Branching Paths: A Novel Teacher Evaluation Model for Faculty Development

James P. Bavis and Ahn G. Nu

Department of English, Purdue University

ENGL 101: Course Name

Dr. Richard Teeth

Jan. 30, 2020

2

Abstract

A large body of assessment literature suggests that students’ evaluations of their teachers

(SETs) can fail to measure the construct of teaching in a variety of contexts. This can

compromise faculty development efforts that rely on information from SETs. The disconnect

between SET results and faculty development efforts is exacerbated in educational contexts

that demand particular teaching skills that SETs do not value in proportion to their local

importance (or do not measure at all). This paper responds to these challenges by proposing an

instrument for the assessment of teaching that allows institutional stakeholders to define the

teaching construct in a way they determine to suit the local context. The main innovation of this

instrument relative to traditional SETs is that it employs a branching “tree” structure populated

by binary-choice items based on the Empirically derived, Binary-choice, Boundary-definition

(EBB) scale developed by Turner and Upshur for ESL writing assessment. The paper argues

that this structure can allow stakeholders to define the teaching construct by changing the order

and sensitivity of the nodes in the tree of possible outcomes, each of which corresponds to a

specific teaching skill. The paper concludes by outlining a pilot study that will examine the

differences between the proposed EBB instrument and a traditional SET employing series of

multiple-choice questions (MCQs) that correspond to Likert scale values.

Keywords: college teaching, student evaluations of teaching, scale development, EBB

scale, pedagogies, educational assessment, faculty development

3

Branching Paths: A Novel Teacher Evaluation Model for Faculty Development

According to Theall (2017), “Faculty evaluation and development cannot be considered

separately … evaluation without development is punitive, and development without evaluation is

guesswork” (p. 91). As the practices that constitute modern programmatic faculty development

have evolved from their humble beginnings to become a commonplace feature of university life

(Lewis, 1996), a variety of tactics to evaluate the proficiency of teaching faculty for development

purposes have likewise become commonplace. These include measures as diverse as peer

observations, the development of teaching portfolios, and student evaluations.

One such measure, the student evaluation of teacher (SET), has been virtually

ubiquitous since at least the 1990s (Wilson, 1998). Though records of SET-like instruments can

be traced to work at Purdue University in the 1920s (Remmers & Brandenburg, 1927), most

modern histories of faculty development suggest that their rise to widespread popularity went

hand-in-hand with the birth of modern faculty development programs in the 1970s, when

universities began to adopt them in response to student protest movements criticizing

mainstream university curricula and approaches to instruction (Gaff & Simpson, 1994; Lewis,

1996; McKeachie, 1996). By the mid-2000s, researchers had begun to characterize SETs in

terms like “…the predominant measure of university teacher performance […] worldwide”

(Pounder, 2007, p. 178). Today, SETs play an important role in teacher assessment and faculty

development at most universities (Davis, 2009). Recent SET research practically takes the

presence of some form of this assessment on most campuses as a given. Spooren et al.

(2017), for instance, merely note that that SETs can be found at “almost every institution of

higher education throughout the world” (p. 130). Similarly, Darwin (2012) refers to teacher

evaluation as an established orthodoxy, labeling it a “venerated,” “axiomatic” institutional

practice (p. 733).

Moreover, SETs do not only help universities direct their faculty development efforts.

They have also come to occupy a place of considerable institutional importance for their role in

4

personnel considerations, informing important decisions like hiring, firing, tenure, and

promotion. Seldin (1993, as cited in Pounder, 2007) finds that 86% of higher educational

institutions use SETs as important factors in personnel decisions. A 1991 survey of department

chairs found 97% used student evaluations to assess teaching performance (US Department of

Education). Since the mid-late 1990s, a general trend towards comprehensive methods of

teacher evaluation that include multiple forms of assessment has been observed

(Berk, 2005). However, recent research suggests the usage of SETs in personnel decisions is

still overwhelmingly common, though hard percentages are hard to come by, perhaps owing to

the multifaceted nature of these decisions (Boring et al., 2017; Galbraith et al., 2012). In certain

contexts, student evaluations can also have ramifications beyond the level of individual

instructors. Particularly as public schools have experienced pressure in recent decades to adopt

neoliberal, market-based approaches to self-assessment and adopt a student-as-consumer

mindset (Darwin, 2012; Marginson, 2009), information from evaluations can even feature in

department- or school-wide funding decisions (see, for instance, the Obama Administration’s

Race to the Top initiative, which awarded grants to K-12 institutions that adopted value-added

models for teacher evaluation).

