Table 3.

ANCOVA for the effects of context relevance and context familiarity on posttest performance (H1-H4)

Effect	B (SE)	β	t (139)	p-value	95 % confidence interval
					Lower	Upper
Intercept	20.015 (0.768)		26.061	< 0.001	18.497	21.534
Contexta	-3.803 (1.027)	-0.301	-3.704	< 0.001	-5.834	-1.773
Relevantb	0.942 (0.784)	0.091	1.201	0.232	-0.608	2.492
Familiarc	1.769 (0.789)	0.169	2.244	0.026	0.210	3.329
Self-perceived learningd	1.018 (0.354)	0.196	2.874	0.005	0.318	1.719
Preteste	0.830 (0.113)	0.507	7.333	< 0.001	0.606	1.054

β-values around 0.10, 0.25 and 0.40 are indicative of small, medium, and large effects, respectively

^acontext (1) vs. no context (0); Hypothesis 1 ‘learning with a paper-patient context leads to better performance than learning without context’ could not be confirmed

^brelevant (1) vs. irrelevant (0); Hypothesis 2 ‘relevant context leads to better performance than irrelevant context’ is not supported convincingly

^cfamiliar (1) vs. unfamiliar (0); Hypothesis 3 ‘familiar context leads to better performance than unfamiliar context’ is confirmed

^dmean centered; Hypothesis 4 ‘higher scores on the self-perceived learning scale predict higher performance’ is confirmed

^emean centered; Even in the pretest, participants scored significantly better on the test questions with context than on test questions without context