Daryl Siedentop’s Epistemological Lens, and His Influence on the Fields of Physical Education and Teacher Education

ABSTRACT

Daryl Siedentop has been viewed as one of the most influential scholars in the fields of physical education and physical education teacher education. In this article, we examine his epistemological lens, behavior analysis that he used to underpin his research and practice. After an introduction to behavior analysis, we discuss his approach to conducting research, and then we discuss some of his more significant contributions to the fields, providing attention to backstories of his work, what the work was, and how it impacted the fields of physical education and teacher education. Throughout, we use examples to highlight his contributions to the two fields.

KEYWORDS:

• Legacy
• Mentorship
• Leadership
• Program Culture

Daryl Siedentop was a behavior analyst. While he was influenced by a variety of philosophical positions (see 2023 et al., this issue), Radical Behaviorism (Skinner, 1953, 1974), the philosophy of behavior analysis, was very influential in both his professional and personal life (Siedentop, 2002a; Personal communication, Bobbie Siedentop, July 16, 2022). Our goal in this article is to describe the lens through which Siedentop approached research questions, research, and interpretations in research on teaching in physical education. We begin with a brief overview of behavior analysis. Next, we discuss his approach to conducting research and then we discuss some of his more significant contributions to the fields, providing attention to the backstories of his work, what the work was, and how it impacted the fields of physical education and teacher education. Throughout, we use examples to highlight his contributions to the two fields.

Behavior analysis is a natural science and can be contrasted with social science principally by its sole focus on the behavior of the individual and not mentalism, and its focus on experimentation over theorizing (Johnson, 1993). It can be situated between biology, its root science, that explains the evolution of a species and cultural materialism that explains the practices of cultures. In all three sciences the primary mechanism is selection, which depends on variability of genes and behavior. In biology, the selection process is natural selection principally at the genetic level; in behavior analysis, the selection process for behavior is reinforcement¹. In cultural materialism, the selection process is via cultural contingencies that select cultural practices (for more details on these selection contingencies see Dawkins, 1986; Glen, 1988; Harris, 1979). In behavior analysis, the analytic unit is the three-term contingency that describes the relationship between a behavior (e.g., student performance), its consequence (success or teacher or praise that serves as a reinforcer), and the environmental context that occasions the behavior (e.g., teacher instructions).

All students introduced to behavior analysis learn that it is an applied science in the sense that it focuses on (a) socially significant behaviors that improve the life of the individual; (b) the behavior of the individual and not mental constructs; (c) demonstrating functional (i.e., causal) relationships between environment and behavior; (d) developing interventions (i.e., procedures) that can be replicated, preferably with multiple behaviors, and across different persons and settings; (e) being conceptually systematic, meaning that the procedures (e.g., teaching practices) used to improve behavior should be identified and interpreted using behavioral principles, (f) being effective and eschewing procedures that are not, and (g) establishing generality of behavior to different settings and behaviors (Cooper et al., 2020).

Siedentop first became interested in behavior analysis as a graduate student at Indiana University (IU) where he met Brent Rushall who was also studying at IU and who was a behavior analyst (Siedentop, 2002a). During their time at IU, Rushall and Siedentop studied behavior analysis (Brent Rushall personal communication, July 11, 2022). Following graduation, as newly minted professors, both went on to write the book The development and control of behavior in sport and physical education (Rushall & Siedentop, 1972). During Siedentop’s career at Ohio State, he worked with professors John Cooper, Tim Heron, and Bill Heward who are influential applied behaviors analysts, and who had masters and doctoral programs in behavior analysis. Many of Siedentop’s masters and doctoral students took those classes. The collaborations with Cooper, Heron, Heward, and Rushall created a reinforcing environment for Siedentop to advance behavior analysis in physical education. This influence extended to some of his doctoral students who were trained in behavior analysis at Ohio State and who used behavior analysis in their professional work throughout their careers. These students included Ken Alexander, Shiri Ayvazo, Eitan Eldar, Andrew Hawkins, Thom McKenzie, Amos Rolider, Hans van der Mars, and Phillip Ward. One of the most impactful examples of this carry forward by Siedentop’s students was the design of the SPARK curriculum, a behaviorally and evidence-based curriculum by McKenzie et al. (2009) that has impacted physical education for 30 years. Siedentop’s epistemological lenses were inherited not only by his mentees but also by his mentees, mentees, placing him in the role of an “academic grandfather.”

