Validity: how much does your tool measure what it should be?

Validity: how much does your tool measure what it should be?

1. external: he extent to which a study’s findings are applicable or relevant to a group outside the study (also called generalizability).

   • can research be duplicated?
   • how were participants selected?
   • sample size is important

2. internal: how confident the researcher can be about the intervention truly causing a change in the dependent variable.

   • THREATS to internal validity
     • extraneous events
     • mortality of participants
     • selection bias
     • testing effect
     • passage of time
     • selection bias
     • lack of causal direction

3. Subcategories of Validity

   • FACE: never use this type of validity alone. The idea that what you're doing 'looks' like it makes sense.
   • content: do items in tool represent the concept
   • criterion: does measure make accurate predictions
   • concurrent: does measure correlate w/ other valid measures
   • construct: was measure developed to test single concept?