Problem-Solving Followed by Instruction¶

Theory · All theories
Evidence · 55 claims (21 for, 24 mixed, 10 against) · 29 studies (21 causal, 4 theoretical, 3 review, 1 quant-synthesis), q1–q4 · 12 of 29 report an effect size · 54 claims rest on one study

Description¶

Problem-solving followed by instruction (PS-I, closely related to "productive failure") reverses the conventional tell-then-practice sequence: students first attempt a problem — typically without succeeding at a canonical solution — and only afterward receive direct instruction on the correct method. The prior attempt activates relevant prior knowledge, surfaces students' own (often suboptimal) representations, and raises motivation to learn the canonical method by making its advantages concrete rather than abstract. A meta-analysis across 166 comparisons found PS-I outperforms instruction-first sequencing with Hedge's g=0.36 [95% CI 0.20, 0.51] (Sinha & Kapur, 2021). The underlying attempt can be left unguided ("productive failure" proper, where students generate as many solution representations as they can) or shaped by scaffolds that deliberately steer students toward specific representations before instruction corrects them — see Failure-Driven Scaffolding.

Implications¶

Context¶

Requires a problem rich enough to admit multiple representations or partial solutions, so the pre-instruction attempt genuinely differentiates students rather than producing uniform failure
Best suited to domains with a clear canonical method the later instruction phase can deliver (e.g., statistical concepts, data visualization, physics formulas), so the contrast between the student's attempt and the canonical method is legible

Target Learners¶

Novices to the targeted concept, who have enough general reasoning ability to attempt the problem but lack the specific canonical method

Target Learning Objectives¶

Deep conceptual understanding and transfer, which PS-I designs typically outperform instruction-first sequencing on, more than immediate procedural fluency

