Academic journal article Fuzzy Economic Review

Advantages of Using Self-Organizing Maps to Analyse Student Evaluations of Teaching

Academic journal article Fuzzy Economic Review

Advantages of Using Self-Organizing Maps to Analyse Student Evaluations of Teaching

Article excerpt

Surveys to evaluate teaching performance are one of the most widely used instruments for assessing teaching capabilities and, consequently, the quality of teaching. Their success is largely due to how they are designed and the way in which they process information. The aim of this paper is to simplify the design of the student evaluations by removing the most correlated items, and to propose that Kohonen's self-organizing Kohonen maps be used to group teachers in accordance with all the characteristics surveyed. The methodology is applied to the particular case of the Rovira i Virgili University.

Keywords: surveys to evaluate teaching performance, higher education, kohonen's selforganizing maps

JEL Classification: I21, I23, C45

(ProQuest: ... denotes formulae omitted.)


Surveys to evaluate teaching performance were first used in America in the 1920s. One of the first publications based on the results of these surveys can be found in Remmers and Brandenburg (1927). They have subsequently been used in most universities in the world. Kulik (2001) notes that in 1973, 29% of American universities used student evaluations as a tool to assess teaching, while in 1993, they were used in 86%.

Student evaluations have two types of users: teachers and educational institutions. In this respect, Catano and Harvey (2011) point out that surveys are often used to provide remunerative supplements, to decide on continuity of teachers and to support faculty promotion decisions.

Theall and Franklin (2001) state that teaching staffare extremely sensitive to this issue. Some teachers, especially new ones, consider evaluations to be positive as they help improve the effectiveness of teaching. Kulik (2001), however, notes that some teachers are reluctant because they fear that students can turn them into a sort of personality contest.

Student evaluations can also be understood as a form of accountability to society on the efficiency and effectiveness of teaching conducted by a university.

Among the existing literature, Arthur (2009) has studied the validity of using the views of students as a means of evaluating the performance of a teacher. Other studies analyze the factors that affect the evaluation. For example, Remedios and Lieberman (2007) studied the impact on outcomes of qualifications and workload during the course, and Griffin (2004) studied the effect of gender.

Although the content of evaluations is open to criticism and some items are of questionable relevance, Penny (2003) argues that universities continue to use them as a central element in assessing the effectiveness of teaching. It is therefore important that results be interpreted correctly.

This study has two objectives: first, to consider whether student evaluations of teaching can be simplified by reducing the number of items; and, second, to examine whether overall evaluations take into account all those aspects of teaching that need to be measured.

Although this study has been carried out in the particular case of the Rovira i Virgili University (URV), the methodology and the procedure can be generalized to any type of questionnaire.

The paper is organized as follows. Section 2 describes introduces the main methodological tool used, unsupervised Kohonen's self-organizing maps, a particular type of neural network. Section 3 describes the data used and shows the results obtained from the application of the Kohonen maps. Section 4 states presents the conclusions and there are some annexes at the end.


To evaluate the data obtained in the survey we used a particular type of artificial neural network called a Kohonen's self-organizing map (SOM) (1989), which contains all the features of overall teacher evaluation and does not limit its analysis to the mean of the valuation (or average and standard deviation).

These networks were first developed by Dr. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.