Psychology 3580F/G Lecture Notes - Lecture 2: Extraversion And Introversion, Factor Analysis, Drivespace
22 May 2019
School
Department
Course
Professor
Thursday, September 20th — Lab 1
•basic vocabulary
•construct — something to be measured
•an individual difference variable/thing
•variance/variability — information, individual differences
•represented mathematically as an individual’s difference from a mean or average score
•variance = individual score - average score of group
•correlation — quantitative measure of relatedness (r)
•item — a question on a test or survey
•scale — a test or survey designed to measure one individual difference variable or construct
•sub-scale — a component of a scale (e.g. sociability and social boldness are 2 sub-scales
or facets of extraversion
•reliability
•refers to an estimate of the stability or consistency of measurement
•freedom from random error variation
•theoretical concept that scores are composed of true and random (error) variation
•X = T + e
•true score is what a person would get in a “perfect environment” (i.e. no random error)
•error variance assumptions
•mean of error measurement = 0
•T and e are uncorrelated
•errors on different measures are uncorrelated
•the goal is to determine how mush of the variability in test scores is due to variability in true
scores vs. measurement error since minimizing error can help improve tests
•X = T + Me + Re
•can ultimately be thought of as the proportion of systematic variance in the observed score
that is attributable to the true score
•4 basic forms of testing reliability
•test-retest
•alternate forms
•split-half
•internal consistency
•coefficient alpha (estimate of reliability)
•a function of the number of items in the test, the extent to which each item is measuring
the same thing as the other items, and sample characteristics
•true reliability exists but depending on the characteristics of the group the estimate
may be low
•standard error of measurement (SEM)
•an absolute measure of the amount of variability in test scores due to measurement errors
using confidence intervals
•a function of reliability and variability of test scores
•the more reliable the test the smaller the confidence interval as a function of error of
measurement will be
•can be used to create confidence intervals around a score
•as reliability decreases SEM increases
•SEM = (standard deviation)(the square root of 1 - the reliability)
•validity looks at whether the test tests what it’s supposed to test
•reliability is a necessary pre-condition of validity and serves as an upper bound for validity
