform of the 1990s (Danielson, 1996; Fosnot, 1996; Prawat, 1992; Richardson, 1997).

The domains are presented in Table 1 in chronological order based on the earliest measures’ publication year: progressivism, teacher control and student autonomy, centered and oriented, developmental focus, and constructivism. We provide historical background for each domain before reviewing measures within the domain.

Progressivism

Reform touched nearly every aspect of American society during the Progressive Era, including education. Spanning the late 1800s through the mid-1900s (Cremin, 1961), the progressive education movement questioned assumptions about the purpose of education, instructional practices, and beliefs about children and learning. Traditionally, education aimed to transmit a fixed body of knowledge through strict discipline and subject-based curricula (Cremin, 1959; Dewey, 1938). John Dewey (and other progressives) challenged this view, arguing that children learn through experience and should engage with curricula reflecting their interests and development (Dewey and Dewey, 1915). Progressive schools sought to connect learning to life by incorporating activities such as physical education, industrial training, and nature studies (Cremin, 1959; Dewey and Dewey, 1915). Dewey believed that practices should support student learning, rather than blindly adhering to a particular ideology. He argued that, taken to extremes, neither traditionalism nor progressivism supports optimal learning and development (Dewey, 1938). Thus, toward the latter years of the progressive movement, Dewey sought to find a synergistic balance between organizational structure (e.g., curriculum and behavioral expectations) and student individuality (Cremin, 1959; Tyack and Cuban, 1995).

The progressive educational movement was coming to a close just as researchers began developing measures assessing LOs in the late 1940s. Our search found three Likert scale measures related to progressivism. The Minnesota Teacher Attitude Inventory was developed during the progressive era but does not explicitly label LO in terms of progressivism (Cook et al., 1951; Cook and Leeds, 1947). Alternatively, the Education Scale (Kerlinger and Kaya, 1959a) and the Modernity Scale (Schaefer and Edgerton, 1985) explicitly assess traditional and progressive LOs as orthogonal constructs.

Measures

Minnesota Teacher Attitude Inventory

The earliest LO measure we found was the Minnesota Teacher Attitude Inventory (MTAI; Cook et al., 1951). This 150-item measure aimed to predict career satisfaction and teacher-student relationship quality by assessing teacher beliefs. In addition to research application, the authors suggested using MTAI scores to vet candidates for teacher preparation and teaching positions. According to Cook et al. (1951), the MTAI distinguished between two LOs, conceptualized as opposite ends of a unidimensional continuum. Superior teachers would maintain positive relationships with students, enjoy schoolwork, and prioritize cooperative learning, student autonomy, and student interest. In contrast, inferior teachers would distrust students, dislike schoolwork, and prioritize controlling teaching practices.

MTAI items did not equally represent both viewpoints, as most items reflected negative, traditional teacher beliefs. In a factor analysis of data from 306 pre-service teachers, Horn and Morrison (1965) identified a factor labeled “traditionalistic versus modern beliefs about child control” (p. 121), in which all items aligned with a traditional perspective (e.g., “As a rule, teachers are too lenient with their pupils”; “Most pupils have too easy a time of it and do not learn to do real work”). Yee and Fruchter (1971) replicated these findings with 368 elementary teachers, labeling the factor “children's irresponsible tendencies and lack of self-discipline” (p. 122). They interpreted disagreement with such items as indicating a permissive, warm, and positive view of children. Across both analyses, participant agreement or disagreement with traditional items was treated as reflecting two opposing LOs, though explicit representation of progressive beliefs remained minimal (Cook et al., 1951; Horn and Morrison, 1965; Yee and Fruchter, 1971). Given its emphasis on teacher control and discipline, the MTAI was likely developed with conventional classroom contexts in mind; this aligns with the era in which it emerged (Zilversmit, 1993).

The MTAI was used across several decades of research. Measure development preceded the MTAI's initial publication by at least four years (Cook and Leeds, 1947; Leeds, 1950). Researchers continued to explore the MTAI's validity (e.g., Brody, 1970; Leeds, 1952; Scott and Brinkley, 1960; Stein and Hardy, 1957; Yee, 1967) and factor structure (Bell, 1977; Ferguson et al., 1954; Horn and Morrison, 1965; Yee and Fruchter, 1971) for nearly three decades.

Education Scale (ES-I)

Kerlinger and Kaya (1959a) developed the 20-item Education Scale (ES) following two Q-methodology studies designed to identify salient learning orientations (Kerlinger, 1958); these revealed progressivism and traditionalism as primary LOs. Kerlinger (1956) characterized traditionalism as restrictive and curriculum-focused, whereas progressivism emphasized permissiveness and teacher–student relationships. The preliminary ES consisted of 40 statements from these prior Q-sort studies (20 for each LO; Kerlinger and Kaya, 1959a). Items were pilot tested among “about 200 subjects, approximately half [education graduate] students and half people outside the university” (p. 17). Participants indicated their level of agreement with each statement on a 7-point Likert scale. Based on factor analysis of this preliminary data, the authors narrowed the ES down to ten items per LO.

Next, Kerlinger and Kaya (1959a) experimented with two response formats using the same 20 items. The ES-I retained the Likert scale format from the pilot study, whereas the ES-II presented 10 four-item sets—one traditional item, two neutral buffer items, and one progressive item—for forced-choice ranking. The authors suspected that forced-choice ranking could reduce social desirability bias but introduce interdependence among items. A sample of 598 participants (136 undergraduate and 157 graduate education students from the same university and 305 from outside) completed both forms. Factor analysis of ES-I responses revealed two orthogonal factors replicating traditional and progressive LOs. In contrast, ES-II responses showed a strong correlation between traditional and progressive beliefs, producing a single continuum factor. Kerlinger (1958) argued that these beliefs are not necessarily opposites, and orthogonal factors replicated across Q-sort and Likert formats (Kerlinger, 1956, 1958, 1967; Kerlinger and Kaya, 1959a). Thus, Kerlinger and Kaya (1959a) adopted the Likert scale format.

The authors also assessed predictive validity with data from this sample (Kerlinger and Kaya, 1959b), and found that students were significantly and consistently more progressive than traditional in LO. In contrast, participants outside of the university were significantly more traditional, but showed higher levels of inconsistency. These results were consistent across the ES-I and ES-II, supporting the authors’ overall argument for predictive validity.

With this larger sample, the ES subscales (i.e., traditionalism and progressivism) demonstrated sufficient internal reliability (Spearman-Brown corrected r = 0.83 and r = 0.75, respectively). The authors suggested that doubling the scale length could raise internal reliability. Retest reliability was also assessed among a subgroup of 106 students 3-4 months after initial testing. Retest reliability for the ES-I was low but sufficient (r = 0.71 and r = 0.70, respectively), while ES-II scores were insufficient (r = 0.68 and r = 0.64, respectively). ES-I internal and retest reliability were consistently higher than ES-II, supporting the decision to retain a Likert scale response format.

Kerlinger continued developing and honing the ES. By 1967, the ES-VI consisted of 46 items (i.e., 23 per LO; Kerlinger, 1967). A factor analysis of ES-VI data from 1304 education graduate students and teachers identified four factors associated with traditionalism: criticism of schools, learning as knowledge storage, educational conservatism, and an unnamed factor referring to teacher freedoms. An additional four factors were associated with progressivism: experimentalism, reconstructionism, life adjustment, and romantic naturalism. Table 2 shows the ES-VI factors and example items. Kerlinger (1967) describes the romantic naturalism factor (i.e., beliefs focusing on students' needs and interests) as child-centered; to our knowledge, the first use of a -centered term in reference to LO assessment. Internal reliability improved as scale length increased: the progressive and traditional scales averaged 0.82 and 0.84 (i.e., Cronbach's α), respectively, across the three participant sub-groups. Thus, the ES-VI demonstrated satisfactory internal reliability in assessing orthogonal, multi-dimensional LOs.