However, while SETs play a crucial role in faulty development and personnel decisions

for many education institutions, current approaches to SET administration are not as well-suited

to these purposes as they could be. This paper argues that a formative, empirical approach to

teacher evaluation developed in response to the demands of the local context is better-suited

for helping institutions improve their teachers. It proposes the Heavilon Evaluation of Teacher,

or HET, a new teacher assessment instrument that can strengthen current approaches to

faculty development by making them more responsive to teachers’ local contexts. It also

proposes a pilot study that will clarify the differences between this new instrument and the

Introductory Composition at Purdue (ICaP) SET, a more traditional instrument used for similar

purposes. The results of this study will direct future efforts to refine the proposed instrument.

6

Methods section, which follows, will propose a pilot study that compares the results of the

proposed instrument to the results of a traditional SET (and will also provide necessary

background information on both of these evaluations). The paper will conclude with a discussion

of how the results of the pilot study will inform future iterations of the proposed instrument and,

more broadly, how universities should argue for local development of assessments.

Literature Review

Effective Teaching: A Contextual Construct

The validity of the instrument this paper proposes is contingent on the idea that it is

possible to systematically measure a teacher’s ability to teach. Indeed, the same could be said

for virtually all teacher evaluations. Yet despite the exceeding commonness of SETs and the

faculty development programs that depend on their input, there is little scholarly consensus on

precisely what constitutes “good” or “effective” teaching. It would be impossible to review the

entire history of the debate surrounding teaching effectiveness, owing to its sheer scope—such

a summary might need to begin with, for instance, Cicero and Quintilian. However, a cursory

overview of important recent developments (particularly those revealed in meta-analyses of

empirical studies of teaching) can help situate the instrument this paper proposes in relevant

academic conversations.

Meta-analysis 1. One core assumption that undergirds many of these conversations is

the notion that good teaching has effects that can be observed in terms of student achievement.

A meta-analysis of 167 empirical studies that investigated the effects of various teaching factors

on student achievement (Kyriakides et al., 2013) supported the effectiveness of a set of

teaching factors that the authors group together under the label of the “dynamic model” of

teaching. Seven of the eight factors (Orientation, Structuring, Modeling, Questioning,

Assessment, Time Management, and Classroom as Learning Environment) corresponded to

moderate average effect sizes (of between 0.34–0.41 standard deviations) in measures of

7

student achievement. The eighth factor, Application (defined as seatwork and small-group tasks

oriented toward practice of course concepts), corresponded to only a small yet still significant

effect size of 0.18. The lack of any single decisive factor in the meta-analysis supports the idea

that effective teaching is likely a multivariate construct. However, the authors also note the

context-dependent nature of effective teaching. Application, the least-important teaching factor

overall, proved more important in studies examining young students (p. 148). Modeling, by

contrast, was especially important for older students.

Meta-analysis 2. A different meta-analysis that argues for the importance of factors like

clarity and setting challenging goals (Hattie, 2009) nevertheless also finds that the effect sizes

of various teaching factors can be highly context-dependent. For example, effect sizes for

homework range from 0.15 (a small effect) to 0.64 (a moderately large effect) based on the level

of education examined. Similar ranges are observed for differences in academic subject (e.g.,

math vs. English) and student ability level. As Snook et al. (2009) note in their critical response

to Hattie, while it is possible to produce a figure for the average effect size of a particular

teaching factor, such averages obscure the importance of context.

Meta-analysis 3. A final meta-analysis (Seidel & Shavelson, 2007) found generally

small average effect sizes for most teaching factors—organization and academic domain-

specific learning activities showed the biggest cognitive effects (0.33 and 0.25, respectively).

Here, again, however, effectiveness varied considerably due to contextual factors like domain of

study and level of education in ways that average effect sizes do not indicate.

