Recycling the same past papers (Dawrat) year after year trains students to recognize questions instead of solving them — and the fix isn't writing a whole new exam from scratch, it's generating one fresh, same-difficulty "sibling" exercise for every past-paper question you'd otherwise reuse unchanged.
The "Dawrat" Dilemma
In Lebanon, the reliance on as'ilat dawrat (past official exams, published by CRDP for every branch of the Bac) is both a blessing and a curse. While it prepares them for the format, it completely destroys the core purpose of an assessment: measuring actual understanding.
There's a well-established reason this backfires. Cognitive scientists Henry Roediger and Jeffrey Karpicke found that repeated testing of material produces dramatically better long-term retention than repeated re-reading or re-solving the same problem — a finding known as the testing effect. In their experiments, one week after studying, students who had been repeatedly tested on new recall attempts remembered 61% of the material, compared to 40% for students who had simply re-read the same passage. A student solving a Dawrat question they've already seen isn't retrieving — they're recognizing.
Teacher's Tip
“Mixing one freshly generated scenario in among past-paper questions is a quick way to tell whether students actually understand the method, rather than having memorized the steps to one recycled problem.”
The Teacher's Burnout
So, what is the alternative? Writing an exam entirely from scratch. For a standard 4-exercise exam, this takes upwards of 3 to 4 hours — time most teachers don't have between correcting the last exam and preparing the next.
Signs Your Exams Are Too Recycled
- Students finish unusually fast, but only on questions that match a past paper almost exactly.
- Scores on past-paper-style questions are consistently higher than on anything phrased differently.
- You reuse the same 2–3 past sessions every year because rewriting from scratch takes too long.
- Corrections feel like checking for a memorized sequence of steps, not reasoning.
The Imtihan Solution
Within seconds, Imtihan generates fresh, unique exercises at matching difficulty and mark structure to the Dawrat question you started from. It doesn't just give you the questions; it generates the full step-by-step corrigé alongside them, and lets you regenerate, edit, or adjust the difficulty of any single exercise without rebuilding the whole exam.
Stop Recycling. Start Teaching.
When students face an exam they can't Google, they are forced to actually engage with the material. They use critical thinking instead of their memory of a 2018 past paper.
This doesn't mean retiring Dawrat from your teaching entirely — they're still the best reference for what the official exam's format and command-term vocabulary actually look like. The change is in what you hand students under exam conditions: a fresh, same-difficulty exercise for the actual test, with the familiar past paper kept for earlier practice and revision instead.
Frequently Asked Questions
Should I stop using official past papers (Dawrat) altogether?
No — past papers stay the ground truth for format, command-term vocabulary, and mark distribution. The issue is using them unchanged as the actual test. Imtihan is built to generate a fresh 'sibling' exercise in the same style, not to replace past papers as a study resource.
What is the 'testing effect' and why does it matter here?
It's the well-documented finding that actively recalling information (being tested on it) produces far better long-term retention than re-reading or re-solving a familiar problem. A student who has already seen a past-paper question is doing recall, not retrieval practice — the fresh scenario is what actually tests understanding.
Does a fresh AI-generated exercise still match the official exam's difficulty and format?
Yes — Imtihan generates the new scenario at matching difficulty and mark structure, and produces a full step-by-step corrigé alongside it, so the exam stays comparable to a real Dawrat paper in rigor even though the numbers and context are new.