Table 2. Education Scale (ES-VI) factors and example items.

See full factor analysis results in Kerlinger (1967).

Modernity Scale

The third scale related to progressivism was initiated much later. Schaefer and Edgerton (1985) developed the Parental Modernity Scale, containing three subscales, two assessing traditional and progressive beliefs about education, and one assessing parental modernity beliefs. The separation of subscales for traditional and progressive beliefs suggests that these orientations were viewed as independent rather than opposing constructs. Although this scale was originally developed to analyze parent beliefs, it is now used most often to assess early childhood educators. Schaefer and Edgerton (1985) characterized traditional teachers as rigid and authoritarian, viewing learning as knowledge acquisition and students as needing extrinsic motivation to learn. Alternatively, progressive teachers were characterized as flexible and permissive, viewing learning as inquiry and students as intrinsically motivated to learn. Table 3 shows scale factors, topics, and example items.

Table 3. Modernity Scale factors, topics, and example items.

See Schaefer and Edgerton (1985) for full scale.

Schaefer and Edgerton reported on the Modernity Scale's development at the 1981 Biennial Meeting of the Society for Research in Child Development (Schaefer and Edgerton, 1981b); an abstract summarizing the work appeared in Resources in Education (Schaefer and Edgerton, 1981a). Items were derived from interviews with mothers (Parent as Educator Interview), and data were collected from three mother–child samples using the Modernity Scale alongside measures of classroom behavior (Classroom Behavior Inventory) and intelligence (Wechsler Preschool and Primary Scale of Intelligence). Parental modernity correlated positively with education, income, child intelligence, and academic competence, negatively with child dependency and distractibility, and showed no relation to social or emotional behaviors.

The Modernity Scale remains widely used with early childhood educators (e.g., Kelly et al., 2026) and employs child-based terms rather than student- or learner-based phrasing. Recent studies describe the traditional factor as adult-centered and the progressive factor as child-centered (Ansari and Pianta, 2019; Hu et al., 2017, 2021; Hur et al., 2015, 2016; Lu et al., 2023; Pianta et al., 2005). Its focus on child-centered constructs and frequent application in early childhood research suggest that it may extend beyond conventional classroom models and could align with alternative educational pedagogies. Its continued use, coupled with evolving terminology, suggests the construct remains relevant even as associated descriptors have shifted over time.

Summary

The earliest Likert scale measures assessing LOs were developed during the latter years of the progressive education movement. These early measures explored aspects of measure design: unidimensional vs. orthogonal factors and forced-choice vs. Likert scale response format. Subsequent measures leaned toward a Likert scale response format, but the debate over unidimensional vs. orthogonal LO factors persists across the following LO domains.

Teacher control and student autonomy

Rather than exploring teachers’ LO relative to an educational movement, measures in this domain explore teachers’ orientation towards control and student autonomy. Researchers approached these LOs from three different frameworks: control ideology, locus of control, and self-determination theory.

Control ideology originated from studies of interactions within organizations characterized by compulsory participation of an unselected clientele, such as psychiatric hospitals and prisons (Hoy, 1969; Willower et al., 1967). Given mandatory participation, clients may or may not desire to participate in the provided services. Organization employees, therefore, must exercise some degree of control over clients. Control ideology classifies the underlying beliefs and strategies used to justify and maintain control in such organizations. While there are notable differences between schools, psychiatric hospitals, and prisons, control ideology can provide a useful framework for understanding interactions between school staff and students (Hoy, 1969; Willower et al., 1967).

As discussed earlier, locus of control refers to individuals’ perceptions of what factors determine life events and outcomes (Rotter, 1966). Based in social learning theory, Rotter (1966) proposed a continuum of internal and external locus of control. Individuals with an internal control orientation tend to believe that outcomes are determined by their own efforts, ability, and choices. Individuals with an external control orientation tend to believe that outcomes are determined by factors outside of their control, such as chance, luck, or other individuals. This theory was applied to teachers’ perceptions of what factors determine events and outcomes in the classroom, such as student performance and classroom management (Rose and Medway, 1981). In terms of teacher LO, we consider locus of control relative to the learning process, distinguishing between teacher control and student autonomy.

Self-determination theory (SDT) proposes that individuals thrive in conditions that support the natural tendencies to engage in interesting, enjoyable, and, often, challenging endeavors (i.e., intrinsic motivation) and to internalize social practices (i.e., integration; Ryan and Deci, 2016). According to SDT, three psychological needs must be met to support these natural tendencies: competence, relatedness, and autonomy. Teachers’ orientation toward control and student autonomy, therefore, may be seen as integral to meeting students’ psychological needs and providing conditions necessary for students to flourish.

Here we discuss three LO measures that explore teacher control and student autonomy, each from a different theoretical framework. Willower et al. (1967) applied control ideology to schools when designing the Pupil Control Ideology Form. Wehling and Charters (1969) devised a subscale assessing locus of control in the Teacher Conceptions of the Educative Process Questionnaire. Finally, Deci et al. (1981) tested hypotheses related to self-determination theory with the Problems in Schools Questionnaire. These measures are discussed in turn.

Measures

Pupil Control Ideology Form

Willower et al. (1967) drew on Gilbert and Levinson’s (1957) study of psychiatric hospital staff-patient relationships and applied their control ideology to teacher-student relationships. They developed the 20-item Pupil Control Ideology Form (PCI) to assess teachers’ control ideology on a continuum ranging from custodial to humanistic rather than distinct orthogonal dimensions. The authors associate the custodial LO with traditional school. Custodial teachers are characterized as controlling, impersonal, pessimistic, and mistrusting of students. Teachers exercise power and impart knowledge while students are expected to demonstrate unquestioning obedience and learn by absorbing facts. At the humanistic end of the LO continuum, schools are seen as communities. Humanistic teachers are characterized as optimistic and flexible. Teachers and students cultivate mutual respect through open communication. They view students as self-disciplined and as learning through active engagement and interaction.

The PCI contains 20 items expressing a range of beliefs from custodial (e.g., Too much pupil time is spent on guidance and activities and too little on academic preparation), to humanistic (e.g., Pupils can be trusted to work together without supervision) LOs. Teachers indicate their level of agreement with each item on a 5-point Likert scale. In Willower et al.’s (1967) initial studies, the PCI demonstrated strong internal reliability (i.e., Spearman-Brown corrected r = 0.95 and 0.91 from two samples, N = 170 and N = 55, respectively). Construct validity was tested indirectly in relation to dogmatism. Data from 1306 teachers, principals, and counselors showed that while participant groups differed significantly in pupil control, no such differences were found in dogmatism. Based on these results, the authors inferred independence between the two constructs. However, no direct correlation analysis between constructs was reported and the lack of mean differences in dogmatism was determined by visual inspection rather than statistical hypothesis testing.

Graham et al. (1985) later conducted exploratory and confirmatory factor analyses on PCI data from 362 pre-service teachers. The exploratory factor analysis identified two factors. Six items were removed for low loading or low communality (i.e., only a small proportion of the item's variance could be explained by the latent variable). In a subsequent confirmatory factor analysis of the 14-item PCI, the two-factor model demonstrated better model fit compared to a one-factor model. However, the two factors were highly correlated (0.86), suggesting that they were tapping into one latent construct. Thus, Graham et al. (1985) concluded that these data supported the theoretical unidimensional model proposed by Willower et al. (1967). Four additional items were dropped from the one-factor model, resulting in a 10-item version of the PCI. Refining the PCI based on factor analysis improved internal reliability. Cronbach's α for the 20- and 10-item PCI, respectively, increased from 0.90 to 0.94. Thus, Graham et al. (1985) provided supporting evidence for the validity and reliability of a PCI short version.