These pieces of evidence suggest that there are multiple teaching factors that produce

measurable gains in student achievement and that the relative importance of individual factors

can be highly dependent on contextual factors like student identity. This is in line with a well-

documented phenomenon in educational research that complicates attempts to measure

teaching effectiveness purely in terms of student achievement. This is that “the largest source of

variation in student learning is attributable to differences in what students bring to school – their

8

abilities and attitudes, and family and community” (McKenzie et al., 2005, p. 2). Student

achievement varies greatly due to non-teacher factors like socio-economic status and home life

(Snook et al., 2009). This means that, even to the extent that it is possible to observe the

effectiveness of certain teaching behaviors in terms of student achievement, it is difficult to set

generalizable benchmarks or standards for student achievement. Thus is it also difficult to make

true apples-to-apples comparisons about teaching effectiveness between different educational

contexts: due to vast differences between different kinds of students, a notion of what

constitutes highly effective teaching in one context may not in another. This difficulty has

featured in criticism of certain meta-analyses that have purported to make generalizable claims

about what teaching factors produce the biggest effects (Hattie, 2009). A variety of other

commentators have also made similar claims about the importance of contextual factors in

teaching effectiveness for decades (see, e.g., Bloom et al., 1956; Cashin, 1990; Theall, 2017).

The studies described above mainly measure teaching effectiveness in terms of

academic achievement. It should certainly be noted that these quantifiable measures are not

generally regarded as the only outcomes of effective teaching worth pursuing. Qualitative

outcomes like increased affinity for learning and greater sense of self-efficacy are also important

learning goals. Here, also, local context plays a large role.

SETs: Imperfect Measures of Teaching

As noted in this paper’s introduction, SETs are commonly used to assess teaching

performance and inform faculty development efforts. Typically, these take the form of an end-of-

term summative evaluation comprised of multiple-choice questions (MCQs) that allow students

to rate statements about their teachers on Likert scales. These are often accompanied with

short-answer responses which may or may not be optional.

SETs serve important institutional purposes. While commentators have noted that there

are crucial aspects of instruction that students are not equipped to judge (Benton & Young,

2018), SETs nevertheless give students a rare institutional voice. They represent an opportunity

9

to offer anonymous feedback on their teaching experience and potentially address what they

deem to be their teacher’s successes or failures. Students are also uniquely positioned to offer

meaningful feedback on an instructors’ teaching because they typically have much more

extensive firsthand experience of it than any other educational stakeholder. Even peer

observers only witness a small fraction of the instructional sessions during a given semester.

Students with perfect attendance, by contrast, witness all of them. Thus, in a certain sense, a

student can theoretically assess a teacher’s ability more authoritatively than even peer mentors

can.

While historical attempts to validate SETs have produced mixed results, some studies

have demonstrated their promise. Howard (1985), for instance, finds that SET are significantly

more predictive of teaching effectiveness than self-report, peer, and trained-observer

assessments. A review of several decades of literature on teaching evaluations (Watchel, 1998)

found that a majority of researchers believe SETs to be generally valid and reliable, despite

occasional misgivings. This review notes that even scholars who support SETs frequently argue

that they alone cannot direct efforts to improve teaching and that multiple avenues of feedback

are necessary (L’hommedieu et al., 1990; Seldin, 1993).

Finally, SETs also serve purposes secondary to the ostensible goal of improving

instruction that nonetheless matter. They can be used to bolster faculty CVs and assign

departmental awards, for instance. SETs can also provide valuable information unrelated to

teaching. It would be hard to argue that it not is useful for a teacher to learn, for example, that a

student finds the class unbearably boring, or that a student finds the teacher’s personality so

unpleasant as to hinder her learning. In short, there is real value in understanding students’

affective experience of a particular class, even in cases when that value does not necessarily

lend itself to firm conclusions about the teacher’s professional abilities.

