Contingent scaffolding improves learning more than fixed or absent support.¶

Claim · All claims
Evidence · 5 studies · 2 causal r1, 1 quant-synthesis r?, 1 review r2, 1 qualitative r? · q2–q4

Contingent scaffolding adjusts the type and level of support in real time based on the learner's demonstrated understanding, rather than providing a fixed or predetermined sequence of help. Evidence supports this approach over non-contingent or no support, though the evidence base is small and the mechanism is contested.

Subclaims¶

q3 i? Students receiving fully contingent tutoring (support level adjusted responsively to each response) outperform students receiving fixed, moderate, or no support on both immediate and delayed tests of the same skill. → Pratt and Savoy-Levine 1998

q2 i? Contingent, interactive tutoring (questioning and hints rather than explaining and feeding back) produces similar immediate learning outcomes to non-interactive tutoring but significantly better transfer and greater student responsibility for learning. → Chi et al. 2001

q4 i? Among forms of dynamic assessment, scaffolding (contingent prompting and mediation) is more effective than coaching but less effective than explicit strategy training for cognitive tasks. → Swanson and Lussier 2001

q3 i? Contingency, fading, and transfer of responsibility are the three defining characteristics of scaffolding in the literature; they are closely intertwined, and the review describes responsibility as being gradually transferred to the learner via contingent fading; it does not test whether contingent support produces fading or transfer. → van de Pol et al. 2010

Evidence¶

Primary evidence link: https://doi.org/10.1016/s0193-3973(99)80041-0

Pratt and Savoy-Levine 1998¶

Pratt, M. W., & Savoy-Levine, K. M. (1998). Contingent tutoring of long-division skills in fourth and fifth graders: Experimental tests of some hypotheses about scaffolding. Journal of Applied Developmental Psychology, 19(2), 287–304. doi:10.1016/s0193-3973(99)80041-0

q3 · peer-reviewed quasi-experiment with intervention fidelity check · i? · no source text available to check; the entry prints no effect size · n=40 (N=8 per condition) · causal · r?

Fourth and fifth grade students were assigned to one of five one-to-one tutoring conditions for long-division: fully contingent support, moderate support, high support, partly contingent support, or no support. Nine levels of support were defined; conditions differed in which levels tutors could use and whether application was contingent on student responses. Audio-taped sessions were coded to verify intervention fidelity. Students in the fully contingent condition solved significantly more long-division problems than all other conditions at both immediate and 1-month follow-up assessment. A small but important limitation: N=8 per condition, one-to-one tutoring setting only; generalizability to classrooms is unknown.

Chi et al. 2001¶

Chi, M. T. H., Siler, S. A., Jeong, H., Yamauchi, T., & Hausmann, R. G. (2001). Learning from human tutoring. Cognitive Science, 25(4), 471–533. doi:10.1207/s15516709cog2504_1

q2 · quasi-experimental comparison: the same tutors' interactive sessions against their earlier naturalistic sessions, not randomized · i? · the abstract prints no effect size; the full text may · n=11 tutor-student pairs · causal · r1

Inexperienced tutors working one-to-one with eighth grade students on the human circulatory system were then instructed to shift to a more interactive/contingent style — more questioning and hints, less explaining and feeding back. More scaffolding episodes were observed in the interactive condition (verified by coding). Immediate learning outcomes were similar across both styles. However, students in the interactive/contingent tutoring condition performed significantly better on transfer questions (topics not directly covered in tutoring) and took greater responsibility for their own learning, shown by higher frequency of self-initiated reading aloud.

Swanson and Lussier 2001¶

Swanson, H. L., & Lussier, C. M. (2001). A selective synthesis of the experimental literature on dynamic assessment. Review of Educational Research, 71(2), 321–363. doi:10.3102/00346543071002321

q4 · meta-analysis of 30 studies · i? · the abstract prints no effect size; the full text may · n=30 studies · quant-synthesis · r?

