HED 373 Lecture Notes - Lecture 14: Pearson Product-Moment Correlation Coefficient, Internal Consistency, Inverse Relation
3 Apr 2017
School
Department
Course
Professor
Document Summary
Test-retest reliability: measure at 2 different points in time, assess reliability by calculating correlation coefficient, physiological measures should be 0. 9, psychological measures should be 0. 8, below 0. 7 problematic. Internal consistency: split-half reliability: treat a test given to a group of subjects as 2 tests by scoring even- numbered items separately or give half the survey & use spearman-brown prophecy. Formula to estimate what reliability would be if given whole test; should be +0. 8. Spearman-brown prophecy formula: r(est)= 2r/1+r where r+reliability: internal consistency: extent to which each item on survey or test is measuring same thing, use coefficient alpha, should be +0. 8. Pearson correlation coefficient: examines relationship between 2 sets of scores, symbol is r, values between -1 and 1: direct relationship: positive, inverse relationship: negative. Factors affecting validity of instrument: guinea pig effect, role selection or impression management , response sets, interviewer effects, population restrictions.
