The ‘goodness-of-fit’ of fit models: creating a multidimensional survey for person-organisation and person-group fit in health care

Table 3 Comparison of the goodness-of-fit of P-O and P-G fit models

Model	df	χ²/df^a	RMSEA	RFI	TLI	SRMR	AIC^b
*Accepted Values*	N.A.	2–4	≤0.05	0.9–0.95	0.9–0.95	≤0.05	N.A.
P-O CFA results
Model 1	120	2.695	0.0892	0.856	0.905	0.052	6647.587
Model 2	129	2.709	0.090	0.856	0.904	0.063	6655.734
Model 3	126	2.971	0.096	0.842	0.889	0.121	6686.598
Model 4	84	4.011	0.119	0.832	0.869	0.070	5932.777^g
Model 5	128	2.613	0.087	0.861	0.087	0.056	6642.653
Modified Model 5	124	2.045	0.071	0.890	0.940	0.051	6569.868
P-G CFA results
Model A	48	2.909	0.099	0.900	0.932	0.059	3798.971
Model B	50	3.037	0.102	0.895	0.927	0.064	3807.188
Model C	50	3.037	0.102	0.895	0.927	0.064	3807.188
Modified Model B	48	2.635	0.092	0.909	0.942	0.058	3785.819
Modified Model C	49	2.854	0.098	0.902	0.934	0.063	3797.192

Note. RMSEA Root Mean Square Error of Approximation, RFI Relative Fit Index, TLI Tucker-Lewis Index; χ² = chi-square; SRMR = Standardized Root Mean Square Residual; AIC = Akaike Information Criterion, which compares second-order non-nested models, lower scores indicate better fit
^aModel 1 χ² = 323.34; Model 2 χ² = 349.48; Model 3 χ² = 374.35; Model 4 χ² = 336.89; Model 5 χ² = 334.40; Modified Model 5 χ² = 251.46; ^†Model A χ² = 139.61; Model B χ² = 151.83; Model C χ² = 151.83; Modified Model C χ² = 139.83
^bThe AIC of Model 4 cannot be compared to the other models as there is one less first-order latent variable. AIC of Models A-C were added for completeness, but are not compared

ISSN: 1471-2288