THE UTILITY OF STUDENT SELF‐RATINGS AND TEACHER RATINGS FOR ASSESSING STUDENTS’ GENERALIZABLE SKILLS

Abstract

A research design procedure was developed to test for the construct validity of the student self‐rating and teacher rating generalizable skills instruments developed by Greenan and Smith (1981). Construct validity was operationally defined as the sensitivity of the student self‐ratings and teacher ratings to (1) the type of vocational program, (2) sex, and (3) standardized mathematics and reading aptitude test scores of each student. A purposive sample of five vocational programs was selected from a postsecondary vocational school: auto mechanics and machinist (male‐dominated), office occupations and interior design (female‐dominated), and accounting (balanced). The instruments asked students to rate themselves and teachers to rate their students’ generalizable skills in mathematics, communications, interpersonal relations, and reasoning on a five‐point Likert‐type scale based on “degree of proficiency.” Analysis of variance was used to test for significant differences on the student self‐ratings and teacher ratings. In general, the results of the study suggested that the mean score ratings of students and teachers on the total generalizable skills instruments were statistically significant for students in male‐dominated, female‐dominated, and the vocational programs balanced by sex distribution. Thus we conclude that male students rated their generalizable skills consistently lower than females or students in the program with sex balance. In addition, the generalizable mathematics scale can produce ratings that are sensitive to differences in programs and consistent with students known to possess high, average, and low aptitudes in mathematics. The student self‐ratings and/or teacher ratings of the generalizable communications skills of students are not highly sensitive to differences by type of vocational program as defined by sex enrollment or by levels of reading aptitude. The construct validity of these ratings is therefore questionable since they measure a construct seemingly different from reading aptitude.