measurement scales/validity

Nominal= name/category

no meaningful order, ex: Majors-Psychology, Biology, English

Ordinal=Order/Rank

no meaningful order but exact distance between ranks isnt equal, ex: Unsatisfied —> Neutral —> Satisfied

Interval=Equal Distances, no true zero

numbers have equal intervals, zero doesn’t mean complete absence of construct

Ratio=Equal Distance + True Zero

ex: hours student studied —> 0hrs rlly means no studying occurred

Variable: Both continuous Interval/ratio

Use: Pearson R

Variable: One/both ordinal

Use: Spearman’s Rho

Variable: True binary + Continuous

Use: Point Biserial

Variable: Forced Binary + Continous

Use: Biserial

Variable: Both Binary

Use: Phi

Pearson R

Ex: Height + Weight (continuous interval/ratio)

Spearman Rho

Ex: Class Rank + Satisfaction Rank (one/both ordinal)

Point Biserial

Ex: Smoker Yes/No + # of Cancer Cells (true binary + continuous)

Biserial

Ex: Smoker Yes/No + Cancer Yes/No (forced binary + continuous)

Face Validity

does measure look like its measuring what its supposed to measure?

Criterion Vailidity

how well a tests results match up w/ a trusted, real world outcome

Concurrent Validitty

compare it to established measure @ same time

Predictive Validity

does measure successfully predict future/relevant behavior

Convergent Validity

diff measurements of same construct should correlates

Divergent Validity

measures of a diff constructs should have little/no relationship when theory says they should be distinct

Construct Validity

whether the study’s operationalization properly represents the thoerteitcal construct