However, a wealth of scholarly research has demonstrated that SETs are prone to fail in

certain contexts. A common criticism is that SETs can frequently be confounded by factors

10

external to the teaching construct. The best introduction to the research that serves as the basis

for this claim is probably Neath (1996), who performs something of a meta-analysis by

presenting these external confounds in the form of twenty sarcastic suggestions to teaching

faculty. Among these are the instructions to “grade leniently,” “administer ratings before tests”

(p. 1365), and “not teach required courses” (#11) (p. 1367). Most of Neath’s advice reflects an

overriding observation that teaching evaluations tend to document students’ affective feelings

toward a class, rather than their teachers’ abilities, even when the evaluations explicitly ask

students to judge the latter.

Beyond Neath, much of the available research paints a similar picture. For example, a

study of over 30,000 economics students concluded that “the poorer the student considered his

teacher to be [on an SET], the more economics he understood” (Attiyeh & Lumsden, 1972). A

1998 meta-analysis argued that “there is no evidence that the use of teacher ratings improves

learning in the long run” (Armstrong, 1998, p. 1223). A 2010 National Bureau of Economic

Research study found that high SET scores for a course’s instructor correlated with “high

contemporaneous course achievement,” but “low follow-on achievement” (in other words, the

students would tend to do well in the course, but poor in future courses in the same field of

study. Others observing this effect have suggested SETs reward a pandering, “soft-ball”

teaching style in the initial course (Carrell & West, 2010). More recent research suggests that

course topic can have a significant effect on SET scores as well: teachers of “quantitative

courses” (i.e., math-focused classes) tend to receive lower evaluations from students than their

humanities peers (Uttl & Smibert, 2017).

Several modern SET studies have also demonstrated bias on the basis of gender

(Anderson & Miller, 1997; Basow, 1995), physical appearance/sexiness (Ambady & Rosenthal,

1993), and other identity markers that do not affect teaching quality. Gender, in particular, has

attracted significant attention. One recent study examined two online classes: one in which

instructors identified themselves to students as male, and another in which they identified as

11

female (regardless of the instructor’s actual gender) (Macnell et al., 2015). The classes were

identical in structure and content, and the instructors’ true identities were concealed from

students. The study found that students rated the male identity higher on average. However, a

few studies have demonstrated the reverse of the gender bias mentioned above (that is, women

received higher scores) (Bachen et al., 1999) while others have registered no gender bias one

way or another (Centra & Gaubatz, 2000).

The goal of presenting these criticisms is not necessarily to diminish the institutional

importance of SETs. Of course, insofar as institutions value the instruction of their students, it is

important that those students have some say in the content and character of that instruction.

Rather, the goal here is simply to demonstrate that using SETs for faculty development

purposes—much less for personnel decisions—can present problems. It is also to make the

case that, despite the abundance of literature on SETs, there is still plenty of room for scholarly

attempts to make these instruments more useful.

Empirical Scales and Locally-Relevant Evaluation

One way to ensure that teaching assessments are more responsive to the demands of

teachers’ local contexts is to develop those assessments locally, ideally via a process that

involves the input of a variety of local stakeholders. Here, writing assessment literature offers a

promising path forward: empirical scale development, the process of structuring and calibrating

instruments in response to local input and data (e.g., in the context of writing assessment,

student writing samples and performance information). This practice contrasts, for instance, with

deductive approaches to scale development that attempt to represent predetermined theoretical

constructs so that results can be generalized.

Supporters of the empirical process argue that empirical scales have several

advantages. They are frequently posited as potential solutions to well-documented reliability and

validity issues that can occur with theoretical or intuitive scale development (Brindley, 1998;

Turner & Upshur, 1995, 2002). Empirical scales can also help researchers avoid issues caused

12

by subjective or vaguely-worded standards in other kinds of scales (Brindley, 1998) because

they require buy-in from local stakeholders who must agree on these standards based on

their understanding of the local context. Fulcher et al. (2011) note the following, for instance:

Measurement-driven scales suffer from descriptional inadequacy. They are not sensitive

to the communicative context or the interactional complexities of language use. The level

of abstraction is too great, creating a gulf between the score and its meaning. Only with

a richer description of contextually based performance, can we strengthen the meaning

of the score, and hence the validity of score-based inferences. (pp. 8–9)

There is also some evidence that the branching structure of the EBB scale specifically

can allow for more reliable and valid assessments, even if it is typically easier to calibrate and

use conventional scales (Hirai & Koizumi, 2013). Finally, scholars have also argued that

theory-based approaches to scale development do not always result in instruments that

realistically capture ordinary classroom situations (Knoch, 2007, 2009).

