Table 3. Confirmatory factor analyses of the Fatigue Assessment Scale: overall model fit.

Model	${\chi }_{\mathrm{S}-\mathrm{B}}^2$	df	CFI	GFI	SRMR	RMSEA
Subsample 1, N = 406
Model 1a	108.8	35	0.98	0.99	0.064	0.072
Model 1b	83.8	34	0.98	0.99	0.048	0.060
Model 2a	87.5	34	0.98	0.99	0.059	0.062
Model 2b	61.44	33	0.99	0.99	0.040	0.046
Model 3a	84.99	33	0.98	0.99	0.048	0.062
Model 3b	50.94	25	0.99	1.00	0.035	0.051
Model 4a	–	–	–	–	–	–
Model 4b	38.5	24	1.00	1.00	0.032	0.039
Subsample 2, N = 397
Model 1a	119.1	35	0.97	0.98	0.064	0.078
Model 1b	114.1	34	0.97	0.99	0.064	0.077
Model 2a	73.0	34	0.99	0.99	0.052	0.054
Model 2b	62.4	33	0.99	0.99	0.045	0.047
Model 3a	112.7	33	0.97	0.99	0.060	0.078
Model 3b	50.94	25	0.99	1.00	0.035	0.051
Model 4a	–	–	–	–	–	–
Model 4b	41.63	24	0.99	0.99	0.033	0.043

Notes.

${\chi }_{\mathrm{S}-\mathrm{B}}^2$

Chisquare Satorra-Bentler

CFI

Comparative Fit Index

GFI

Goodness of Fit Index

RMSEA

Root Mean Square Error of Approximation

SRMR

Standarized Root Mean Square Residual

Model 1a

one-dimensional model

Model 1b

model 1a with correlated errors between items with positive wording, items #4 and #10

Model 2a

two correlated factors: physical fatigue (five items) and mental fatigue (five items)

Model 2b

model 2a with correlated errors between items #4 and #10

Model 3a

bifactorial model of a general factor and a specific factor with items with positive wording

Model 3b

bifactorial model of a general factor and two specific factors, one with items with positive wording and the other with items with negative wording

Model 4a

bifactorial model of a general factor and two specific factors, one for the items of physical fatigue and the other for those of mental fatigue. It did not converge in any of the two subsamples

Model 4b

model 4a with correlated errors between items #4 and #10