methodology · 15 min read
How we generate and review LET practice questions
How LETReviewer writes and reviews LET questions: ToS competency, validators, automated gates, and editor review. Original practice, not leaked papers.
Why a pipeline exists at all
A LET practice item is easy to publish and hard to stand behind. Anyone can paste a stem into a site and call it a reviewer. The cost shows up later: a wrong key, a duplicated idea, a competency that does not match the Table of Specifications, or a “recalled” question that should never have been typed. LETReviewer treats publishing as a chain, not a folder of PDFs.
The public version of that chain lives on the editorial policy page. This essay is the longer walkthrough: what we write against, who answers without seeing the key, which gates are automatic, when an editor is required, and what we will never put in the bank. It is also a list of claims we refuse, because methodology without refusals is just production volume.
Practice content is aligned to the PRC Table of Specifications. Practice content aligned to the PRC Table of Specifications. Not affiliated with, endorsed by, or a product of the Professional Regulation Commission. PRC does not release past board papers. PRC does not release past board papers, and we never claim to have them. The sitting you are training for is still 150 items per paper. The official sitting is still 150 items per paper: BEEd without specialization is 300 items, and BSEd (or BEEd with specialization) is 450. Confirm application dates and fees on prc.gov.ph. This page will not invent a schedule to look current.
Start from a ToS competency and a verbatim learning outcome
Each published item is written to a named competency and a verbatim Table of Specifications learning outcome. The spec also carries difficulty and cognitive level. Drafts are generated from that specification, not from a scavenged booklet. Official public references — DepEd, CHED, statute — can inform the stem. A competitor’s item cannot.
Alignment here means something narrow and checkable. This item is supposed to test this competency at this demand. It does not mean “this will appear on the next LET.” We cannot guarantee that these exact questions will appear. We can help you sharpen preparation. The how we pick questions guide is the shorter methodology note if you want the chain without the essay.
Source, author, reviewer, program, and Table of Specifications version stay on each question's record. That is how we answer “where did this question come from?” without claiming a leaked sitting. Old attempts keep their program id so history does not silently rewrite when a ToS version branches.
Independent validators who never see the key
After candidates exist, independent validators answer the item without seeing the intended key. If a competent reader cannot land on the published answer from the stem and options alone, the item is not ready. A beautiful rationale that only makes sense after you already know the letter is a teaching note, not a fair multiple-choice question.
Validators are not a rubber stamp on the generator. They are a second job: can this be solved from what is on the page? Attractive distractors should fail for a reason a teacher can explain. They should not fail because the stem hid a trick word or because two options are both defensible. Ambiguity is a reject, not a feature.
We do not ask validators to compare our stem to a recalled paper. There is no honest recalled paper to compare against. PRC does not release past board papers. A group-chat screenshot is not a source. If a draft looks like it was typed from someone else’s copyrighted reviewer, it does not enter the bank.
Automated gates before anything goes live
Automated checks catch the boring failures that still ruin trust: missing keys, weak or duplicate distractors, grammar issues, and near-duplicates of an item already in the bank. They also catch structural gaps — a spec without a verbatim outcome, a blank rationale, a program that does not match the competency row.
High-confidence items may publish after those gates. That sentence is not “the model is the editor.” It is “some items clear a deterministic bar.” Borderline items do not. Sampled items do not. A disputed key never publishes. If validators split, or if the key cannot be defended from the stem, the item stays out of the draw.
Those gates are why we will not quote a giant inventory number to impress a search page. Published means it survived the chain. A draft in a queue is not a practice question. A PDF on a disk is not a practice question. Limited bank on a set means we shrink N rather than pad from another subject.
Editor review for borderline and sampled items
Borderline and sampled items wait for editor review. Sampling exists so high-confidence automation does not become a closed loop. An editor can still reject a clean-looking item if the competency is wrong, the cognitive demand is mislabeled, or the local context is off.
User flags reopen the same queue. Reasons we accept include incorrect answer, ambiguous, outdated, grammar, duplicate, or other. A cluster of flags takes the item out of the draw until it is checked. Flag anything that looks wrong instead of silently distrusting the whole bank. A single miss is not proof the pipeline failed. A pattern of flags is a reason to halt.