Teacher Conceptions of the Educative Process Questionnaire

Wehling and Charters (1969) designed the 86-item Teacher Conceptions of the Educative Process Questionnaire (TCEP) to assess a wide range of teacher belief topics (e.g., classroom management, subject-matter, motivation) rather than LOs (e.g., “pupil-centered versus teacher-centered,” p. 8). Measure development began in 1962. Multiple items were borrowed from the MTAI and Education Scale (Cook et al., 1951; Kerlinger and Kaya, 1959a). Like Kerlinger and Kaya (1959a), Wehling and Charters (1969) experimented with forced-choice before adopting a Likert scale response format. Despite the deliberate conceptual design, factor analysis revealed eight factors, one of which was dichotomous: student autonomy (e.g., When given a choice of activity, pupils generally select what is best for them) vs. teacher direction (e.g., Pupils must see clearly that it is the teacher, not they, who has charge of classroom learning). Given this single revealing factor, it appears that the TCEP conceptualized teacher beliefs along a continuum from student autonomy to teacher direction, reflecting conventional power dynamics and making it less applicable to alternative pedagogies. The authors suggested that this factor assessed locus of control. Though the authors do not elaborate, this description implies that the authors considered locus of control relative to the learning experience, rather than assessing the teacher's view of internal/external control over aspects of the teaching experience (e.g., administrative support, relationships with parents, lesson planning).

Wehling and Charters (1969) used exploratory factor analysis to propose these eight factors, and subsequent studies have attempted to replicate this structure (Victor and Harootunian, 1971). However, the original publication did not report standard psychometric indices, and later applications of the measure typically treated the dimensions as theoretically meaningful rather than psychometrically established.

Problems in Schools Questionnaire

The Problems in Schools (PIS) Questionnaire assesses teachers’ LO towards control and autonomy-support (Deci et al., 1981). Other measures reviewed here ask teachers to evaluate belief statements and provide responses on a Likert scale; the PIS differs slightly. Instead of evaluating belief statements, teachers read a series of eight vignettes about difficult situations that commonly occur in school. Accompanying each scenario are four solutions, one each that is highly controlling, moderately controlling, moderately autonomous and highly autonomous. Table 4 provides an example scenario and accompanying solutions. Teachers rate each solution's appropriateness on a 7-point Likert scale. Subscale mean scores are weighted (i.e., -2, −1, 1, 2, respectively) and summed to produce teachers’ overall score. Thus, PIS item design and scoring treat control and autonomy-support LOs as opposite ends of a continuum, rather than orthogonal constructs.

Table 4. Problems in Schools Questionnaire example scenario and accompanying solutions.

Deci et al. (1981) collected data from 68 teachers; their scores spanned a wide range. The four subscales demonstrated moderate to good internal reliability (i.e., Cronbach's α = 0.73, 0.71, 0.63, 0.80 for highly controlling, moderately controlling, moderately autonomous, and highly autonomous, respectively). Because the moderately autonomous subscale was positively and significantly correlated with the control-oriented subscales, the authors revised their conceptualization of this subscale to slightly controlling. Data collected from 19 of the 68 teachers two months after initial testing showed low but acceptable test-retest reliability (i.e., 0.70 for the overall scale).

Deci et al. (1981) used these data to inform two validity studies. The first considered data from 35 of the initial 68 teachers and 610 students. In late October and mid-May, students completed measures evaluating their classroom climate (controlling or autonomy-supporting), intrinsic motivation, and perceived competence. Students' October classroom climate ratings and teachers’ PIS scores were positively and significantly correlated. Furthermore, students with autonomy-supporting teachers scored higher on intrinsic motivation and self-esteem relative to students with controlling teachers. Unexpectedly, these relationships were established within the first two months of school and remained stable for the duration of the school year.

The second validation study analyzed data from 6 of the original 68 teachers (3 control-oriented and 3 autonomy-oriented) and their students, all from different schools (Deci et al., 1981). Students’ intrinsic motivation and perceived competence were measured at the start of school and seven weeks later. Students’ change scores were weighted by teachers’ PIS z-scores to test whether stronger orientations produced greater change. Analysis of variance showed that teachers’ LO significantly predicted change only in students’ perceived cognitive competence: those with autonomy-oriented teachers improved more than those with control-oriented teachers. Change in intrinsic motivation reached significance only when comparing students of teachers with the highest and lowest PIS z-scores. These findings supported the PIS's external validity and replicated earlier evidence that autonomy-supporting teachers foster greater intrinsic motivation and perceived competence.

Summary

From control ideology to student motivation, researchers have taken various approaches to understanding teachers’ orientation toward control and student autonomy. Unlike other LO domains, all three measures assessed LO as a bipolar, unidimensional construct, even when the design was unintentional, as was the case with the TCEP (Wehling and Charters, 1969). No subsequent measures used the terms teacher control and student autonomy as LO labels, though the topics were certainly addressed in measure items across domains.

Centered and oriented

Centered and oriented terms have been used to describe education for nearly a century. In The Child-Centered School, Rugg and Shumaker (1928) characterized the child-centered approach as promoting student freedom, interest, initiative, creative self-expression, and personality development. In The Learner-Centered Classroom and School: Strategies for Increasing Student Motivation and Achievement, McCombs and Whisler (1997) draw a distinction between the terms student- and child- vs. learner-centered, clarifying that learner- refers to learning across the lifespan, though the terms are often used interchangeably (e.g., Petrie, 1981). The child-centered literature tends to involve younger learners (e.g., Hur et al., 2015; Pianta et al., 2005; Spidell Rusher et al., 1992), whereas the student-centered literature ranges up through post-secondary education (e.g., De Souza et al., 2021; Garrett, 2008; Li et al., 2019; Roehrig and Kruse, 2005; Simmons et al., 1999; Thomas et al., 2001). Additionally, centered and oriented terms can focus on teachers, subject-matter, and learning process (e.g., Bolhuis and Voeten, 2004; Garrett, 2008; Meirink et al., 2009). Across the other four LO domains discussed here, authors use centered and oriented terms to describe other LOs. Thus, rather than originating from an educational movement or psychological theories, measures in this LO domain applied widely used descriptive terms to LO constructs.

Here we discuss four measures that use centered or oriented LO terms, representing the largest timespan of the LO domains. Bunting (1985) was the first to use a centered LO term, comparing teachers’ student-centered and directive LOs. McCombs and Whisler (1997) focused on assessing learner- and non-learner-centered LOs. Most recently, Meirink et al. (2009) and de Vries et al. (2013) developed measures to explore student- and subject-matter-oriented LOs.

Measures

Educational Attitudes Inventory

Though the Educational Attitudes Inventory (EAI) was not widely used in peer-reviewed empirical studies, it is worth mentioning as the first LO measure to use the term student-centered as a factor label (Bunting, 1985). Taking inspiration from the MTAI and TCEP, Bunting surveyed 290 K-6th grade teachers, asking them to evaluate a series of belief statements related to learning and teaching on a 5-point Likert scale (Bunting, 1981, 1984; Cook et al., 1951; Wehling and Charters, 1969). Factor analysis extracted four scales (i.e., affective, cognitive, directive, and interpretive) containing 35 items in total.

However, the four-factor structure did not replicate in a subsequent study (Bunting, 1985). An independent sample of 320 K–6 teachers rated 35 original EAI items and 46 new statements on a 5-point Likert scale. Factor analysis extracted two factors: student-centered (19 items) and directive (15 items). Bunting (1985) did not define the student-centered LO explicitly but noted alignment with the original affective, cognitive, and interpretive factors (e.g., “The student feels more positively toward school when open and effective communication exists between himself and the teacher”; “Instruction should provide opportunities for the student to exchange ideas with the teacher”). The directive LO reflected the view that students lack experience and motivation to direct their own learning (e.g., “Provisions for independent learning often produce little of real benefit to the student”). Bunting (1985) justified the two-factor structure by referencing Kerlinger and Kaya's (1959a) claim that LOs are orthogonal rather than opposite ends of a single continuum. While the directive factor aligns with conventional pedagogies, the student-centered factor reflects alternative approaches that emphasize autonomy and choice.

