MEASUREMENT

Operationalize

Translate definition from abstract to concrete forms of observation/manipulation

Self-report measures

People report on the variable

• Limited by self-awareness and honesty

Behavioral measures

• Observations

• Could be naturally occurring/ lab induced

Physiological measures

• Assessment of bodily states

• Accurate and objective

• Expensive

• May not always reflect psychological processes

Quantitiative

Variables are assigned a numerical value

Qualitative

Verbal description

True score

The actual score - not something you can measure

Obtained score

Score the measure gives

Measurement error

Difference between the true score and obtained score

• Something researchers want to minimize

Reliabiltity

Does the test give consistent measurements?

Test-retest reliability

• Same test is given twice with a delay

• Not useful for unstable variables like mood

Parallel-form reliability

Different forms of the same test

• E.g., GRE questions

Interrator reliability

2+ people record what they observe

Validity

Measures can be reliable but inaccurate

• E.g., a scale always 5 pounds off

Construct validity

Does it sample relevant behavior?

• E.g., employment test should ask about job related skills

Predictive validity

Does it predict expected outcomes?

• E.g., measure of stress predicts illness

Convergent validity (broad)

Does it correlate with related variables?

• E.g., shyness measure with social anxiety

Discriminant validity (divergent)

Does it not correlate with unrelated variables?

• E.g., extraversion & neuroticism

Instrument threat

Instrument we developed no longer applies

• E.g., measures of life satisfaction that included item “I normally feel very gay”

Sensitivity (potential problems with adequacy)

Can it detect subtle differences?

Range effects (potential problems with adequacy)

Min or max score is reached too frequently

• Ceiling effect

• Floor effect

Ceiling effect

Everyone scores upper limit

• E.g., memory test with only 5 words

Floor effect

Everyone scores low limit

• E.g., giving 8 yr olds calc exam

Reactivity

When observing someone changes their behavior

• Threatens validity

Subject-expectancy effects

When participants anticipate the hypothesis & respond differently

Demand characteristics

Cues that signal purpose of a study

Good subject

Participants want to help researcher affirm hypothesis

Negative subject

Wants to act contrary to hypothesis

Apprehensive subject

Wants to present themselves in a desirable manner

Experimenter-expectancy effects

When researchers knowledge of the predictions affects how they treat participants

• reducing reactivity

Blind design

Participant does not know their condition

Double blind

Experimenter & participants do not know condition