The ore samples were very well fit, as indicated by average fit values of 105%, as were air-exposed sediments (average = 99 ± 21%).
It is noteworthy that the exponential growth models did not fit the data sets, as indicated by effective sample size (ESS) scores <200 (data not shown).
Both instruments revealed good indices of fit as indicated by the high reliability coefficients, showing good internal consistency.
A 3-factor structure was found to be a better fit, as indicated by sizeable reductions in the values of goodness of fit indices, and a CFA was conducted with the domain specification shown in Table 2.
The two-factor solution had a superior fit as indicated by the significant factor loadings (p <.05) and the goodness-of-fit indices (χ2 = 8.43; df = 8; Root Mean Square Error of Approximation (RMSEA) = 0.011; Taka-Lewis Index (TLI) = 0.997; Adjusted Goodness-of-Fit Index (AGFI) = 0.984 relative to the null (one-factor solution) model (χ2 = 39.83; df = 9).
The predicted two-factor solution had a superior fit as indicated by the significant factor loadings (p <.05) and the goodness-of-fit indices (χ2 = 17.82; df = 8; RMSEA = 0.05; Taka-Lewis Index 0.971; AGFI = 0.968) relative to the null model (χ2 = 179.55; df = 9), which specified a single factor solution.
However, all models met the criteria for good model fit as indicated by a CFI > .97.97
The final model, Model 3, had the best overall model fit as indicated by the log likelihood ratio, AIC, and BIC.
In this study, we considered all models with (partial) strong MI restrictions for countries in which this MI assumption holds to show a good model fit as indicated by the RMSEA and CFI.
The retained variables in the final model were based both on goodness-of-fit statistics (such as their significance, as indicated by the t-test, and their contribution to the model fit, as indicated by the Akaike Information Criterion, AIC, [1]), as well as on the sign and relative magnitude of the estimated coefficients.
Since the autoregressive models provide superior fit (as indicated by both the summary goodness of fit statistics), as well as satisfy the assumption of independent residuals (as indicated by the graphical diagnostics), it may be concluded that the "ordinary" non-linear models underestimate the standard errors.
