Analyses were performed in Mplus version 6.1.1 (28) using maximum likelihood estimation and bootstrapping. Bootstrapping is advantageous with small samples as it derives an approximation of the sampling distribution via repeated resampling of the available data to yield bias-corrected 95% confidence intervals. Significant associations were only presented if they survived bootstrapped confidence intervals. Model fit was first established using the chi-square statistic, which tests the difference between observed and expected covariance matrices, producing a non-significant value if this difference is close to zero (29). In the event of a significant chi-square value, we further examined relative fit indices that also test the discrepancy between the estimated model and the data, including the mean square error of approximation (RMSEA; acceptable fit =< .08), the Comparative Fit Index and Tucker-Lewis Index (CFI & TLI; acceptable fit => .90) (30). Nested model comparisons – using the Satorra-Bentler Scaled Chi-Square test (31) – were used to assess differences between the INT− and INT+.
