Faculty Evaluations: An Alternative Approach Based on Classroom Observations

Department chairs and deans face many challenges in their roles. One of the most difficult is the evaluation of faculty regarding teaching effectiveness. This is particularly challenging for two reasons: (1) lack of formal preparation for instructors concerning teaching, and (2) limited choice of evaluation tools. One tool, classroom observation, can help address both of these issues and provide an objective measure of teaching effectiveness.

Classroom observation—either for the purpose of promotion and tenure or through the assistance of a teaching and learning center for those faculty seeking self-improvement in instruction—is not that common of a methodology for helping higher education instructors with information that they might use to improve their teaching. This is due, in part, to the infrequency of classroom visits by either colleagues or supervisors, the general unfamiliarity with what measures could be used to evaluate effective face-to-face instruction, and the time involved in documenting an observation.

However, one classroom observation tool, Flanders’ Interaction Analysis, can minimize the potential drawbacks and objections because it focuses on one aspect of a classroom: the interaction between students and their instructor. The focus on interaction as one aspect of an effective classroom and the identification of a set of predefined variables to measure interactions makes the Flanders model a viable method for analyzing teaching.

Hilda Taba, Edmund Amidon, and J.B. Hough were among the first proponents of examining classroom interactions as a way of describing what occurred during the teaching and learning process. Working with Amidon, Ned Flanders developed a system of interaction analysis to give a picture of what occurred during opportunities to observe classroom activities. This system was presented in The Role of the Teacher in the Classroom (Amidon & Flanders, 1967).

While the work of Flanders and others was designed for the precollegiate arena, it has applicability for the collegiate classroom. One is able to observe the behavior (Flanders calls it “talk”) of the instructor as being direct (lecturing, giving directions, or criticizing or justifying authority) or indirect (accepting feelings, praising or encouraging, accepting or using ideas of students, or asking questions). Student “talk” is categorized as either response (usually a convergent answer to a question posed by the teacher) or initiation (a divergent question or statement that departs from the flow established by the instructor). Flanders also allowed that some activity may be seen as confusion, with different communication occurring simultaneously, or silence.

To make use of these observations, codes are recorded every three to five seconds by the observer. The juncture between two coded responses is considered the interaction. For example, if a teacher asks a question (4) and a student responds appropriately (8), this would be called a 4-8 interaction. When the observation period (a full or partial teaching session) is completed, the interactions are entered into a matrix. This matrix can be examined for how much of the time the teacher talked (with direct or indirect influence) and how much the time students talked (either convergently or divergently). This data can then be used for a conference between the instructor and the observer.

In addition to providing department chairs and directors of centers for teaching and learning with an objective evaluation tool for assessing teaching, research conducted by Flanders and others indicates that the use of the tool by instructors for self-improvement purposes results in improved teaching. Many articles concerning the Flanders Interaction Analysis focus on the usefulness of the tool to provide immediate feedback to instructors regarding their interactions, but also report that it is easy to use, understand, and implement.

Recently, a pilot study was conducted at Central Michigan University to assess the feasibility of using Flanders’ Interaction Analysis system in providing faculty and administrators with better feedback regarding teaching effectiveness. After approximately 10 hours of training on use of the tool, the pilot study was conducted in the observation of university faculty. Three independent observers reviewed video recordings of several classroom sessions to verify consistent coding. The inter-rater reliability was very high (>.90) between the video analysis as well as between the video analysis and the data collected during the live classroom sessions. The next phase of this study includes establishing a debriefing format and obtaining usefulness and satisfaction feedback from instructors.

Michael B. Gilbert is professor and chairperson of the Department of Educational Leadership at Central Michigan University. Alicia Haley is a doctoral student and graduate research assistant in the same department.

Excerpted from Filling a Need: Classroom Observations for Higher Education Faculty, January 2009, Academic Leader.