The most prevalent criticism of empirical scale development in the literature is that the

local, contingent nature of empirical scales basically discards any notion of their results’

generalizability. Fulcher (2003), for instance, makes this basic criticism of the EBB scale even

as he subsequently argues that “the explicitness of the design methodology for EBBs is

impressive, and their usefulness in pedagogic settings is attractive” (p. 107). In the context of

this particular paper’s aims, there is also the fact that the literature supporting empirical scale

development originates in the field of writing assessment, rather than teaching assessment.

Moreover, there is little extant research into the applications of empirical scale development for

the latter purpose. Thus, there is no guarantee that the benefits of empirical development

approaches can be realized in the realm of teaching assessment. There is also no guarantee

that they cannot. In taking a tentative step towards a better understanding of how these

assessment schema function in a new context, then, the study described in the next section

13

asks whether the principles that guide some of the most promising practices for assessing

students cannot be put to productive use in assessing teachers.

Materials and Methods

This section proposes a pilot study that will compare the ICaP SET to the Heavilon

Evaluation of Teacher (HET), an instrument designed to combat the statistical ceiling effect

described above. In this section, the format and composition of the HET is described, with

special attention paid to its branching scale design. Following this, the procedure for the study is

outlined, and planned interpretations of the data are discussed.

The Purdue ICaP SET

The SET employed by Introductory Composition at Purdue (ICaP) program as of

January 2019 serves as an example of many of the prevailing trends in current SET

administration. The evaluation is administered digitally: ICaP students receive an invitation to

complete the evaluation via email near the end of the semester, and must complete it before

finals week (i.e., the week that follows the normal sixteen-week term) for their responses to be

counted. The evaluation is entirely optional: teachers may not require their students to complete

it, nor may they offer incentives like extra credit as motivation. However, some instructors opt to

devote a small amount of in-class time for the evaluations. In these cases, it is common practice

for instructors to leave the room so as not to coerce high scores.

The ICaP SET mostly takes the form of a simple multiple-choice survey. Thirty-four

MCQs appear on the survey. Of these, the first four relate to demographics: students must

indicate their year of instruction, their expected grade, their area of study, and whether they are

taking the course as a requirement or as an elective. Following these are two questions related

to the overall quality of the course and the instructor (students must rate each from “very poor”

to “excellent” on a five-point scale). These are “university core” questions that must appear on

every SET administered at Purdue, regardless of school, major, or course. The Students are

14

also invited to respond to two short-answer prompts: “What specific suggestions do you have for

improving the course or the way it is taught?” and “what is something that the professor does

well?” Responses to these questions are optional.

The remainder of the MCQs (thirty in total) are chosen from a list of 646 possible

questions provided by the Purdue Instructor Course Evaluation Service (PICES) by department

administrators. Each of these PICES questions requires students to respond to a statement

about the course on a five-point Likert scale. Likert scales are simple scales used to indicate

degrees of agreement. In the case of the ICaP SET, students must indicate whether they

strongly agree, agree, disagree, strongly disagree, or are undecided. These thirty Likert scale

questions assess a wide variety of the course and instructor’s qualities. Examples include “My

instructor seems well-prepared for class,” “This course helps me analyze my own and other

students’ writing,” and “When I have a question or comment I know it will be respected,” for

example.

One important consequence of the ICaP SET within the Purdue English department is

the Excellence in Teaching Award (which, prior to Fall 2018, was named the Quintilian or,

colloquially, “Q” Award). This is a symbolic prize given every semester to graduate instructors

who score highly on their evaluations. According to the ICaP site, “ICaP instructors whose

teaching evaluations achieve a certain threshold earn [the award], recognizing the top 10% of

teaching evaluations at Purdue.” While this description is misleading—the award actually goes

to instructors whose SET scores rank in the top decile in the range of possible outcomes, but

not necessarily ones who scored better than 90% of other instructors—the award nevertheless

provides an opportunity for departmental instructors to distinguish their CVs and teaching

portfolios.

