Syntes: When and how seductive details harm learning. A study using cued retrospective reporting (2021)

Källa: https://doi.org/10.1002/acp.3822 (Applied Cognitive Psychology 35(4), 948–959; open access)

Verified content

• [full text] 113 German university students (mean age 23.7), randomised to a lightning-formation lesson (537 words, four explanatory drawings) without (n = 54) or with (n = 59) four seductive details (photo plus caption, 141 extra words, 20.5% of the text). Fixed study time 6.5 minutes in both conditions; free recall (max 9) and a nine-item multiple-choice transfer test (ω = 0.65 after dropping one item) right after learning. Power analysis for d = 0.5 asked for 102.
• [full text] Overall: no significant seductive-details effect on recall (F(1,110) = 0.47, p = .496) or transfer (F(1,110) = 2.00, p = .160), controlling for spatial ability, which differed between groups by chance; without that covariate, transfer p = .055, d = −0.37, in the direction of a harm. Means: transfer 6.17 vs 4.86, recall 3.31 vs 2.81.
• [full text] Moderation (hypothesised in advance: weaker cognitive prerequisites, more harm): final school grade moderated the transfer effect (interaction b = 2.41, p = .027). In the half with weaker school-leaving grades (n = 55) extras lowered transfer (d = −0.60, p = .030); in the half with better grades no significant effect (d = −0.21, p = .446). Prior knowledge and spatial ability did not moderate recall or transfer (p = .162 to .808); grade did not moderate recall (p = .655). One of six moderator tests was significant.
• [full text] Think-aloud with gaze replay: diversion (building a mental model around the extras) was the most frequent process (M = 2.08 of 4) and predicted worse recall (b = −0.50, p = .008), not transfer.

Limits

• [full text] One lab experiment, university students only (the authors call for a wider range of prerequisites), eye-tracking and chin rest may have changed reading, closed-format transfer items, immediate test only; the grade split is a median split within a narrow, high-achieving range.
• [assessment] Bears on C8 (transfer harm, same study time): an overall null restates C8's consistency bound; the grade moderation raises who bears the harm, held as open question 12 (2026-10-19), not a narrowing.

Added 2026-10-19 by the project (nightly attack run, widened search for C8). [abstract] = from the published abstract; [full text] = read in the full text; [assessment] = the project's own assessment.