How to Perform Item Analysis After an Assessment

1. Validate the assessment data

Confirm the answer key, scoring rules, omitted items, duplicate responses, and scan exceptions before calculating or interpreting statistics. A keying or capture error can make a sound question appear defective.

Record the examinee group and assessment purpose because item behavior depends on who took the test and why it was administered.

2. Review difficulty and discrimination together

Use item facility to understand how frequently the group answered correctly. Then inspect discrimination or item-total evidence to see how the item relates to overall performance.

Very easy or difficult items are not automatically poor. They may be appropriate when they represent essential knowledge, screening thresholds, or intentionally demanding outcomes.

3. Inspect content and responses

Review distractor selection, common misconceptions, wording, alignment, and whether more than one answer could reasonably be defended. Compare the statistics with the intended cognitive demand.

Classify the item as retain, revise, investigate, or retire, document the reason, and validate revised items in future administrations.

Frequently asked questions

How large should a class be for item analysis?

Larger samples generally produce more stable estimates, but small-class analysis can still support cautious formative review when educators avoid rigid cutoffs.

Should every low-discrimination item be removed?

No. First check the key, scoring, wording, alignment, sample, instruction, and purpose. Statistics identify questions for review rather than replacing professional judgment.