Problem Sets

Problem Sets is a learning activity in which participants work through a curated sequence of problems that systematically build capability in a specific skill or domain. Each problem is designed to target a particular move, pattern, or decision the participant needs to master, and the difficulty progresses across the set so that working through it produces a durable skill rather than isolated practice on disconnected problems. The curation is the method — which problems are chosen, in what order, and how the review happens determines whether the problem set actually builds the capability it was designed to build or just gives participants reps on work they could already do.

The approach has deep roots in mathematics and physics education. George Polya's How to Solve It (first published 1945, still in print) codified problem-solving as a teachable discipline with named heuristics — understand the problem, devise a plan, carry out the plan, look back — that a practitioner builds by working through increasingly demanding problem sets. Alan Schoenfeld's Mathematical Problem Solving (1985) extended the research to show how expert problem-solvers differ from novices, and how deliberate work on curated problem sets moves a learner along that spectrum. Anders Ericsson and Robert Pool's Peak (2016) generalized the finding into the research on deliberate practice, which is the mechanism problem sets operationalize — purposeful, targeted, feedback-rich work on specific skills, repeated across a progression of difficulty.

Problem Sets is the first activity in the problem-and-inquiry family, and it sits structurally as the individual-practice counterpart to the collaborative activities in the preceding family. Where Think-Pair-Share and Peer Review build capability through paired exchange, Problem Sets builds capability through sustained individual work on a deliberately sequenced series of problems. The two families are complementary — many workshops combine collaborative activities for shared framing with problem sets for individual skill consolidation.

What It Is

Four structural components define Problem Sets.

• A precisely defined target skill or decision. The problem set is curated to build one specific capability — a named skill, a named decision-making pattern, a named diagnostic move. Vague targets produce vague problem sets; a problem set without a precise target is just a pile of exercises. The target is named before any problems are chosen.
• A curated sequence of problems with progressive difficulty. The problems are chosen deliberately, not assembled randomly. Each problem either isolates a specific element of the target skill or combines elements in a new way, and the difficulty progresses so that later problems assume mastery of earlier ones. The sequence is the activity's load-bearing element.
• Individual work on each problem. Problems are worked through individually by default. This is different from most activities in the collaborative family — Problem Sets is a deliberate-practice activity, and the research on deliberate practice is clear that the core work happens when the learner is wrestling with a problem alone, without help. Paired or group variations exist but are variations, not the default.
• Review of reasoning, not just answer. After each problem (or after a small group of problems), the workshop leader runs a review focused on how participants reasoned through the problem, not only on whether they arrived at the right answer. Two participants can reach the same answer through very different reasoning, and the reasoning is what transfers to the next problem.

When to Use It

Problem Sets fits workshops where the learning goal is a specific, identifiable skill or decision-making pattern that can be practiced through concrete problems. It is especially strong when participants need to build the skill to transfer to real work they will do after the session, and when the skill has component moves that can be isolated and sequenced.

Good candidates:

• Skills with identifiable component moves — diagnostic decisions, framing choices, pattern recognition, structured judgments — where problems can isolate each component and then combine them.
• Content where difficulty progression is possible — a clear "easier" version and "harder" version of the same problem type.
• Participants who learn by working through problems (most adult learners do, especially in applied domains).
• Sessions where individual deliberate practice time is valuable, usually 60 minutes or longer.

Less suitable:

• Skills that do not have a clear "problem" format — relational skills, presence-based work, improvisational capability — where the unit of practice is not a discrete problem.
• Content where collaboration is the point of the learning, not the individual skill.
• Very short sessions where there is no time for sustained individual work plus reasoning review.
• Topics where participants do not yet have enough foundation to engage meaningfully with even an entry-level problem.

In-Person and Virtual Delivery

Problem Sets runs cleanly in both in-person and virtual delivery, and both modalities have distinct advantages — virtual delivery excels at asynchronous completion plus synchronous review; in-person delivery excels at real-time observation of participant struggle. The activity's core mechanism, individual work on curated problems, is modality-neutral.

In person. Participants work at their own pace on printed problem sheets or workbooks. The workshop leader circulates during the work, observing where participants get stuck and what reasoning paths they try. In-person delivery has a strong advantage in the review stage: the leader has seen the struggle in real time and can reference specific participants' approaches during the group review, which makes the reasoning-focused review concrete rather than abstract. Printed artifacts also give participants something to take with them.

Virtual. Problems are distributed through a shared document, a worksheet, or a dedicated platform (quiz platforms, exercise platforms). Participants can work through problems in advance of the session — asynchronously — and use the live time for the reasoning review, which is often the higher-value part. This split uses both time budgets well: asynchronous work is what deliberate practice research most strongly supports, and synchronous review is where the reasoning-focused teaching lands. Virtual delivery also handles self-paced progression better than in-person does: participants who finish quickly can move ahead without waiting for the group, and participants who need more time get it without holding the session up.

