Validity and Reliability

Question 1

Reliability

Answer 1

Reliability refers to whether something is consistent. Any tool used to measure something must be reliable. For example, if you measure something with a ruler, you expect it to have the same measurement if you measure it today or tomorrow. Also if 2 people measure the same thing with the same ruler, you expect the 2 or more measures to be consistent with each other.

This also applies to psychological tests measuring any psychological event. The key word to emphasize when asked about reliability is consistency. For example, the reliability of an experiment can be determined through replication. Similarly, inter-observer reliability refers to the extent to which there is agreement between 2 observers involved in observing a behaviour.

Question 2

Validity

Answer 2

The results of research may be reliable but this does not mean that they are valid. Valid results are true and accurate. Validity includes internal and external validity.

Question 3

Internal validity

Answer 3

Refers to whether the study actually measures what it claims to be measuring. This includes the issues of whether the researcher tested what it intended to test and the control of relevant extraneous variables in an experimental design so they don’t become confounding variables. Internal validity can be established by repeated measures design, controlling situational and investigator variables and attempting to reduce demand characteristics, investigator effects and social desirability effects.

Question 4

External validity

Answer 4

Simply because a research study has internal validity does not mean its results can be generalised. External validity refers to the extent to which the findings can be generalised. External validity includes population validity (ability to generalise to other people), historical or temporal validity (ability to generalise to other times) and ecological validity (ability to generalise to other settings). Ecological validity includes: generalisability (ability to generalise findings outside the lab) and representativeness or ‘mundane realism’ (extent to which the set-up mirrors real life). External validity can be established by the sampling, partial replication, realism and triangulation.

Question 5

Measuring validity

Answer 5

1. Face validity
2. Concurrent validity
3. Predictive validity

Question 6

Face validity

Answer 6

This is the most basic measure and it simply means to decide whether the test looks like it is testing what it claims.

Question 7

Concurrent validity

Answer 7

This involves comparing the results obtained by a new test with those of an older test known to have good validity.

Question 8

Predictive validity

Answer 8

This is the ability of the test to predict performance on future tests.