A Monte Carlo simulation study was conducted to investigate the effects of sample size, estimation method, and model specification on structural equation modeling (SEM) fit indices. Based on a balanced 3x2x5 design, a total of 6,000 samples were generated from a prespecified population covariance matrix, and eight popular SEM fit indices were studied. Two primary conclusions were suggested. First, for misspecified models, some fit indices appear to be noncomparable in terms of the information they provide about model fit; some fit indices also seem to be more sensitive to model misspecification. Second, estimation method strongly influenced almost all the fit indices examined, especially for misspecified models. These two issues do not appear to have been well documented in the previous literature. Perhaps the focus of most previous simulation studies on correctly specified models may have failed to detect these dynamics. It is further suggested that future research should study not only different models relative to model complexity, but also a wider range of model specification conditions, including correctly specified models as well as models specified incorrectly to varying degrees. (Contains 2 figures, 6 tables, and 26 references.) (Author/SLD)