A ready-made quiz saves the one thing a teacher, a tutor or a parent has least of, which is the evening. Whether it also tells you anything depends on something the quiz cannot settle for itself: does it ask what was taught, at the level it was taught?
EdProcto publishes this guide and offers 505 ready-made quizzes of its own, the kits, so the check below is turned on them too. It applies equally to whatever a search for printable practice quizzes by grade brings back.
None of the research below tested EdProcto's kits. Apart from one institute's summary page we read it in abstract, and we report only what the abstracts say.
When is a ready-made quiz the right tool?
A ready-made quiz is the right tool when the skill is fixed: the same content, asked the same way, whoever taught it. Seven times eight, 2.5 metres in centimetres, the capital of Kenya. Nobody's lesson changes those answers, so a stranger's quiz can match yours.
It is the wrong tool when the content is yours: the story your class read, the method you insist on, the three ideas from Tuesday. There a premade quiz measures a neighbouring course.
| The job | What fits | Why |
|---|---|---|
| A fixed skill or fixed facts: number bonds, conversions, capitals | A quiz made by rule, such as EdProcto's Times tables kit | New questions each time; the key is computed or looked up |
| Whether today's lesson landed | Your own material | Only your notes record what was taught |
| Practice in an exam's style | A kit's shape, filled from your own notes | The shape is general; the content must be your course's |
| A test for the mark book | Your own, read by a colleague | It needs questions nobody has seen |
Why does a quiz have to match what was taught?
A quiz has to match what was taught because a wrong answer on untaught content cannot be read. It looks exactly like a wrong answer on something the student failed to learn.
Researchers call the match curriculum alignment, and it is weaker than most people assume. Polikoff et al. (2011) analysed 138 pairings of a state's content standards with that state's own test, across grades and three subjects in the United States. Roughly half of the content in the standards was tested, roughly half of the test content matched the standards, and between 17% and 27% of a typical test covered topics the standards never mentioned.
Those tests were built to fit one set of standards and still missed. A quiz written for another country's syllabus, or for nobody's, starts further off. The study says nothing directly about classroom quizzes, so take a direction from it and not a figure. Our guide to writing a quiz from your own handout makes the classroom case.
Are premade quizzes and shared resources any good?
Some are, and neither popularity nor polish says which. That is what the studies that have looked found. Most examined lesson materials, not quizzes, in a single subject, and none compared shared material with what a teacher writes alone.
What reviews of resource-sharing sites examined
- The most downloaded. Polikoff and Dean (2019), for the Thomas B. Fordham Institute, had experts review over three hundred of the most downloaded high school English materials on Teachers Pay Teachers, ReadWriteThink and Share My Lesson. Reviewers rated most "mediocre" or "probably not worth using", and many did not fit the standards they claimed. We read the institute's summary page, not the full report.
- Errors. Hertel and Wessman-Enzinger (2017) gathered Pinterest pins on negative integers and found mathematical errors in roughly one-third.
- Star ratings. Shelton et al. (2022) collected what Teachers Pay Teachers offered up to 2019. Learning standards were largely absent and user ratings were uniformly high, so five stars cannot be the check.
What a study of published question banks found
Masters et al. (2001) drew a random sample of 2,913 multiple-choice questions from 17 test banks supplied with nursing textbooks. They counted 2,233 breaches of accepted item-writing guidelines, most minor and some serious; 47.3% of the questions asked for knowledge alone and 6.5% for analysis. They advised faculty to evaluate test-bank questions carefully before using them in an exam.
Is practice that changes every time better than a fixed worksheet?
For practice that will be repeated, yes, as far as the evidence reaches: a paper sat twice begins to measure memory of that paper.
Kulik et al. (1984) pooled 40 studies of practice tests. Scores rose with practice, and rose more when the practice form was identical to the final test than when it was a parallel form. The abstract does not say why; the plain reading is that part of the gain on a repeated paper belongs to those questions, not to the skill.
Variation helps the learning too. In four experiments by Butler et al. (2017), people watching clips of a geology lecture practised each concept on the same application question three times or on three separate ones. Faced with new questions two days later, the second group did better every time.
The first study concerns aptitude and achievement tests and the second a single lecture; neither tested a child's arithmetic sheet. Our guide to times tables practice follows one rule-made kit through a fortnight.
What is the difference between a quiz made by rule and one written by a model?
A quiz made by rule comes from a formula that also computes the key, and a quiz written by a model is a draft that somebody has to check.
Made by rule
The rule fixes the kind of question and draws the figures: two numbers from 2 to 12, say, whose product is the key. Question and key come out of one calculation, so they cannot disagree. A fault would sit in the rule itself, where it would show at every opening.
