Three types of measurement
Self-report
Observational
Physiological measures
Reliability, consistency, and measurement across
Time (test retrst)
Individuals (interrater )
Items (Internal)
Three things test retest and interrater have in common
Precisely two sets of values
Both derived from some measure
Can use common correlation, coefficient, and scatterplot
Test, retest reliability
Give your instrument to part participants at two different times and see how answers are related
Interrater reliability
The amount two raiders agree on something they have the same observation by two people, the more agreement, the more reliable
Dichotomous
Only takes on two values A or B
Pearson R value only tells you
Strength and direction
Direction has to be linear
Strength this how far are is away from zero
Pearson or ranges and correlation
-1 r 1
Ends are strong
-# is negitive correlation
+# Positive correlation
Correlation coefficient only tell you also validity
Direction and strength
Statistical validity
Two quantitative measures for internal reliability, three or more items
Average inter item correlation AIC
chronbachs alpha
AIC
Calculated by taking the average of correlation between different variables or items
Reasonable .15 AIC .5
Chronbachs alpha
Range from 0-1
Zero is weak relation
One is a strong relation
Looks at internal consistency, how closely group questions are related to one other
Five sub types of construct validity
Face
Content
Criterion
Convergent
Discriminate
Which of the five sub types are subjective
Face and content
Which of the five sub types are empirical
Criterion, convergent and discriminate
Face validity
A measurement looks like what you would want it to measure
Content validity
The extent that your measure contains all the parts that your theory says it should contain
Criterion validity
If test scores are able to align with direct real world measures behavioral
Can use scatterplots
Known groups paradigm
Taking people you know belong to a group, but you want to see how the assessment differentiates between the two
Convergent validity
Your test is strongly associated with another test that has already been validated
Discriminate validity
Your test is less strongly associated with test measures of dissimilar constructs
Want low correlation
Self-report operationalize
A variable by collecting people’s answers to questions by questionnaires or interview
Observational operationalized
A variable by tracking observable behaviors or physical actions
Physiological measures operationalized
A variable by recording biological data