A meta-analysis evaluating three forms of dynamic assessment: scaffolding (contingent prompting and verbal mediation), coaching (training and intervening), and strategy training (modeling and general feedback). Effect sizes were highest for strategy training, followed by scaffolding, followed by coaching. This result shows that contingent scaffolding outperforms coaching but does not outperform explicit strategy instruction, a meaningful qualifier for instructional designers choosing between approaches.

van de Pol et al. 2010¶

van de Pol, J., Volman, M., & Beishuizen, J. (2010). Scaffolding in teacher–student interaction: A decade of research. Educational Psychology Review, 22(3), 271–296. doi:10.1007/s10648-010-9127-6

q3 · systematic review of 66 articles (8 effectiveness studies) · i? · the abstract prints no effect size; the full text may · n=66 articles · review · r2

A comprehensive review of scaffolding research from 1998–2009. Identifies contingency, fading, and transfer of responsibility as the three defining characteristics of scaffolding — not as independently ranked causal variables, but as interlocking features: responsibility is gradually transferred to the learner via contingent fading. The review finds that scaffolding is generally effective, but cautions that the evidence base is thin (only eight effectiveness studies, of mixed experimental, quasi-experimental and correlational designs, plus one meta-analysis), measurement is inconsistent across studies, and most evidence comes from one-to-one tutoring rather than classroom settings. Importantly, the review does not claim contingency is more effective than fading or transfer of responsibility — all three are treated as necessary features of the same process.

Wood et al. 1976¶

Wood, D., Bruner, J. S., & Ross, G. (1976). The role of tutoring in problem solving. Journal of Child Psychology and Psychiatry, 17(2), 89–100. doi:10.1111/j.1469-7610.1976.tb00381.x

q3 · foundational observational study · i? · no source text available to check; the entry prints no effect size · n=30 · qualitative · r?

The study that introduced the scaffolding metaphor. Researchers observed 3-, 4-, and 5-year-olds attempting a block-building task with a tutor and identified six key scaffolding functions (recruitment, reduction in degrees of freedom, direction maintenance, marking critical features, frustration control, demonstration). Established the contingent-shift principle: increase support when the learner fails, decrease it when the learner succeeds. Foundational for the field; not an effectiveness study in the causal sense.

Studies bearing on this claim¶

Must support be contingent on the learner's last response, rather than fixed? This table gathers the studies behind this claim and the 27 narrower claims that bear on it: 33 distinct studies (11 causal, 6 review, 5 quant-synthesis, 5 qualitative, 4 associational, 1 theoretical, 1 design), q1–q4, 5 of 33 with an effect size coded. One row per study, codes as each entry records them; no effect is averaged and no verdict is drawn. Built by scripts/synthesize_claim.py from eval/synthesis/; a row that looks wrong is fixed on the claim it comes from.

Discussion¶

The evidence that contingent scaffolding outperforms non-contingent or absent support is real but narrower than it appears. Most effectiveness studies (Pratt & Savoy-Levine, Chi et al., and Murphy & Messer, unsourced on this page) use one-to-one tutoring settings; generalizability to whole-class instruction is largely undemonstrated. Van de Pol et al. explicitly note that genuine contingency is scarce in practice — Myhill and Warren (2005, cited therein) observe that teaching contingently across 30 students simultaneously is exceptionally difficult. This implementation gap is a major design constraint: contingent scaffolding may work best when delivered through one-to-one human tutoring, small-group coaching, or adaptive intelligent tutoring systems.

A second qualification is the Swanson and Lussier finding: explicit strategy training outperforms contingent scaffolding for cognitive tasks. For instructional designers choosing between an approach built on contingent coaching versus one built on explicit strategy instruction, the evidence slightly favors the latter, at least for structured cognitive skills. Contingent scaffolding may have an advantage for transfer and self-regulation (per Chi et al.), even when immediate performance gains are similar.