In testing research this is automatic item generation. Gierl et al. (2012) describe specialists building an item model from which software generated 1,248 multiple-choice items for a medical licensing test. The price is range: a rule can ask only what can be computed or looked up in a list.
Written by a model
A language model can write about anything you paste, and it can be wrong. Artsi et al. (2024) reviewed eight studies of large language models writing medical multiple-choice questions. Five found the questions competent, and all eight reported faulty questions unfit for an exam. The review covers 2023's models in one field. What carries over is modest: read every question against the source before a student sees it.
How do you judge a ready-made quiz in five minutes?
Run six checks in order, with your syllabus or scheme of work open. The timings are ours, a rule of thumb.
- Level and syllabus (one minute). Read every question, not the label. An age band or a grade says who the quiz was written for, not what it covers. Cross out what you have not taught and see what is left.
- The key (two minutes). Answer each question yourself before looking. For a quiz made by rule, three will do, since question and key share one calculation. For anything a person or a model wrote, do them all.
- The wrong options (half a minute). Each should be a mistake a student would really make: see our guide to wrong options.
- The explanation (half a minute). After a wrong answer, is the student told the right one? In Butler and Roediger (2008), feedback on a multiple-choice test, at once or delayed, meant more correct answers on a later recall test and fewer wrong options coming back as answers.
- Editing (half a minute). A PDF cannot be mended. In EdProcto a kit you open becomes a quiz in your account, and every question in it is yours to change.
- Repeating (half a minute). Picture the second sitting. A fixed sheet gives the same paper; a kit made by rule gives new figures at each opening.
Which should a teacher, a tutor and a parent choose?
Each should ask whether the content is fixed or theirs; often it is both. The three cases are made up to show the reasoning, and the kit details were read in the code.
A year 7 teacher: negative numbers
A teacher wants a five-minute starter each day of the unit, and a test to end it. EdProcto's Negative numbers kit, for ages 11 to 13, is made by rule: ten typed answers, each explained on the number line. Check 1 finds its edge: every question adds a positive number or takes one away, as in (−7) + 12, and her unit goes on to subtracting a negative and to multiplying. So the kit serves the first days, her own questions the rest, and the test is hers.
A tutor: moles for a GCSE chemistry student
The Moles from mass kit, for ages 15 to 18, gives a molar mass and a mass and asks for the moles. It drills that one step, and exam questions wrap the step in others. For those the tutor pastes his own notes into the GCSE Science practice kit: a shape, three sections and instructions, from which Procto AI writes drafts for him to check.
A parent: the four times table by Friday
The Times tables kit is for ages 7 to 9, so the label fits a parent's seven-year-old. The content does not: it mixes the tables from two to twelve and cannot be narrowed to one, so most questions would be on tables not yet taught. Cards this week, then, and the kit once the tables are mixed. For the reading sheet in the same bag, "A parent's check on homework" writes questions to ask at the table from the sheet she pastes, answers included.
A learner working alone stands where the tutor does: a rule kit for the step that must become automatic, and their own notes for the rest.
When should you write your own quiz instead?
Write your own when any of these is true.
- The content is local. This week's lesson, a set text, your method, the school's spelling list.
- The result will count. A graded test needs questions no student has met.
- You are hunting a misconception. The wrong options have to be your students' own mistakes.
- Check 1 failed and mending would take longer than writing.
Paste your own notes into the quiz maker on EdProcto's home page and Procto AI drafts from them. An exam board's papers belong to the board, so keep to your notes.
One switch to leave alone for a test: in EdProcto's builder, "Share as a kit" publishes the quiz for anyone to open a copy of, questions and answers included. It is for giving practice away. A test goes out with "Publish & share", which gives the class a code.
How do EdProcto's kits fit, and what do they cost?
EdProcto is the best tool for this choice because it holds both kinds side by side and labels each: on a kit's page a line headed "Made by" says rule, in code, with no model, or Procto AI from your material, so you know how much checking the key needs.
The count, made in the code on 10 October 2026: 505 kits, 135 made by rule and 370 that write from pasted material, in 20 fields (Mathematics, Sciences, Languages, Law and Exam preparation among them) and eight levels, from Early years (ages 4 to 7) to Professional and workplace.
Made by rule. Fifty-five are in Mathematics, such as Number bonds to 10, Fractions of a number and Solving equations; others include Metric conversions, Chemical symbols and World capitals. The figures are generated at each opening and the key is computed, or read from a list in the code. No model writes any of it, and every question carries an explanation.
