Does quizzing yourself beat rereading? Memory research you can run at school
Rereading feels like studying, but in a 2006 study recall tests worked better a week later. Two word lists and a group of classmates are enough to test it.
"I study all night and still forget everything." It comes up in the week before every exam, often with a worse thought attached: maybe my memory is just bad. A psychologist would check something else first: the way you study may matter more than the hours, and you can test that yourself.
Below is one well-known study, and a plan you could carry out with classmates.
Why rereading feels like learning
You read the chapter. Then you read it again, maybe five times. Each pass feels easier. On Friday, in front of a blank answer sheet, much of it has faded.
Two skills are at work. Recognizing means seeing something and knowing you've seen it before. Recalling means producing it from memory with nothing in front of you. Rereading gives you a lot of practice at the first and little at the second, but an essay or short-answer question asks for the second. Feeling familiar is not the same as remembering.
That's a testable claim: the method that feels best in the moment may not be the one that lasts. In the study below, students who spent every study period rereading were the most confident they'd remember the passage a week later, and they remembered the least.
A late night at the desk can feel productive, but how familiar your notes feel can be a poor guide to what you'll recall a week later.
Two study methods, one change, a test later
Memory researchers test "what makes memory stick?" with a controlled experiment: two study methods, only one change between them, and a test later.
The one change is the study method, called the independent variable. Some material is reread. Other material is practiced with a recall test, where you write down everything you remember on a blank page.
Everything else stays the same, including the time spent and the final test.
Have a question you want to investigate?
Explore our faculty-guided programs, or tell admissions what you are curious about.
The outcome is how much people recall on that final test, called the dependent variable.
"Later" matters. Five minutes after studying, rereading can look better because the material is still fresh. What you care about is what's left days later.
A good design also rules out easy explanations. Taking a quiz is one more pass through the material, so maybe testing helps only because of that extra look. To check, Roediger and Karpicke gave their rereaders the full passage again each time the other students took a test. Test-takers saw only what they could pull from memory, with no feedback. If anything, that favored rereading.
The 2006 experiment where rereading won, then lost
Henry L. Roediger III and Jeffrey D. Karpicke of Washington University in St. Louis published "Test-Enhanced Learning" in Psychological Science in March 2006. In their second experiment, 180 undergraduates aged 18 to 24 learned one short science passage ("The Sun" or "Sea Otters," from a TOEFL preparation book). Over four back-to-back periods, each student either spent all four reading the passage, read it for three and then took one recall test, or read it for one and then took three recall tests. Reading periods lasted 5 minutes. Half took the final test 5 minutes later and half came back after a week, making six groups of 30.
After 5 minutes, rereading won: 83% of the passage's ideas recalled, against 71% for the read-once, test-three-times group. After one week the order flipped. The repeated-testing group recalled 61% and the repeated-reading group 40%, even though rereaders had read through the passage 14.2 times on average and the tested group only 3.4 times.
Confidence went the other way. On a 7-point scale, rereaders rated how well they'd remember the passage in a week at 4.8, against 4.0 for the repeated-testing group.
The limits are real. These were students at one American university, tested in small groups on two short passages, with 30 people per group. The final test was the same kind of recall test the tested group had practiced, which the authors see as part of why testing helped. The study can't tell you how big the effect would be for a semester of high school chemistry or for younger students.
A recall test can be as simple as a blank card or sheet: the work is in pulling the words out of memory, not in looking at them again.
A word-list experiment for your classmates
Your version differs from the 61% versus 40% result in two ways: word lists instead of prose, and one quiz instead of three. The closer match is the paper's first experiment, where each student used both methods, one per passage, and a single recall test gave 56% recall after a week against 42% for rereading. Most earlier testing-effect studies used word lists, the authors note, but don't expect your numbers to match theirs. Your result is what you'll report.
Step 1: Make two lists of 20 words (about 2 hours). If one list is easier, you'll be measuring the lists, not the methods. Use common, concrete nouns of similar length that aren't related to each other. For Korean-speaking classmates, the National Institute of Korean Language's free vocabulary list for learners of Korean (2003, in Korean) has 5,965 words with a frequency rank, a part of speech and a level from A to C, so you can take all 40 nouns from one level. For English, Ghent University's SUBTLEX-US gives frequencies counted from about 51 million words of film and TV subtitles. After approval (Step 3), a short pilot with two or three volunteers can show whether one list is clearly easier.
