ERIC Number: ED223692
Record Type: Non-Journal
Publication Date: 1982-Mar
Reference Count: N/A
Matching Factors Derived from the Same Set of Items but Different Groups of Respondents: A Review of Methods with an Example.
Bowers, John J.
Procedures devised for computing factor matching statistics in questionnaire construction are described. In a study of the impact of vocational education on social, affective and nontechnical development of students, an attitude questionnaire to measure outcomes was developed. Each item in the final section had a six-point Likert type response scale ranging from "strongly agree" to "strongly disagree." Using a factor analysis of data from occupational and nonoccupational students' responses, the high-loading items on each factor formed a scale score. Factor analysis brought up the question of whether factor loading patterns for respondent groups were congruent. To evaluate the congruence, Burt's congruence coefficient (the sum of cross-products of the loadings being compared divided by the square root of the cross-product of the sum of the squared loadings) was compared with Cattell's salient variable similarity index (classifying all variables loading on a factor as salient or nonsalient). Application of the methods in comparing and matching factors derived for the two respondent groups on the same variables showed the congruence coefficient to be easier to compute and interpret and a conservative test. The salient index was practical with relatively few variables and factors. (CM)
Publication Type: Speeches/Meeting Papers; Reports - Research
Education Level: N/A
Sponsor: New York State Education Dept., Albany.
Authoring Institution: N/A