PSYC 212 Research Methods Exam 2

Psychometrics

Is measuring tool dependable and measuring the right thing (quality of measurment)

Reliability

Consistency by reducing error

Test-retest reliability

Same people do measures more than ones should get similar scores

Inter-rater reliability

Two observers should score behaviors similarly

Parallel forms reliability

different versions of tests should give similar forms (ex. ACT)

Internal Consistency

Questions are measuring the same construct

Cronbach’s Alpha

Value greater than 0.8 means that items fit together well, creates reliability

Homogenous questions

can add together to create composite score (ex. all positive)

Heterogenous questions

can not add together to create composite score some may need to be reveresed

Validity

Accuracy of measurements, does the test measure what it should?

Convergent Construct Validity

strong positve correlation with similar concepts

Discriminant construct validity

weak correlation between dissimilar concepts

Criterion-related validity

strong positive correlation between measure and relevant criterion (relation to real world outcome)

Content validity

measure covers content it is meant to assess

Face validity

measure looks like what it claims to be (survey about anxiety looks like it is about anxiety)

Selection effects

Created when not using random assignment- confound

Order effect

How you respond to one IV level can influence the next level (order could effect results) (WI subject design)

Practice effects

Improvement/fatigue from too much repetition (WI subject design)

Carryover effects

how you experience first condition changes in next (WI subject designs)

To solve order effect of WI subject designs…

Counterbalancing- IVs in different sequences

Quasi Experiments

Control and experimental groups but groups were not randomly assigned

Multi-factor experiment term for IV

Factor

Factorial Design…

written blankxblank or number of IVs times level of IVs

n in n-way ANOVA

number of IVs