Tímarit um menntarannsóknir - 01.01.2007, Blaðsíða 37

35
Tímarit um menntarannsóknir, 4. árgangur 2007
Þáttabygging kennslukönnunar við Háskóla Íslands
Method
Subjects. Individual student ratings and
course means from 614 undergraduate
courses at the University of Iceland, taught
by 728 teachers, were both used as units
of analysis. A total of 12,282 electronically
administered questionnaires were answered
by undergraduate students. The number of
returned questionnaires was higher than the
number of students since students answered
the questionnaires for all courses they were
registered in. In some courses two or more
teachers were evaluated. A total of 61% of the
courses were evaluated once, 24% twice and
15% three times or more. Most teachers (66%)
were evaluated once, 20% in two courses
and 14% in three to six courses. Prior to data
analysis courses with missing data on teachers
as well as courses with fewer than five students
were excluded. In cases where the same teacher
was evaluated in two or more courses one
course was randomly selected for each teacher.
Thus, in the complete data set each teacher was
only evaluated once. The complete data set
consisted of 9,075 questionnaires, 518 courses
that were taught by 518 teachers.
Instrument. A total of 37 variables are in
the data set. Of those, 25 variables are used to
evaluate various aspects of teaching, teachers
and courses, 10 are background variables
on students´ and courses characteristics, and
2 variables contain written comments about
teachers, courses and curriculum.
Data analysis. First, individual student
ratings were analysed in two random samples
of the data set. The purpose of these analyses
was to cross-validate the factor solution in
one sample with that in another independent
sample. Following this analysis the data was
factor analysed for each of 8 faculties. The
unit of analysis was individual student ratings.
Finally, the course means were factor analysed.
Principal axes factor analysis was used to
explore the factor structure of student ratings
and parallel analysis to determine the number
of factors in all analyses.
Results
Three factors, with identical content, emerged
in the two samples. The three factors explained
59% (sample A) and 60% (sample B) of the total
variance of 21 questions in the questionnaire.
Factor analysis of class means resulted in
three factors with the same content as the three
factors that resulted from factor analyses of
individual students’ ratings. The three factors
explained 71% of the total variance of the
21 questions. Factor analysis of individual
students’ ratings in eight university faculties
resulted in three factors in all faculties with
the same content as in previous analyses. The
three factors explained 59.21% (Faculty of
Law) to 66.97% (Faculty of Nursing) of the
21 questions´ total variance. The reliability
of the three factors is satisfactory, alpha
coefficients ranging from 0.81 to 0.97. The
main result of this study is that student ratings
of undergraduate instruction at the University
of Iceland consists of three stable factors. The
first factor, teaching, consists of 14 questions
on teaching, course organization and planning,
student teacher interaction, grading and
examination and student learning. The second
factor consists of five questions on laboratory
facilities. The third factor, workload, consists
of two questions on course workload and
difficulty.
Discussion
The results indicate that the instrument
used at the University of Iceland to evaluate
instructional quality should be interpreted in
terms of three factors rather than individual
questions. The factor structure of the instrument
is stable in the sense that the same factors
emerge in two independent samples and eight
faculties. In addition, the same three factors
emerge with different units of analysis, i.e.
individual student ratings and course means.