Siedentop’s research lens

As a behavior analyst, Siedentop was mostly an inductive scientist being led by the data to models, rather than using data to fit models. In the tradition of behavior analytic research, his search was for robust variables, and he placed the highest value on the internal validity of a study. He saw teacher behavior, such as a teacher’s instructional skills, as a dependent variable and teaching training either at the preservice level or as professional development as the independent variable. And, student behavior such as successful trials or engagement in the lesson as a dependent variable and teacher behavior such as instructional skills or curricular approaches as an independent variable.

He was a strong proponent of experimental research noting that, “if you really want to understand something try to change it” (1982, p. 48). In behavioral terms, he was making the case that if you are unclear about the controlling variables of a behavior, changing it will bring the behavior under the control of variables that you do know. Similarly, if you believe that you know the controlling variables for a behavior, changing the controlling variable will validate or invalidate your assumptions. Of course, it might also demonstrate the power of the new controlling variable. He also supported experimental research, because as a behavior analyst he was focused on the social significance of the teaching settings (e.g., effective management, student learning, safety). Experimental research is one way to demonstrate improvement of desired behavior or the cessation of teacher and student behaviors that are undesired or irresponsible. Finally, he was committed to ecologically valid research, working with teachers in intact classes eschewing laboratory-based studies of teaching effectiveness and studies that were not focused on behavior. Siedentop’s modeling of quality teaching research served to carry forward as Shiri Ayvazo (personal communication, July 22, 2022) observes:

I can identify several components of my work that can be traced back to Siedentop’s philosophical, conceptual, and experimental legacy: studying teaching and learning phenomena as behaviors of interest, subject to operational definitions, direct and systematic observation, and experimental manipulation; deriving conceptualizations from other fields of study and applying it in physical education.

Siedentop’s behaviorizing of other perspectives

As many of the papers in this special issue will attest, Siedentop was known for reading outside his field, but he did so by following the rule, “get out of your own paradigmatic space, but do so without leaving your paradigm behind” (Siedentop, 2002a, p. 455). He described how he did this “Having become accustomed to reading widely outside my own paradigm, I was also learning to recognize good behavioral approaches that were masquerading under other paradigms, most notably, in the late 70s, the emerging cognitive paradigm” (Siedentop, 2002a, p. 455). In conversations with the lead author of this article, he called this the “behaviorizing” of research traditions that were not framed from a behavioral perspective. He describes how this occurred with his view of the ecological paradigm:

My “Eureka” experience came when I read Walter Doyle’s (1979) chapter in Peterson and Walberg’s book on research on teaching. Doyle starts that chapter by arguing that the reinforcement view of teaching was bankrupt. Now, when I see an argument begin with negative references to a reinforcement view or, even worse, a stimulus-response model, I know that the person making the argument doesn’t know much about what he or she is condemning, so I’m ready for the worst. So, I was ready for the worst from Doyle, but much to my amazement, the following explanation of his ecological model followed a page or two thereafter. “A second feature of an ecological viewpoint is, appropriately, that there is a direct focus on environment—behavior relationships. The model postulates that environments establish limits on the range of behavioral options and that observed behavior is in large measure a response to the demand characteristics of classroom environments (Doyle,1979, p.189).”

(Siedentop, 2002a, pp. 455–456).

He went on to say,

Of course, that was a perfectly behavioral description; Indeed, nearly every aspect of his emerging ecological model made as much sense from my set of assumptions as it did from his information processing, cognitive explanatory system. So, I used the Doyle model for 20 years and have used it with a mixture of methodologies but always with behavioral explanations

(Siedentop, 2002a, p. 456).

These comments show an important perspective of Siedentop: Always look for the commonalities in different traditions, come to know the different perspectives, and to behaviorize their commonalities (Siedentop, 2002a). This also carried forward into successive generations of behavioral researchers:

It wasn’t until recently I realized how much of the stuff I do in research, teacher education, and possibly the supervision of doctoral students I owe to Siedentop and his grad students’ scholarly work. As a mentee of his first-generation grad students, I have been conducting research and developing a teacher education program informed by behavior analysis, of which most of the spadework was done at OSU. I have also, learned to stay open for other epistemologies and seek the commonalities (Peter Iserbyt, personal communication, 25 July 2022):

Examples of Siedentop’s impact on the fields of teacher education and physical education

Academic learning time in physical education

Early on, Siedentop was influenced by the process-product paradigm in general education (Dunkin & Biddle, 1974). But his subsequent research and that of his students focused on the mediating-process paradigm, and in particular on the construct of academic learning time (ALT). Academic learning time was a variable developed for use in the Beginning teacher evaluation study (BTES) conducted in classrooms in the late 70s (Fisher et al., 1981). Their search was for a variable that would allow for a more immediate indication of the effects of teaching rather than waiting for the more distal end-of-year achievement test (i.e., the mediating-process paradigm). The BTES identified three important variables, namely, allocated time—the time scheduled by the teacher for student learning; engaged time—the time students are actively engaged in the lesson; ALT—that portion of lesson time where the student is engaged with the content matter (i.e., the task set by the teacher) at a success rate of 80% or more (Fisher et al., 1981).

