The Poverty-Related Stress Scale: Development and Validation of a Multidimensional Measure Assessing Poverty-Related Stressors
Table 4
Model fit statistics.
Fit indices
Cut-off criterion
Sensitive to
Penalty for model complexity
Absolute fit indices
Chi-square ()
(i) Lowest comparative value between measurement models (ii) Nonsignificant chi-square () (iii) Significant difference in chi-square between models (iv) For model comparison: retain the model with the lowest chi-square
Yes
Yes
Approximate fit indices
Root mean square error of approximation (RMSEA)
(i) 0.06 to 0.08 (marginally acceptable); 0.01 to 0.05 (excellent) (ii) Not-significant () (iii) 90% confidence interval Rande should not include zero (iv) For invariance: retain model where
Yes
Yes
Standardized root mean square residual (SRMR)
(i) 0.06 to 0.08 (marginally acceptable); 0.01 to 0.05 (excellent)
Yes
No
Incremental fit indices
Comparative fit index (CFI)
(i) 0.90 to 0.95 (marginally acceptable fit); 0.96 to 0.99 (excellent) (ii) For invariance: retain model with highest CFI value ()
No
Yes
Tucker-Lewis Index (TLI)
(i) 0.90 to 0.95 (marginally acceptable fit); 0.96 to 0.99 (excellent) (ii) For invariance: retain model with highest TLI value ()
No
Yes
Akaike information criterion (AIC)
(i) For model comparison: retain the model with the lowest value
No
No
Bayes information criterion (BIC)
(i) For model comparison: retain the model with the lowest value
No
No
Sample-size adjusted BIC (aBIC)
(i) For model comparison: retain the model with the lowest value
No
No
Note. Table adapted from Van Zyl & Ten Klooster [48].