- Foundation
- About 13 min
- 5 practice questions
Kinds of intelligence tests and what makes a test good
Learn how intelligence tests are grouped as individual or group and verbal or non-verbal, name the best-known tests, and see what standardisation, reliability and validity mean.
If something looks wrong, please tell us.
An intelligence test is a set of tasks, given under fixed conditions, that is designed to measure mental ability. Teachers meet test results in school records and in guidance work, and exam questions ask which kind of test suits which situation. This lesson sorts tests into kinds, names the famous ones and sets out the qualities of a good test.
Individual and group tests
An individual test is given to one person at a time by a trained examiner. The examiner can watch how the person works, put the person at ease and note effort and attention. Such tests are best for young children and for those with special difficulties, but they take much time. The Binet-Simon scale, its American revision the Stanford-Binet, and the Wechsler scales are individual tests. A group test is given to many people at once and usually marked by a key. It is quick and cheap, though it shows less about each person. The Army Alpha and Army Beta tests, developed in the United States during the First World War to sort large numbers of recruits, were early group tests.
Verbal, non-verbal and performance tests
A verbal test uses language: vocabulary, sentence completion, verbal analogies, reading. It is unfair to people who do not know the language well. A non-verbal test uses pictures, shapes and patterns instead of words. An example is Raven's Progressive Matrices, in which the person chooses the piece that completes a pattern. A performance test asks the person to do something with objects, such as arranging blocks to match a design. Army Alpha was written for those who could read, and Army Beta for those who could not. Wechsler's scales were divided into a verbal part and a performance part (newer editions report index scores instead). Non-verbal tests reduce, but do not fully remove, the influence of language, culture and experience.
| Kind | Given to | Example task | Good for | Limit |
|---|---|---|---|---|
| Individual | One person | Wechsler or Stanford-Binet tasks | Young children, detailed diagnosis | Slow and costly |
| Group | Many at once | Paper test marked by key | Screening large numbers | Less detail on each person |
| Verbal | Language-based | Vocabulary, analogies | Those who read well | Unfair to weak readers |
| Non-verbal | Pictures and patterns | Matrices | Those with limited language | May still depend on experience |
| Performance | Using objects | Block design | Practical ability | Needs materials |
Qualities of a good test
- Standardised: same instructions, time limits and scoring for everyone, so scores can be compared.
- Normed: the scores are compared with those of a large, well-chosen sample, called the norm group.
- Reliable: it gives consistent results when repeated or marked by different people.
- Valid: it measures what it says it measures, intelligence and not just reading speed or school knowledge.
- Objective: scoring does not depend on the marker's opinion.
Worked example 1: choosing a test
(a) A school wants to screen 300 new pupils quickly. (b) A five-year-old with a speech delay needs assessment. (c) A group of pupils who have just joined from another language medium need an ability check. Working: (a) a group test, because many pupils must be tested quickly. (b) an individual test, ideally a performance one, since an examiner can adapt to the child and language is not needed. (c) a non-verbal group test, since the children's language skill would otherwise lower their scores.
Worked example 2: using norms
A made-up norm table for a test shows that a raw score of 30 corresponds to a mental age of 8 years, which is 96 months. Ravi, aged 10 years (120 months), gets a raw score of 30. Mental age: 96 months. IQ: 96 / 120 x 100 = 80. The norm table is what lets us turn a raw score into a mental age. Without norms from a large sample, the score of 30 would have no meaning. (The table is fictional; real tests have their own published norms.)
Individual for depth, group for numbers. Verbal tests need language; non-verbal and performance tests need less. A good test is standardised, normed, reliable, valid and objective.
Key takeaways
- Tests are individual or group, and verbal, non-verbal or performance.
- Binet-Simon, Stanford-Binet and Wechsler are individual tests; Army Alpha and Beta were group tests; Raven's matrices are non-verbal.
- A good test is standardised, normed, reliable, valid and objective.
Check your understanding
-
1. Which kind of test is most suitable for quickly screening 500 pupils?
- An individual test
- A group test
- A performance test given one by one
- A clinical interview
Show answer
Answer: B. A group test can be given to many at once and marked by a key.
-
2. Raven's Progressive Matrices is an example of a:
- Verbal test
- Personality inventory
- Achievement test in language
- Non-verbal test
Show answer
Answer: D. It uses patterns and shapes instead of words.
-
3. A test gives similar scores when repeated. This shows that it is:
- Reliable
- Valid
- Verbal
- Individual
Show answer
Answer: A. Consistency of results is reliability.
-
4. Ravi, aged 10 years, has a mental age of 8 years on a test. What is his IQ?
- 125
- 90
- 80
- 110
Show answer
Answer: C. IQ = 8 / 10 x 100 = 80.
-
5. Why are norms needed for a test?
- They make the test a good deal shorter
- They remove the need for any reliability
- They show the pupil the answers in advance
- They let a raw score be compared with a large sample
Show answer
Answer: D. A raw score has meaning only when compared with the norm group.
Sign in to keep track of the lessons you finish.