The findings for each of these variables are important for teaching effectiveness and teacher education. First, the amount of allocated time that teachers devote to instruction is positively associated with learning (Fisher et al., 1981). Second, the proportion of the time that students are engaged with the content is positively associated with learning Fisher et al. (1981). But the major conclusion of the BTES was that students who spend more time actively engaged with a task at a high success rate (ALT variable) learn more than those who do not (Fisher et al., 2015). As such, the ALT variable served as proxy variable for learning. Teachers who arrange their lessons to produce more ALT for their students create greater learning gains for these students and are more effective (Rosenshine & Stevens, 1986). In the late 1970 and 80s, the majority of the research had been conducted in elementary math and reading classrooms, though since then there are few educational subjects and grade levels K-16 where the ALT variable has not been used as a measure of teaching effectiveness and student learning.

So, it was not surprising that ALT found its way to physical education. It was a common occurrence for Siedentop that his graduate students often led him in particular directions. In the case of academic learning time, Mike Metzler (personal communication, July 12, 2022) describes how the ALT variable came to be considered by Siedentop:

I was taking a research on teaching course from Don Cruickshank, where he showed us the work from the Far West Laboratory and introduced the construct of ALT. I became immediately interested in it as a dissertation topic—to measure ALT in PE … I walked right from my class with Cruickshank to Daryl’s office and told him I had found my dissertation topic. He did some reading from my course materials, and immediately got on board.

What followed was a process of category adaptation for physical education settings and measurement refinement (see supervision section) and several iterations of the instrument assessed in physical education settings (ALT-PE; Metzler, 1989; Siedentop et al., 1979; Siedentop et al., 1982), the development a short version for use in The Ohio State University physical education supervision program (n.d.), the first computer application for any measurement instrument in physical education (Metzler, 1989), followed later by more sophisticated software (Lund & Gaylor, 2002). From this start, physical education researchers throughout the world began using the academic learning time instrument and its ALT variable counterpart, the motor appropriate category, as dependent variables in descriptive, correlational, and experimental research (for reviews of these studies see: Metzler, 1989; van der Mars, 2006).

As was the case in classroom subject areas, the allocated time, engaged time, and ALT transformed judgments about teacher effectiveness, moving from checklist, rating scales, and other high inference, and often context-free judgments, to judgments based on what teachers and students actually did in lessons. These were observable and measurable behaviors that formed the basis for judgments about the work of teachers and students, and these data served as a basis for the improvement of both. Importantly, the ALT-PE research at Ohio State like most of the research that Siedentop did was programmatic, grounded in direct and systematic replication (Cooper et al., 2020), with a low level of inference, and ecological validity. This served as a model for many researchers in the field.

Ecological framework and tasks systems

As we have noted previously, Siedentop had been influenced strongly by teaching effectiveness researchers who studied classroom settings (e.g., Rosenshine & Stevens, 1986) using the process-product model (Dunkin & Biddle, 1974) and the mediating process model (Fisher et al., 1981). However, the most influential of the teaching effectiveness perspectives for him was the ecological model proposed by Doyle (1983; 1986). Central to Doyle’s model is the task, which represents the primary organizing unit of the curriculum. Students learn, in both the narrow and broad sense of that word, principally through the instructional, managerial, and social tasks that teachers present to their charges (Hastie & Siedentop, 2006). Tasks exist in systems within the classroom ecology that serve to define the conduct for participants (i.e., teacher and students). Each task system overlaps and interacts, thus creating the ecology of the gymnasium. Within each task system, tasks are defined and refined by both the effectiveness and the accountability used by the teacher.

