Explicit contextualization instruction improves procedural aspects of historical writing without improving contextualization scores themselves¶

Claim · All claims
Evidence · 1 study · causal r2 · q3 · i2 medium · n=140 undergraduate L2 English learners (B2 CEFR level) at a university in Istanbul; experimental n=60, control n=80
Currency · newest study 2021 · not yet checked against its sources · not reviewed by a person

In a quasi-experiment with 140 university L2 history students, explicit instruction in historical contextualization did not raise contextualization rubric scores relative to a content-matched active control, but it unexpectedly improved claim-writing and shifted how (not how much) students used and placed contextualization in their essays.

Subclaims¶

q3 i2 Both the experimental and control groups improved significantly on total historical-writing rubric score from pretest to posttest (Mdn 5→13 and Mdn 4→12 respectively, T=1, p<.05, r=-.61 for both), indicating the shared CLIL course content drove most of the general improvement. → Sendur et al. 2021

q3 i0 The historical contextualization rubric score did not differ significantly between the experimental group (Mdn=3.00) and control group (Mdn=3.00) on the posttest (U=2146.50, z=-1.11, ns, r=.09) — the targeted instruction did not move the outcome it targeted. → Sendur et al. 2021

q3 i2 The experimental group scored significantly higher than control on claim quality (Mdn 3.00 vs. 2.00, U=1704.50, z=-3.08, p<.01, r=.26); 75% of the experimental group formulated an arguable claim versus 44% of control — an unexpected, unplanned transfer effect. → Sendur et al. 2021

q3 i? Both groups scored similarly on a closed-book historical-knowledge test (Mdn 11.00 vs. 11.75, U=2167.50, z=-.75, ns), ruling out unequal background knowledge as the explanation for any contextualization differences. → Sendur et al. 2021

q3 i? Among instances of contextualization present in posttest essays, 61% in the experimental group included an explicit connection or conclusion linking background to evidence, versus 43% in control (descriptive frequencies, not hypothesis-tested); offset (unintegrated) placement of contextualization was rarer in the experimental group (2% vs. 10%). → Sendur et al. 2021

Evidence¶

Sendur et al. 2021¶

Sendur, K. A., van Drie, J., & van Boxtel, C. (2021). Historical contextualization in students' writing. Journal of the Learning Sciences, 30(4-5), 797-836. https://doi.org/10.1080/10508406.2021.1939029

q3 · quasi-experimental pre/post design with an active, content-matched control group (13 intact classes, 11 instructors), non-parametric hypothesis testing throughout; not randomized at the student level, not pre-registered · i2-i3 · large, well-quantified overall improvement in both conditions (r=-.61); the targeted contextualization-score comparison is a clear, precisely reported null (r=.09); the claim-writing transfer effect is small-to-medium (r=.26) but well quantified; the connection-frequency and placement findings are descriptive only · n=140 undergraduate L2 English learners (B2 CEFR level) at a university in Istanbul; experimental n=60, control n=80 · causal · r2

140 undergraduate L2 English learners across 13 intact classes received a 28-hour CLIL history course; 5 classes (n=60) received Explicit Instruction in Historical Contextualization targeting the Historical Reasoning Framework's contextualization aspect specifically, while 8 classes (n=80) received the same CLIL course without that focus. Pre/post document-based-question essays were scored on a 5-point rubric across five reasoning aspects (Cohen's kappa .64-1.0 pretest, .66-.82 posttest). The central, well-powered comparison — contextualization rubric score — is a clean null result; the paper's positive findings are either an unplanned transfer effect (claim quality) or descriptive qualitative patterns (connection frequency, placement) without inferential statistics or reported inter-rater reliability for those specific codes.

Discussion¶

This is a genuine, informative null result on the paper's primary target and should not be overstated into a general endorsement of the instructional pattern. The unexpected claim-writing transfer effect and the descriptive placement/connection patterns are plausible but weaker-tier evidence (q1-q2, no inferential statistics for the qualitative codes) that the intervention changed how students write even when the primary rubric score didn't move — a genuine dissociation between procedural/discourse-level change and holistic rubric scoring that is itself worth documenting, but which the study cannot explain mechanistically. Findings are specific to L2 learners at B2 English proficiency; generalization to L1 history students is untested.

Learner Variables¶

• Writing Skills — outcome: instruction changes it

Related Claims¶

• Disciplinary Literacy Instruction Improves Comprehension — the general claim this finding bears on, with the studies behind it
• (none yet linked)
• Both procedural and descriptive writing scores improved significantly after GBSRI — related
• Interdisciplinary humanities units improve students' interdisciplinary competences — related
• Self-assessing against a rubric improved the quality of student work in two non-randomised studies, one of which also found more stress and performance-avoidance — related