validity

what is validity

• Refers to whether a psychological investigation matches up with reality

• Refers to accurately measuring what the study sets out to

• Refers to what extent findings can be generalised to wider population and outside of the research setting

• Validity can be split into two main branches: internal and external

  • It is possible for something to be reliable - e.g. very consistent - but not valid such as broken scales or equipment may produce the wrong value consistently

what is internal validity

• Measure of whether results obtained are due to the manipulation of the independent variable and no other extraneous variables

  • Demand characteristics are the main threat to this

• Can be improved by reducing investigator effects and demand characteristics

what is external validity

• Measure whether the results can be generalised beyond the research setting

• Ecological validity - whether it can be generalised to other settings

  • Lab settings do not automatically mean something has low ecological validity as it is more to do with artificial tasks and low mundane realism - this can occur in field experiments too

• Population validity - whether it can be generalised to other people

• Temporal validity - whether it can be generalised over time (whether it holds true over time)

  • Asch's study was conducted during conformist era and Freud's concept of penis envy may be outdated, sexist and a reflection of patriarchal Victorian society

how to assess validity

• Predictive validity - assesses validity by predicting how well a test does at predicting future outcomes/behaviour

• Concurrent validity - assesses through correlation, correlating scores from research/test already existing and known to be valid (+0.8 is the benchmark for correlation to be significant in this case)

  • Takes your test, takes established test, compare results to see if they correlate, if there is positive correlation - has concurrent validity

• Face validity - assesses whether something is what it looks like, to what extent does the item look like what the test measures (intuitive)

reduce social desirability bias and improve validity in questionnaires

• Assure anonymity and confidentiality so that students feel safe giving honest responses

• Triangulate with other measures e.g. combine self-report with observations or interviews with friends/relatives/teachers

• Put random questions in that are about another topic to prevent ppts from guessing the aim of the questionnaire

• Repeat questions but slightly differently to check if answers are in line with previous responses

• Ask reverse or negative questions (I often feel confident and I often doubt myself in same questionnaire)

how to improve validity - general

• Experiments

  • Single/double blind - increases internal validity by reducing demand characteristics and investigator effects in experiments so less ppt reactivity

  • Anonymity - increases internal validity in questionnaires

    • Assure anonymity and confidentiality, so people feel safe giving honest responses

  • Standardised procedures - increases internal validity by reducing investigator effect in most methods, less ppt reactivity

  • Control group used - better able to determine whether changes in behaviour were due to effect of independent variable or not

• Questionnaires

  • Triangulation with other data/methods - increases internal validity by gathering data from different sources

    • Combine self-report with observations or interviews with friends/teachers/relatives 

• Observations

  • Behavioural categories - increases internal validity in observational research and content analysis by minimising ambiguity of data collected

  • Covert observation - increases internal validity in observational research

• Interviews

  • Direct quotes - increase the interpretive validity (extent to which researcher's interpretation of events matches ppts) of conclusions in self-report data, direct quotes should be coherent with researcher narrative

    • Qualitative data has higher ecological validity due to more detail and depth