The methods
Recall practice
StrongWhat it is. Pulling an answer out of memory strengthens it more than reading it again.
How we use it. Every session starts with a few recall items. Phrases must be used in a sentence, not just recognised. Model answers appear after you try.
What the research found. Students who practised recall remembered 61% of a passage a week later, against 40% for students who reread it [1]. A later review of many studies found practice tests helped more than restudying [2].
Limits. Five minutes after studying, rereading looked better. The benefit shows up days later, so early scores in a session can mislead.
Spacing set by your test date
Strong for spacing, Moderate for the exact gapsWhat it is. Spreading reviews out beats cramming, and the best gap depends on how long you have before the test.
How we use it. We ask for your test date and keep review gaps shorter as it gets closer.
What the research found. In a study of more than 1,350 people, the best gap was about 20 to 40% of the time left when the test was a week away, and about 5 to 10% when it was a year away [3].
Limits. The study taught facts, not speaking or writing skills. Our exact gaps are our own estimate from it.
Reviews when you are about to forget
ModerateWhat it is. A model predicts when you are likely to forget a phrase and brings it back at that moment.
How we use it. Each phrase you learn has a predicted recall level. Phrases that are fading come back first.
What the research found. A model of this kind, fitted on 12.9 million practice records from a language app, cut prediction error by 45% or more compared with simpler schedules [4].
Limits. That study measured prediction accuracy and engagement, not test scores.
Focused feedback
Moderate to StrongWhat it is. Feedback works best when it explains what to change, why and how, in amounts you can act on.
How we use it. Up to three fixes per answer, ranked by how much they matter, each with an example. The rest are there if you want them.
What the research found. A review of 435 studies found an average effect of d = 0.48 for feedback. Feedback with more information reached d = 0.99. Simple corrective feedback reached d = 0.46 [5].
Limits. Results varied a lot between studies. The limit of three is our choice, to keep feedback usable. No study tested that exact number.
Practice at the right difficulty
EarlyWhat it is. Learning is fastest when practice is neither too easy nor too hard.
How we use it. We aim for you to get most objective questions right, but not all, and adjust as you go.
What the research found. A mathematical analysis found the best training accuracy for a broad class of learning algorithms is about 85% [6].
Limits. That result is about machine learning models, not people. We use it as a starting point and tune it.
Doing the task again
Moderate for writingWhat it is. Repeating a task soon after feedback improves the next attempt.
How we use it. You retry the same prompt straight after feedback, then the same task type with a new prompt a day and a week later.
What the research found. A review of 31 studies found that repeating a writing task improved accuracy, vocabulary and fluency, with smaller effects when the gap between attempts was longer [7].
Limits. The strongest results we have checked are for writing. Evidence for speaking is still being reviewed.
Practice in the real test format
Strong for practice effects, Early for language testsWhat it is. Practising on the same kind of test tends to raise scores on that test.
How we use it. Full mock tests follow the official section order, timing, number of listening plays and answer types.
What the research found. A review covering 134,436 test takers found an average gain of d = 0.26 on retesting, larger when the practice used identical forms [8].
Limits. Those were cognitive ability tests used in hiring and admissions, not language tests.
Speaking in small steps
ModerateWhat it is. Facing a feared speaking situation gradually makes it less frightening.
How we use it. You can record privately with no score, then get feedback on one area, then full scoring, then real exam timing. You move up when you choose.
What the research found. A review of 11 studies with 508 participants found large drops in public speaking anxiety after gradual exposure [9].
Limits. Those studies were about public speaking anxiety in clinical settings. PR Language Test is a practice tool, not a treatment.
Honest score ranges
ModerateWhat it is. Showing a likely range is more honest than a single number, and people still trust it.
How we use it. We show your practice estimate as a range of levels, never as an official score.
What the research found. In five experiments with 5,780 people, showing uncertainty in numbers lowered trust only slightly [10].
Limits. The study did not look at test preparation or test anxiety.
If then plans
Strong for following throughWhat it is. Deciding in advance when and where you will study makes you more likely to do it.
How we use it. Your study plan is built from your test date, and each session is saved as a plan such as "If it is Tuesday at 8 pm and I am home, then I do one listening section."
What the research found. Across 94 tests, plans of this kind had an effect of d = .65 on reaching goals [11].
Limits. The research measured following through on goals in general, not language test scores.
A calm start before mocks
Moderate, small effectWhat it is. Treating a racing heart as your body getting ready, and slow breathing, can take the edge off nerves.
How we use it. An optional two minute routine before each mock, also available on your phone on test day. You can always skip it.
What the research found. Across randomized trials, reframing stress gave a small gain in performance, d = 0.23, a little larger for speaking in front of others than for written tasks [12]. A separate review of 12 trials found breathing exercises lowered how stressed people felt [13].
Limits. The performance gain is small and shrank after a correction for missing studies. There is no evidence that breathing exercises raise test scores. This routine is not a treatment for anxiety. If you feel severe distress, please speak to a health professional.
Feedback about your answer, not about you
StrongWhat it is. Feedback helps when it points at the work. It often backfires when it points at the person, through rankings, labels or praise.
How we use it. Feedback describes your answer, for example "two verbs in the wrong tense, here is the fix", never "you are weak at grammar". We compare you only with your own earlier attempts.
What the research found. A review of 607 effects found feedback helped on average, but in over a third of cases it made performance worse, especially when it drew attention to the self [14].
Limits. The review covered many kinds of tasks, mostly outside language learning.