[管理学]Measurement and Construct Validation.pptVIP

* EFA creates a more parsimonious representation of the original set of observations providing evidence of construct validity. Because the principal-components method of analysis mixes common, specific, and random error variances, a common factoring method such as principal axis is recommended. Before EFA, check inter-item correlations of the variables? any variable that correlates at less than .4 with all other variables may be deleted from the analysis. If the items have been carefully developed, the number of factors that emerge on both Kaiser and scree criteria should equal the number of scales being developed. The intent is to develop scales that are reasonably independent of one another, so an orthogonal rotation is recommended for this analysis. * * Reliability is the accuracy or precision of a measuring instrument and is a necessary condition for validity. 0.7 来自nunnally 1871 cited * CFA affords a stricter interpretation of uni-dimensionality than does EFA (on fit) Compare fit of a single factor model with a multitrait model The chi-square statistic permits the assessment of fit of a specific model as well as the comparison between two models. The smaller the chi-square the better the fit of the model. It has been suggested that a chi-square two or thre times as large as the degrees of freedom is acceptable (Carmines McIver, 1981), but the fit is considered better the closer the chi-square value is to the degrees of freedom for a model. A nonsignificant chi-square is desirable, indicating that differences between the model-implied variance and covariance and the observed variance and covariance are small enough to be due to sampling fluctuation. A model with a large chi-square may still be a good fit if the fit indices are high, as this measure is particularly dependent on sample size. It is desirable to have a significantly smaller chi-square for the specified model than for competing models. Comparative Fit Index (CFI) and the Relative Noncentral