A teacher's guide to spiral, blocked, and interleaved math practice: the research on why interleaving beats blocked practice, how to build a differentiated set using the 60/30/10 rule, a full worked example on grade 7 proportional relationships, and a free math worksheet generator.
Math practice is one of the most researched instructional activities in K to 12 education, and one of the most under thought. Most teachers pick between two default patterns: photocopy the odd numbers from the textbook (unintentionally blocked), or hand out a mixed review sheet the week before the test (unintentionally interleaved). Both work. Neither is optimal. And the gap between them shows up months later, on the standardized test students actually need to pass.
This post is the working teacher's version of the last twenty years of practice research. Blocked, spiral, and interleaved are three different structures for the same set of problems, and each teaches different skills. Interleaving in particular has one of the strongest evidence bases in cognitive science for durable math learning, and it is also the pattern students hate most in the moment, which is why most classrooms drift back to blocked. The rest of this guide is the how: how to build a spiral set that scaffolds students into interleaved practice without triggering the mutiny, how to combine it with retrieval practice, and how to include enough worked examples that no student is left staring at a page they cannot begin.
Generate a spiral math practice set on any topic — Pick a topic, grade band, and DOK mix. AssessmentWiz produces a worksheet with interleaved problem types, a full answer key, and worked solutions on a chosen subset. Free, no signup to try.
The three terms often get used interchangeably in teacher rooms and even in curriculum documents. They describe genuinely different structures, and treating them as synonyms is how classrooms end up thinking they are doing one thing while actually doing another.
| Structure | Order of problems | What it feels like during practice | What it actually teaches |
|---|---|---|---|
| Blocked | AAA BBB CCC (all of one type, then all of the next) | Efficient. Fluency climbs fast. Students report "I get it now." | The procedure once you know which procedure to apply. Removes the recognition step. |
| Spiral | Weekly sets that revisit prior units alongside the current one (typically 60% current, 30% recent, 10% cumulative) | Familiar. Feels like a review sheet. Individual sessions can still be mostly blocked within each topic. | Retention of prior content and light discrimination between related topics. Bridge between blocked and interleaved. |
| Interleaved | ABC ABC ABC (types shuffled inside every set, no cue about which strategy applies) | Slow. Frustrating. Students feel less confident during practice. | Discrimination between problem types. Recognition of when a procedure applies. Durable memory. |
Blocked practice is what happens by default: a page of the textbook, a worksheet on today's lesson, a warm up on the current skill. Spiral is what most published K to 8 curricula call their weekly practice pages: some current, some earlier. Interleaved is the least common in classroom materials and the strongest research bet for long term retention.
The clearest finding in the practice literature comes from Doug Rohrer and colleagues, who have run interleaving experiments in middle school math classrooms since 2007. Students get the same problems, the same total practice time, and the same instruction. The only variable is the order. On the practice worksheets themselves, blocked students score higher. On a delayed test one to four weeks later, interleaved students score dramatically higher.
| Study | Content | Blocked score on delayed test | Interleaved score on delayed test |
|---|---|---|---|
| Rohrer & Taylor (2007) | Volume of solids (college) | 20% correct | 63% correct |
| Rohrer, Dedrick, Stershic (2015) | Grade 7 math (slope, graphs, expressions) | 38% correct after one month | 72% correct after one month |
| Taylor & Rohrer (2010) | Grade 4 geometry | 38% correct | 77% correct |
The pattern holds across ages and content. Practice performance during blocked sessions overstates what students have learned, because the cue for which strategy to use is baked into the practice context: all volume problems, so use the volume formula. Delayed testing strips that cue away. Interleaved practice forces the discrimination step (which strategy fits this problem?) during learning, which is the same step required on the test.
A related finding matters just as much: students report feeling worse during interleaved practice and often prefer blocked practice, even after they see the delayed test results. This preference for the pattern that produces worse learning has a name, the fluency illusion, and it is the single largest reason interleaving is underused in classrooms. If a teacher listens to the "we are not getting it" energy in the room and switches back to blocked, the gains disappear.