Teacher Beliefs Survey

Based on the American Psychological Association's Learner-Centered Psychological Principles (APA Task Force on Psychology in Education, 1993; APA Work Group of the Board of Educational Affairs, 1997), McCombs et al. (1997) developed the Learner-Centered Battery to promote teacher self-assessment and reflection. Among this series of self-assessment scales, the Teacher Beliefs Survey (TBS-M) assessed learner- and non-learner-centered beliefs. According to McCombs (2003), learner- and non-learner-centered practices, respectively, promote and hinder students' intrinsic motivation, learning, and success. Additionally, the learner-centered orientation emphasizes individual learners and an understanding of learning and cognition (McCombs and Whisler, 1997). The TBS-M conceptualizes learner- and non-learner-centered beliefs as opposite ends of a continuum, with the learner-centered orientation aligning more closely with alternative pedagogies. McCombs (2003) proposed that both student and teacher perspectives on learner-centered practices should be considered, and that education could improve by addressing discrepancies between the two.

McCombs et al. (1997), with a team of educators, generated 180 belief statements reflecting learner- or teacher-centered orientations. Two pilot tests with teachers yielded 87 items. In the first validation study, 672 teachers across four states rated these items on a 4-point Likert scale. Factor analysis produced three factors: [1] learner-centered beliefs about learners, learning, and teaching; [2] non-learner-centered beliefs about learners; and [3] non-learner-centered beliefs about learning and teaching. Items loading ≥0.40 were retained, resulting in a 35-item scale. A second validation with 236 teachers across five states replicated the structure and showed good internal reliability (α = 0.84, 0.73, 0.80). Student reports of learner-centered teaching were positively correlated with teachers’ learner-centered beliefs, negatively with non-learner-centered beliefs about learners, and unrelated to non-learner-centered beliefs about learning and teaching. McCombs et al. (1997) thus provided consistent support across all five validation stages.

McCombs and Whisler (1997) published the TBS-M for teachers’ self-reflection and provided teachers with directions for score interpretation. Again, teachers responded to each item on a 4-point Likert scale. Scores above 3.4 on [1] and below 2.0 on [2] and [3] indicate learner-centered beliefs. Scores below 2.8 on [1] and above 2.4 on [2] and [3] indicate non-learner-centered beliefs. Thus, contrary to initial reasoning for the development of separate scales for each LO, teachers are instructed to interpret scale scores along a unidimensional continuum ranging from non-learner- to learner-centered.

Beliefs about Teaching and Learning Questionnaire

Meirink et al. (2009) developed the Beliefs about Teaching and Learning Questionnaire (BTLQ) using a 5-point Likert scale response format instead of forced-choice. The BTLQ explores two LOs (i.e., student- and subject-matter-oriented), each as a multidimensional construct with subscales for cognitive regulation, affective regulation, knowledge reproduction/construction, and individual/collaborative learning. Table 5 shows the BTLQ scales and example items.

Table 5. Beliefs about Teaching and Learning Questionnaire scales and example items.

See Meirink et al. (2009).

The authors characterize subject-matter-orientation as traditional and teacher-oriented: teachers are responsible for imparting subject matter and regulating the learning process, while students work individually to reproduce knowledge. Alternatively, student-orientation is characterized as progressive: students regulate the process of constructing knowledge and work cooperatively, while teachers show students how to learn. These definitions reflect the fact that, across the teacher beliefs literature, LOs are represented by multiple terms.

Belief statements about teaching and learning were included in each of the eight scales. Item stems were consistent across beliefs about teaching (i.e., It is important that a teacher…) and learning (i.e., Students learn better when…). Thus, beliefs about teaching focus on the role of the teacher, rather than the student, in the learning process, even in the student-oriented scales. Replacing teacher with student in the student-oriented item stems may have resulted in greater distinction between student- and subject-matter orientations. For example, “It's important that the teacher allows students to relate the different aspects of the subject matter themselves” could be reworded as “It's important for students to relate the different aspects of the subject matter themselves.” A factor analysis of data from 34 teachers showed that beliefs about teaching and learning loaded onto one factor, rather than two orthogonal factors; there was no explanation of the factor analysis method and results. This sample is small for the measure's complexity, and exploratory factor analysis is error-prone at small sample sizes; Costello and Osborne (2005) caution that such analyses often fail to recover the correct factor structure and advise against drawing substantive conclusions from them without adequate data. The single-factor result therefore warrants cautious interpretation. Additionally, no support was offered for the validity of four subscales for each LO (Meirink et al., 2009). Factor analysis using data from a larger sample could provide supportive evidence of the BTLQ's validity.

Belo et al. (2014) expressed concerns about the BTLQ's length. They developed a short version (28 items) and conducted a factor analysis on data from 126 teachers. Three factors were extracted: [1] student-regulated learning and knowledge construction, [2] knowledge reproduction, and [3] teacher-regulated learning. It is interesting to note that, like the Teacher Beliefs Surveys (McCombs and Whisler, 1997; Woolley et al., 2004), the teacher-oriented beliefs constituted two scales [i.e., [2] knowledge reproduction and [3] teacher-regulated learning], whereas the student-oriented beliefs constituted only one [i.e., (1) student-regulated learning and knowledge construction].

Beliefs about Learning and Teaching

Similar to Meirink et al. (2009), de Vries et al. (2013) developed a measure to assess subject-matter- and student-oriented teacher beliefs. However, unlike Meirink et al. (2009), de Vries et al. (2013) were not interested in exploring the LO multidimensionality, resulting in a shorter measure containing one scale for each LO. The authors described subject-matter LO as traditional, with a focus on transmission of knowledge from teacher/expert to student. Subject-matter-oriented teachers view students as a whole-class collective and are responsible for maintaining a calm classroom environment. Alternatively, the student-oriented LO takes a constructivist view of learning, focusing on development, active engagement, and social interaction. The student-oriented teacher views students individually, is an expert in subject-matter and pedagogical knowledge and is responsible for activating and guiding student learning. The authors note that the current culture in education leans toward student-orientation and that multiple terms have been used to describe LOs throughout the literature. The Beliefs about Learning and Teaching measure views student- and subject-matter orientations as distinct yet connected, with the student-oriented dimension reflecting the autonomy-supportive teaching found in alternative pedagogies.

De Vries et al.'s (2013) 14-item measure prompted teachers to evaluate how much each teacher belief statement applied to them using a 4-point Likert scale. All belief statements began with, “In my teaching it is important…” and ended with either a subject-matter-oriented (e.g., that I pass the subject matter to the students) or student-oriented (e.g., that students, where relevant, learn cooperatively in groups) phrase. The authors proposed that first-person wording and item randomization would help participants report based on their beliefs rather than social desirability bias, though no citations were provided to support these strategies. Exploratory factor analysis of data from 260 Dutch secondary teachers supported the theoretical two-factor structure. Both scales, student- and subject-matter-oriented, demonstrated acceptable reliability (i.e., Cronbach's α = 0.80 and 0.84, respectively). Finally, de Vries et al. (2013) compared teachers’ participation in continued professional development relative to LO. Survey results indicated that student-oriented teachers participated more in continued professional development. No relation was found between subject-matter orientation and continued professional development participation.

Summary

Though centered and oriented terms were used to describe education as early as 1928, they did not surface in Likert scale LO measures until 1985. Even then, two of the measures, the Educational Attitudes Inventory and the Teacher Beliefs Survey, were not widely used in peer-reviewed empirical studies (Bunting, 1985; McCombs and Whisler, 1997). Instead, the TBS-M gained traction as a self-reflection tool among educators. Centered and oriented terms are still widely used in the literature today, reflected in the more recent measure development assessing student- and subject-matter-oriented LOs (de Vries et al., 2013; Meirink et al., 2009). In contrast to the teacher control and student autonomy domain, all the measures in this domain assess LOs as orthogonal constructs.