Step 2: Reread one list, quiz the other (about 25 minutes per session). Your classmates study List A for 2 minutes, then reread it for 2 minutes. They study List B for 2 minutes, then spend 2 minutes writing every word they remember on a blank sheet, with no answers shown afterward. Between periods, give 2 minutes of simple math (the original study used multiplication problems) so people can't silently repeat the words. Make four versions of the session so that each list is the quizzed one for half your classmates, and the quizzed list comes first for half. That's counterbalancing, and it keeps an easier list or the order from deciding your result.
Step 3: Test classmates, with school approval (2 to 4 weeks to arrange). Get approval before you run any session, including the pilot. Ask your teacher and school, have your plan reviewed by an IRB (an ethics review board) where required, and collect written consent, including a parent's consent for anyone under 18. Let anyone stop at any time, and record scores under code numbers, not names. Aim for 20 or more classmates; with fewer, you can easily miss a real difference because of chance variation. Ask everyone not to practice the words during the week.
Step 4: Compare scores one week later (15 minutes of testing, then an afternoon of analysis). Ask each classmate to write every word they remember from both lists on one blank sheet. When you score, sort the words by list and give one point per correct word. In Google Sheets, put one classmate per row, one column per list, and use AVERAGE on each column. =T.TEST(B2:B25, C2:C25, 2, 1) runs a paired t-test: 2 means two-tailed, and 1 means paired, because the same people produced both scores. You get a p-value, roughly the chance of seeing a difference at least this large if both methods worked equally well. JASP, free software supported by the University of Amsterdam, runs the same test under T-Tests.
You'll end up with a bar chart of two averages, a p-value, and notes on what might have affected the result. That's enough for a research poster or a short paper of four to six pages. A small or reversed difference still counts. Explaining it is part of the work.
Your main result can be one chart: the average score for the reread list next to the average for the quizzed list.
What to ask instead of "how should I study?"
Spell out the people, the two conditions, what you'll measure, and when.
Too big
Workable
What is the best way to study?
Do 10th graders recall more words from a list they were quizzed on than from a list they reread, one week later?
Does testing help memory?
With word lists, does rereading win on a test 5 minutes later while quizzing wins after a week, as Roediger and Karpicke found with prose?
Why do students trust rereading?
Right after studying, do classmates predict higher scores for the reread list, and how close are those guesses to their scores a week later?
For the second question, you'll need an extra group who take the final test after 5 minutes. For the third, you'll add one question at the end of your session: "How many words from each list will you remember next week?"
The testing effect suits a literature review or a research proposal as well as it suits an experiment.
A literature review answers a question from published studies. Did quizzing help as much in middle and high school classrooms, and did those studies measure it with free recall, multiple choice or unit exams? Search ERIC, the US Department of Education's education research library, or Google Scholar for "retrieval practice," researchers' term for studying by recalling.
A research proposal plans a semester-long classroom study. Would high school chemistry classes that open each lesson with a short ungraded quiz score higher on the semester final than classes that reread? Your proposal would pose that question and lay out the method (classes at several schools assigned at random to quizzing or rereading), the data (final exam scores and earlier grades), and what has to be approved first: an IRB review, school permission, and consent from students and parents.
Your forgetting was the question. Now research it.
CRI students research with professors in many fields, such as biology, economics, chemistry and history, and write their own paper, from original research to literature reviews and research proposals. If a question like this one keeps coming back to you, CRI's research programs show how students turn it into a finished project.
Sources
Roediger, H. L., III, & Karpicke, J. D. (2006). Test-enhanced learning: Taking memory tests improves long-term retention. Psychological Science, 17(3), 249-255. link (free copy: PDF)
Everding, G. (2006, March 1). Repeated test-taking better for retention than repeated studying, research shows. The Source, Washington University in St. Louis. link
National Institute of Korean Language (2003). Vocabulary list for learners of Korean (in Korean). link
Brysbaert, M., & New, B. (2009). Moving beyond Kučera and Francis: A critical evaluation of current word frequency norms and the introduction of a new and improved word frequency measure for American English. Behavior Research Methods, 41(4), 977-990. Word lists: SUBTLEX-US, Ghent University. link