Validity and reliability

• Validity is a measure of truth, determining if a test, experiment, or observation measures what it says it's measuring.
• Internal validity is the extent to which the findings of the dependent variable are due to the independent variable and not other factors, especially in laboratory experiments.
• External validity is the extent to which results can be generalised beyond the study.
• Population validity is whether results can be generalised to the general population.
• Ecological validity is whether results can be generalised to the real world and to real life.
• Mundane realism is to what extent do the study’s tasks replicate real life.
• Face validity is the degree to which a procedure, especially a psychological test or assessment, appears effective in terms of its stated aims.
• Face validity is the simplest technique to test for validity.
• A face validity test is completed by an independent expert who assesses whether a measure will measure what it sets out to measure.
• Concurrent validity is comparing the new procedure with a similar previous procedure, where validity has already been established.
• If scores correlate as a strong positive correlation (+.80) then the test is deemed as valid.
• The correlation coefficient has to be +.80 for two data sets to have concurrent validity.
• Temporal validity is whether the results of the study or concepts within a theory remain true and valid over a period of time.
• What are two was of assessing validity?

  • Face validity
  • Concurrent validity
• What are some ways in which validity can be improved in experimental research?

  • Use a control group
  • Standardise procedures
  • Use single-blind or double-blind procedures
• How does the use of a control group improve validity of an experimental study?

  Researcher can better assess if changes to DV were due to IV alone.
• How does standardising procedures improve the validity of experimental research?

  Minimises influence of participant reactivity and investigator effects on validity of outcome.
• What is a single-blind study?


  Participants not aware of aim - to reduce demand characteristics.
• What is a double-blind study?


  A study where neither the participants nor the researcher knows the true aim of the study. A third party conducts the investigation without knowing its purpose.
• How is validity improved in questionnaires?

  Incorporate a lie scale which assesses the consistency of responses or control effects of social desirability bias. Or ensure respondent's data remains anonymous.
• How is validity improved in observational research?

  Covert observations minimise demand characteristics and increase ecological validity. Also, behavioural categories should not be broad or ambiguous.
• How can the validity of qualitative methods be improved?

  Demonstrate interpretive validity, the extent to which investigator interpretations match those of the ppts. Or through triangulation, using a number of different sources of evidence.