Reliability and Validity

Question 1

What is Reliability ?

Answer 1

• Reliability refers to how consistent something is

Question 2

What are the 3 types of Reliability?

Answer 2

• Internal Reliability
• External Reliability
• Inter-Observer Reliability

Question 3

Internal Reliability

Answer 3

• Whether a test is consistent within itself
  > Relevant to research studies that have used tests to measure a person’s ability
• High internal reliability if all questions are of an equal difficulty
• The Split-Half Method: Testing the internal reliability of a test is comparing a person’s scores from the first half of the test to the second.
• Improved by conducting pilot studies

Question 4

External Reliability

Answer 4

• Whether results are consistently found when the study is repeated.
• The Test-Retest Method: Doing the study again to see if similar results are obtained
• Controlling extraneous variables, ensuring all participants have the same standardised instructions and trying to eliminate investigator effects.

Question 5

Inter-Observer Reliability

Answer 5

• Consistency between different researchers
  > If more than one researcher is collecting data, are the different researchers in agreement?
  > If they are reliability is high.
• To check agreement researchers will observe the same situation and use the same method to record the data
• Results will then be compared to see if agreement is high
• To improve agreement between observers behavioural categories can be tightened so that they are not ambiguous and observers could receive training in observational techniques.

Question 6

Validity

Answer 6

• Whether a researcher is measuring what they’re setting out to measure

Question 7

Internal Validity

Answer 7

Research setting – is the researcher measuring what they’re trying to measure within the research setting?

• In Experiments - demand characteristics or extraneous variables the researcher might not end up measuring what they are trying to
• In Questionnaires/Interviews – questions are not valid if they don’t help the researcher achieve their aim
• In Observations – if the researcher doesn’t end up observing the behaviours they intended to their measures lack validit

Question 8

External Validity

Answer 8

• Whether what was measured in the research setting is representative of behaviours in the real world.
• Ecological Validity - extent to which the research setting represents the real life setting
• Population Validity - extent to which the findings apply to groups of people other than the participants
• Temporal Validity - extent to which the findings apply to other time periods

Question 9

Face Validity

Answer 9

• Looking at the questions or measure and making a judgement.
• At face value do the questions or measure seem to be measuring what they’re supposed to?

Question 10

Content Validity

Answer 10

• Experts are sought out to check the questions.

- eg. For instance, a psychiatrist might be used to check the validity of questions designed to measure anxiety levels.

Question 11

Construct Validity

Answer 11

• Looking at how well the measure or questions link to the key underlying constructs.
• eg. For example, if a questionnaire was developed to measure loneliness, we would be looking at how well the questions specifically relate to the construct of loneliness.

Question 12

Predictive Validity

Answer 12

• Seeing whether a score on a test predicts later behaviour.
• eg. For example, if GCSE results were a valid measure of academic ability they should predict results on later tests such as A levels

Question 13

Concurrent Validity

Answer 13

• Assessing the validity of a new test or measure by comparing it to an already established test/measure of the same concept.
• eg. For example, if a researcher developed a new test to measure the severity of depression, a person’s score on this could be compared to their score on an already established test for measuring depression. If the outcome is similar on both tests then the new measure can be said to be valid.