Analyze test results by checking the answer key, counting responses to each question and mapping errors to learning objectives. Use item difficulty and distractor choices to decide what to investigate next. A low score alone cannot tell you whether the problem was understanding, wording, timing or access.
Start with the paper, not the ranking
Independently solve the questions and compare the saved answer key with the displayed options. Confirm marking rules, missing responses and any interrupted attempts. Keep “absent,” “not submitted” and “incorrect” separate. If a faulty item needs rescoring, document the correction and apply the same rule to all affected attempts.
Download the item-analysis worksheet. This is an editable, ungated worksheet from ThePrepLab. Examples are illustrative, not customer results. Use anonymous IDs instead of student personal information.
A worked example with ten fictional students
For a single-correct-answer item, one simple facility measure is correct responses divided by students who were presented with the item. State how omitted answers are counted. In this example all ten fictional students saw each question; an omitted response counts as not correct. A larger proportion means the item was easier for this group, despite the common name “difficulty index.”
| Item and key | A | B | C | D | Omitted | Proportion correct |
|---|---|---|---|---|---|---|
| Q1: key B | 1 | 6 | 2 | 1 | 0 | 6/10 = 0.60 |
| Q2: key A | 3 | 4 | 1 | 1 | 1 | 3/10 = 0.30 |
| Q3: key D | 0 | 0 | 1 | 9 | 0 | 9/10 = 0.90 |
Q2 deserves a closer look because more students selected B than the correct A. First check whether B is also defensible or the saved key is wrong. If the question is sound, inspect the working of students who selected B. Their common choice is a clue to investigate, not proof they share one misconception. Q3 may be suitable as a prerequisite check even though most students answered correctly.
What is distractor analysis?
Distractor analysis examines which incorrect options students selected. An option nobody chooses may be implausible, but with ten students that observation is weak evidence. A popular wrong option may reflect a taught misconception, ambiguous wording or a scoring error. The University of Washington’s item-analysis guidance provides background on these measures and their interpretation.
Why small classes need caution
In this ten-person example, one response changes the proportion correct by 0.10. Do not turn a single small-class result into a permanent “hard” tag, a teacher ranking or a student ability diagnosis. Compare later comparable attempts, read written working and ask a short follow-up question. Different cohorts, hints and retakes change what a number means.
Turn a finding into a teaching decision
| Finding | Next check | Possible response |
|---|---|---|
| Several errors on inverse operations | Ask students to explain their first step | Model one example, then use a new equation |
| Many omissions near the end | Review timing and access interruptions | Adjust working time if speed was not the objective |
| High marks but weak explanations | Use a short reasoning task | Practise explaining, not just selecting an answer |
Write the objective, evidence, proposed response and date of a fresh check in the worksheet. If you reteach 3x + 2 = 14, do not use the identical item as your only evidence of improvement: use a new example such as 5x + 1 = 31, whose answer is x = 6, and ask for the reasoning.
Does a dashboard replace teacher judgment?
No. A report can organise responses, but the teacher must inspect the item, context and work. This guide does not promise that every measure described is automatically available in every PrepLab package. Ask for a demonstration using an anonymous sample. For general assessment design, see Carnegie Mellon’s exam-design guidance.
Use the test blueprint before the next paper, exit tickets for a short recheck and the remedial teaching record for follow-through. The practice module is a delivery option, not a substitute for diagnosis.