Does retrieval practice depend on semantic cues? Assessing the fuzzy trace account of the testing effect

Figures & data

Table 1. Mean proportion of correctly recognised targets in Experiment 1 by surface features overlap and study method.

Study method	Surface features overlap
	Words	Scrambled	Background	Background scrambled
Restudy	.79 (.02)	.72 (.02)	.82 (.02)	.60 (.02)
Testing	.81 (.02)	.72 (.02)	.79 (.03)	.62 (.02)

Note: Standard errors are between brackets.

Table 2. Mean proportion of correctly recognised targets in Experiment 2 by surface features overlap and study method.

Study method	Surface features overlap
	Words	Synonyms
Restudy	.78 (.03)	.69 (.03)
Testing	.77 (.03)	.65 (.03)

Note: Standard errors are between brackets.

Table 3. Mean proportion of correctly recognised targets in Experiment 3 by surface features overlap and study method.

Study method	Surface features overlap
	Words	Images
Restudy	.82 (.01)	.72 (.02)
Testing	.83 (.01)	.75 (.01)

Note: Standard errors are between brackets.

Table 4. Mean proportion of correctly recognised targets in Experiment 4a by surface features overlap and study method.

Study method	Surface features overlap
	Words	Images
Restudy	.84 (.02)	.77 (.03)
Testing	.82 (.03)	.84 (.02)

Note: Standard errors are between brackets.

Table 5. Mean proportion of correctly recognised targets in Experiment 4b by surface features overlap and study method.

Study method	Surface features overlap
	Words	Images
Restudy	.88 (.02)	.77 (.03)
Testing	.85 (.02)	.79 (.03)

Note: Standard errors are between brackets.

Figure 1. The regression coefficients, with their 95% confidence intervals, corresponding to the log of the odds ratio between restudy and testing with in the twelve different surface features overlap conditions of the five experiments. The white squares correspond to the five word-cue conditions, and the black squares correspond to the conditions where the surface features of the final test cues were altered as compared to the learning phase.