Skip to the content
Peabody TeachFor teachers and grown-ups

Differentiated instruction: what it is, which forms work, and what the research says

Adjusting teaching, tasks, support or grouping within a class so students at different starting points can all make progress toward the goal.

Evidence MixedArea How students learn and remember4 sources, about 7 minutes to readChecked October 2026

What the research says

  1. Strong
  2. Promising
  3. Mixed
  4. Thin
  5. Contradicted

Mixed, because "differentiation" covers many different practices. In primary schools, differentiation built into computer systems or a broader program had small to moderate benefits, while ability grouping alone had a small negative effect for low-ability students. Across a century of research, grouping within a class helped but sorting students into separate classes by ability didn't. In secondary schools, too few rigorous studies exist to draw firm conclusions. The forms worth using are flexible, specific and tied to assessment.

Rated Mixed. No rigorous study supports it. Two rigorous studies found it works only in some forms or for some students. One study couldn't tell either way. One review or weaker study adds context but doesn't move the rating. How we rate the evidence

Checked against each source in October 2026.

The studies behind this rating

  1. Deunk et al. (2018)Meta-analysisMixed

    Differentiation through computer systems, and differentiation built into a broader program or reform, had small to moderate positive effects on language and math. Grouping students by ability, between or within classes, had a small negative effect for low-ability students and no effect for others.

    Sample 21 studies in primary education since 1995, 78 effect sizes, language and mathWho School-age childrenRead it at the source (opens in a new tab)
  2. Steenbergen-Hu et al. (2016)Meta-analysisMixed

    For this card: grouping within a class (0.19 ≤ g ≤ 0.30) and grouping across grades for one subject helped, but sorting students into separate classes by ability didn't (0.04 ≤ g ≤ 0.06).

    Sample Two second-order meta-analyses: 13 ability grouping meta-analyses and 6 acceleration meta-analysesWho School-age childrenEffect within-class grouping 0.19 ≤ g ≤ 0.30; cross-grade subject grouping g = 0.26; between-class grouping 0.04 ≤ g ≤ 0.06Read it at the source (opens in a new tab)
  3. Smale-Jacobse et al. (2019)Systematic reviewCan't tell

    Only eleven studies qualified, and they defined differentiated instruction in many different ways. Most showed small to moderate positive effects on achievement, which the authors read as some indication of benefit, but they conclude more research is needed before drawing convincing conclusions.

    Sample 11 empirical studies of within-class differentiation in secondary schools, 2006–2016Who School-age childrenRead it at the source (opens in a new tab)
  4. Gersten et al. (2009)Practice guide (adds context, doesn't move the rating)Supports

    Recommendation 2: Provide time for differentiated reading instruction for all students based on assessments of students’ current reading level. Panel's rating: Minimal.

    Recommends time for differentiated reading instruction for all students based on assessments of their current reading level, and rates the evidence Minimal: sensible, but not tested rigorously on its own.

    Who School-age childrenRead it at the source (opens in a new tab)

What it is

Differentiated instruction means planning so that students who start in different places can all learn. It isn't one technique; it's a family of decisions about what to change for whom. Teachers usually differentiate one or more of:

  • Content or level: the same goal, with different starting points or amounts of practice (more worked examples for some, extension problems for others).
  • Support: scaffolds such as sentence starters, manipulatives, a word bank, or a reread passage, which are removed as students grow.
  • Grouping: small groups for targeted teaching, which change as students progress.
  • Product: different ways to show the same learning.
  • Pace: some students move on sooner, others get more time.

Two things are often confused with differentiation. Tracking, sorting students into separate classes by ability for the year, is a school structure, not classroom differentiation, and the research on it (below) is discouraging. Learning styles, teaching visual learners visually and so on, isn't differentiation either, and rigorous tests of it have found no benefit; see the learning styles myth page.

