Table 5.

Fixed effects from best-fitting multilevel linear models of location of first Manner mention, Experiment 1

Effect	Estimate	S.E.	z-value
Manner in Subject
Intercept	−1.80	0.58	−3.10*
Condition: Control vs. Verb + Event Type	2.68	0.51	5.26**
Condition: Control vs. Event-Type Only	0.70	0.51	1.38
Condition: Control vs. No-Overlap	0.46	0.51	0.91
Manner in Verb
Intercept	0.68	0.64	1.05
Condition: Control vs. Verb + Event Type	−1.85	0.55	−3.34**
Condition: Control vs. Event-Type Only	−0.06	0.52	−0.12
Condition: Control vs. No-Overlap	0.27	0.51	0.55

Z-values are significant at

^*p<0.01 and

^**p<0.001.

Formula in R for both models: DepVar ~ Condition + (1 | Subject) + (1 + Condition | Event). For Manners in Post-verbal positions, the model including Condition did not provide a better fit than an empty model with no fixed effects.