Different types of validity
VALIDITY
No single measure of personality is perfect. Instead, its better to use multiple techniques/principles of “triangulation”
Internal reliability: participants to give consistent responses, no matter how the question is phrased
In developing a good questionnaire, we recognize that people that might answer questions differently because of slight differences in phrasing (there are 2 factors that influence someone’s response to a questionnaire item:
Their true score on some construct
Random error
How do we account for this?
Ask a lot of questions slightly different and average together
Usually one or two item measures are not considered enough to accurately measure a construct
How do we assess internal reliability?
Use statistic called Croncbach’s alpha
Calculated like a correlation (ranges from 0-1) but can be negative
Higher numbers indicate greater reliability
Usually want an alpha>.70
Inter-rater reliability: measures of consistency used to evaluate the extent to which different judges agree in their assessment decisions
Only applies to behavioral/observational measures
Construct Validity: Does your measure accurately assess what you think it does (and only what you think it does)(is your operational definition accurately measuring your concept)
Face validity: your measures appears “on its face” to measure what it says it does (face value). In other words, it just looks like what you want to measure
Is considered “weak evidence” for validity of a measure, and it is NOT a necessary component of a good measure
In fact, sometimes LOW face validity may be a good thing
Criterion (also predictive) validity: the extent to which a measure is correlated with a relevant observable outcome or behavior
Criterion- a standard on which a judgment or decision may be based (singular of criteria)
Considered the best evidence for validity
What is an “observable outcome or behavior”: gpa, x number of times doing blank, brain activity, things that are observable in general, something OUTSIDE of yourself
What does not count as concrete outcome or behavior: Another self report measure, self-esteem, internal feelings/subjectiveness
Convergent validity: When a scale correlates with similar self report scales (measuring the same/similar constructs)
Discriminant validity: A measure does NOT correlate (too highly) with unrelated constructs (very little confounding variables)(measure only ONE thing and ONLY one thing)
Content validity: Covering ALL parts of a construct (as defined by the theory) (DID YOU COVER ALL THE CONTENT)