Table 3. Fixed effects of optimal GLMM (Model 2) for the probability of skipping the test words.

	Estimate	Std. Error	z value	Pr(>|z|)
(Intercept)	-1.19934	0.14841	-8.08123	< 0.00001
FREQ	-0.00302	0.04398	-0.06872	0.94521
PRED	-0.00839	0.02461	-0.34099	0.73311
LENGTH	-0.46003	0.05747	-8.00517	< 0.00001
LAUNCH	0.61519	0.04106	14.98372	< 0.00001
FREQ:LENGTH	-0.04980	0.03213	-1.54996	0.12115
FREQ:LAUNCH	0.00729	0.01811	0.40250	0.68732
PRED:LENGTH	-0.02576	0.01987	-1.29634	0.19486
PRED:LAUNCH	0.01514	0.01428	1.06002	0.28914
LENGTH:LAUNCH	-0.00224	0.01905	-0.11767	0.90633
PRED:LENGTH:LAUNCH	0.01968	0.01150	1.71096	0.08709
FREQ:LENGTH:LAUNCH	0.02751	0.01339	2.05526	0.03985

The fixed structure included the effects of word length (“LENGTH”; 4–8 letters), word frequency (“FREQ”; 0.20–5.93 log units), word predictability (“PRED”; between -2.60 and 2.60 logit units), and saccadic launch-site distance (“LAUNCH”; between -6.00 and -0.002 letters from the space in front of the test words), the three-way interactions between word frequency, word length and launch-site distance and between word predictability, word length and launch-site distance, as well as corresponding two-way interactions; the random structure included a random intercept by participant and by sentence pair, as well as by-participant random effects of word length and launch-site distance, but without the correlation between random effects (see S1 Table). The model's estimates and standard errors are expressed in logit units. The intercept estimate (logit: -1.19934) indicates that test words were skipped in about 23% of the cases, when all variables were at their reference, mean, value (Word Length: 5.82 letters; Launch Site: -2.93 letters; Word Frequency: 3.06 log units; Word Predictability: -0.96 logit units). Colon stands for interaction. See S3 Table for the corresponding minimalist optimal GLMM.