In the research, the word has been used for so many practices that reviews struggle to pool them. Annemieke Smale-Jacobse and colleagues found that the eleven secondary-school studies they could include defined differentiated instruction in many different ways: teacher training programs, ability grouping and tiering, individualization, mastery learning, mixed-ability grouping and more.

Why it works

Students in one class can be years apart in what they know. Teaching only to the middle leaves some lost and others bored. Differentiation is the attempt to teach the same goal to students who need different things to reach it.

The best evidence comes from Marjolein Deunk and colleagues' 2018 meta-analysis of 21 studies of differentiation in primary schools, covering language and math. Two forms had small to moderate positive effects: differentiation through computer systems (software that adjusts practice to each student), and differentiation built into a broader program or school reform. Grouping students by ability, whether between or within classes, had a small negative effect for low-ability students and no effect for others. The authors conclude that grouping alone isn't enough to count as differentiation.

Saiying Steenbergen-Hu and colleagues looked at a century of research on ability grouping and acceleration, pooling earlier meta-analyses. Grouping within a class for particular lessons helped, as did grouping students across grades for one subject, and special grouping for gifted students. Sorting students into separate classes by ability (tracking) didn't help, and the effects were similar for high-, middle- and low-ability students. Accelerated students outperformed same-age peers who weren't accelerated.

The two reviews look different on within-class grouping (Deunk found a small negative effect for low-ability students; Steenbergen-Hu found within-class grouping helped overall). One way to square them: grouping helps when the teaching in each group is targeted and the groups change; it hurts low achievers when it means a permanently slower track.

The practical rule Differentiate by assessment, not by label. Group students by what a recent check shows they need for this skill, teach that group directly, and regroup when the next check shows they've moved.

How to start, by grade

Grades K–2

  1. Use quick checks (letter sounds, counting, a short reading) to form small groups by specific skill, not by an overall label.
  2. Teach small groups while others work at independent centers; a centers rotation board on the projector keeps rotations running.
  3. Regroup every few weeks based on new checks, so no child stays in the "low group" by default.
  4. Use manipulatives and picture supports for children who need them, and fade them as children get faster.

Grades 3–5

  1. Start each unit with a short pre-check and plan two or three entry points to the same goal.
  2. Use tiered practice: everyone works on the same skill, with a core set, more scaffolded problems for some and extension problems for others.
  3. Pull small groups for reteaching after an exit ticket shows who missed the idea.
  4. Let students who already know the material move to extension work or accelerate, rather than repeating it.
  5. Keep whole-class teaching for shared content and discussion; differentiate the practice, not everything.

Grades 6–8

  1. Keep mixed-ability classes, and differentiate within them through flexible small groups and tiered tasks.
  2. Use adaptive practice software or leveled practice sets for skills students are at very different points in.
  3. Give scaffolds on access to grade-level texts (audio, vocabulary previews, guiding questions) rather than giving struggling readers easier content only.
  4. Track who's in which group and make sure groups change across units.

Grades 9–12

  1. Differentiate support and pace within courses, and offer acceleration for students ready to move on.
  2. Use frequent short checks to decide who needs reteaching before the unit test.
  3. Be wary of permanent tracks that lock students into lower courses; the research on between-class grouping is discouraging.

Examples

  • Flexible math groups from an exit ticketGrades 2–6

    Monday's exit ticket on regrouping shows three groups: eight students have it, twelve nearly have it, six don't. Tuesday: the first group does challenge problems, the middle group does guided practice in pairs, and the teacher works with the six using base-ten blocks. Thursday's check reshuffles the groups.

  • Tiered practice setGrades 3–8

    All students practice adding fractions with unlike denominators. Set A starts with fraction bar pictures beside each problem. Set B is standard problems. Set C adds word problems and three-fraction sums. Students start where the pre-check suggests and move up when they get four in a row right.

