ch 5: good measurement

conceptual definition

a researchers definition of a variable at the theoretical level. AKA construct/conceptual variable

self-report measure

method of measuring a variable where people answer questions about themselves

observational measure

method of measuring a variable by recording observable behaviours

physiological measure

method of measuring a variable by observing biological data. can be confounded

quantitative variable

variable thats values can be recorded as meaningful numbers

categorical variable

a variable thats levels are categories. AKA nominal variable. levels dont vary quantifiably

ordinal scale

type of quantitative variable where number represents rank order but they aren’t equal intervals between them

interval scale

type of quantitative variable where numbers represents rank order equal distances and there is no true zero (eg temperature)

ratio scale

type of quantitative variable where numbers represents rank order equal distances and there is a true zero (eg weight, grades on a test)

construct validity

how well are the variables operationalized

reliability

consistency of results when study is repeated

test-retest reliability

when a participant gets pretty much the same score anytime they retake the experiment

internal consistency/reliability

participant gives pretty much the same answers no matter how the question is phrased

inter-rater reliability

scores are pretty consistent whomever performs the test

correlation coefficient ( r )

quantifies reliability. indicates strength of relationship and direction of correlation (pos/neg)

face validity

subjective way of assessing construct validity. does the measurement look like its measuring what you want it to

content validity

subjective way of assessing construct validity. does the measurement contain all the parts of the theory it claims it does

criterion validity

empirical way of assessing construct validity. does the measure correlate with a well established behavioural outcome?

convergent validity

empirical way of assessing construct validity. self-report measures correlate strongly with other measures/wording of the same thing

discriminant validity

empirical way of assessing construct validity. self-report measures do not correlate with measures of different things. shows two concepts are distinct