Insofar as it is distributed digitally, it is composed of MCQs (plus a few short-answer

responses), and it is intended as end-of-term summative assessment, the ICaP SET embodies

15

the current prevailing trends in university-level SET administration. In this pilot study, it serves

as a stand-in for current SET administration practices (as generally conceived).

The HET

Like the ICaP SET, the HET uses student responses to questions to produce a score

that purports to represent their teacher’s pedagogical ability. It has a similar number of items

(28, as opposed to the ICaP SET’s 34). However, despite these superficial similarities, the

instrument’s structure and content differ substantially from the ICaP SET’s.

The most notable differences are the construction of the items on the text and the way

that responses to these items determine the teacher’s final score. Items on the HET do not use

the typical Likert scale, but instead prompt students to respond to a question with a simple

“yes/no” binary choice. By answering “yes” and “no” to these questions, student responders

navigate a branching “tree” map of possibilities whose endpoints correspond to points on a 33-

point ordinal scale.

The items on the HET are grouped into six suites according to their relevance to six

different aspects of the teaching construct (described below). The suites of questions

correspond to directional nodes on the scale—branching paths where an instructor can move

either “up” or “down” based on the student’s responses. If a student awards a set number of

“yes” responses to questions in a given suite (signifying a positive perception of the instructor’s

teaching), the instructor moves up on the scale. If a student does not award enough “yes”

responses, the instructor moves down. Thus, after the student has answered all of the

questions, the instructor’s “end position” on the branching tree of possibilities corresponds to a

point on the 33-point scale. A visualization of this structure is presented in Table 1.

16

Figure 1

Illustration of HET’s Branching Structure

Note. Each node in this diagram corresponds to a suite of HET/ICALT items, rather than to a single item.

The questions on the HET derive from the International Comparative Analysis of

Learning and Teaching (ICALT), an instrument that measures observable teaching behaviors for

17

the purpose of international pedagogical research within the European Union. The most recent

version of the ICALT contains 32 items across six topic domains that correspond to six broad

teaching skills. For each item, students rate a statement about the teacher on a four-point Likert

scale. The main advantage of using ICALT items in the HET is that they have been

independently tested for reliability and validity numerous times over 17 years of development

(see, e.g., Van de Grift, 2007). Thus, their results lend themselves to meaningful comparisons

between teachers (as well as providing administrators a reasonable level of confidence in their

ability to model the teaching construct itself).

The six “suites” of questions on the HET, which correspond to the six topic domains on

the ICALT, are presented in Table 1.

Table 1

HET Question Suites

Suite # of Items Description

Safe learning environment 4 Whether the teacher is able to

maintain positive, nonthreatening

relationships with students (and to

foster these sorts of relationships

among students).

Classroom management 4 Whether the teacher is able to

maintain an orderly, predictable

environment.

Clear instruction 7 Whether the teacher is able to

explain class topics

comprehensibly, provide clear sets

of goals for assignments, and

articulate the connections between

the assignments and the class

topics in helpful ways.

18

Suite # of Items Description

Activating teaching methods 7 Whether the teacher uses strategies

that motivate students to think about

the class’s topics.

Learning strategies 6 Whether teachers take explicit steps

to teach students how to learn (as

opposed to merely providing

students informational content).

Differentiation 4 Whether teachers can successfully

adjust their behavior to meet the

diverse learning needs of individual

students.

Note. Item numbers are derived from original ICALT item suites.

The items on the HET are modified from the ICALT items only insofar as they are phrased

as binary choices, rather than as invitations to rate the teacher. Usually, this means the addition

of the word “does” and a question mark at the end of the sentence. For example, the second

safe learning climate item on the ICALT is presented as “The teacher maintains a relaxed

atmosphere.” On the HET, this item is rephrased as, “Does the teacher maintain a relaxed

atmosphere?” See Appendix for additional sample items.

As will be discussed below, the ordering of item suites plays a decisive role in the teacher’s

final score because the branching scale rates earlier suites more powerfully. So too does the

“sensitivity” of each suite of items (i.e., the number of positive responses required to progress

upward at each branching node). This means that it is important for local stakeholders to

participate in the development of the scale. In other words, these stakeholders must be involved

in decisions about how to order the item suites and adjust the sensitivity of each node. This is

described in more detail below.