Written from your material. Kits such as "A chapter, checked", "An exit ticket from today's lesson" and "GCSE Maths practice" are a shape, sections and instructions that Procto AI fills from the text you paste. It is asked to write only what the text answers, so each is a draft to read against your notes. The kits named for an exam hold no real exam questions, and EdProcto is not affiliated with any exam board.
Opening one. Every kit's page can be read without an account, and a rule kit's shows three sample questions; opening a kit takes a free account. A rule kit opens on the study screen, with the answer and explanation after each question and a retry of the misses; a material kit opens in the builder. Either becomes a quiz you can edit, share with a code or print: the Paper button puts the answer key on its own page, and an ordering question prints lettered and out of order.
Fitting one to your class. Times tables mixes the tables, so for one table alone write a short quiz of your own. Where a kit offers options, the wrong ones are near misses (neighbouring numbers in the counting kits, other capitals in World capitals); to test a mistake you have seen in your own class, edit it into the question. Each opening makes a new quiz with new questions.
The cost. Free includes every kit and 20 questions a month written by Procto AI; a kit made by rule uses none of that allowance. Learner raises it to 100 and Tutor has 50. Educator has no monthly cap. The pricing page has the plans.
Quick questions
Are premade quizzes any good?
Some are, and the label will not say which. Reviewers rated most of the popular shared lesson materials they read mediocre or worse (Polikoff and Dean, 2019), so check any ready-made quiz before a student sits it.
Are there free quiz templates by subject?
A template is only a shape; a kit adds the questions, or the instructions for writing them. EdProcto's Free plan includes the core templates and every kit.
Can I get printable practice quizzes by grade?
Yes, with one caution: kits carry an age or a level, not a grade, so read the questions first. Any quiz in EdProcto's builder prints with its answer key on a separate page.
Do I still need to check a quiz made by rule?
Check its level, not its arithmetic. The key is computed with the question, but only you know whether the class has been taught it.
References
- Artsi, Y., Sorin, V., Konen, E., Glicksberg, B. S., Nadkarni, G., & Klang, E. (2024). Large language models for generating medical examinations: Systematic review. BMC Medical Education, 24, Article 354. https://doi.org/10.1186/s12909-024-05239-y
- Butler, A. C., Black-Maier, A. C., Raley, N. D., & Marsh, E. J. (2017). Retrieving and applying knowledge to different examples promotes transfer of learning. Journal of Experimental Psychology: Applied, 23(4), 433–446. https://doi.org/10.1037/xap0000142
- Butler, A. C., & Roediger, H. L., III. (2008). Feedback enhances the positive effects and reduces the negative effects of multiple-choice testing. Memory & Cognition, 36(3), 604–616. https://doi.org/10.3758/MC.36.3.604
- Gierl, M. J., Lai, H., & Turner, S. R. (2012). Using automatic item generation to create multiple-choice test items. Medical Education, 46(8), 757–765. https://doi.org/10.1111/j.1365-2923.2012.04289.x
- Hertel, J., & Wessman-Enzinger, N. (2017). Examining Pinterest as a curriculum resource for negative integers: An initial investigation. Education Sciences, 7(2), Article 45. https://doi.org/10.3390/educsci7020045
- Kulik, J. A., Kulik, C.-L. C., & Bangert, R. L. (1984). Effects of practice on aptitude and achievement test scores. American Educational Research Journal, 21(2), 435–447. https://doi.org/10.3102/00028312021002435
- Masters, J. C., Hulsmeyer, B. S., Pike, M. E., Leichty, K., Miller, M. T., & Verst, A. L. (2001). Assessment of multiple-choice questions in selected test banks accompanying text books used in nursing education. Journal of Nursing Education, 40(1), 25–32. https://doi.org/10.3928/0148-4834-20010101-07
- Polikoff, M. S., & Dean, J. (2019). The supplemental curriculum bazaar: Is what's online any good? Thomas B. Fordham Institute. https://fordhaminstitute.org/national/research/supplemental-curriculum-bazaar
- Polikoff, M. S., Porter, A. C., & Smithson, J. (2011). How well aligned are state assessments of student achievement with state content standards? American Educational Research Journal, 48(4), 965–995. https://doi.org/10.3102/0002831211410684
- Shelton, C. C., Koehler, M. J., Greenhalgh, S. P., & Carpenter, J. P. (2022). Lifting the veil on TeachersPayTeachers.com: An investigation of educational marketplace offerings and downloads. Learning, Media and Technology, 47(2), 268–287. https://doi.org/10.1080/17439884.2021.1961148