Comparing Contrasting Cases¶

Strategy · All strategies
Evidence · 5 claims (3 for, 2 mixed) · 14 studies (6 causal, 5 review, 2 quant-synthesis, 1 theoretical), q1–q4 · 1 of 14 report an effect size

Description¶

Learners examine two or more cases that differ on one or more critical dimensions and are prompted to explain why they differ. The comparison surfaces deep features that any single case leaves invisible — variation across cases isolates what matters. The strategy is often used as a preparation phase: learners struggle productively with the contrast before being told the canonical explanation.

Design Implications¶

Comparison is one of the most consistently supported learning strategies: contrasting cases improve both immediate learning and transfer by directing attention to structural rather than surface features Comparing contrasting cases improves learning. [+S]. The mechanism is discriminative: without variation, learners cannot tell which features of a case are essential and which are incidental Multiple varied cases support flexible, transferable knowledge. [+M]. Comparison works best when learners generate the differences themselves rather than reading them pre-listed Comparing contrasting cases improves learning. [~M].

Context¶

Requirements¶

• Cases that are minimally different — varying on the target dimension while holding surface features constant (or vice versa, depending on the discrimination goal)
• An explicit comparison prompt ("How are these the same? How are they different? Why?") rather than mere juxtaposition
• A telling or consolidation phase after comparison, in which the instructor names and formalizes what the contrast revealed (Schwartz & Bransford's [+M] "a time for telling")
• Sufficient working-memory headroom; two cases plus a prompt is usually the ceiling for novices (Cognitive Load Management)
• Two or more cases that share deep structure but differ in surface features (or, for discrimination learning, differ in structure)
• Spatial or visual alignment of the cases so corresponding elements can be mapped directly
• Comparison prompts that direct attention to relations ("How are these solution methods alike?") rather than leaving comparison to chance
• Sufficient time; comparison is effortful and truncated comparison yields surface-level noticing

Constraints¶

• Comparison fails when cases differ on too many dimensions at once — novices compare surface features and miss the deep structure Comparing contrasting cases improves learning. [-M]
• Learners with very low prior knowledge may lack the vocabulary to articulate differences; unstructured comparison can produce frustration rather than preparation for future learning [~M]
• High-similarity pairs can be worse than no comparison when the critical difference is subtle and learners conflate the cases [-W]
• Comparison imposes extraneous load when cases must be held in mind simultaneously and cannot be viewed side by side (e.g., sequential presentation) [-M]
• Sequential presentation (studying cases one at a time) largely eliminates the benefit; simultaneous side-by-side viewing is what drives structural alignment [-M]
• Novices may compare on surface features unless prompts direct attention to relations, producing spurious generalizations [-M]
• Poorly chosen cases — too dissimilar, or differing in too many ways — overload working memory and prevent mapping Chunking reduces working memory load. [~M]
• For learners with strong prior knowledge, comparison of near-identical cases adds little beyond studying one case well Guidance becomes less effective as learner expertise increases. [~M]

Implementation Variability¶

• Problem-solution comparison: two worked solutions to the same problem, compared for efficiency or generality (Rittle-Johnson & Star's approach)
• Preparation for future learning: contrasting cases explored before instruction, to make learners ready to understand the lecture that follows
• Non-example contrast: one correct and one flawed case, sharpening discrimination of error conditions (Non-Examples)
• Multiple varied cases: three or more cases across contexts, supporting flexible abstraction (Cognitive Flexibility Theory)
• Different-solution comparison: two correct methods for the same problem (e.g., algebraic vs. graphical), building flexible procedural knowledge
• Correct–incorrect comparison: a worked solution paired with an erroneous one, sharpening discrimination of the critical feature Erroneous examples build conceptual knowledge. [+M]
• Multiple surface variants: several isomorphic problems with different cover stories, supporting transfer to novel contexts
• Invention before comparison: learners attempt to invent a solution first, then compare cases — the "time for telling" sequence that prepares learners to benefit from instruction

Target Learners¶

• Novices who cannot yet distinguish deep from surface features Comparing contrasting cases improves learning. [+S]
• Intermediate learners consolidating multiple procedures — comparing solution methods improves flexibility in strategy choice [~M]
• Less effective as pure discovery for learners with minimal prior knowledge; pairing with brief instructional support restores benefit [~M]
• Learners prone to overgeneralizing from one instance or anchoring to a single solution method