Once the scale has been developed, the assessment has been administered, and the

teacher’s endpoint score has been obtained, the student rater is prompted to offer any textual

19

feedback that s/he feels summarizes the course experience, good or bad. Like the short

response items in the ICaP SET, this item is optional. The short-response item is as follows:

• What would you say about this instructor, good or bad, to another student considering

taking this course?

The final four items are demographic questions. For these, students indicate their grade

level, their expected grade for the course, their school/college (e.g., College of Liberal Arts,

School of Agriculture, etc.), and whether they are taking the course as an elective or as a

degree requirement. These questions are identical to the demographic items on the ICaP SET.

To summarize, the items on the HET are presented as follows:

• Branching binary questions (32 different items; six branches)

o These questions provide the teacher’s numerical score

• Short response prompt (one item)

• Demographic questions (four items)

Scoring

The main data for this instrument are derived from the endpoints on a branching ordinal

scale with 33 points. Because each question is presented as a binary yes/no choice (with “yes”

suggesting a better teacher), and because paths on the branching scale are decided in terms of

whether the teacher receives all “yes” responses in a given suite, 32 possible outcomes are

possible from the first five suites of items. For example, the worst possible outcome would be

five successive “down” branches, the second-worst possible outcome would be four “down”

branches followed by an “up,” and so on. The sixth suite is a tie-breaker: instructors receive a

single additional point if they receive all “yes” responses on this suite.

By positioning certain suites of items early in the branching sequence, the HET gives

them more weight. For example, the first suite is the most important of all: an “up” here

automatically places the teacher above 16 on the scale, while a “down” precludes all scores

26

References

Ambady, N., & Rosenthal, R. (1993). Half a minute: Predicting teacher evaluations from thin

faculty in US higher ed. AAUP Updates. https://www.aaup.org/news/data-snapshot-

contingent-faculty-us-higher-ed#.Xfpdmy2ZNR4

of Educational Psychology, 87(4), 656–665. http://dx.doi.org/10.1037/0022-

0663.87.4.656

Becker, W. (2000). Teaching economics in the 21st century. Journal of Economic Perspectives,

14(1), 109–120. http://dx.doi.org/10.1257/jep.14.1.109

Benton, S., & Young, S. (2018). Best practices in the evaluation of teaching. Idea paper, 69.

American Association of University Professors. (2018, October 11). Data snapshot: Contingent

Anderson, K., & Miller, E. D. (1997). Gender and student evaluations of teaching. PS: Political

Science and Politics, 30(2), 216–219. https://doi.org/10.2307/420499

Armstrong, J. S. (1998). Are student ratings of instruction useful? American Psychologist,

53(11), 1223–1224. http://dx.doi.org/10.1037/0003-066X.53.11.1223

Attiyeh, R., & Lumsden, K. G. (1972). Some modern myths in teaching economics: The U.K.

experience. American Economic Review, 62(1), 429–443.

https://www.jstor.org/stable/1821578

Bachen, C. M., McLoughlin, M. M., & Garcia, S. S. (1999). Assessing the role of gender in

college students’ evaluations of faculty. Communication Education, 48(3), 193–210.

http://doi.org/cqcgsr

Basow, S. A. (1995). Student evaluations of college professors: When gender matters. Journal

slices of nonverbal behavior and physical attractiveness. Journal of Personality and

Social Psychology, 64(3), 431–441. http://dx.doi.org/10.1037/0022-3514.64.3.431

American Association of University Professors. (n.d.) Background facts on contingent faculty

positions. https://www.aaup.org/issues/contingency/background-facts

27

Berk, R. A. (2005). Survey of 12 strategies to measure teaching effectiveness. International

Bloom, B. S., Englehart, M. D., Furst, E. J., Hill, W. H., & Krathwohl, D. R. (1956). Taxonomy of

educational objectives: The classification of educational goals. Addison-Wesley

Longman Ltd.