Contribute-a-question uses the same queue. There is no side door where a helpful email becomes a live stem. Courtesy is not a publish gate. The About page is the shorter trust story if you want affiliation and Readiness in one place.
PDFs are candidates only
PDFs are ingested as candidates only. Nothing from a file auto-publishes. That includes files we made, files a user uploaded, and files that look like a Table of Specifications printout with items in the margins. A candidate still needs a competency, a verbatim outcome, validators, and gates.
We do not scrape other LET sites. We do not copy copyrighted reviewers. We do not ingest a file whose license is unclear. Original practice is slower than a scrape. It is also the only inventory we will defend on a legal page and on a result screen.
If you brought a booklet from a review center, use it as your notes if you have the right to. Do not ask us to type it in. Do not ask us to “just add the leaked ones.” There are no leaked ones we will publish. A recalled stem is a copyright and ethics problem, not a growth channel. Candidate status is not a preview of a live item, and it is not a promise that the file will ever clear the gates.
What we will never publish
We will not scrape competitors. We will not sell leaked or recalled PRC items. We will not call a Readiness Score a probability of passing. The number we display is a Readiness Score based on your practice — not a probability of passing. We will not unlock specialization practice because someone bought Plus.
Specialization practice stays locked while majorship banks grow. Plus does not unlock it. Timed majorship exams use published items — we do not invent Spec inventory. The diagnostic stays Gen Ed and Prof Ed only for every track — 30 items, no majorship sample, no fake third paper. Timed Gen Ed and Prof Ed mocks still use Annex A length: 150 items, two hours and three hours. A specialization timer, if shown, is 3.5 hours.
We will not promise that a published practice stem is an official paper. We will not hide the 50% floor. We will not pad a thin bank from another subject to hide a Limited state. We will not invent application dates. Check prc.gov.ph. The LET Exam Reviewer loop exists so the next session has a reason that is not “do more items.”
How a published item shows up in your study
Once an item is published, it can appear in the diagnostic, in daily practice, and in a timed mock. Those surfaces share the bank. They do not share the job. A diagnostic is a 30-item weighted read. Practice is a named competency set with instant rationales. A mock is one 150-item paper with rationales after submit.
Today’s Mission is the only learner-facing next-action engine. It will send you to diagnose, remediate, Remember, or a focus drill from that same published bank. Plus adds volume — no Manila-day practice cap, Deep 50, more remediation starts. Plus does not unlock a second item factory. Free and Plus study the same stems.
Wrong answers stay open until a remediation retest passes. Remember cards can keep a repaired competency from fading. Grading a card does not rewrite the key and does not clear the miss. If the stem later fails a flag review, it leaves the draw. Your history does not get a silent rewrite of what you already answered.
Flag the item, then go back to the named job
A pipeline is only as good as the next person who refuses to shrug. If a question looks wrong, ambiguous, outdated, or duplicated, flag it. If you only screenshot it into a group chat, the bank cannot change. If you treat one bad stem as proof that every item is leaked, you will leave the only honest loop we offer.
Then return to the work the record supports. Know your track and the 300-item or 450-item sitting that comes with it. Keep a diagnostic baseline. Follow Today’s Mission. Sit a timed paper when you want pressure. Read Readiness as a band, not a prophecy.
Share this methodology with a review partner if you want accountability, but do not outsource the miss. Explaining why a distractor almost worked is the study. Copying someone else’s answer key is not. The bank is original practice aligned to the Table of Specifications — not a leaked sitting, and not a promise that the next LET will look like last night’s set. If you want the product tour after the pipeline, start with a diagnostic, then let Today’s Mission name the next job.
Questions
They are original practice items written to a Table of Specifications competency and a verbatim learning outcome, then run through independent validators and automated gates. They are not leaked or recalled PRC items.
Read the short policy, then try a diagnostic
Editorial policy is the one-page version. The diagnostic is thirty Gen Ed + Prof Ed items from the published bank.
- ToS-aligned items
- Independent — not PRC
- 75% + 50% floor
- Readiness, not pass chance
Practice content aligned to the PRC Table of Specifications. Not affiliated with, endorsed by, or a product of the Professional Regulation Commission. PRC does not release past board papers.