Quizzes are an established fact of the school experience. The pop quiz is a trope that occurs again and again in movies, always — literally always — accompanied by a groan from students. It would be easy to dismiss quizzes as either a form of teacher retribution for bad behavior or low-level recall that kids memorize information for and then immediately forget. But research has long demonstrated that frequent quizzing and testing can have a marked impact on how much learning a student retains. What we don’t know is whether that improvement is dependent on certain givens, or whether it has the undesirable side effect of increased anxiety, which actually decreases achievement.
A proposed study that began in 2022* is looking into this phenomenon. The researchers are interested in three effects of frequent testing/quizzing: achievement, text anxiety**, and how both of those might be further affected by subject, grade level, gender (girls have more anxiety than boys), type of test (summative or formative), and the duration of the intervention (how long were students subjected to the more frequent testing).
Researchers are conducting an international review of existing studies to look for effects. They have established some parameters for which studies will be included:
- Results had to be observed in a school setting (no laboratory studies)
- The intervention was specifically tailored for either primary or secondary, defined as K-12, and Kindergarten had to be a grade level (in some countries it isn’t)
- Only the testing frequency could be manipulated (i.e., they couldn’t have included other strategies)
- There had to be at least two “units” (schools or classes) in each portion of the intervention to control for something called the unit effect, where kids’ achievement changes not because of the intervention but because they are placed in a group, and,
- The follow up test to measure results had to be different than anything the students had already done to control for students who memorized the specific questions.
They cite some fascinating earlier research of the “testing effect” — the two main theories around why more frequent testing appears to help students retain and recall information. The first is “amount of processing:” the sheer frequency of exposure to the content aids in better correct recall of the content. Testing is one more avenue of exposure to content and therefore improves learning. The second theory involves the retrieval process itself and posits that retrieving content from memory strengthens the pathways to retrieval in the brain or creates new pathways for retrieval and it is these pathways that increase the probability of correct recall***. This last one — actively processing and retrieving content specifically in testing — has been proven to strengthen long term recall better than a more passive recall activity such as re-studying.
They also cite some important — really important — research demonstrating that there’s a sweet spot with frequent testing and to go outside that sweet spot results in over-testing which decreases student achievement. Like a lot of things in this life, you can have too much of a good thing.
We already know a few other things about testing. We know that somewhere between 10-40%**** of students report test anxiety. We also know students report more test anxiety with state tests than with classroom tests, perhaps not surprising given the high stakes of state tests and how some schools and even whole districts respond to them. A study out of the United Kingdom on middle school students showed that anxiety accounted for about 7% the negative variance in achievement — a considerable amount. A further U.S. study on high school students found that anxiety accounted for 2 – 15%*^ of the negative variance. More evidence in this area would greatly add to our understanding of how testing of any kind impacts students across a range of measures.
No results have been published yet. I am very interested in their eventual findings, particularly how the various givens — gender, subject, test type, grade level, and duration — are affected by this type of intervention. This sort of research is important because “frequent testing” is something some well-intended person in education might hear and utterly run amok with before fully understanding all its effects and implications, especially the ones that negatively impact learning. I will certainly provide an update when I have one.
_________________________________________________________________
*Thomsen MK, Seerup JK, Dietrichson J, Bondebjerg A, Viinholt BCA. PROTOCOL: Testing frequency and student achievement: A systematic review. Campbell Syst Rev. 2022 Jan 7;18(1):e1212. doi: 10.1002/cl2.1212. PMID: 36908658; PMCID: PMC8742135.
**This dovetails neatly with brain research showing that a brain under threat (and testing could definitely be perceived as a threat) experiences cognitive suppression — it literally stops thinking analytically and devolves into fight/flight/freeze/appease mode. One study found that reading comprehension went down measurably in the presence of test anxiety.
***My money is on this theory. It also tallies perfectly with existing brain research confirming that using pathways repeatedly results in stronger, faster connections between areas of the brain and faster, more accurate recall of information. Conversely, not using repeated, active recall causes the brain to “prune” connections that aren’t being used. This is why cognitively undemanding learning such as memorizing dates often evaporates after a test — the student no longer needs the information so the brain prunes those connections.
***I really wanted this number to be more precise, but when I looked into the study it discussed at length the vast differences in anxiety experienced by children during testing. That 40% is the anxiety experienced by minority students. Interestingly, though, while ethnicity was a predictor of anxiety it was not a predictor of negative variance. Perhaps more surprising was that high SES was a predictor of lower anxiety but also had little effect on the negative variance. There’s a lot more here — like gifted kids experiencing more anxiety when testing against their gifted peers and girls experiencing more anxiety than boys — but suffice it to say that kids experience widely different rates of text anxiety, hence the big range.
*^See above for why the range is so broad. It’s interesting to note, though, that the UK study’s finding on negative variance is smack in the middle of the US study’s finding. This is doubly interesting since the UK finding was among what we could call 6th graders and the US study was of 11th graders. It makes me wonder if what we see between the two studies is the development of anxiety as kids begin to navigate those secondary spaces.