  • Same text, different supportsGrades 4–8

    The whole class reads the same grade-level article. Some students hear it read aloud first; some get a vocabulary preview of five key words; some get guiding questions in the margin. Everyone joins the same discussion and answers the same final question.

  • Reading groups that changeGrades K–3

    Small reading groups are formed from a phonics check, not from an overall reading level, and reformed every four to six weeks. A child who masters short vowels moves to the next group the following week, instead of waiting for the end of the term.

  • Acceleration in one subjectGrades 3–8

    A fourth grader who tests well above grade level in math joins the fifth-grade class for math only, or works through the next grade's unit with a different set of problems. Steenbergen-Hu's review found accelerated students outperformed same-age peers who weren't accelerated.

Common mistakes

Fixed ability groups.
Groups that never change become tracks, and ability grouping alone had a small negative effect for low-ability students in Deunk's meta-analysis. Regroup often, based on recent evidence of what each student needs.
Lowering the goal for some students.
Differentiating the support is different from giving struggling students less content. Keep the goal and change the route to it wherever you can.
Differentiating by learning style.
Sorting students into visual, auditory and kinesthetic groups isn't supported: experiments that matched teaching to students' preferred style found no benefit. Differentiate by what students know and can do instead; see the learning styles page.
Writing three lesson plans for every lesson.
That isn't sustainable. Teach the core whole-class, differentiate the practice, and use small groups for the students who need reteaching or extension.
Differentiating without checking.
Without a recent check, groups are based on guesses and old labels. Use short, frequent checks to decide who needs what.

Time and cost

Time: planning time is the main cost: pre-checks, tiered tasks and small-group lessons. Starting with practice sets and small-group reteaching keeps it manageable.

Money: low for teacher-made differentiation; adaptive software costs money, and its quality varies.

Training: moderate to high. Managing small groups while others work independently is a skill, and so is designing tiered tasks that keep the goal the same.

Debates and open questions

Is the evidence weak because differentiation doesn't work, or because it's hard to study? Mostly the second. Smale-Jacobse and colleagues found few rigorous studies and many definitions. When reviews pull out specific practices, some clearly help (within-class grouping, acceleration, differentiation in broader programs) and some don't (tracking).

Within-class ability grouping: helpful or harmful? Steenbergen-Hu and colleagues found within-class grouping helped across ability levels; Deunk and colleagues found homogeneous ability grouping had a small negative effect for low-ability primary students. The difference may lie in what happens inside the groups, and whether they're flexible.

Technology. Deunk and colleagues found computer-based differentiation had positive effects, which earlier reviews didn't cover. Software varies a lot in quality, so look for products with evidence of their own.

Tools that do it

Free, on this site, nothing to sign up for. Class lists stay in your browser.

Questions teachers ask

Does differentiated instruction work?
Some forms do, some don't. Within-class grouping, acceleration and differentiation built into a broader program or computer system have positive evidence; sorting students into separate classes by ability doesn't. The term covers so much that the overall rating is Mixed.
Is differentiation the same as ability grouping?
No. Ability grouping is one tool, and used alone and kept fixed it can hurt low-achieving students. Differentiation is broader: adjusting support, pace, tasks and groups based on what students need, with groups that change.
How do I differentiate without burning out?
Teach the core lesson to everyone, differentiate the practice with tiered sets, and use short small-group sessions for reteaching and extension. Plan two or three levels, not one per student.
What about gifted students?
Steenbergen-Hu and colleagues found special grouping for gifted students and acceleration both had positive effects on achievement. Letting students who already know the material move on is a form of differentiation with good support.

Cite this card

Peabody Teach. (2026, October 29). Differentiated instruction: what it is, which forms work, and what the research says. https://teach.peabody.kids/library/practices/differentiated-instruction

Written for teachers by Peabody Teach from the sources listed above, which we read on the publisher's or clearinghouse's own page. We summarize and link; we don't copy. No ads, no affiliate links. Something wrong or out of date? Tell us.