Table 5.

Experiment 3 omnibus analysis results and effect sizes derived from a mixed-effects model (N trials = 4,231).

	β	Approximate effect size (ms)	SE	t
Intercept	−13.41	745.63	0.39	−34.43 *
Trial number	0.0048	0.27	0.0025	1.94 †
Contextn (Sentence)	−3.78	−215.32	0.31	−12.27 *
Bare ordinal position	0.21	11.66	0.072	2.87 *
Bare ordinal position * Contextn	0.12	0.30	0.21	0.57
Sentence ordinal position	0.19	10.69	0.072	2.64 *
Sentence ordinal position * Contextn	0.056	−2.95	0.22	0.25
Contextn-1 (Sentence)	0.11	6.02	0.25	0.44
Contextn-1 * Contextn	−0.42	−28.49	0.49	−0.85

Note. Naming latencies were inversely transformed and multiplied by −10,000. All fixed effects except trial number were allowed to vary by all random factors (subjects, semantic categories, pictures). Approximate effect sizes represent slopes computed on back-transformed condition means, and may not always have the same sign as their corresponding beta estimates for non-significant effects. Words in parentheses denote factor levels for which positive beta estimates would indicate slower naming latencies.

^†p < .10.

^*p < .05.