Psych 3301 Exam 3

Multivariate Designs

involve more than two variables

Covariance

are the variables related

Temporal Precedence

does one variable cause the other variable

Internal Validity

is there a 3rd variable that might explain the relationship of the first two variables

Developmental designs

how do the relationships between variables change over time

Multiple regression designs

how do other variables affect the relationship between the variables of interest

Cohort effects

People who are born at different times may experience very different environmental influences

Longitudinal

Same person tested over prolonged period of time

Cross-sectional correlations

whether two variables measured at the same point at the same point in time are correlated

Auto-correlations

the correlation of each variable with itself across time - but cannot examine between variable influences

Cross-lag correlations

the degree to which an earlier measure of one variable is correlated with a later measure with the other variable

Predictor variable

Independent variable and 3rd variable

Criterion variable

dependent variable

Regression Coefficient

Called Beta and Indicates how well it predicts the criterion variable controlling for all other predictors

Mediation

What specific aspect of a variables causes it to be associated with another variable

Moderation

The presence of subgroups has implication for external validity

Confound

Alternative explanations that threaten internal validity

Systematic

Internal validity only threatened

Unsystematic

random, makes it difficult to see DV, but not a confound

Moderator

3rd variable that affects the correlation between the variables of interest

Mediator

any variables that explains the indirect relationship between two other variables

Experimental research

correlation does not equal causation

Placebo effect

When control condition is so convincing, it has an effect on DV

Selection effects

Participants in one group systematically differ from those in other

Between subject/groups

each different person/group of people has a different version of the IV

Within subject/groups

responses from a single group are compared across all versions of the IV

Post-test only

DV is measured only after exposure to the IV

Pretest - post-test

DV is measured both before and after exposure to the IV

Concurrent measures design

DV measured after all versions of the IV are presented simultaneously

Repeated measures design

DV measures after successive exposure to all versions of the IVs

Maturation threats

when participants development or passage time causes a change in their behavior (not IV)

History threats

when an external event affects the DV happens at the same time as the IV

Regression threats

when scores become less extreme over time

Time interval threats

Maturation

History

Regression

Repetition threats

Attrition threats

Instrumentation threats

Testing threat

Attrition threats

when participants drop out of a study in a non-random fashion; change due to changes in the composition group

Instrumentation threats

when scores change because the DVs change at each measurement/over time; people change because of change in the DV not IV

Testing threats

when scores change because of previous experience with the DV

Null-effects

effects not found:

not enough between groups variance

too much within groups variance

IV has no effect

Null effects for between-groups variance

less is bad

Null effects for within-groups variance

more is bad