ELEMENTS OF A RESEARCH PROTOCOL

Because SEM has the ability to model complex relationships between multivariate data, sample size is an important (but unfortunately underemphasized) issue. Two popular assumptions are that you need more than 200 observations, or at least 50 more than 8 times the number of variables in the model. A larger sample size is always desired for SEM. ................
................