How to Make Two Versions of the Same Test (Version A and Version B)

The short answer: To make two versions of the same test, keep the blueprint identical (same topics, same number of items, same difficulty balance) and vary the surface. The cheapest method is scrambling item and answer order. The stronger method is parallel forms, where each version has different questions testing the same objectives. Scrambling deters copying; parallel forms deter question sharing between sections.

Two versions solve two different problems, and picking the wrong method for your problem is the usual mistake. If students sit next to each other in one room, you need their neighbors' answer sheets to be useless, and scrambling does that. If one section takes the test Tuesday and another Thursday, scrambling helps not at all, because the questions are the same and the first group can simply tell the second what they were.

Which method fits your situation

Most instructors overbuild here. For a single-room, single-sitting quiz, scrambling both item order and answer positions is enough and costs almost nothing. Reserve parallel forms for staggered sittings, make-up exams, and anything you plan to reuse next term.

What must stay identical

This is the part that determines whether your two versions are fair, and it is where the real work is. Both forms must match on:

Blueprint coverage. If version A has six items on chapter four and version B has three, the two groups took different tests regardless of how similar the questions look. Map both forms against your topic list before you finalize either.

Item count and point total. Obvious, and still worth checking, because swapping items late often changes the count by one.

Difficulty balance. Equal numbers of recall, application, and analysis items. A parallel form that unintentionally loads the harder cognitive level penalizes whoever received it, and students notice.

Question types. Short answer takes longer than multiple choice. If version B has three short answer items and version A has one, version B is a longer test at the same time limit.

Time and instructions. Same window, same rules, same allowed materials.

Building parallel forms without writing everything twice

The traditional objection to parallel forms is straightforward: writing forty defensible items instead of twenty is twice the work, and nobody has that evening. That objection is what changed. Generating a larger pool from your source material and splitting it into two forms costs about the same as producing one form used to.

The workflow is: generate forty to fifty items from the chapter, deck, or notes the test covers; tag each one by objective; then deal them into two forms so each objective is represented equally on both. Anything left over becomes make-up material or next term's starting point. The question bank generator is built for this, since the point of a bank is having more items than any single test needs.

One caution: do not build version B purely from leftovers. The items you liked least will cluster on one form, and that form will be measurably worse. Alternate as you assign, best item to A, next best to B, and so on down each objective.

Does scrambling questions actually prevent cheating?

It prevents the specific behavior of reading a neighbor's sheet and transcribing letters, which is the most common form of in-room copying. It does nothing about phones, notes, shared answers between class sections, or a student who photographs the test. Treat scrambling as one cheap layer rather than as a solution, and pair it with seating, device rules, and a make-up form for anyone testing later.

There is also a small cost worth knowing. Scrambled item order makes it harder to walk through the test afterward in class, since "question seven" means different things to different students. Numbering by objective in your key, or reviewing by topic instead of by number, avoids the confusion.

Should I scramble answer choices too?

Yes for multiple choice, with two exceptions. Leave numerical options in ascending order, because scrambling them adds cognitive load unrelated to what you are testing. And leave "all of the above" or "none of the above" in the final position, since moving those creates genuinely confusing items. Everything else can move freely.

Keeping the two answer keys straight

This is where version A and version B go wrong most often in practice, and the failure is entirely administrative. Grading version B against version A's key produces a room full of wrongly failed students and a very bad afternoon.

Three habits prevent it. Put the version letter in the header of every page, not just the first, since pages separate. Print each version's key on paper of a different color, or label it in large type at the top. And spot-check by grading one known-good sheet from each version before you grade the stack.

If you are delivering online, most platforms handle randomization and per-student keys automatically, which removes the problem, and a training platform that tracks completion and scores across a whole cohort handles the same requirement for corporate assessments where two sittings are the norm rather than the exception.

How many versions do I actually need?

Two covers almost every classroom situation. Four is worth it for large lecture halls with tight seating, or when the same exam runs across multiple days. Beyond four you are usually better served by drawing each student's test randomly from a bank, which is what online delivery does natively and which makes the number of distinct forms effectively unlimited.

A short pre-flight check

Before printing, confirm both forms cover the same objectives in the same proportions, carry the same item count and point total, use the same mix of question types, and are labeled clearly on every page. Then confirm each key is labeled and matches its form. Five minutes here saves the regrade email thread that otherwise follows.

To build the pool both forms come from, start with the material you already teach from: the PDF to quiz converter, the Word document to quiz converter, or the PowerPoint to quiz converter. For the wider classroom workflow, see the quiz maker for teachers.