In lieu of an abstract, here is a brief excerpt of the content:

Appendix 3 Statistical Tests Applied the Study • III Reliability of a scale refers to its consistency of performance internally, item by item, which should not fluctuate as a result of random error or chance factors. Reliabilities above .60 are acceptable. ANOVA is a short term for analysis of variance, a test which assesses significant differences when more than one comparison is made. Variance is the sum of standard deviation squared and provides information on the spread of the scores. The F test compares the spread of different distributions (for example adjusted and non-adjusted spouses) to determine whether the two differ significantly on particular variables. Correlation matrix measures specifically the extend of association between variables. The DAS subscale affectional expression correlates with consensus subscale at .69 p<.OOl. Thus scores from affectional expression is able to explain consensus at .69 squared = 48%. Significance at p<.OOl indicate that the role of chance factors is one in a thousand. Regressive equation estimates or predicts the score on a variable from, a known score of another variable. Logistic regression analysis demonstrate the extent scores on a particular item, or subscale or scale can estimate or predict the group membership of spouses who are adjusted or who have considered divorce. / Thus in Table A2.4, the three DAS subscales - consensus, cohesion and satisfaction - are able to correctly classify 107, and misclassify 7 of the adjusted spouses, and correctly classifY 83, and misclassify 7 of the non-adjusted spouses. The overall percentage of spouses correctly classified is 93.14%. Appendix 3 185 Chi-square (X2) is a measure of differences in frequency or association. The higher the value of X2, the more likely the distribution differs. The degree of freedom, df, indicates the number of cells free to vary in frequency. Factor analysis clusters items with certain properties (i.e. homogeneity) into factor groups which reflects particular constructs. Factor loading denotes the correlation between the item and the construct. The higher the factor loading, the higher its power to explain that construct. Variance indicates the percentage the total variance (total individual differences) explained by the factors extracted. ...

Share