Brandenburg, D., Slinde, C., & Batista, J. (1977). Student ratings of instruction: Validity and

normative interpretations. Research in Higher Education, 7(1), 67–78.

http://dx.doi.org/10.1007/BF00991945

Carrell, S., & West, J. (2010). Does professor quality matter? Evidence from random

assignment of students to professors. Journal of Political Economy, 118(3), 409–432.

https://doi.org/10.1086/653808

Cashin, W. E. (1990). Students do rate different academic fields differently. In M. Theall, & J. L.

Franklin (Eds.), Student ratings of instruction: Issues for improving practice. New

Directions for Teaching and Learning (pp. 113–121).

Centra, J., & Gaubatz, N. (2000). Is there gender bias in student evaluations of

teaching? The Journal of Higher Education, 71(1), 17–33.

https://doi.org/10.1080/00221546.2000.11780814

Davis, B. G. (2009). Tools for teaching (2nd ed.). Jossey-Bass.

Denton, D. (2013). Responding to edTPA: Transforming practice or applying

shortcuts? AILACTE Journal, 10(1), 19–36.

Dizney, H., & Brickell, J. (1984). Effects of administrative scheduling and directions upon

student ratings of instruction. Contemporary Educational Psychology, 9(1), 1–7.

https://doi.org/10.1016/0361-476X(84)90001-8

DuCette, J., & Kenney, J. (1982). Do grading standards affect student evaluations of teaching?

Some new evidence on an old question. Journal of Educational Psychology, 74(3), 308–

314. https://doi.org/10.1037/0022-0663.74.3.308

Journal of Teaching and Learning in Higher Education, 17(1), 48–62.

28

Edwards, J. E., & Waters, L. K. (1984). Halo and leniency control in ratings as influenced by

format, training, and rater characteristic differences. Managerial Psychology, 5(1), 1–16.

Fink, L. D. (2013). The current status of faculty development internationally. International

Journal for the Scholarship of Teaching and Learning, 7(2).

https://doi.org/10.20429/ijsotl.2013.070204

Fulcher, G. (2003). Testing second language speaking. Pearson Education.

Fulcher, G., Davidson, F., & Kemp, J. (2011). Effective rating scale development for speaking

tests: Performance decision trees. Language Testing, 28(1), 5–29.

https://doi.org/10.1177/0265532209359514

Gaff, J. G., & Simpson, R. D. (1994). Faculty development in the United States. Innovative

Higher Education, 18(3), 167–76. https://doi.org/10.1007/BF01191111

Hattie, J. (2008). Visible learning: A synthesis of over 800 meta-analyses relating to

achievement. Routledge.

Hoffman, R. A. (1983). Grade inflation and student evaluations of college courses. Educational

and Psychological Research, 3(3), 51–160. https://doi.org/10.1023/A:101557981

Howard, G., Conway, C., & Maxwell, S. (1985). Construct validity of measures of college

teaching effectiveness. Journal of Educational Psychology, 77(2), 187–96.

Kane, M. T. (2013) Validating interpretations and uses of test scores. Journal of Educational

Measurement, 50(1), 1–73.

Kelley, T. (1927) Interpretation of educational measurements. World Book Co.

Knoch, U. (2007). Do empirically developed rating scales function differently to conventional

rating scales for academic writing? Spaan Fellow Working Papers in Second or Foreign

Language Assessment, 5, 1–36. English Language Institute, University of Michigan.

Knoch, U. (2009). Diagnostic assessment of writing: A comparison of two rating

scales. Language Testing, 26(2), 275-304.

http://dx.doi.org/10.1037/0022-0663.77.2.187

32

Appendix

Sample ICALT Items Rephrased for HET

Suite Sample ICALT Item HET Phrasing

Safe learning environment The teacher promotes mutual

respect.

Does the teacher promote mutual

respect?

Classroom management The teacher uses learning time

efficiently.

Does the teacher use learning time

efficiently?

Clear instruction The teacher gives feedback to

pupils.

Does the teacher give feedback to

pupils?

Activating teaching methods The teacher provides interactive

instruction and activities.

Does the teacher provide interactive

instruction and activities?

Learning strategies The teacher provides interactive

instruction and activities.

Does the teacher provide interactive

instruction and activities?

Differentiation The teacher adapts the instruction

to the relevant differences between

pupils.

Does the teacher adapt the

instruction to the relevant

differences between pupils?