Date of Award

Spring 1980

Project Type

Dissertation

Program or Major

Sociology

Degree Name

Doctor of Philosophy

Abstract

The final say decision measure of marital power has fared poorly in several studies which have examined its validity. This research was designed to explore two possible explanations for these findings: (1) inadequate or inappropriate procedures in previous validity studies, and (2) weaknesses of the final say measure itself, including insensitivity to item saliency and the discrepancies in responses between family members.

Several US samples and one from India provided the data for this study. These samples had been gathered in other studies in which husbands', wives' and/or children had responded to the final say decision index. Up to three versions of the final say index were computed for each respondent. In the first version, decision items were unweighted (FSD index). In the other two versions, items were weighted by their relative importance (WFSD index) or by the amount of conflict associated with the particular decision area (CFSD index).

The reliability and validity of these measures and of measures based on the responses of husbands, wives and children was assessed through analytical procedures such as (1) analysis of item-removed alpha coefficients and (2) external criterion correlation analysis. The major findings of the study are: (1) The final say decision measure has cross-cultural validity, evidenced by consistent patterns among the validity coefficients across samples (though the coefficients were generally low). (2) Weighting the final say decision measure by importance or conflict does not improve validity and reliability. (3) The reliability of power measures based on wives' reports is higher than measures using husbands' reports. (4) The validity of measures derived from husbands' and wives' reports is basically equivalent. (5) Although there is some indication that power measures based on children's reports have the highest reliability and validity, the small sample size and other problems make such comparisons tenuous.

The results indicate that the simple unweighted final say measure is a more valid and reliable instrument than previous research has suggested.

Download

Share

COinS