Reliability and validity

RELIABILITY

• Results are reliable if they are consistent. If the study was repeated, then the results should be the same.
• INTERNAL: If a test or a set of scores is consistent within itself.
• EXTERNAL: If a test is consistent over time - the same results can be found on future occasions.

TESTING THE RELIABILITY

• INTERNAL: Split half: the items in the test are split in half and responses to the first half are compared to responses in the second half. If they are consistnt, the test is reliable. 
• EXTERNAL: Test re-test: the test is repeated with the same participant on future occasions. If it is consistent, the results are reliable.
• OBSERVATIONAL STUDIES: Inter-observer reliability: the results of at least 2 observers are compared. If the total number of agreements divided by the total number of observations less than 0.8, data = reliable. Observers should be trained to ensure reliable observations.

IMPROVING RELIABILITY

• 1) Pilot studies - check the measurement works correctly. 
• 2) Take more than 1 measure from…