Please enable JavaScript.

Coggle requires JavaScript to display documents.

Jennifer's Lectures - Coggle Diagram

- - - - The stability or consistency of the measure
        
        A measure is reliable when : the same people measured under the same circumstances, produce nearly identical measurements over different time points
        
        Different raters produce the same or near same measurements
        
        In a reliable measure, error shouldn't fluctuate
    - - Successive measurements
        
        Test-Retest ability: Same measure administered to same people at different times
        
        Alternative forms: Different forms of the same measure administered to the same people at different times
        
        Simultaneous measurements
        
        Inter-rater agreement: Agreement between observers coding the subjects behaviour.
        
        Inter-rater reliability: Observers consistently distinguish between different subjects behaviour.
        
        Internal consistency
        
        Split-half reliability: high correlation between 2 halves of a measure
        
        Cronbach's Alpha: (q) Average of all possible split halves for internal ratio scale measures. >. 70 = adequate. >. 80 desirable
        
        Kuder-Richardson: Chronbachs alpha for dichotomous measures
        
        Reliability coefficient: True score variance/ Observed score variance
  - - - Face validity: Does the measurement appear to measure what it sets out to measure?
        Content Validity: Is the measurement capturing all representative aspects of the construction in question?
    - - Convergent validity: Are multiple Measurement methods for the same construct highly correlated?
        Divergent Validity: Are multiple measures of different constructs not highly correlated?
        Concurrent Validity: does the measure relate to other previously validated measures and theoretically predictable behaviors surrounding the same construct.
        Predictive validty Does the measure accurately predict a theoretically expected behaviour (in a casual sense)?
        ((incremental validity: Predictive validity above and beyond existing measures))