Psyc 203 1st Edition Lecture 10Outline of Last Lecture I. Review II. Reliability & ValidityIII. ErrorIV. Fundamental Equation of Classical Test Theory V. Internal Consistencya. Alpha Outline of Current Lecture I. Review II. Alpha III. Test- Retest Reliability IV. Parallel Forms V. Inter-Rater Reliability VI. Validity Content I. Review a. Definitions of Reliability and Validity– Reliability: Consistency and Precision– Validity: Capturing the construct with accuracyb. Classical Test Theory Equation– Measurement errors diminish reliabilityc. Coefficient Alpha These notes represent a detailed interpretation of the professor’s lecture. GradeBuddy is best used as a supplement to your own notes, not as a substitute.i. Lengthii. UnidimensionalityII. Alpha a. Doesn’t index unidimensionalityb. Ways to increase average inter-item correlation and increase length i. Statistical – Increase Average Inter-Item Correlation and Increase Lengthii. Focus on “good” itemsiii. SPSS can tell us if removing an item improves/reduces alphaiv. Don’t get carried away...too many items can frustrate survey takers and their answers can become hindered because of this v. Design Issues 1. Have clear instructions for participants2. Standardize assessment situationa. Make sure all assessments asses everyone equally III. Test- Retest Reliability a. The extent to which scores at one time point do not perfectly correlate with scores at another time point is another indicator of measurement error b. This assumes the underlying construct is stable.c. How to do this? i. Maybe wait a two week period to administer the same test again ii. The time waited is up to the testers’ discretion d. Sometimes called Dependability (Cattell) IV. Parallel Forms a. Different specific types of questions, but all measuring same ability or psychological construct V. Inter-Rater Reliability a. Inter-rater reliability: How consistent are the observations made by two people? (AKA Agreement!)b. Computed by calculating the number of agreements between the raters out of possible agreementsc. Proportion of Agreement= Number of agreements/ number of possible agreements d. For example: # of agreements/# of doctors when it comesto diagnosing a patient VI. Validity a. 3 types of validity:i. content : 1. conceptual2. property of a test such that the test items sample the universe of the items for which thetest is designed (most often used on achievement testst)ii. criterion: assesses whether a test reflects a set of abilities in a current or future setting 1. if a criterion is taking place her and now: concurrent2. if in the future: predictive iii. construct (not on this
View Full Document
Unlocking...