A developmental focus

One common concern in early childhood education is that pressure to prepare children for rigorous academic experiences at younger and younger ages might erode their joy for learning (Charlesworth et al., 1993; Hirsh-Pasek, 1991). As instructional practices commonly employed during the elementary years began to surface in early childhood classrooms, multiple organizations published guidelines for developmentally appropriate practices (DAP; e.g., Bredekamp, 1987; International Reading Association and National Association for the Education of Young Children, 1998; Moyer, 2001; Moyer et al., 1987; National Association for the Education of Young Children, 1986; National Association of State Boards of Education, 1988). The National Association for the Education of Young Children's (NAEYC) most recent position statement addresses ages birth through 8 (National Association for the Education of Young Children, 2020) and continues its tradition of privileging self-initiated and independent play as developmentally appropriate educational practice, or DAP.

The educational movement towards DAP in early childhood education in the late 1980s catalyzed a flurry of measure development (e.g., Bryant et al., 1991; Charlesworth et al., 1991, 1993; Smith, 1993; Spidell Rusher et al., 1992). Researchers wanted to know whether teachers' beliefs and practices aligned with these principles of learning and development. Here we discuss three measures aimed at answering this question by assessing LOs related to development. The first two measures were directly inspired by NAEYC's DAP position statements (Charlesworth et al., 1991, 1993; Smith, 1993). The third measure uses the term developmentally-based more broadly, encapsulating progressive, constructivist, and child-centered teacher beliefs (Hermans et al., 2008).

Measures

Teacher Beliefs Scale

Charlesworth et al. (1991, 1993) developed a series of measures assessing teacher LO and practices based on NAEYC's DAP guidelines (Bredekamp, 1987; National Association for the Education of Young Children, 1986). Accordingly, these measures assessed two LOs (i.e., developmentally appropriate and inappropriate) among teachers working with young children. Developmentally inappropriate items addressed beliefs about practices such as workbooks, drills, and whole-class instruction. Developmentally appropriate items addressed beliefs about practices such as individualized instruction, play, using manipulatives, and exploration. The TBS conceptualizes developmentally appropriate and inappropriate beliefs as distinct but co-occurring orientations, with the former reflecting aspects of alternative, child-centered pedagogies.

Rather than assessing the degree to which teachers agree or disagree with various belief statements, the 36-item Teacher Beliefs Scale (TBS) prompts teachers to rate the importance of developmentally appropriate and inappropriate practices. Each item begins with, “It is _______ for/that/to…” Teachers then select a Likert-response from 1 (i.e., not important at all) to 5 (i.e., extremely important). Concurrently, Charlesworth et al. (1991, 1993) developed a self-report Instructional Activities Scale (IAS, 34-item) in which teachers report practice frequency. An accompanying 24-item observation checklist, Rating Developmentally Appropriate Practice in Kindergarten Classrooms, was developed to assess the validity of IAS self-report results.

The TBS has been modified multiple times to reflect updates to NAEYC's DAP guidelines. Initial principal components analysis of data from 204 kindergarten teachers extracted two inappropriate belief factors [i.e., [1] activities and materials, and [6] inappropriate structure] and four appropriate belief factors (i.e., [2] social, [3] individualization, [4] literacy activities, and [5] integrated curriculum; Charlesworth et al., 1993). However, the six subscales demonstrated only moderate internal reliability (i.e., Cronbach's α = [1] 0.84, [2] 0.77, [3] 0.70, [4] 0.60, [5] 0.66, and [6] 0.58). Buchanan et al. (1998) modified the TBS to reflect NAEYC's 1987 guidelines for children up to age 8 (Bredekamp, 1987). A factor analysis based on data from 277 first- through third-grade teachers extracted only two factors (i.e., appropriate and inappropriate beliefs), again with moderate to fairly low reliability (i.e., 0.84 and 0.71, respectively). Table 6 shows the TBS factors and example items. Finally, Kim and Buchanan (2009) modified the TBS to reflect NAEYC's 1997 position statement, which added consideration for the child's linguistic, cultural, and family background (Bredekamp and Copple, 1997). Factor analysis of data from 375 kindergarten teachers extracted three factors from this scale version: developmentally appropriate, developmentally inappropriate, and context appropriate practices, with relatively higher reliability (i.e., Cronbach's α = 0.85, 0.82, 0.81, respectively). Thus, while modifications to the TBS resulted in an inconsistent factor structure across versions, internal reliability improved with each series of modifications.

Table 6. Teacher Beliefs Scale factors and example items.

All items begin with stem, “It is ______ for/that/to…”.

Primary Teacher Questionnaire

Concurrently with Charlesworth et al. (1993), Smith (1993) developed the Primary Teacher Questionnaire (PTQ) based on NAEYC's position statement for DAP in the primary grades (child ages 5-8; Bredekamp, 1987). Smith (1993) noted that NAEYC's position statement presents paired developmentally appropriate and inappropriate practices as alternatives rather than opposites. Viewing LO as two orthogonal factors, the PTQ contains a separate scale for each LO, with DAP capturing child-centered approaches and traditionally-based (TRAD) representing conventional instruction. One 24-item scale represents TRAD practices such as early formal reading instruction, whole-group instruction, and a high degree of teacher control and structure (e.g., Curriculum should primarily facilitate the child's meeting of group expectations as defined by grade level). The other 18-item scale represents developmentally appropriate practices (DAP) such as play, hands-on activities, and child work choice (e.g., Teacher preparation time should be used primarily to prepare the physical learning environment for hands-on activities). Smith (1993) asked for teachers’ degree of agreement or disagreement with items using a 4-point Likert scale. An even-point Likert scale offers no neutral option, creating a forced-choice between agreement and disagreement. Unlike the MTAI, all items were positively worded, using “should do” or “should be” when referring to practices (Cook et al., 1951).

A field test among 144 early childhood and elementary pre- and in-service teachers showed significant differences in LO; early childhood teachers scored significantly higher on DAP and significantly lower on TRAD, relative to elementary teachers. The PTQ also obtained acceptable internal reliability estimates (i.e., Cronbach's α = 0.80 on DAP and 0.87 on TRAD). Interestingly, the DAP and TRAD scales were significantly and negatively correlated. The correlation between LOs may be an artifact of the item generation method (i.e., developing paired appropriate and inappropriate belief statements). While Smith (1993) originally viewed DAP and TRAD as orthogonal constructs, this correlation suggested that DAP and TRAD scales captured beliefs at opposite ends of a unidimensional continuum. Smith (1993) provided guidance for computing a composite TRAD and DAP score. However, subsequent studies maintained the PTQ's two-factor structure (e.g., Ketner et al., 1997; McMullen, 2001; Smith and Croom, 2000).

Beliefs about Primary Education Scale

In 2008, Hermans et al. developed the 18-item Beliefs about Primary Education Scale (BPES), hoping to tap into general teacher beliefs about the nature of good education. The BPES assesses two LOs: transmissive and developmental. Strict adherence to curriculum, emphasis on child performance and product, and the belief that knowledge is acquired via transmission characterize the transmissive LO. Alternatively, flexible curriculum, a focus on individual development, and the belief that knowledge is constructed characterize the developmental LO.

Hermans et al. (2008) conducted multiple studies to assess factor structure and content validity. A series of exploratory and confirmatory factor analyses were conducted on data from two samples. The first sample consisted of 166 undergraduate education students and 186 student-teachers. The second sample included BPES data from 377 primary school teachers. Analyses of these data supported a two-factor structure aligned with transmissive and developmental beliefs and sufficient internal reliability (i.e., Cronbach's α = 0.77 and 0.73, respectively). In a third study exploring construct validity, a sample of 380 primary school teachers completed both the BPES and the Teacher Beliefs Survey (Woolley et al., 2004; see discussion in section on constructivism below). According to Hermans et al. (2008), the Teacher Beliefs Survey was designed to assess traditional and constructivist beliefs about specific teaching practices, as opposed to the general statements about good teaching included in the BPES. Before assessing the correlation between like factors, a confirmatory factor analysis established the two-factor structure for each measure. Though fit for the two-factor model was superior relative to a one-factor model (i.e., two orthogonal factors vs. one bipolar continuum) for both measures, fit indices indicate only moderately acceptable model fit. Nonetheless, as expected, the model indicated moderate correlation between like factors across measures (i.e., transmissive/traditional and developmental/constructivist). Thus, both the factor analyses and correlational study provide supporting evidence for BPES measurement validity.

