Table 1.
Fit statistics for the IRT models for SDE and IM dichotomous scoring.
| Model | Dichotomous 2PL | |||
|---|---|---|---|---|
| SDE | IM | |||
| One factor | Two factors | One factor | Two factors | |
| −2LL | −6160.042 | −6119.080 | −6088.515 | −6068.787 |
| M2 (df) | 441.965353 (170)* | 315.047632 (169)* | 356.40856 (170)* | 321.025991 (169)* |
| RMSEA2 (CI.90%) | .053 (.047, .059) | .039 (.032, .045) | .044 (.037, .050) | .040 (.032, .046) |
| SRMSR | .059 | .052 | .052 | .049 |
| AIC | 12400.08 | 12320.16 | 12257.03 | 12219.57 |
| BIC | 12574.40 | 12498.83 | 12431.34 | 12398.25 |
| saBIC | 12447.41 | 12368.67 | 12304.36 | 12268.09 |
| Estimated parameters | 40 | 41 | 40 | 41 |
| −2LLchange | 40.962 (1), p < .001 | 19.728 (1), p < .001 | ||
−2LL = log-likelihood test. RMSEA2, bivariate root mean square error of approximation; SRMSR, Standardized Root Mean Square Residual; AIC, Akaike information criterion, BIC, Bayesian information criterion; saBIC, sample size adjusted BIC. The one-factor scale represents the unidimensional construct SDE or IM, two factors indicate items of SDE and IM divided in Denial and Enhancement based on reversed and non-reversed items, Modified two-factor model is a two-factor model after correcting for possible local dependence and item misfit (S-X2). *Statistically significant model misfit at p < .05. Bold indicates when model fit index indicates the better fitting model.