PSY 380 Lecture Notes - Lecture 6: Inter-Rater Reliability, Observational Error

• School: Drexel University
• Department: Psychology
• Course: PSY 380
• Professor: Gallo Jennifer

Document Summary

Any time you use an instrument, the reliability will vary. To minimize error that is naturally inherent in any sampling of behavior. More concerned with random than systematic error. X= observed score t = true score e= error. Error: factors that have nothing to do with true score but that are affecting t. Variance (sd^2) tells us something about the amount of variability in previous equation. score. Overall variance of x = variance of true scores + variance of expected. The smaller variance of expected the score, the closer the reliability score will. Error variance goes down, reliability coefficient goes up. Random error vs. systematic error be to 1. Ex: a scale that is off by 5 pounds. Relative standing is always the same with systematic error. Items created (valid and consistent), how many items on instrument (affects reliability), Test environment, test taker variables, examiner related. Types of reliability and methods for evaluating it.