Evidence file: mixed math practice

Study label	Meaning
Randomized trial	Students, classes, or schools assigned by chance. The strongest single-study design.
Lottery study	Admission by lottery, which works like random assignment for those who applied.
Meta-analysis	A combined analysis of many studies; only as strong as the studies inside it.
Comparison study	Groups compared without random assignment. Useful, but differences may come from who chose what.
Lab study	Controlled experiment outside a regular classroom. Explains why; doesn't prove it works at scale.
Correlation	Two things move together. Shows a link, never cause. Used only for "why it matters."

Study label

Meaning

Randomized trial

Students, classes, or schools assigned by chance. The strongest single-study design.

Lottery study

Admission by lottery, which works like random assignment for those who applied.

Meta-analysis

A combined analysis of many studies; only as strong as the studies inside it.

Comparison study

Groups compared without random assignment. Useful, but differences may come from who chose what.

Lab study

Controlled experiment outside a regular classroom. Explains why; doesn't prove it works at scale.

Correlation

Two things move together. Shows a link, never cause. Used only for "why it matters."

Study	Design	Who	Compared with	Found	Limits
Rohrer et al. 2020(opens in a new tab)	Preregistered cluster randomized trial; meets WWC standards without reservations (WWC(opens in a new tab))	787 seventh graders, 54 classes, 14 teachers, 5 schools, one Florida district	Mostly blocked practice, same problems	61% vs 38% on an unannounced test a month later (d = 0.83)	Researcher-made test; one district
Rohrer, Dedrick & Stershic 2015(opens in a new tab)	Classroom experiment over 3 months	126 seventh graders	Blocked practice, same problems	Higher on immediate and delayed tests (d = 0.42 and 0.79)	Within-student design
Rohrer, Dedrick & Burgess 2014(opens in a new tab)	Classroom experiment over 9 weeks	7th graders	Blocked practice	72% vs 38% on interleaved vs blocked problem types, even for dissimilar types	One school
Taylor & Rohrer 2010(opens in a new tab)	Experiment	4th graders, prism problems	Blocked practice	77% vs 38% on a test one day later	Short delay; lab-like setting
Rohrer et al. 2020, textbooks(opens in a new tab)	Count of textbook problems	13,505 problems in six texts	—	Under 10% were mixed	Sample of six texts

Counter-evidence and cautions

Blocked practice often looks better during practice and on immediate tests. The benefit shows up at a delay (Taylor & Rohrer 2010, via AFT(opens in a new tab)).
Nearly all the strong evidence is in middle school math, from one research group. An independent replication with a representative East Coast 7th grade sample and state test outcomes runs through 2027 (IES(opens in a new tab)).