In Siedentop’s behaviorizing of Doyle’s model (Siedentop, 2002a; Hastie & Siedentop, 2006), he viewed instructions on how to perform the tasks as antecedent conditions, the task as the behavior that students performed, and the presence or absence of accountability as a consequence (Siedentop, 2002a; Hastie & Siedentop, 2006). Doyle (1983) had observed that when there is no accountability, or when there is an inappropriate task, student accomplishment of the task may be incomplete. Research conducted by Alexander (1982) provided an empirical demonstration of how this functionally occurs in teaching. Tousignant & Siedentop (1983) identified a unique student behavior in physical education, they termed “competent bystanders” referring to students who avoid participation but stayed within the boundaries of the managerial task system and thus often escaped accountability and were inactively engaged in the content of the lesson. Tousignant and Siedentop (1983) also added the transitional task system to the model of physical education to account for logistical movement that is considerably less frequent in classroom settings but commonplace in physical education. Several studies by Siedentop’s students validated the model in physical education (see studies reported in Siedentop, 2002a; Hastie & Siedentop, 2006).

As was the case of the ALT-PE variable, Siedentop’s investigation of Doyle’s model was impacted by the interests of his graduate students (e.g., Tinning & Siedentop, 1985, Romar & Siedentop, 1995) which led to a more comprehensive and diverse methodological analysis than merely behavioral analyses. As Siedentop (2002a) described:

First, we began to more fully utilize ethnographic observational methods in addition to low-inference observation systems … Second, we began to examine more closely the teachers’ knowledge of the activities they were teaching … Third, we began to investigate student understandings and evaluations of their physical education experiences.

Fourth, we began to assess what teachers believed about physical education in general, specifically their beliefs related to the unit they were about to teach, what is typically referred to as an espoused theory of instruction, and then to compare those assessments with data from observations, that is, their enacted theories of instruction. Fifth, I decided that we would … intentionally seek out teachers who were thought to be effective

(p. 432).

Today the ecological paradigm is one of the core underpinnings for teaching effectiveness research in general and physical education (Cohen et al., 2003; City et al., Ward, 2013) and for what has become known as adaptive teaching that is a teacher’s ability to adjust their planning and teaching to the individual needs of students (Xie et al., 2021). The model moved Siedentop’s and the field’s focus from linear models such as process-product and mediating process-product models to a more comprehensive understanding of teaching and learning in physical education existing in an ecological context.

Systematic observation

Other authors in the special issue have emphasized the impact of systematic observation on the field (2023 et al., this issue). Siedentop (1981) began the systematic observation methods section of his Developing teaching skills in physical education text with:

The systematic observation of teachers has revolutionized teaching research and has led to important discoveries about the nature of effective teaching … Systematic observation is the foundation on which teaching research has been built, it should also be the foundation on which teaching skills are developed

(p. 252).

Although systematic observation instruments have been used in a variety of contexts and by a variety of sciences, the methods developed in behavior analysis are considered to be the gold standard for observation of behavior (Cooper et al., 2020; Johnson, 1993; McKenzie, 2016). Among the assumptions underpinning behavioral methods of measurement are:

• Direct observational measurement of the events under investigation versus indirect measures requiring a high inference to be made (e.g., a questionnaire).

• Systematic observation methods where there are operational definitions of the behavior/s being investigated (e.g., defining slow transitions as longer than 5 seconds).

• Measurement instruments that are inclusive of variations of the behavior (e.g., moderate versus vigorous physical activity).

• The sampling procedures track the behaviors across the context and time (e.g., a student’s motor-appropriate behavior across a lesson or lessons).

• Measurement is conducted continuously in the setting using repeated sampling (e.g., every trial, short intervals such as 6 or 10 seconds).

We previously noted that the discussion of the ALT variable occurred in physical education through the initial work of Metzler and Siedentop (Mike Metzler, personal communication, July 12, 2022). The development of the ALT-PE instrument and its variations previously reported led to many other instruments being developed. For example, there were two editions of Analyzing physical education and sport instruction (Darst et al., 1989) in which there were 31 direct and systematic observation instruments for physical education and sport settings, and two chapters on ethnographic analysis. McKenzie used the ALT-PE system as the basis to develop a host on measurement systems of physical activity in different contexts such as System for Observing Fitness Instruction Time (SOFIT), System for Observing Play and Leisure in Youth (SOPLAY), System for Observing Play and Active Recreation in Communities (SOPARC; McKenzie, 2016).

Many teacher educators and teachers who used these instruments often knew little about the behavioral tactics and strategies underlying direct systematic observation, but they used them to produce accurate and reliable data because they followed the procedures outlined in the original ALT-PE instruments. Today, using systematic observation in physical education, preservice teacher education is commonplace around the world.

