ReliCheck helps teachers, instructional designers, and certification authors check whether each test item is doing the work it should. Item difficulty, discrimination, distractor patterns, skill alignment, and reliability appear automatically, so the scores you report are defensible to learners, families, administrators, and certifying bodies.
Every test score gets questioned eventually, by the learner who missed an item, the family asking for clarity, the administrator preparing a report, or the certifying body reviewing the program. The questions are predictable. The answer lives in the item analysis.
ReliCheck answers
Skill tag rollups show per-standard performance, so you see which objectives the test actually measured and which students mastered each one. The score becomes diagnostic, not just summative.
ReliCheck answers
Per-item analysis shows whether the missed item was hard for everyone (a sign it may be poorly worded) or specifically hard for that student (a sign of where to focus). Distractor analysis surfaces if a confusing wrong answer pulled students who actually knew the material.
ReliCheck answers
Pre/post item analysis with effect sizes and the Reliable Change Index labels the score lift with the evidence behind it: likely learning gain, possible measurement drift, or interpret with caution. The interpretation comes with the numbers that support it.
ReliCheck answers
Difficulty and discrimination for scored items, with distractor utilization for multiple-choice items. Flags identify items that are too easy, too hard, miskeyed, or have dead distractors. The report exports in a format external reviewers expect.
The best test next year is the one this year's items just told you to build.
Most platforms produce a score and stop. ReliCheck reads the score the way a measurement specialist would and surfaces the items that earned a place on the next administration, the items that need revision, and the items that should be retired.
A test gets administered. Items are scored against the answer key. The class score arrives, but without item analysis the score is just a number, with no story about which items did the measuring.
Per-item difficulty, discrimination, and distractor usage are computed automatically. Skill tag rollups show per-standard performance. Reliability across the test is reported with KR-20 alongside alpha.
Each item is flagged keep, revise, or drop. The next administration starts from a stronger item bank, and the score becomes evidence you can explain, not just a number someone might dispute.
The checks a measurement specialist would run by hand, computed automatically the moment scored responses arrive. The reports come out in the formats teachers, departments, districts, and certifying bodies expect.
Difficulty (proportion correct) labeled easy, medium, or hard. Discrimination (point-biserial correlation) flags items where stronger students did worse than weaker students, which usually points to a miskey or a confusing prompt.
For each multiple-choice item, ReliCheck reports how many students chose each option. Distractors no one picked are flagged as candidates to rewrite or drop, since they are not contributing to measurement.
Reliability per test reported across the standard estimators, with item-if-deleted analyses that flag the items pulling reliability down. Useful when a department, district, or certifying body asks for the number.
Tag each item to a skill, standard, or objective inside the builder. Reports group performance by skill, so the conversation shifts from "what was the score?" to "which standards did students master?"
Paired-sample analyses with Cohen's d and the Reliable Change Index. A score that moved from pre to post is labeled with the evidence behind it: likely learning gain, possible measurement drift, or interpret with caution.
Publish the test to a shareable URL. Students take it in the browser. Responses score against the answer key automatically. Useful for online administrations, makeup sittings, and certification studies.
Every analysis produces an artifact ready for the parent conference, the department meeting, the principal's report, or the certifying-body packet. The math is exposed, the writing is editable, and the item-level evidence is right there.
| Item | Difficulty | Discrimination | Distractors | Flag |
|---|---|---|---|---|
| Q1 | 0.78 | 0.41 | All used | Good |
| Q2 | 0.92 | 0.05 | B unused | Too easy |
| Q3 | 0.54 | 0.48 | All used | Good |
| Q4 | 0.18 | -0.12 | All used | Likely miskeyed |
| Q5 | 0.64 | 0.36 | All used | Good |
One per learner, broken down by skill tag. Item-level performance, missed items in context, and a plain-language paragraph instructors can share with learners, families, or advisors.
Group-level distribution, per-skill mastery, and the items most learners missed. Useful for the next instructional block, department meeting, or certification review.
Each flagged item comes with a recommended action: keep, revise the prompt, replace a distractor, double-check the key, or drop. The next test starts from those recommendations.
Effect sizes, Reliable Change Index, and per-learner tier counts. The score lift is labeled with the evidence behind it: likely learning gain, possible measurement drift, or interpret with caution.
Whether you already gave the test and have scores in hand, want to build a new test with item analysis in mind, want to run a pre/post study, or want to see what a teacher-ready report looks like, ReliCheck gives the next administration a clear starting point.
Upload scored responses and an answer key. Item analysis appears immediately, with flags identifying items that earned their place and items that need a second look.
Path 02The test builder lets you write items, tag them to skills or standards, and set the answer key. The next administration runs item analysis automatically against the responses that arrive.
Path 03Administer the same test before and after instruction. ReliCheck computes pre/post change with effect sizes and the Reliable Change Index, so the lift is interpreted with the evidence behind it instead of stated as fact.
Path 04See the kind of item analysis, skill rollup, and revision recommendations ReliCheck produces on a real classroom test, end to end.
Start free, upload a test you already gave, or preview a sample item-health report and see how item evidence helps instructors, administrators, and certification teams explain the score.