Example Files

All of the binary distributions of GENBLIS include four example model specification files (example1 through example4), and the resulting list (i.e., summary output) files (example1.lst through example4.lst). Each of these files has been included in this manual.

example1 is a confirmatory factor analysis which uses the data contained in the example1.dat file. The bootstrap goodness-of-fit test results (presented in the example1.lst file) show that the model fits the data at both $\alpha=.05$ and $\alpha=.1$ test levels. In the output, ``The ML Fit'' labels the statistic denoted above by $\tilde{F}_{\text{ML}}$. The bootstrap estimate of the critical value for a test at level $\alpha=.05$, previously denoted $F^*_{\text{ML},1-\alpha}$, is labeled ``The 95 Percent Bound.'' The hypothesis that the model fits the data is not rejected at level $\alpha=.05$ because ``The ML Fit'' value is less than ``The 95 Percent Bound'' value, that is, $\tilde{F}_{\text{ML}}<F^*_{\text{ML},1-\alpha}$.

example2 is a confirmatory factor analysis which uses the data contained in the example2.dat file. The bootstrap goodness-of-fit test results (presented in the example2.lst file) show that the model does not fit the data. Here ``The ML Fit'' value is greater than both ``The 95 Percent Bound'' and``The 90 Percent Bound'' values, that is, $\tilde{F}_{\text{ML}}>F^*_{\text{ML},1-\alpha}$ for both the $\alpha=.05$ and the $\alpha=.1$ test levels.

example3 estimates a model in which the data from example1.dat and example2.dat are treated as coming from two different groups. No constraints are imposed across groups, so there is no advantage compared to analyzing the datasets separately. The model does not fit the data.

example4 estimates a model in which the example1.dat and example2.dat data are treated as coming from two different groups. The factor loadings and the variance of the common factor are constrained to be equal across the groups. The model does not fit the data, but with the unconstrained model of example3 also not fitting the data, this is not terribly interesting.