Table 5. Final model predicting the first individual look of a habituation trial.

	Fixed Effects					Random Effects
Model Parameters	Β	SE	t	df	p	Variance	SD	χ2	df	p
Intercept	0.98	0.02	58.14	126	< .001	0.03	0.16	450.26	126	< .001
Log10 Look Count	-0.62	0.04	-14.53	126	< .001	0.07	0.27	178.42	126	.002
Lag 1 Duration	0.07	0.03	2.55	126	.012	0.03	0.17	162.01	126	.017
Model Fit Statistics	Deviance	Number of Parameters			Level 1 Observations (Individual Looks)		χ2	p	AIC	BIC
	595.16	10			1215		146.26	< .001	615.16	626.01

The duration of the first individual look of a trial is uniquely predicted by log10 look count (habituation) and lag 1 duration (temporal dependency). The model uses the last look of one trial to predict the first look of the next trial in the habituation protocol. The χ² describes a comparison to a model with no temporal dependency (Lag 1) effects (see text). The inclusion of Log10 Look Count as a predictor accounted for 40.8% more variance than a model with the intercept alone. The inclusion of Lag 1 Duration as a predictor accounted for 10.7% more variance than a model with Log10 Look Count. For the model equation and unstructured (full) covariance matrix see S1 Text.