ASSESSING GOODNESS-OF-FIT - IS PARSIMONY ALWAYS DESIRABLE

Many mechanistic rules of thumb for evaluating the goodness of fit of structural equation models (SEM) emphasize model parsimony; all other things being equal, a simpler, more parsimonious model with fewer esti mated parameters is better than a more complex model, Although this is usually good advice, in the present article a heuristic counterexampl e is demonstrated in which parsimony as typically operationalized in i ndices of fit may be undesirable. Specifically, in simplex models of l ongitudinal data, the failure to include correlated uniquenesses relat ing the same indicators administered on different occasions will typic ally lead to systematically inflated estimates of stability, Although simplex models with correlated uniquenesses are substantially less par simonious and may be unacceptable according to mechanistic decision ru les that penalize model complexity, it can he argued a priori that the se additional parameter estimates should be included. Simulated data a re used to support this claim and to evaluate the behavior of a variet y of fit indices and decision rules, The results demonstrate the valid ity of Bollen and Long's (1993) conclusion that ''test statistics and fit indices are very beneficial, but they are no replacement for sound judgment and substantive expertise'' (p. 8).