Summary

In reaction to a movement focusing on developmentally appropriate practice in early childhood education, researchers developed measures assessing how well teachers’ LO and practices aligned with DAP principles. Appropriately, use of these measures was limited to educators working with young children. The three Likert scale measures discussed here treated teachers’ LO (e.g., beliefs about developmentally appropriate and inappropriate practices) as orthogonal constructs, despite Smith's correlational evidence supporting the argument that LOs represent opposite ends along a unidimensional continuum (Charlesworth et al., 1991, 1993; Hermans et al., 2008; Smith, 1993). After the initial DAP movement, researchers shifted away from development-based LO labels. In fact, Smith (1993) described DAP as child-centered and constructivist, reflecting the previously discussed and following LO domains.

Constructivism

Constructivism is currently a widely accepted theory about how people learn (Bransford et al., 2000; Piaget, 1952; Piaget and Elkind, 1967). According to constructivist theory, individuals construct understanding as new information interacts with prior knowledge and beliefs (Fosnot, 1996; Richardson, 1997). Given that every individual possesses a unique set of prior knowledge and beliefs, new information may be perceived and interpreted differently across individuals (Danielson, 1996). Theorists have taken different approaches to understanding the mechanisms underlying and influencing knowledge construction (Fosnot, 1996; Richardson, 1997). According to the Piagetian or cognitive approach, meaning-making is an individual process as new information is either assimilated into or used to modify existing mental representations (i.e., prior knowledge and beliefs-accommodation). Alternatively, the situated cognition approach takes a more social view of learning, suggesting that both the learner and the environment are influenced by the learning process. Therefore, knowledge is socially constructed through activity and language. Yet another approach (i.e., sociocultural) was taken by Vygotsky (1978). The sociocultural approach suggests that individual learning is dependent on social experiences; the individual internalizes culturally-derived meaning through social interactions (Vygotsky, 1978). Thus, even though the theory is widely accepted, there are different approaches to understanding constructivism.

Constructivist theories of learning inspired a new wave of educational reform in the 1990s (Danielson, 1996; Fosnot, 1996; Prawat, 1992; Richardson, 1997). However, given the differing approaches to constructivism, there was no universal agreement on what constituted constructivist education policy and practices. Teachers were encouraged to provide opportunities for students to explore, discovering patterns and relationships through hands-on, cooperative activity. However, Danielson (1996) cautioned that hands-on does not necessarily mean minds-on. Teachers were encouraged to design activities that engaged student thinking centered on the learning objective, rather than activity for the sake of activity. Danielson (1996) also cautioned that constructivism does not mean that teachers relinquish all control and anything goes. Thus, echoing Dewey's cautions during the progressive movement, Danielson (1996) encouraged educators to establish mutually supportive relationships between students, teachers, and the classroom environment (Dewey, 1938).

As is the pattern across multiple LO domains, this new educational movement spurred on measurement development, in this case, to address how well teachers’ beliefs align with constructivist theories of learning. Here we discuss two measures that explicitly assess traditional and constructivist LOs as orthogonal constructs.

Measures

Teacher Beliefs Survey

Woolley et al. (2004) published the 21-item Teacher Beliefs Survey (TBS-W) to assess teachers’ constructivist and traditional LOs. Despite referring to constructivist teaching, approaches, and theories of learning, the authors did not explicitly define the terms constructivist and traditional. Instead, they referenced the emerging emphasis on constructivist approaches in education (e.g., Danielson, 1996; Richardson, 1997) and tension between the constructivist (i.e., Dewey, Piaget, Vygotsky) and behaviorist (i.e., Skinner and Bereiter) approaches to cognition and learning.

Woolley et al. (2004) generated 34 survey items based on a series of interviews probing in-service teachers’ beliefs about teaching. Similar to the Modernity Scale, the TBS-W initially included a subscale on constructivist parenting beliefs in addition to subscales assessing teachers' LO (Schaefer and Edgerton, 1985). Teachers responded on a 6-point Likert scale ranging from strongly disagree to strongly agree, providing no neutral option. The authors piloted the initial survey among 198 pre- and in-service teachers. Exploratory factor analysis of the pilot data extracted four factors: traditional management, traditional teaching, constructivist teaching, and constructivist parenting. Subsequently, survey data from 896 elementary, secondary, and K-12 special education pre-service teachers failed to replicate this 4-factor model in a confirmatory factor analysis. After dropping low loading and constructivist parenting items, model fit improved to moderately acceptable, and the authors settled on a 3-factor model. The final TBS-W included two scales for the traditional LO (i.e., management and teaching), but only one scale (i.e., teaching) for the constructivist LO. Internal reliability was also relatively low across the three factors (Cronbach's α = 0.52, 0.78, and 0.73, respectively). Thus, while Woolley et al. (2004) provided evidence for four of the five stages of measurement validity, factor analysis model fit and internal reliability were relatively low, and predictive validity was not addressed. Because the TBS-W scores traditional and constructivist beliefs on separate scales, the two orientations can vary independently, with the constructivist orientation aligning with alternative, child-centered pedagogies.

Teaching and Learning Conceptions Questionnaire

Concurrently, Chan and Elliott (2004) developed the 30-item Teaching and Learning Conceptions Questionnaire (TLCQ) to assess teachers’ traditional and constructivist LOs. According to the authors, the traditional LO views learning as the transfer of knowledge from teacher to student (e.g., Learning means remembering what the teacher has taught). Alternatively, the constructivist LO views the learner as employing reason and justification to create knowledge while the teacher guides and facilitates the process (e.g., Learning means students have ample opportunity to explore, discuss, and express their ideas). Chan (2001) characterizes the traditional LO as teacher-centered and the constructivist LO as student-centered, recognizing that multiple terms are used across the literature to characterize LOs.

The TLCQ represents traditional and constructivist beliefs on separate scales to be evaluated by teachers on a 5-point Likert scale. Chan and Elliott (2004) refined the initial measure through a series of pilot studies, factor analyses, and pre-service teacher interviews. Exploratory and confirmatory factor analyses of data from 385 pre-service teachers in Hong Kong extracted two factors, as expected, aligning with traditional and constructivist LO. The confirmatory analysis demonstrated good model fit. Additionally, the resulting scales and overall measure demonstrated sufficient reliability (i.e., Cronbach's α = 0.84, 0.84, and 0.86 for the traditional subscale, constructivist subscale and overall scale, respectively). Thus, Chan and Elliott (2004) employed measure development strategies that provide supportive evidence for the validity and reliability of the TLCQ. However, only pre-service teachers were assessed during measure development (Chan, 2001). Input from both in- and pre-service teachers would have provided evidence that the TLCQ taps into teacher beliefs across multiple career stages.

Chan (2001) proposed further measure development to facilitate cross-cultural comparisons. As it stands, TLCQ item wording may trigger social desirability bias in certain cultures. For example, one traditional LO item states, “During the lesson, it is important to keep students confined to the textbooks and the desks” (Chan and Elliott, 2004, p. 822). Bunting (1984) included a similar item (i.e., Teachers tend to keep the student too confined to the textbook and the desk). Teachers from certain cultures may be less likely to endorse this statement because it would be seen as socially undesirable to “keep students confined.” However, teachers may actually believe that lesson content should be textbook-based or students should remain seated during instruction. Thus, teacher response to this item may be biased and provide an inaccurate representation of teachers’ LO in certain cultures. Indeed, further development assessing cross-cultural social desirability and measurement invariance would provide supportive evidence for the efficacy of cross-cultural comparisons (André et al., 2020). So far, the TLCQ has been examined in Singapore (Teo and Chai, 2008), Turkey (Aypay, 2011; Şahin and Yılmaz, 2011), Ethiopia (Mengiste, 2025) and Myanmar (Pyae Kyaw, 2023).

