Designing Formal Tests and Exams

Now that you understand the principles of good testing, this lesson shows you how to apply them when designing your own formal tests and exams.

Formal tests are summative — they measure what students have achieved at a specific point in time. The test design process Designing a good test is not just writing questions — it is a systematic process. [IMG:tesol160_18-2.png] Step 1: Define what you are testing Every test should be built around a clear test specification — a document that outlines: - The purpose of the test — what decisions will be made based on the results? - The target population — who is being tested? - The skills and language being tested — reading, writing, grammar, vocabulary? - The level — what CEFR level is the test targeting? - The format — how long is the test? How many sections? What question types? Step 2: Choose question types Different question types test different things and have different strengths and weaknesses. [IMG:tesol160_18-3.png] Step 3: Write clear, unambiguous test items Common mistakes in test writing: - Trick questions — designed to confuse rather than assess - Ambiguous questions — more than one correct answer is possible - Culturally biased questions — assume knowledge of a specific cultural context - Questions that test knowledge, not language — testing facts about the topic rather than language ability

Weak test item: "Do you think technology is good or bad for society?" — This tests opinion, not language ability. Any answer could be correct.

Strong test item: "Read the article and answer: What two reasons does the writer give for the negative effects of social media? Use information from the text." — This tests reading comprehension with a clear, verifiable answer. Step 4: Design a marking scheme Every formal test needs a clear marking scheme before it is administered — not after.

For objective items (multiple choice, gap-fill): - List the correct answers with acceptable alternatives

For subjective items (essays, speaking tasks): - Use a detailed rubric with clear descriptors for each score level - Train all markers to apply the rubric consistently Analyzing test results After marking, analyze the results to evaluate the test itself: - Item difficulty — if everyone gets a question right (or wrong), it may be too easy or too hard - Item discrimination — do stronger students perform better on each item than weaker students? - Overall score distribution — is the spread of scores reasonable? A good test teaches you something — not just about your students, but about your own teaching. If most students fail a particular item, ask yourself: Did I teach this clearly enough?