If you teach welding, auto, electrical, manufacturing, healthcare, IT, or construction, you probably run practice the same way every other CTE teacher does. You drill weld joint A until it looks right, then B, then C. That is blocked practice. It feels efficient, and your students leave confident. The same students often struggle a week later on a credentialing assessment. The research is consistent: interleaved practice (mixing problem types in one session) produces better long-term retention than blocked. Blocked practice feels easier, even though it works less well (Rohrer & Taylor, 2007; see Sources for the correct journal).
The problem with how we usually practice
Blocked practice has been the CTE default since shop classes moved into scheduled blocks. It lets students build momentum, rack up correct reps, and leave the lab confident. The trouble shows up later: on a delayed test, the same students do significantly worse than students who mixed problem types in the same window. Rohrer and Taylor (2007) first showed this with college math students. On a delayed test, interleaved students roughly doubled the blocked-group score. They scored slightly worse on the same-day test. The immediate-failure feedback is what makes the technique feel hard. The delayed payoff is what makes it work.
Kornell and Bjork (2008) extended the finding to category learning. Adult learners who studied paintings in mixed order learned to discriminate styles better than learners in blocked order, but only on a delayed test. Rohrer, Dedrick, and Burgess (2014) showed the same pattern for conditional-rule learning. Students must first recognize the problem type and then apply the right rule. That is exactly the structure of most CTE lab work. Hattie’s meta-meta-analysis of more than 800 studies puts interleaving in the top tier of instructional strategies. Its effect size sits well above the 0.40 hinge that marks an average year of instruction (Hattie, 2009). Brunmair and Richter (2019) confirmed the effect and its moderators: the gain is largest when problem categories are highly similar and the test is delayed.
PhillyCTE’s read: the absence of direct CTE studies is a gap in the literature, not a gap in the principle. CTE labs are one of the most natural settings in education for interleaving. The categories students must discriminate (weld types, fault codes, PLC ladder patterns, dosage routes, soil horizons, setup sequences) are precisely where the effect is largest.
The method (in plain English)
Interleaving is not a new curriculum; it is a reshuffle of practice time. Instead of five welds of type A then five of type B, build a 16- to 20-rep set that cycles randomly through A, B, C, and D. Same total reps, same equipment, same rubric — only the order changes. The win is in a same-minute reallocation, not in adding more practice time.
The hardest part is the conversation with students. Mixed practice feels harder, and students will mistake that for bad teaching. Tell them, in plain language, that mixed practice is supposed to feel harder and that same-day scores will look worse, while the win shows up on a one-week delayed test. If you do not explain this, half your students will quietly switch back to blocked practice the next day. You will lose the benefit (Learning Scientists, 2016; ACTE, 2022).
Six steps to run interleaved practice this week
Step 1. Pick a unit with 3 or 4 distinguishable categories. Welding: T-joint, lap-joint, butt-joint, corner-joint. Auto: misfire, fuel-trim, O2-sensor, and ignition DTC families. Electrical: open, short, and ground fault. Manufacturing/CNC: tool-offset, work-offset, tool-length-comp, spindle-orientation. Health science: PO, IV push, IV piggyback, IM, subcutaneous. IT: for-loop, while-loop, recursion, list-comprehension.
Step 2. Build a 16- to 20-rep mixed set. Aim for an even split across categories. One repeatable sequence is the every-other-other pattern: A-C-B-D-A-D-C-B-D-B-C-A-D-B-A-C-D. The order matters less than the mix. The point is that a student cannot predict the next problem type, so they have to identify it each time.
Step 3. Run the same total reps you would have run anyway. No new equipment time, no new prep, no new curriculum.
Step 4. Tell students why. Mixed practice is supposed to feel harder; the win is on the delayed test. If they expect the difficulty, they will not blame the strategy.
Step 5. Delay the test by at least one week. A same-day quiz will favor blocked practice; a one-week delayed quiz, performance assessment, or credentialing pre-test will favor interleaved. Build the delay into your unit plan first.
Step 6. Document what you did. A one-page note on the practice sequence, the categories mixed, the quiz delay, and the score difference from your prior blocked-practice cohort is enough. Bring it to your next program advisory meeting.
Classroom and lab implementation
Interleaving works best when categories are highly similar, when students must judge between them, and when the test is delayed (Brunmair & Richter, 2019). All three are usually true in a CTE lab. The two places teachers lose the gain are quiz timing and front-loading. Build the one-week delay into the unit plan first, and limit each mixed set to 3 or 4 categories students have already met. Interleaving pairs naturally with retrieval practice and spaced review (ACTE, 2022). Within one session, mix problem types. Across sessions, distribute practice over time.
What this means for your program
Action 1. Pick one unit this week and switch only its practice sequence. Pick a unit with 3-4 clearly distinguishable categories. Build a 16-rep mixed set. Run it on Monday. Document the result.
Action 2. Build a one-week delay into the unit quiz or performance assessment. Same-day quizzes will make interleaving look like it failed; delayed assessments are where it pays off.
Action 3. Tell students, in plain language, why the lab feels harder. Mixed practice is supposed to feel harder; if they know that, they will not quietly switch back to blocked practice.
Action 4. Bring the result to your next program advisory meeting — one paragraph on what you changed, what the data showed, and whether you would run it again.
Source: https://link.springer.com/article/10.1007/s11251-007-9015-8