Supervision

The paper The Ohio State University supervision research program summary report (Siedentop, 1981) was a seminal paper talking about conducting supervision and doing research on supervision. Siedentop began with an extensive rationale as to why teacher education should pay attention to the act of supervision. Siedentop then discussed how the supervision research program was developed, describing eight steps such as deciding what to change, dealing with complexity of measurement, and developing self-management skills in PSTs. All of which were designed to integrate the practice of supervision with the measurement of effective teaching. But these were not prescriptive. Siedentop emphasized in the paper that PST should become “students of their own teaching.” That is, they can and should track their own teaching performance and adjust to it. This reflects the notion that teachers should be adaptive (Xie et al., 2021), and adjust their teaching based on the effects of their teaching behavior. In the article, he also advocated for PSTs to engage in a behavior change project, “The student teacher was responsible for deciding what to change, how to observe it, the observations, the implementation of the change, and the verification of its efficacy” (p.34). Notwithstanding these directions and outcomes, one of the most important recommendations in this paper was the development of university training of mentor teachers that began with these teachers having their own teacher behavior systematically observed and then being trained to supervise PSTs.

The cooperating teacher was the responsible agent, providing all instructions, in charge of the goal setting, coding, feedback, and other aspects of supervision. There was no university supervisor assigned. This approach was successful in terms of all of our experimental criteria: i.e., reliability of the data, the magnitude of the changes, the maintenance of changes, etc. (Siedentop, 1981, p. 33).

In this extract, Siedentop lays out a solution to the single largest issue of school-based supervision both then and now. That is a recognition that the person power involved in the supervision is the largest barrier in practicum experiences. The financial costs to the universities and their outcomes reduce the supervisory scope of university supervisors to a cursory effort at best because of the time constraints placed on them relative by the university. This framework was expanded upon by faculty and graduates at Ohio State (e.g., van der Mars et al., 1994; O’Sullivan, 2003) and by researchers around the world in the decades that followed (e.g., Behets and Vergauwen, 2006).

Content knowledge

Siedentop’s (2002b) paper Content knowledge for physical education was an article derived from his 1989 keynote address at the Curriculum and Instructional Academy of the then called American Alliance for Health, Physical Education, Recreation and Dance. It was the latest effort by him to make the case that the content of K-12 physical education should be movement aka behavior. He was further arguing that sport should be a central part of the curriculum. The paper was situated with the ongoing debate regarding disciplinary versus profession debate (Henry, 1964). There are several implications arising from this debate, but the core argument in this article was that teachers of physical education should study how to teach the content that they teach. Siedentop’s argument was that movement is the core subject matter of physical education and sport in particular, should play a central role and that most of what occurs in teacher education under the heading of content knowledge for teaching is less than useful to teachers. He went further articulating that movement was considered by some in the field as unworthy of academic study:

This, then, is the root problem—the direct study of sport skill and strategy through experiential learning is not considered to be of sufficient academic quality to form the core of an undergraduate degree program. Learning basketball, volleyball, and gymnastics—and all the associated issues of training, technique, performance, and strategy—are not worthy of formal academic credit as the central foci of a pre-professional program

(p. 372).

Siedentop went on to specify what that content should look like, referencing technical and tactical development, training implications, and situating the movement in the culture both historically and in the present. He used the dance curriculum at Ohio State to further his case that performance of the content was a core element of teacher education. Siedentop strongly believed in experiential learning in teacher education (Siedentop, 1972, 2002b). Though there have been some challenges to how Siedentop conceptualized the importance of pedagogy and content across his career (O’Sullivan, 1996), and challenges to his emphasis on performance leading to understanding of content knowledge for teachers (Ward, 2006), the paper made a strong stand for the field. It provided a clear response to both the profession and disciplinary debate and the most importantly made clear the central role of content knowledge training in physical education teacher education. Today, his arguments would find a home in the practice-based teacher education movement (Forzani, 2014). His paper served as one foundation for the content knowledge research findings that have made clear his central argument on the importance of content knowledge (Ward et al., 2020).

Conclusion

There is considerably more we could discuss such as role of behavior analysis in Siedentop’s curriculum and instructional model, Sport education (Siedentop et al., 2020). Though the roots of the sport education curriculum were driven by Siedentop’s graduate work on play at the University of Indiana, the instructional framework is grounded in pedagogy including aspects of accountability, how tasks are presented to students, and content knowledge that reflect his behavior analytic perspective. His four editions of his textbook Developing teaching skills in physical education (Siedentop & Tannehill, 2000) framed the knowledge base for physical education, teacher education and successive generations of teacher educators. Our hope is that this article highlights not just Siedentop’s epistemological lens and contributions from a behavior analytic perspective, but also his approach tendencies as a professor that allowed him to be led to good work by the literature, data, and his graduate students.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. There are other behavioral processes such as extinction and punishment, but the selection is driven by reinforcement.