Summary

As educators placed a new-found emphasis on constructivist theories of learning, researchers designed measures to assess teachers’ traditional and constructivist LOs; these may well map onto teacher-centered and student-centered beliefs. Woolley et al.’s (2004) TBS-W is the most widely used constructivist LO measure in peer-reviewed literature (e.g., Bonner and Chen, 2021), but its internal reliabilities are not impressive. Chan and Elliott (2004) aimed to develop a constructivist LO measure appropriate for cross-cultural comparison. As a result, the TLCQ is well-represented in the international literature.

Item-level analysis

An item-level analysis of LO measures identified multiple themes underlying the five domains. Subscales aligned with traditional beliefs (i.e., teacher control, developmentally inappropriate, directive, subject-matter-oriented, and non-learner-centered) consistently included belief statements about teacher control, subject-matter/curriculum, learning via transmission or acquisition, behavior management, and obedience/discipline. These items reference aspects of the learning experience that the teacher designs, implements, or manages, such as a curriculum, instruction, rules, and expectations. While the end goal may be to facilitate student learning, these beliefs prioritize the teacher's role, responsibilities, or importance over the role students play in their own learning.

From a locus-of-control perspective, these beliefs position the locus of learning agency outside the learner. Echoing Rotter’s (1966) external orientation, in which outcomes are seen as determined by forces beyond the individual, the student is positioned as acted upon rather than acting, with the teacher as the agent who causes learning. Thus, these beliefs can be seen as teacher-oriented. Table 7 lists teacher-oriented topics and example items across LO measure domains.

Table 7. Teacher-oriented topics and example items across learning orientation domains.

Only included topics spanning a minimum of four LO domains. Provided only one example item, even when multiple example items were available.

Alternatively, subscales aligned with progressive beliefs (i.e., student autonomy, student-centered, learner-centered, developmentally appropriate, and constructivist) consistently referenced student autonomy, interest, choice, individualized education, intrinsic motivation, a constructivist view of learning, and cooperative learning. These belief statements prioritize students’ role in shaping the learning experience, viewing the teacher as one of many resources that students leverage to support their learning. Considering Rotter’s (1966) definition of internal control orientation (i.e., individuals believe that outcomes are determined by their efforts, abilities, and choices), this LO can be seen as student-oriented, as the locus of agency is positioned within the learner. However, even in these student-oriented items, the teacher often still brokers the autonomy, implying that agency is something granted by the teacher as opposed to embedded in the classroom environment (e.g., I make it a priority in my classroom to give students time to work together when I am not directing them, from Woolley et al., 2004). Table 8 lists student-oriented topics and example items across LO measure domains.

Table 8. Student-oriented topics and example items across learning orientation domains.

Only included topics spanning a minimum of four LO domains. Provided only one example item, even when multiple example items were available.

Quality, use and correlates of the reviewed measures

To facilitate comparison across instruments, Table 9 summarizes the reliability, validity, and factor-analytic evidence reported for each of the 15 measures, drawn from the sources reviewed above.

Table 9. Summary of reliability, validity, and factor-analytic evidence across learning orientation measures.

“Not reported” entries reflect gaps in the original or reviewed sources, not omissions by the present authors.

The evidence in Table 9 leads to several observations. On quality, the picture is uneven and incompletely reported: internal reliability is absent for several instruments and, where reported, sometimes falls below conventional thresholds (e.g., TBS-W, α = 0.52). The intended two-orientation structure is frequently not recovered; factor analyses collapsed the orientations into a single continuum or dominant factor (e.g., TCEP, PCI), produced theoretically orthogonal scales that were negatively correlated (e.g., PTQ), or failed to replicate developers’ proposed solutions (EAI; TBS-W). Criterion-validity evidence is likewise absent for several measures.

A second observation is that independent use is concentrated rather than evenly distributed. A few instruments have been reused and revalidated well beyond their original studies (e.g., the TLCQ; Chan and Elliott, 2004). On the criteria to which measures relate, validation has targeted a scattered, instrument-specific set of external criteria rather than any shared outcome: respondent background (ES; Modernity Scale), student intrinsic motivation and perceived competence (PIS), teacher group membership (PTQ), student-reported teaching behavior (TBS-M), and professional-development participation (BLT). No common criterion has been used consistently, and for several measures none was located.

Discussion

Reviewing measures developed over the past 75 years makes clear that education's long-standing major LO tension (between teacher-oriented and student-oriented) has been repeatedly operationalized but rarely reconciled. Across decades, researchers have attempted to capture where teachers stand on this instructional divide using Likert-style measures, yet these efforts often reflect different theoretical commitments about what LOs actually represent. As a result, researchers studying “teacher beliefs” have not always been measuring the same construct, even when using superficially similar language. Because teacher beliefs are a multidimensional and value-laden construct (Fives and Buehl, 2012; Pajares, 1992), sound LO measurement design requires not only psychometrically valid and reliable measures but also conceptual clarity about the relation between teacher- and student-oriented beliefs (Schraw and Olafson, 2014).

Five dominant perspectives (progressivism, teacher control and student autonomy, developmental, constructivism, and centered/oriented terminology) emerged across this review, each reflecting broader cultural and theoretical movements in education. The centered/oriented domain transcends the others, capturing the enduring appeal of “teacher-centered” and “student-centered” as shorthand for different educational ideologies. Mapping how these frameworks overlap reveals that while the vocabulary has evolved, the fundamental question of who drives learning, the teacher or the learner, has persisted across eras of educational reform.

Two overarching patterns emerge. First, the same instructional dilemma has been reframed repeatedly across the decades: as authority vs. autonomy in the teacher-control tradition (Deci et al., 1981; Wehling and Charters, 1969; Willower et al., 1967), as readiness and developmental appropriateness/inappropriateness in early childhood and primary education (Charlesworth et al., 1993; Hermans et al., 2008; Smith, 1993), and as knowledge transmission vs. construction in constructivist frameworks (Chan and Elliott, 2004; Woolley et al., 2004). Terminology aside, the debate essentially revolves around the distribution of agency, or instructional decision-making power, in the classroom.

Second, measures differ in how they conceptualize the relation between the two main constructs. Some treat teacher- and student-oriented beliefs as opposite ends of a single continuum, suggesting that valuing one requires rejecting the other (e.g., Deci et al., 1981; Willower et al., 1967), whereas others treat them as orthogonal, allowing for the possibility that a teacher may endorse both strong structure and strong autonomy depending on context (e.g., Bunting, 1985; Kerlinger and Kaya, 1959a). This distinction is critical: continuum models reinforce what we consider a false binary of student-oriented vs. teacher-oriented beliefs, whereas orthogonal models can better capture the nuanced, context-dependent belief systems that characterize real-world teaching. This matters because continuum-style measures cannot capture teachers who intentionally combine explicit instruction with opportunities for student autonomy (e.g., guided exploration); orthogonal models can.

The item-level analyses of scales represent two coherent systems of belief about agency and responsibility in learning. The teacher-oriented and student-oriented beliefs have been labeled in different ways as educational philosophy has changed, but their essential positions recur with striking stability across time (Loveless, 2001) and across theoretically distinct measures. Future LO measures may benefit from building upon these recurring dimensions while testing their structure against diverse pedagogical contexts. However, we also distinguish between the dimensions themselves and their expression. While the dimensions have remained remarkably stable across these instruments, a teacher's expression of learning orientation may be context-dependent, varying with the instructional setting, subject or age group. An orthogonal model is well suited to this view, because it allows a teacher to hold both orientations and weight them differently across contexts.