The published research on interleaving is compelling but rarely gives teachers a practical recipe. The default that has held up across grades and topics is the 60/30/10 split. It is a spiral set that ramps into interleaving without dumping students into the deep end.
| Share of the set | Content pool | Purpose |
|---|---|---|
| 60% | Current unit | Establishes fluency on the topic being taught this week. Can be mostly blocked within this bucket in week 1 of a unit, mostly interleaved by week 3. |
| 30% | Prior 2 to 4 weeks | Interleaves recent units with the current one. This is where the discrimination practice happens. |
| 10% | Anything from earlier this year | Retrieval practice on content students think they are done with. Cheap early warning on decay. |
A concrete example. A grade 8 class starting week 3 of a systems of linear equations unit might get a Friday practice set of 20 problems: 12 on systems (this unit), 6 that mix single variable equations, slope, and graphing lines (recent units), and 2 on integer operations or fraction arithmetic (cumulative). The 12 current problems can be blocked or mini blocked (three of each subtype in a row, then switch). The 6 recent problems should be shuffled, not grouped by topic. The 2 cumulative problems function as retrieval practice.
The 60/30/10 numbers are a starting point, not a rule. In the first week of a brand new topic, tilt to 80/15/5, since students need consolidation before they can discriminate. Before a benchmark or state test, flip to 30/30/40 for cumulative review. The key is that no practice set is 100% current: every set should touch at least two prior topics.
Full interleaving on day one of a new topic is malpractice. Students have not yet consolidated a single procedure; asking them to discriminate between three procedures they cannot yet execute produces frustration, not learning. The path from blocked to interleaved has to be built over the arc of a unit.
A four week progression that works for most middle school and high school math units:
| Week | Practice pattern | Rationale |
|---|---|---|
| Week 1 (new topic) | Fully blocked. Same subtype for each set. Worked examples on 30% of problems. | Students consolidate one procedure. Fluency needs a base before discrimination becomes possible. |
| Week 2 | Mini blocked. Sets of 3 to 4 same subtype problems in a row, then switch to another subtype. Worked examples on 20%. | First discrimination practice. The transitions between blocks are where the discrimination work happens. |
| Week 3 | Interleaved within topic. Shuffled subtypes, no grouping. Worked examples on 10%. | Full within topic interleaving. Students learn to recognize which subtype they are looking at before choosing a strategy. |
| Week 4 (spiral) | 60/30/10 interleaved across topics. Worked examples on the newest content only. | Bridge to cumulative practice. This is the pattern that transfers to state tests and benchmarks. |
Three supports make this progression bearable. First, worked examples of decreasing sparseness. In week 1, a student who is stuck can look at a fully worked solution on the same subtype three problems above. By week 4, worked examples are reserved for the newest content. Second, categorization prompts. Before every mixed set, students spend about 60 seconds writing next to each problem which type it is ("this looks like a substitution problem," "this is factoring by grouping"). The category label is the discrimination step made explicit. Third, spacing. Two 20 minute interleaved sessions across the week beat one 40 minute session on Friday; the research on spacing is even stronger than the research on interleaving, and the two effects stack.
Generate a differentiated math practice set with any blocked, spiral, or interleaved mix →
The set below is a week 3 practice sheet from a proportional relationships unit. Students have seen unit rate, tables, graphs, and equations of proportional relationships, in that order. The goal of this sheet is within topic interleaving with a light spiral to slope and integer operations.