Claims¶

Emotion dynamics during problem-solving predict learning outcomes in a manner that depends on scaffolding design [~M]
Productive failure, problem-solving before instruction, shows a measured benefit of about d = 0.36 rising to about 0.58 at high design fidelity (as reported) [+W]
PF design fidelity predicts larger PS-I effects, with four fidelity criteria showing significant subgroup differences [+W] — attached 2026-10-13 from Tanmay Sinha et al. (2021), which proposed "Productive Failure as a high-fidelity subset of PS-I designs within preparation for future learning"; tests this page's relationship.
PS-I fosters learning only if specific design features (contrasting cases or building instruction on student solutions) are implemented [+W] — attached 2026-10-14 from Loibl et al. (2017), which proposed "Initial theory of PS-I: a set of interconnected cognitive mechanisms accounting for when and how problem solving followed by instruction supports learning".
Productive failure students significantly outperformed lecture-practice and facilitated complex problem solving students on well-structured and higher-order application post-test problems [+W] — attached 2026-10-15 from Kapur (2011), which proposed "Productive failure: complex problem solving in small groups without instructional facilitation until teacher-led consolidation"; tests this page's relationship.
Productive failure students demonstrated greater representation flexibility in working with graphical representations than students in the other two conditions [+W] — attached 2026-10-15 from Kapur (2011), which proposed "Productive failure: complex problem solving in small groups without instructional facilitation until teacher-led consolidation"; tests this page's relationship.
Students exposed to failure-driven scaffolding demonstrated higher quality of constructive reasoning [+W] — attached 2026-10-15 from Sinha et al. (2021), which proposed "Explicit failure-driven scaffolding via problematizing to facilitate problem-space exploration".
Failure-driven and success-driven scaffolding before instruction produce similar posttest conceptual understanding [~W] — attached 2026-10-15 from Sinha et al. (2021), which proposed "Explicit success-driven scaffolding via structuring with low-specificity heuristics".
PF students did not outperform DI students when learning social science research methods in two quasi-experimental studies with 10th graders, failing to replicate the STEM-domain PF effect [-W] — attached 2026-10-18 from Valentina Nachtigall · Katja Serova · Nikol Rummel (2020), which proposed "Productive Failure: design features and three hypothesized learning mechanisms"; tests this page's relationship.
Of eight PF-similar studies of problem solving prior to instruction in non-STEM or domain-general settings, only Schwartz and Bransford (1998) found a positive learning effect [-W] — attached 2026-10-18 from Valentina Nachtigall · Katja Serova · Nikol Rummel (2020), which proposed "Domain-structuredness hypothesis: PF effectiveness may be restricted to structured domains"; tests this page's relationship.
Giftedness interacts significantly with instruction type on the learning and recall outcomes in the 2×2×2 study [~M] — attached 2026-10-18 from Lim et al. (2023), which proposed "Invention-first instruction (Invention as Preparation for Future Learning / productive failure)"; tests this page's relationship.
Merely delaying instruction until after problem solving does not improve conceptual understanding when no additional measures trigger knowledge-gap awareness or deep-feature discovery [~W] — attached 2026-10-18 from Loibl et al. (2020), which proposed "Three-mechanism account of why problem solving before instruction can benefit conceptual understanding"; tests this page's relationship.
In an out-of-school lab for social sciences, the authenticity level of the learning activity (Productive Failure vs. Direct Instruction) did not affect students' perceived authenticity or situational interest [~W] — attached 2026-10-18 from Valentina Nachtigall (2021), which proposed "Productive Failure as an authentic learning activity that emulates scientific inquiry".
In primary school children learning javelin throwing, PS-I did not outperform I-PS on throwing distance or movement-feature execution [-W] — attached 2026-10-18 from Christian Leukel (2025), which proposed "KLI framework for composite instructional designs: intermediate knowledge from the PS phase anchors processing of subsequent instruction"; tests this page's relationship.
Collaborative and individual problem solving in traditional PF do not differ in conceptual knowledge after instruction [~W] — attached 2026-10-18 from Charleen Brand et al. (2023), which proposed "Three preparatory mechanisms account of Productive Failure".
For introductory biology students with low topic-specific prior knowledge, problem-solving-first (PS-I) outperformed instruction-first (I-PS) on a near-transfer test [+W] — attached 2026-10-19 from Cheng-Wen He et al. (2025), which proposed "Competing theoretical accounts: cognitive load theory versus productive failure for instructional sequencing"; tests this page's relationship.
For biochemistry students with relatively higher topic-specific prior knowledge, instruction-first (I-PS) outperformed problem-solving-first (PS-I) on a near-transfer test [-W] — attached 2026-10-19 from Cheng-Wen He et al. (2025), which proposed "Competing theoretical accounts: cognitive load theory versus productive failure for instructional sequencing"; tests this page's relationship.
The relationship between learners' prior knowledge and the optimal instructional sequence may not be straightforward [~W] — attached 2026-10-19 from Cheng-Wen He et al. (2025), which proposed "Competing theoretical accounts: cognitive load theory versus productive failure for instructional sequencing"; tests this page's relationship.
In PS-I, prompting students to generate multiple strategies during the initial phase generally outperformed inventing one strategy or I-PS [~W]
Adding a practice phase after instruction (PS-I + P) did not improve javelin throwing performance over PS-I alone [-W]
The teacher-conference problem met the sweet spot of complexity: 10th graders generated solution ideas containing the targeted errors without solving the problem canonically [+W]
PF students outperformed direct-instruction students on conceptual understanding and transfer without compromising procedural fluency [+W]
Instruction before problem solving leads students to generate fewer but higher-quality solution attempts than problem solving before instruction [-W]
Initial failure did not hurt subsequent learning during instruction: no condition differences on example generation or immediate learning tests [+W]
Error conditions differed in negative affect but not positive affect after the PS-phase learning materials [~W]
PS-I research demonstrates beneficial effects on conceptual knowledge and knowledge transfer relative to I-PS [+W]
No significant interaction between condition and pre-test knowledge on learning outcomes in either topic [~W]
PS-I over a year in a university linear algebra course was followed by significantly higher final exam performance versus baseline cohorts [+W]
No significant differences among PF, VF-high, VF-low, and DI conditions on procedural knowledge acquisition [~W]
PS-I with a virus transmission simulation showed better transfer of multivariable causality to a new context, despite similar end-of-intervention performance to I-PS [+W]
PS-I effectiveness did not depend on the breadth of prior knowledge activated, but on activation of relevant prior knowledge [~W]
Across two studies that differed in the duration of the initial problem-solving phase, the PS-I versus WS-I transfer advantage reversed [~W]
PS-I sequence outperforms I-PS on conceptual knowledge only when instruction builds on student solutions [~W]
The manipulation worked as intended: I-PS students largely succeeded on the problem-solving sheets while PS-I students largely failed [~W]
Solution quality does not mediate the effect of collaboration on conceptual knowledge in PF [~W]
Across studies, high solution rates in the initial problem-solving phase co-occur with beneficial PS-I effects and low rates with no beneficial effects, a cross-study pattern not yet tested in a shared sample [~W]
Sense-making-first preparation supports learning more strongly than fluency-first because conceptual intermediate knowledge alters second-phase learning processes [+W]
Working memory capacity did not moderate learning gains in PS-I and I-PS, but digit recall correlated with declarative knowledge of the features [~W]
The intervention significantly benefited 6th-grade students but showed no significant effects for 7th or 8th graders [~W]
Receiving instruction before problem solving improves procedural fluency more than problem solving before instruction when both phases are otherwise identical [-W]
The proposed cognitive mechanisms explain both positive and negative effects of PS-I [~W]
No significant condition differences remain at posttest: both approaches converge on MVC strategy and reasoning after both phases [~W]
Neither the quantity nor the quality of PF students' own solution attempts correlated with their conceptual knowledge acquisition [-W]
Sequencing of exploration and explicit instruction shows inconsistent results, moderated by task complexity per the expertise reversal account [~W]
Ninth-grade students in a Productive Failure design scored significantly higher than Direct Instruction students on complex systems explanatory knowledge about climate [+W] — attached 2026-10-22 from Jacobson et al. (2017), which proposed "Productive Failure two-phase learning design"; tests this page's relationship.
PF students scored significantly higher than DI students on climate explanatory knowledge [+M] — attached 2026-10-22 from Jacobson et al. (2017), which proposed "Productive Failure two-phase learning design"; tests this page's relationship.
PF students scored significantly higher than DI students on near transfer (butterfly effect problem) [+M] — attached 2026-10-22 from Jacobson et al. (2017), which proposed "Productive Failure two-phase learning design"; tests this page's relationship.
PF students scored significantly higher than DI students on far transfer (robot mining problem), with an independent teacher experience effect [+M] — attached 2026-10-22 from Jacobson et al. (2017), which proposed "Productive Failure two-phase learning design"; tests this page's relationship.
PF and DI groups showed no significant difference in climate declarative knowledge gains [~M] — attached 2026-10-22 from Jacobson et al. (2017), which proposed "Productive Failure two-phase learning design"; tests this page's relationship.
DI students relied on notes rather than exploring the agent-based models, while PF classes showed lively collaborative exploration and questioning [+W]
DI-PS significantly outperformed PS-DI on the two supplier-reasoning problem-solving criteria (d = .60 and d = .64) [-W]
Learners in both sequencing conditions predominantly preferred and were satisfied with DI-PS sequencing [~W]
PS-DI learners consulted the reference book on supplier selection procedure far more often than DI-PS learners (d = .83) [+W]
PS-DI learners reported higher intrinsic in-scenario cognitive load during problem solving (d = .53) [+W]
Neither timing nor learning activity produced significant transfer effects in either experiment [-W]