Our review shows that past measures have often imposed a strict either/or structure on these orientations, positioning them as mutually exclusive ends of a continuum. We argue that this approach is inadequate. Treating teacher-oriented and student-oriented beliefs as orthogonal, potentially coexisting dimensions better reflects how teachers actually describe their work, better accommodates instructional models that intentionally blend structure and agency, and better supports comparison across diverse pedagogical contexts, such as Forest Schools (Egan et al., 2022), Montessori (Lillard, 2017), Waldorf (Taplin, 2024) and Reggio Emilia (Aljabreen, 2020).

We raise these alternative pedagogies not as contexts this review sets out to measure, but to identify an untested boundary on the existing instruments’ demonstrated validity: because these measures were developed and normed in conventional classrooms, whether they function equivalently for teachers in such settings is an open empirical question rather than an established one. None of the reviewed measures accounted for alternative pedagogies, and item wording sometimes did not make sense for other pedagogical approaches. Our examples rely on the alternative pedagogy we know best, Montessori education; we do not assume the specific features we point out generalize to every alternative pedagogy, but they illustrate the kinds of problems that might arise when using the measures in non-conventional educational settings. As a first example, I prefer to cluster students’ desks or use tables so they can work together (Woolley et al., 2004) presupposes that students have desks; in Montessori classrooms, there are tables to work at, but no student has their own table, and there are no desks, per se. As another example, the statement, A teacher's major task is to give students knowledge/information, assign them drill and practice, and test their recall (Chan and Elliott, 2004) assumes drilling and tests, which Montessori lacks. Woolley et al. (2004) discuss thematic units based on student interest as a child-oriented item, but thematic units are not used in Montessori. Items in Wehling and Charters’ (1969) measure assume a teacher-managed class structure in which learning depends on how much control the teacher exerts (e.g., The effective teacher has complete control of the learning situation at all times). Such items assume that autonomy is granted by the teacher, rather than structurally embedded in the environment, as is common in Montessori classrooms. These are just a few examples of how measures developed with traditional educational contexts in mind might not be useful in other educational contexts.

Alternative pedagogies like Montessori and the others cited above often endorse learner-driven agency, fluid teacher roles, and non-hierarchical classroom architecture, conditions less represented in the conventional-classroom-based LO measures reviewed here. We do not claim that existing instruments have never been administered in such settings; some have been applied in early-childhood and child-centered contexts. But these instruments have not been validated across pedagogical contexts; their measurement equivalence between conventional and alternative classrooms has apparently not been formally established (Vandenberg and Lance, 2000). Measurement-invariance testing across groups has become increasingly common for related teaching measures (e.g., André et al., 2020), yet to our knowledge no such analysis has compared LO measures across pedagogical contexts. Without such evidence, a teacher in a Montessori classroom could score “low” on student-centered items not because they resist student agency, but because the items themselves presuppose desks, whole-class pacing or teacher-led behavior management. Given the global spread of alternative educational approaches (Debs et al., 2022), developing LO instruments that function across diverse pedagogical contexts is increasingly important, both to improve cross-pedagogical validity and clarify whether existing constructs generalize beyond traditional schooling models. Further research should address this issue.

A related limitation of the measures reviewed here concerns the decontextualized nature of Likert scale belief assessment. Researchers have long noted that rating generic statements about teaching and learning may not accurately capture the beliefs that guide teachers’ actual practice, as beliefs are often context-dependent (Buehl and Beck, 2015; Fives and Gill, 2015; Lee, 2009; Liu, 2011; Safrudiannur and Rott, 2019; Witter and Hattie, 2026). Even vignette-based approaches, such as the PIS (Deci et al., 1981), have been questioned for their distance from the complexity of teachers’ daily classroom decisions (Skilling and Stylianides, 2020). This concern is particularly relevant for LO measures, where item wording frequently presupposes conventional classroom conditions, such as whole-class instruction, teacher-assigned tasks, and standardized curricula, that may not reflect the contexts in which many teachers actually work. As our analysis of alternative pedagogical contexts illustrates, decontextualized items risk misclassifying teachers whose orientations are structurally embedded in their environment rather than explicitly enacted through individual belief statements. Future LO measures should therefore attend not only to psychometric properties but also to the ecological validity of items across diverse instructional contexts.

A further limitation common to self-report LO measures is social desirability bias; teachers may endorse the belief that they perceive to be favored rather than their actual beliefs. Several measures reviewed here attempted to mitigate this through forced-choice formats (Kerlinger and Kaya, 1959a; Wehling and Charters, 1969) or vignette-based response (Deci et al., 1981), with mixed success; the concern is heightened in cross-cultural application, where the social desirability of particular beliefs varies (Chan and Elliott, 2004).

Limitations

Several limitations qualify this review. First, our search was iterative rather than systematic. For our initial search, we relied on Google Scholar together with reference-list searching, and did not search multiple bibliographic databases under a pre-registered protocol, possibly impacting future reproducibility (Gusenbauer and Haddaway, 2020). However, we ran supplementary searches in three databases (ERIC, Scopus, and Web of Science) in July 2026, and no instrument meeting all five of our inclusion criteria emerged beyond the 15 reviewed here. These supplementary searches, together with a June 2026 reverification of the original Google Scholar search, also surfaced no qualifying instrument published after 2013.

Relatedly, because we reviewed only measures reported in peer-reviewed publications, the set considered here may be subject to publication bias: instruments developed but never published, often owing to weaker psychometric performance, are by definition absent, which could lead a reader to overestimate the typical psychometric quality of measures in this domain.

Second, we did not apply a formal, standardized framework for appraising the psychometric quality of each instrument. We summarized the reliability and, where available, validity evidence reported for each measure, but we did not score instruments against a structured checklist of measurement properties. Third, our organization of measures into five theoretical domains is interpretive. Other defensible groupings are possible, and several measures could reasonably be placed in more than one domain.

A further limitation concerns our inclusion criteria. We elected to only include measures that yielded scores for exactly two LOs, which necessarily excluded the substantial body of multidimensional or unidimensional measures. This includes instruments such as the Teacher Beliefs Survey (von Oppell and Aldridge, 2021), a constructivist-classroom belief survey comprising multiple scales, the Student-Centered Teachers’ Beliefs Survey (Behar-Horenstein and Leite, 2011), which resolves to a single dimension, and an instrument developed by Isikoglu et al. (2009), which validated a four-factor inventory assessing student-oriented beliefs. This was a deliberate scoping decision aligned with our focus on the teacher-oriented and student-oriented LOs, but this review should be understood as a history of these two-orientation LO measures, rather than a comprehensive account of the wide range of instruments used to assess teacher beliefs.

Finally, our observations about the fit between existing measures and alternative pedagogies are conceptual rather than empirical. We argue that measures developed for conventional classrooms may not function equivalently for teachers in settings such as Montessori, Waldorf, or Reggio Emilia, and we illustrate this with item-level examples. These observations are best understood as hypotheses to be tested empirically (e.g., through differential item functioning or measurement-invariance analyses comparing teachers across pedagogical contexts) rather than as established findings. We recommend this as a direction for future work.

In sum, this review provides a synthesized framework and catalog of available Likert scale LO measures, along with their theoretical commitments and psychometric properties, as well as guidance for future work. Future measures must move beyond treating these orientations as two ends of a single scale. Instead, researchers should design LO instruments that acknowledge teacher- and student-oriented beliefs as distinct but potentially co-endorsed orientations and ensure that those instruments are meaningful across a range of pedagogical approaches. Doing so will support more inclusive, theoretically coherent, and context-sensitive research on how teachers understand teaching and learning.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Serafina Pastore, University of Bari Aldo Moro, Italy

Reviewed by: Erkan Hasan Atalmış, Manisa Celal Bayar University, Türkiye

Slavica Šimić, University of Zadar, Croatia