Seatmates with wandering eyes, make-up exams for absentees, two sections taking the test on different days — all classic reasons to prepare parallel forms of a test. But a Form B that’s accidentally harder than Form A is unfair, and one that’s easier invites complaints. Here’s how to build forms that are genuinely equivalent, plus how to prove they are.
What “parallel” really means: for every item number, both forms measure the same competency, at the same cognitive level, at roughly the same difficulty — with different questions. A learner of equal ability should score about the same on either form.
The golden rule: one TOS, two tests
Parallel forms are built from a single Table of Specifications. If item 14 on Form A is Analyzing under competency S5LT-IIb-2, then item 14 on Form B must also be Analyzing under S5LT-IIb-2. The blueprint is shared; only the surface changes. (Haven’t built a TOS yet? Start with our step-by-step TOS guide.)
This is why parallel-form construction is easier when you already have a TOS: the plan tells you exactly what each item slot must be, so you’re just writing two questions per slot instead of one.
Three ways to write Form B (best to weakest)
1. Item shells (recommended)
Strip each Form A item down to its skeleton and refill it with new content:
Form A: “What is the main function of the large intestine?”
Shell: “What is the main function of the ____?”
Form B: “What is the main function of the small intestine?”
The structure, cognitive level, and difficulty carry over almost automatically — this is the safest method for keeping forms equivalent.
2. Value/scenario swaps
For Math and problem-solving: keep the problem structure, change the numbers or context. ₱450 becomes ₱380; a sari-sari store becomes a school canteen. Caution: verify the new numbers don’t accidentally make the arithmetic easier (round answers, no regrouping) or harder. A “parallel” math item with uglier numbers is not parallel.
3. Free rewrites (use sparingly)
Writing a brand-new item from scratch for each slot gives the freshest test, but matching difficulty becomes guesswork. Reserve it for slots where shells feel too repetitive — and double-check equivalence afterward with item analysis.
The equivalence checklist (before printing)
- ☐ Same number of items, same item types, same point values
- ☐ Item n on both forms → same competency and same Bloom’s level
- ☐ Same directions, same time limit, same allowed materials (formula sheets, etc.)
- ☐ Answer keys differ! If both forms’ answers spell B-B-A-D-C…, you’ve defeated the purpose — shuffle the position of the correct option
- ☐ Read Form B aloud once — accidental clues (one option much longer, a grammatical mismatch with the stem) creep into rewrites
- ☐ Same visual layout and spacing, so neither form looks harder
A note on shuffling vs. true parallel forms
A common shortcut is to simply shuffle the question order (or the option order) to make “Form B.” This deters copying a seatmate’s letter answers, but it’s still the same test — same questions, same difficulty. That’s perfectly fine for low-stakes quizzes. For make-up exams, though, a learner who saw the original could memorize answers, so you need genuinely different questions (the shell method), not just a reshuffle.
After the test: verify with item analysis
The honest check of parallelism happens after administration. Run an item analysis on both forms and compare item by item:
| Item | Form A difficulty (p) | Form B difficulty (p) | Verdict |
|---|---|---|---|
| 14 | 0.62 | 0.60 | Parallel ✓ |
| 22 | 0.55 | 0.28 | Not parallel — Form B’s item is too hard |
If a pair diverges sharply (item 22 above), note it and fix that item before the forms are reused. Over a few cycles, this turns your A/B pair into a genuinely calibrated set.
Frequently asked questions
Is shuffling the questions enough to make a “Form B”?
For low-stakes quizzes where the goal is just to stop seatmate copying, yes. For make-up exams, no — you need different questions testing the same skills.
Do both forms need the same TOS attached when I submit?
Yes — one TOS governs both forms; attach it with both papers. The item placement is identical by design.
How different should the questions be?
Different enough that memorizing Form A’s answers doesn’t help — new values, examples, and distractors — while testing the identical skill at the identical level.
Can I use more than two forms?
Yes, the same principles extend to Forms C, D, etc. — one TOS, parallel slots. Just multiply the proof-reading effort accordingly.
What if my two sections have very different abilities?
Keep the forms parallel anyway — the test should be the same challenge for both. Differences in section performance are information, not a reason to make one form easier.
Generate Form B in one click
Skoolari’s TOS-based generator creates Form A and a true parallel Form B automatically — the same competency and cognitive level for every item number, different questions, each with its own answer key, both downloadable as DepEd-format Word test papers.