Finally, van de Pol et al. treat contingency as definitionally necessary for scaffolding — support that is not contingent is not scaffolding, it is instruction. This means some evidence cited for "scaffolding effectiveness" may conflate genuinely contingent support with structured-but-fixed instructional sequences.

Related Claims¶

• In one experiment with 256 undergraduates, computer-based lesson versions that included practice produced significantly higher posttest scores than versions without practice — the general claim this finding bears on, with the studies behind it

• Minimal guidance is less effective for novices than explicit instruction — the general claim this finding bears on, with the studies behind it

• Fading support promotes the transfer of responsibility from instructor to learner.
• Prompting learners to self-explain improves understanding and problem solving on immediate tests by a moderate average amount, with little evidence yet on delayed or classroom gains.
• Intelligent tutoring systems improve achievement over non-individualised instruction — related
• Human tutoring has a medium effect over no tutoring, and in one small study students learned as well when tutors only prompted them as when tutors also explained and gave feedback — related
• A student began interacting with the sim within 10 seconds and verbalized sense-making without explicit guidance — related
• PAIR-C scaffolding shows mixed evidence for deep understanding and reduced misconceptions in emergent-phenomena instruction — a narrower finding that bears on this claim
• A contingent tutor surrendered the full answer in one session in sixteen, against one in six for the question-only tutor — a narrower finding that bears on this claim
• Contingent tutoring held delegation at the level of an answer-on-request assistant while delivering more artifacts, one decision at a time — a narrower finding that bears on this claim
• AI scaffolding differs from human scaffolding in availability, feedback timing, and dynamic adjustment, making its developmental impact conditional — related
• Effective intervention timing in AI-based instruction does not always mean immediate response; the interval between student difficulty and teacher intervention can allow independent problem solving — related
• Adaptive fading with periodic reinforcement yields higher independence than fixed-high or linear-decay scaffolding in simulation — a narrower finding that bears on this claim
• Adaptive fading of worked examples based on current skill level produced better learning outcomes than fixed fading or tutored problem solving — related
• Learning benefits of Cognitive Tutor interactivity are supported to varying extents by a growing body of experimental studies — a broader claim this one bears on
• Teacher and researcher coaches both used a cognitive apprenticeship approach but differed in scaffolding style — related
• When controlling for task effort, low contingent support improved multiple-choice achievement with short independent working time whereas high contingent support improved it with long independent working time — a narrower finding that bears on this claim
• Coach-guided interactive teaching yields higher and more progressive recovery than static relearning, particularly under moderate forgetting levels (10%–30%) — a narrower finding that bears on this claim
• Teachers who completed a scaffolding intervention program increased the contingency of their support far more than non-participating teachers — a narrower finding that bears on this claim
• Teacher-student scaffolding and autonomy support co-evolve into self-reinforcing attractor states over sustained one-on-one instruction — related
• The association between tutor questioning and impasse recovery decays as impasses persist, while addressing the student's error becomes more beneficial — related
• Current serious games give all learners the same unregulated scaffolding, which the authors argue contradicts the notion of scaffolding — related
• Contingency is not treated as a necessary condition of scaffolding in bilingual education research, unlike broader educational research — a narrower finding that bears on this claim
• Dynamic (adaptive) scaffolding is linked to stronger effects of ICT-supported writing interventions — related
• Fading-out prompts yield more cognitive strategy use and higher learning outcomes than permanent prompts — related
• System-controlled adaptive instruction does not necessarily yield higher transfer than nonadaptive static approaches — reports the opposite
• Adaptive fading adjusted to discourse quality is reported as more effective than fixed fading regimes for developing internal script components, but long-term evidence is lacking — a narrower finding that bears on this claim
• Expert modeling, feedback, hints, and multiple strategies improve cognitive outcomes, but question prompts show no evident difference from control — related
• Reflected case-based training combined with explicit strategy instruction and adaptive tutoring yields the best transfer performance, whereas strategy instruction without tutoring adds no benefit over case-based training alone — related