Choosing the modality. Virtual is often slightly stronger for Problem Sets specifically because of the async-plus-sync split. In-person delivery works well for shorter problem sets that can complete in the session, or when the workshop leader wants to observe participant reasoning in real time. For mixed-modality cohorts, run Problem Sets virtually with an async-completion window before the live session — the individual work is already in the platform's structure, and the live time can focus entirely on review.

How to Run It

Before the Session

1. Name the target skill or decision precisely (leader design work). Before choosing any problems, write out the specific capability the problem set will build. Participants should be able to [specific move] in [specific context] by the end of the set. Vague targets produce bad problem sets.
2. Curate the sequence (leader design work, significant time). Select or create four to six problems that together build the named capability. Each problem should either isolate one element of the skill or combine elements in a new way. Order by difficulty — the first problem should be solvable by most participants, the last should stretch even strong participants. Curation is the activity's load-bearing element (see Design Considerations).
3. Optional — distribute problems for asynchronous work. For virtual cohorts, send the problems one to three days before the session with instructions to attempt each and bring the work to the session. Asynchronous attempt + synchronous reasoning review is often the strongest configuration because deliberate practice research supports sustained individual work over short in-session attempts.

During the Session

1. Introduce the target and the set (5 minutes). Name the skill the problem set builds. Tell participants what they should be able to do by the end. Briefly walk the first problem's frame so participants understand the shape of what they are being asked to do.
2. Work through the set, problem by problem (40–60 minutes total, or shorter if pre-solved async). Participants work individually. After each problem (or after every two problems, depending on set length), the workshop leader runs a short review focused on reasoning — how participants decided, what they considered, what they ruled out, what moved their decision.
3. Close with a skill-level reflection (10–12 minutes). The closing reflection is not about specific problems but about the skill the set built. Ask: what reasoning pattern did you use by the fifth problem that you were not using at the first? What is the name of the capability you now have that you did not have at the start of the session?

After the Session

1. Apply the reasoning pattern to a real problem in participants' own work (within one week). The problem set's transferable capability has to meet real work to consolidate. Each participant identifies one situation from their own practice in the next week and applies the reasoning pattern — the specific move or framework they named in the closing reflection.
2. Bring the applied result to the next session (in multi-session cohorts). The cohort compares how the reasoning pattern survived contact with real work, which exposes where the set's problems oversimplified and where the pattern genuinely transfers. This is where deliberate practice becomes durable skill.

An Example in Practice

A 90-minute workshop for eight coaches on rebalancing session content density — the recurring problem of sessions designed with too much content for the time and outcomes available. The workshop leader has built a problem set of five scenarios, each with a session outline and a named learning outcome.

The workshop leader opens by naming the target capability: by the end of this session you should be able to look at any session outline of your own and identify what to cut, what to keep, and why, using a consistent reasoning process.

Problem one, the entry-level case: a sixty-minute session introducing a new coaching offer, with ten teaching segments averaging four minutes each plus Q&A. Outcome — participants can describe the offer's value proposition in their own words. Task — identify which three segments can be cut without compromising the outcome and explain why. Participants work for eight minutes, then the leader runs a five-minute review focused on how they decided what to cut, not which segments they picked.

Problems two through four build in complexity: a ninety-minute session on sales objections, a two-hour session on client discovery, a half-day session on positioning work. Each adds a new dimension — multiple outcomes per session, outcomes that interact, content that is structurally load-bearing even if not directly outcome-producing. After each problem, a short reasoning review.

Problem five, the stretch case: a three-hour flagship session that must be cut by roughly forty percent while preserving its core outcome. By this point, participants are applying a consistent framework — identify the named outcome, rank each segment by its contribution to that outcome, cut from the bottom — and the leader can name the framework back to them at the close.

The closing reflection runs ten minutes. The leader asks what reasoning pattern participants were using by problem five that they were not using at problem one. Most of the cohort names some version of outcome-first ranking, which is the capability the problem set was built to produce.

Variations

• Paired Problem Sets. Participants work each problem in pairs rather than individually. Used when the skill benefits from thinking-aloud — some participants engage more deeply when their reasoning is externalized to a partner — or when the cohort is small enough that paired review becomes richer than individual review.
• Timed Problem Sets. Each problem carries a strict time limit to simulate the real-world time pressure the skill is practiced under. Used when the target capability is specifically about deciding quickly (diagnostic framing under pressure, in-session facilitation moves) rather than about reasoning at leisure.
• Self-Scoring Problem Sets. Participants assess their own work against a provided rubric before the workshop leader's review. The self-assessment then gets compared to the leader's assessment during review, which surfaces where participants' self-perception diverges from external evaluation. Used when metacognitive calibration is itself a target outcome.
• Progressive Disclosure Problem Sets. Problems are presented one at a time without participants seeing what is coming next. Prevents participants from skimming ahead or adjusting their approach based on the remaining set. Used when the workshop leader wants each problem to be engaged fresh, without the framing effect of the full set's shape.
• Tease-and-Reveal Problem Sets. Each problem is presented without the solution, participants work it through to completion, and then the solution is revealed at the end for participants to check their work against. Adapted from Dr. Jason Wright's (founder of Workshop Doctor) canonical Active Learning Activities (ALA-6) as "tease and reveal." Used when the problem's solution space is clear enough to evaluate against, when the reveal serves as satisfying closure of the working, and when participants benefit from the narrative structure of working-toward-reveal rather than working-against-rubric. Distinct from standard Problem Sets' review-of-reasoning close; Tease-and-Reveal closes with outcome comparison rather than reasoning examination.