The evidence by question¶

Should learners attempt a problem before the idea is taught, and what must the teaching afterwards do? Productive Failure Improves Conceptual Learning [+S]
Which features of the attempt and of the instruction after it, and how much time, go with a larger advantage for attempting first? PF design fidelity predicts larger PS-I effects, with four fidelity criteria showing significant subgroup differences [+M]

Self-Regulated Learning — PS-I's pre-instruction attempt phase is a context in which monitoring and affect regulation become consequential for later learning

Examples¶

Failure-Driven Scaffolding

Key Sources¶

Sinha, T. (2022). Enriching problem-solving followed by instruction with explanatory accounts of emotions. Journal of the Learning Sciences, 31(2), 151-198. https://doi.org/10.1080/10508406.2021.1964506
Sinha, T., & Kapur, M. (2021b). When problem solving followed by instruction works: Evidence for productive failure. Review of Educational Research, 91(5), 761-798.
Kapur, M. (2014). Productive failure in learning math. Cognitive Science, 38(5), 1008-1022. doi:10.1111/cogs.12107

Page metadata

Field	Value
`type`	`theory`
`title`	`Problem-Solving Followed by Instruction`
`description`	`Students attempt a problem before receiving instruction on the canonical solution, so that the attempt activates prior knowledge and exposes knowledge gaps that make the later instruction more effective than instruction delivered first.`
`status`	`review`
`canonical`	`True`
`generated`	by `claude/unspecified` on `2026-08-29`
`trust`	unverified — no one has confirmed this page's content
`sources`	1 — `sinha-2022`

Frontmatter is the page's machine-readable record — see the Schema & Guide.