Decision-making inflexibility in a reversal learning task is associated with severity of problem gambling symptoms but not with a diagnosis of substance use disorder

Table 4 Model selection for PRLT performance, including SOGS severity

Model	Fixed factors	df	AIC	χ²	p	Model	Fixed factors	df	AIC	χ²	p
Sat. (0.b)	Model 3 plus SOGS and its interactions with Phase and Log-trial	19	10,418			Sat (0.c)	Model 3 plus Group and Group × Phase plus SOGS and its interactions with Phase and Log-trial	23	10,421
4^b	0.b minus SOGS × Phase × Log-trial	16	10,415	2.81	0.422 (4 ≥ 0.b)	6^b	0.c minus SOGS × Phase × Log-trial	20	10,418	0.82	0.420 (6 ≥ 0.c)
5.1^a,b	Model 4 minus SOGS × Log-trial	15	10,414	0.59	0.443 (5.1 ≥ 4)	7.1^a,b	Model 6 minus SOGS × Log-trial	19	10,417	0.59	0.444 (7.1 ≥ 6)
5.2^b	Model 4 minus Phase × Log-trial	13	10,428	19.05	< 0.001 (4 > 5.2)	7.2^b	Model 6 minus Phase × Log-trial	17	10,431	19.05	< 0.001 (6 > 7.2)
5.3^b	Model 4 minus Phase × SOGS	13	10,418	8.34	0.039 (4 ≥ 5.3)	7.3^b	Model 6 minus Phase × SOGS	17	10,421	8.43	0.038 (6 ≥ 7.3)

Significant p values are in italics
^aBest fitting model
^bAlmost-singular fit: given the risk of overfitting, parameters will be estimated in both best-fitting and saturated (0.b and 0.c) models. See also footnote 2
Sat Saturated

ISSN: 2050-7283