Validity

• What is internal validity?

  Whether a study measures what it intends
• What is a threat to internal validity related to participant behavior?

  Demand characteristics
• How can investigator effects threaten internal validity?

  Researcher unconsciously influences results
• What are confounding variables?

  Uncontrolled factors affecting the DV
• How can extraneous variables be controlled to improve internal validity?

  By using standardized procedures
• What procedure can reduce bias in research?

  Single/double-blind procedures
• What is ecological validity?

  Can findings be applied to real-life settings?
• What is external validity?

  Whether results can be generalized beyond study
• What is temporal validity?

  Are findings still relevant over time?
• What does population validity refer to?

  Generalizability of findings to different people
• How can external validity be improved?

  Use realistic settings with high mundane realism
• What type of samples improve population validity?

  Large, diverse samples
• What is face validity?

  Whether a test appears to measure what it claims
• What is concurrent validity?

  Comparing a new test with an established one
• What is predictive validity?

  Whether test results can predict future behavior
• What is construct validity?

  Whether a test accurately measures a theoretical concept
• What is a threat to validity in experiments related to settings?

  Artificial settings lead to low ecological validity
• How can validity be improved in experiments?

  Use realistic tasks and double-blind procedures
• What is a threat to validity in observations?

  Observer bias
• How can validity be improved in observations?

  Use covert observations and multiple observers
• What is a threat to validity in self-reports?

  Social desirability bias
• What is social desirability bias?

  Participants give socially acceptable answers
• How can validity be improved in self-reports?

  Use anonymous surveys to reduce bias
• What type of questions can avoid leading participants?

  Use open-ended questions