| # | Bucket | Subtype | DOK | Category label student writes |
|---|---|---|---|---|
| 1 | Current | Find the unit rate from a table | 1 | Unit rate from table |
| 2 | Current | Graph a proportional relationship from a table | 2 | Table to graph |
| 3 | Current | Write the equation y = kx from a table | 2 | Equation from table |
| 4 | Current | Word problem: is this ratio proportional? Justify. | 3 | Prop or not, justify |
| 5 | Current | Compare two proportional relationships across different representations (table vs graph) | 3 | Compare across reps |
| 6 | Current | Find k from a graph passing through a marked point | 2 | k from graph |
| 7 | Current | 3 cups flour per 4 servings, how much for 10 servings? | 2 | Scale up recipe |
| 8 | Current | Given equation y = 2.5x, complete the table | 1 | Equation to table |
| 9 | Current | Explain why the graph of a proportional relationship must pass through the origin | 3 | Origin reasoning |
| 10 | Current | Multi step: convert km/h to m/s using unit rates | 3 | Unit conversion |
| 11 | Current | Identify k from a word problem that also contains extra info | 2 | Find k in context |
| 12 | Current | Design a real world scenario for the equation y = 12x | 4 | Create scenario |
| 13 | Recent (slope) | Find the slope between two points | 2 | Slope from points |
| 14 | Recent (slope) | Which line is steeper: y = 4x or y = 3x + 1? Explain. | 3 | Compare slopes |
| 15 | Recent (equations) | Solve 3x + 7 = 22 | 2 | One variable equation |
| 16 | Recent (equations) | Solve 2(x - 5) = 14 | 2 | Distribute then solve |
| 17 | Recent (graphing) | Plot y = 2x + 1 from a table of values | 2 | Graph a line |
| 18 | Recent (mixed) | A line has slope 3 and passes through (0, 5). Is it proportional? Justify. | 3 | Prop discrimination |
| 19 | Cumulative | Simplify: -4 + (-7) + 3 | 1 | Integer arithmetic |
| 20 | Cumulative | Evaluate: (2/3) divided by (1/4) | 2 | Fraction division |
Notice what the set does. Problems 1 to 12 are the current unit at 60%, mostly interleaved but with lighter and heavier examples of the same substructure placed near each other so a struggling student can pattern match up the page. Problems 13 to 18 are the discrimination bomb at 30%: half are related enough to proportional relationships that a student running on autopilot ("apply the proportional strategy!") will get them wrong. Problem 18 in particular is the interleaving payoff, because a student has to notice the line starts at (0, 5) rather than the origin, which means it is not proportional despite looking proportional at a glance. Problems 19 and 20 are pure retrieval at 10%, cheap to include and diagnostic if students have forgotten.
The category label column is not optional. Ask students to write the label before they solve, not after. The act of naming the type is the discrimination practice; solving without labeling is procedural fluency, which is what blocked practice already delivers.
Try the Math Problem Set Generator on your next unit →
Spiral practice revisits prior units alongside the current one across weekly sets, usually with a distribution like 60% current, 30% recent, and 10% cumulative. Interleaved practice shuffles the order of problem types inside a set so students cannot use the context (all volume problems, so use the volume formula) to identify the strategy. Spiral is a bigger unit of organization; interleaving is a within set structure. A strong practice program uses both: interleaved sets inside a spiral schedule.
The strongest evidence is for topics where students have to choose among related but distinct procedures: geometry (volume of different solids), algebra (solving linear systems by different methods), fractions (adding vs multiplying), and probability (independent vs dependent events). Interleaving is less useful when there is only one strategy to apply (basic addition facts, single procedure warm ups). The rule of thumb: if the test question asks the student to choose a strategy, interleaving during practice helps. If the test question just asks the student to execute a procedure, blocked practice is fine.
No, but you should scaffold the transition and set expectations. Student preference in the moment tracks their in the moment fluency, not their learning. Rohrer and colleagues have shown that students consistently rate blocked practice as more effective even when they scored dramatically higher on the interleaved condition's delayed test. Show the class one of the Rohrer studies at the start of the unit ("here is why the practice is going to feel harder for a few weeks") and give them a category labeling prompt for each set so the discrimination work is visible instead of hidden.
A default that works across grades and topics is 60% current unit, 30% content from the last two to four weeks, and 10% cumulative from earlier in the year. Adjust the split by where you are in the unit: 80/15/5 in the first week of a brand new topic (students need consolidation), 60/30/10 in weeks 2 to 4 (standard spiral), and 30/30/40 before a benchmark or state test (cumulative review). Every practice set should touch at least two prior topics, even in the first week of a new unit.
Retrieval practice is the umbrella research (from Roediger, Karpicke, and colleagues) showing that repeatedly pulling information from memory strengthens learning more than reviewing the same information. The 10% cumulative slice of a spiral set is a retrieval practice slot: students have to pull up a topic they think they are done with. Low stakes weekly quizzes work the same way. Interleaving, spiral, and retrieval practice stack cleanly: a spiral worksheet that interleaves within topics and includes a small retrieval slice is combining all three effects in one set.