The text generation effect shrinks with longer texts and is small and nonsignificant for texts over 900 words¶

Claim · All claims
Evidence · 1 study · quant-synthesis r2 · q3 · i1 small
Currency · newest study 2023 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q3 i1 The text generation effect was medium-sized for short texts (0–300 words, g = 0.36), largest for 301–600 words (g = 0.63), smaller for 601–900 words (g = 0.43), and small and nonsignificant for texts over 900 words (g = 0.05). → Julia Schindler · Tobias Richter 2023

Evidence¶

Julia Schindler · Tobias Richter 2023¶

Julia Schindler · Tobias Richter. (2023). Text Generation Benefits Learning: a Meta-Analytic Review. Educational Psychology Review. https://doi.org/10.1007/s10648-023-09758-w

q3 · i1 · quant-synthesis · r2

Subset moderator analysis of text length as a four-category variable in the multi-level meta-analysis. Significant pairwise differences emerged between short vs. 301–600 words and between 900+ words and the shorter categories; the 900+ category had only k = 12 effect sizes.

"The effect was medium-sized for studies based on short texts (0–300 words, g = 0.36, p < 0.001), largest for studies with texts ranging from 301–600 words (g = 0.63, p < 0.001), somewhat smaller in studies with texts ranging from 601–900 words (g = 0.43, p < 0.001), and small and nonsignificant in texts of more than 900 words (g = 0.05, p = 0.738)."

Discussion¶

Related Claims¶

• Interleaving category examples improves inductive category learning for visual and mathematical materials, but not for expository texts or word categories — related
• The TC effect grows with prime time: it is significantly larger at 300 ms than at 60 ms and 150 ms, and does not decrease with longer prime time — related