PSYC 406 Chapter Notes - Chapter 4: 21 Days, Item Response Theory, Intelligence Quotient

• School: McGill University
• Department: Psychology
• Course: PSYC 406
• Professor: Stephen Stotland

Document Summary

Xi=t+e: xi: observed score, t: true score, representative of one"s true skills/abilities/knowledge/ attitudes etc , e: measurement error. You only have one true score on a test, however due to varying in error, you may have multiple different observed scores. Doesn"t account for systematic measurement error (consistent differences in test performance that occurs between groups or individuals. ) as this should not lead to inconsistent/ unstable test scores (ergo not important for reliability analysis. : assumptions of ctt. Mean of error scores= 0 for a population (some have increased scores, others have decreased observed scores because of error, so they balance out. ) Correlation between one"s true score and one"s error score= 0 (no relationship between score on test and the measurement error on test. ) Ctt will allow us to estimate the reliability of test scores, and tells us how much confidence we should have in the. The largest source of error, but also the easiest and most accurately estimated source of error.