Which Method It Serves

Problem Sets is primarily a Problem-Based Learning activity. Where the method's canonical form centers a session on one ill-structured problem, Problem Sets operationalizes a complementary form — many well-structured problems curated to build a specific capability through deliberate practice. Both forms sit inside the method.

Problem Sets also functions inside other methods. In Inquiry-Based Learning, problem sets can structure the investigation by giving participants a sequence of increasingly open questions to investigate. In Project-Based Learning, problem sets serve as skill-building segments between artifact-work sessions — participants practice the specific capabilities the project will demand. In Action Learning, problem sets can be used to build the diagnostic skills participants apply to their own live problems between sessions.

Design Considerations

• The quality of the curation determines the value of the activity. Problem Sets is defined by the sequence of problems chosen — which problems, in what order, building toward what skill. Randomly assembled problems or problems that do not target a specific skill produce busywork, not learning. A well-curated set sequences difficulty so that each problem builds on the last, isolates components of the target skill before combining them, and concentrates on the specific capability the set was designed to build rather than offering general practice. Before running a Problem Sets activity, pressure-test the curation: does each problem target a specific element of the target skill, does the difficulty progression actually build capability rather than just increasing complexity, and could a participant who worked the set arrive at the closing reflection able to name the capability they now have? If any of the three answers is soft, the curation needs revision before the activity is worth running.
• Name the target skill precisely before choosing problems. The discipline is backward — outcome first, problems second. A problem set curated without a precise target ends up as a collection of exercises that feel rigorous but build no specific capability. The named target is what makes the activity more than practice.
• Keep individual work as the default; use paired work as a deliberate variation. Deliberate-practice research supports individual work as the mechanism that builds capability — the learner has to wrestle with the problem without help for the learning to consolidate. Paired work is a valid variation for specific cases (thinking-aloud advantage, small cohorts), but defaulting to paired work turns Problem Sets into a collaborative activity and loses the deliberate-practice effect that distinguishes it.
• Review the reasoning, not only the answer. The workshop leader's review stage is about how participants solved the problem, not whether they got it right. Two participants reaching the same answer through different reasoning are learning different things; one of them may be building the target capability and the other just getting lucky. Reasoning-focused review exposes this and is where the teaching actually consolidates.
• Match problem scope to session time. A four-to-six problem set is the practical range for a 60-to-90-minute session. Fewer problems, and the progression cannot build enough capability; more, and the review stage collapses. For longer sessions or multi-session programs, problem sets can run across meetings — new problems each session, cumulative progression.

Where this goes next: Problem Sets builds a specific capability through deliberate practice on a curated sequence. The next activity in this family moves the work from solving curated problems to reasoning through open questions under a specific questioning discipline. See Socratic Discussion.

Related

Read this next:

• Socratic Discussion — the next activity in the problem-and-inquiry family; moves from solving curated problems to reasoning through open questions under a questioning discipline.

Foundations of this activity:

• Problem-Based Learning — the primary method Problem Sets serves; the curated-problems form is the complement to the canonical ill-structured-problem form.
• Think-Pair-Share — the simplest collaborative activity; Problem Sets is its individual-work counterpart for deliberate practice.
• Peer Review — the collaborative activity that builds the evaluative eye; Problem Sets builds the procedural skill that the eye evaluates.
• The Guiding Principles of Active Learning — the principles Problem Sets activates, especially autonomy and focused outcomes.

Where this leads:

• Which Active Learning Method Fits Your Workshop? — Problem Sets is the strongest sub-activity for workshops where a specific, identifiable skill needs deliberate practice to transfer to real work.
• Action Learning — the primary method Problem Sets most often supports when the skill being practiced will be applied to participants' own live problems between sessions.

References

Ericsson, A., & Pool, R. (2016). Peak: Secrets from the new science of expertise. Houghton Mifflin Harcourt.

Polya, G. (2004). How to solve it: A new aspect of mathematical method (Expanded ed., with a new foreword by John H. Conway). Princeton University Press. (Original work published 1945)

Schoenfeld, A. H. (1985). Mathematical problem solving. Academic Press.