Target Learning Goals¶

• Conceptual discrimination: identifying which features define a concept and which are incidental
• Transfer: preparing learners to recognize the same structure in new contexts
• Strategy evaluation: judging when one method is preferable to another
• Preparation for future learning: building perceptual readiness for subsequent direct instruction
• Conceptual understanding: abstracting principles and relational structure from instances
• Transfer: recognizing when the same structure recurs under different surface features
• Procedural flexibility: knowing multiple methods and when each is advantageous
• Discrimination learning: distinguishing correct from flawed reasoning or designs

Instructions¶

1. Select or author a case set that varies systematically on the target dimension; keep surface features either constant or deliberately varied depending on the discrimination goal.
2. Present cases side by side with a structured comparison prompt; require learners to articulate differences in their own words (Annotating or a comparison table).
3. Let learners attempt to explain the pattern before revealing it — productive struggle with the contrast is the point.
4. Follow with a "telling" phase: instruction that names the deep feature the comparison exposed (Advance Organizers can frame this consolidation).
5. Follow with practice on new cases, asking learners to predict outcomes using the discriminated feature (Practice).
6. Fade the scaffolds: move from guided comparison prompts to independent case analysis as expertise grows (Fading).

Related Strategies¶

• Non-Examples — a two-case comparison where one case is deliberately wrong; the minimal contrast form
• Worked Examples — comparing two worked solutions is the most studied variant of this strategy
• Analogical Reasoning — comparison across cases is the engine of analogical abstraction
• Case-Based Learning — extended case work; contrasting cases is the compressed, comparison-focused form
• Use a contrasting case study design to identify subtle leadership differences between schools with improving and non-improving achievement
• When implementing PS-I, include contrasting cases or build the instruction phase on student solutions
• Use Worked Examples — comparing two worked examples is the natural extension of single-example study
• Erroneous Examples — a correct–incorrect comparison is a special case of case comparison
• Analogical Encoding — comparison as the mechanism of analogical transfer

Examples¶

• Rittle-Johnson & Star's algebra studies — students compared two worked solutions to the same equation (e.g., substitution vs. elimination), improving procedural flexibility and conceptual knowledge relative to studying solutions sequentially.
• Schwartz & Bransford's statistics "A Time for Telling" — students explored contrasting data displays before lecture; the comparison group outperformed lecture-only and invention-only groups on transfer to novel data situations.
• PhET Interactive Simulations — learners vary one parameter at a time across runs, effectively comparing contrasting cases of a physical system.
• Facing History and Ourselves — historical case pairs (e.g., differing responses to the same event) compared to isolate decision points and their consequences.
• Rittle-Johnson & Star's algebra studies — students compared pairs of worked problems using different solution methods (e.g., subtract first vs. distribute first), improving conceptual and procedural flexibility over studying the same examples sequentially.
• Schwartz & Bransford's "A Time for Telling" — statistics students analyzed contrasting datasets before lecture; the comparison phase dramatically improved learning from subsequent instruction.
• Khan Academy — worked example pairs showing alternative solution paths for the same problem, with prompts to identify which method is more efficient and why.
• Case-based business teaching (Harvard Business School method) — discussion sections explicitly juxtapose prior cases with the current one, prompting students to map structural similarities across contexts.

Key Sources¶

• Schwartz, D. L., & Bransford, J. D. (1998). A time for telling. Cognition and Instruction, 16(4), 475–522. doi:10.1207/s1532690xci1604_4
• Rittle-Johnson, B., & Star, J. R. (2007). Does comparing solution methods facilitate conceptual and procedural knowledge? An experimental study on learning to solve equations. Journal of Educational Psychology, 99(3), 561–574. doi:10.1037/0022-0663.99.3.561
• Alfieri, L., Nokes-Malach, T. J., & Schunn, C. D. (2013). Learning through case comparisons: A meta-analytic review. Educational Psychologist, 48(2), 87-113. doi:10.1080/00461520.2013.775712
• Gentner, D., Loewenstein, J., & Thompson, L. (2003). Learning and transfer: A general role for analogical encoding. Journal of Educational Psychology, 95(2), 393–408. doi:10.1037/0022-0663.95.2.393
• Marton, F. (2014). Necessary conditions of learning. New York, NY: Routledge.