Cross-level effects of class characteristics on students' perceptions of teaching quality.

Abstract

Using the multilevel methodology, this study analyzed 3 dimensions of teaching quality perceived by Chinese students in Hong Kong, China: (a) overall satisfaction with lecturing performance, (b) overall satisfaction with course design, and (c) a self-rated measure of effort devoted to studying. The models capture both student-level and class-level effects within the same analytical framework. Seniors and students majoring in the same subject as the course were found to be more satisfied with course design and to work harder than others. The strongest effect on course ratings, however, came from students' judgment of various aspects of course design and lecturing skills. In addition, larger classes tended to record lower ratings on all outcome measures. In terms of cross-level interaction, students' perceptions of teaching attitude affected their rating of course design and effort in studying more markedly in smaller than in larger classes, and appropriate course content played a more critical role in the rating of lecturing performance in heterogeneous than in homogeneous classes.The transformation from elite institutions to public service providers during the past few decades has put teaching performance at universities increasingly under public scrutiny. As an institutional response to the demand for public accountability, course evaluation has become standard practice in most universities in the United States. Having the appeal of being objective and precise, together with the rise of student consumerism, students' ratings based on a standard set of questions have gained popularity over other evaluative procedures such as self-evaluation, peer review, and the department chair's assessment (Dowd, 1988;Hamilton, 1980;Ory, 1991). Recently, student ratings have spread to the East, where economic prosperity and the rise of democracy have begun to challenge the traditional authority and autonomy of higher education institutions. Although course evaluation is a well-established and thoroughly researched practice in North America, few studies have examined whether it can be applied to other countries, particularly in Asia, where such practice is uncommon (Marsh,

Counseling University Instructors Based on Student Evaluations of Their Teaching Effectiveness: A Multilevel Test of its Effectiveness Under Consideration of Bias and Unfairness Variables

“…When investigating the validity of student assessments the choice is usually the course level; in contrast, studies on the effectiveness of feedback or counseling normally use the instructor level (Marsh 2007a;Marsh and Roche 1997). Examples of the few exceptions on this point include studies conducted by Cranton and Smith (1990), Ting (2000) as well as Wendorf and Alexander (2005). Literature on the analysis of multi-level data has documented a series of problems and potential sources of errors which are associated with the neglect of the multilevel structure of data (see Snijders and Bosker 2002).…”

“…Moreover, working only at the course level cannot adequately control student bias variables (including possible interactions across levels, e.g. between prior interest and course format) and thus diminishes test power (see Ting 2000). In order to avoid these problems, the effects of counseling university instructors should be analyzed by means of multi-level analysis.…”

“…it was assumed that mean course prior knowledge influences teaching quality (Rindermann and Heller 2005). In contrast to the variables considered on the student level, these aggregates are to be interpreted as estimates of course characteristics (see Snijders and Bosker 2002;Ting 2000). Independent of the course means of the student characteristics, it is also presumed that the diversity and/or uniformity of specific withincourse student characteristics can exert an influence on assessments of teaching quality.…”

“…In particular, it is expected that a large degree of diversity in prior knowledge and prior interest among course participants will have a negative influence on teaching efficiency (e.g. Ting 2000). In generating indicators for the diversity of prior knowledge and prior interest, the standard deviations for these two variables were calculated for each class in the study.…”

Cross‐cultural comparisons of consumer satisfaction ratings

“…Larger class sizes in Spain may negatively influence satisfaction (Ting, 2000). Conversely, the small class sizes in Colombia that allow for greater level of interaction between students and faculty members will result in high levels of satisfaction.…”