CH 5: MEASURING VARIABLES

EXPERIMENTAL PSYCHOLOGY TEST 2

categorical variables

limited fixed # of values (labels/categories) ex: jury verdict

ordinal measurement

order matters (race positions)

interval measurement

Equal steps & zero is just a point on the scale (tempeature)

ratio measurement

true zero compare & calculate porportions (weight height age)

types of measures

self report

observational (behavior)

physiological (bio response)

conceptual defintion

General description of a variable in theoretical terms

EX: describe stress in general terms

operational defintion

how a variable is measured or manipulated in a study

EX: stress measured by self-report or corisol levels

operational defintion works by

Check past data to make sure things stay the same over time

test reliability

test-restest reliability

measure same ind at diff times get consistent results

ex:take same test month later & get similiar score

inter-rater reliability

diff ppl measure same thing, get similiar results

ex: 2 judges give similiar score to dancer

internal reliability

measure same construct using diff items

ex: teacher test student engagement w/test scores & participation

quanitative measurements

ordianal

interval

ratio

curvilinear association

changes direction at some point (U-shaped curve)

internal reliability: cronbachs alpha

avergae of all-item total correlations

reliability

consistency/stability of a measure

Validity

accurancy of a measure

its measuring what its supposed to measure

face validity

look like it measures what its supposed to

set of math Qs looks like measuring math ability

content validity

measure includes all relevant aspects of the concept

measure of depression should include mood, sleeping, appetite

predicitve validity

measure predicts future outcomes

driving test predicts how well they’ll drive in future

concurrent validity

measure show what it’s supposed to right now

new weight scale matches old scale score

convergent validity

diff measures give similar results

new depression scale show similiar results to exisiting one

discriminant validity

measure stays separate from unrelated measures

stress survey shouldn't relate w/happy survey