Reliability and Validity

Question 1

What does reliability refer to?

Answer 1

How consistent or dependable the test is. A reliable test is carried out in the same circumstances, on the same participants should always give the same results.

Question 2

What is internal reliability?

Answer 2

1. Different parts of the test should give consistent results,
2. In example, if an IQ test contains sections of supposedly equal difficulty, participants should achieve similar scores on all sections.

Question 3

How can internal reliability be measured?

Answer 3

1. Using the split-half method,
2. Splits the test into two halves, e.g. odd and even numbered questions, and the results from each half should produce a high positive correlation.

Question 4

What is external reliability?

Answer 4

1. The test should produce consistent results regardless of when it is used,
2. In example, if you took the same IQ test on two different days you should achieve the same score.

Question 5

How can external reliability be assessed?

Answer 5

1. Using the test-retest method,
2. Involves repeating the test using the same participants. A reliable test should produce a high positive correlation between the two scores,
3. A problem with this is that the participants may have changed in some way since the first test,
4. To avoid this, external reliability can be checked using the equivalent forms test. This compares participants’ scores on two different, but equivalent, versions of the test.

Question 6

What is inter-observer reliability?

Answer 6

1. Test should give consistent results regardless of who administers it,
2. In example, if two researchers observe behaviour and categorise infants as showing signs of a strong attachment or a weak attachment, they should both record the same score.

Question 7

How can inter-observer reliability be assessed?

Answer 7

1. Can be assessed by correlating the scores that each researcher produces for each participant,
2. A high positive correlation should be found.

Question 8

What is validity?

Answer 8

1. How well the test measures what it claims to measure,
2. In example, an IQ test with only maths questions would not be a valid measure of general intelligence.

Question 9

What is face validity?

Answer 9

1. The extent to which the test looks, to the participants, like it will measure what it is supposed to be measuring.

Question 10

What is ecological validity?

Answer 10

1. The extent to which the results of the test reflect real-life.

Question 11

What is concurrent validity?

Answer 11

1. The extent to which the test produces the same results as another established measure, e.g. two different IQ tests should produce the same measure of IQ,
2. Inferential tests can be used to determine whether both measures are highly correlated, and therefore valid.

Question 12

What is temporal validity?

Answer 12

1. The extent to which the test provides results that can be generalised across time.

Question 13

How can validity be measured?

Answer 13

1. A quick method; simply to look at the test and make a judgement on whether it appears to measure what it claims to. In example, an IQ test that just consisted of maths questions could be identified as having low validity by this method,
2. Comparing results of the test with the results of an existing measure (that is already accepted as valid) can help to determine the validity of the test,
3. The results of the test can be used to predict results of future tests. If the initial results correlate with the later results it suggests that the test has some validity and can continue to be used.

Question 14

How can reliability and validity of tests be improved?

Answer 14

1. Standardising research,
2. Operationalising variables.

Question 15

What is meant by standardising research?

Answer 15

1. Involves creating specific procedures which are followed every time the test is carried out,
2. This ensures that all the researchers will test all the participants in exactly the same way, e.g. in the same sequence, or at the same time of day,
3. Reduces the possibility of extraneous variables affecting the research. Therefore, it will help to improve external reliability and inter-observer reliability.

Question 16

What is meant by operationalising variables?

Answer 16

1. Involves clearly defining all of the research variables,
2. For example, in a study of whether watching aggressive TV influences aggressive behaviour, the terms ‘aggressive TV’ and ‘aggressive behaviour’ need to be defined,
3. Clarifying this from the start improves the reliability and validity of the test.