Table 2.

Model comparison of the effects on difficulty ratings. The table reports a Likelihood ratio test between models where one predictor at a time was entered. AIC = Akaike Information Criterion. No.par = number of parameters of the model.

Predictors	No.par	AIC	logLik	LR.stat	df	Pr(>Chisq)
RT	7	16,980	−8483	20.70	1	<.001
Condition	10	16,982	−8481	3.67	3	.300
Cluster	13	16,964	−8469	23.61	3	<.001
Condition × Cluster	22	16,968	−8462	14.08	9	.120