Flashcards
Terms for Final 4
Cards in this deck (31)
- 1Bivariate CorrelationAn association with exactly two variables
- 2Coefficient of DeterminationSquaring correlation to calculate percentage of variance.
- 3Point-serial correlationAssociation between one categorical variable and one quantitative variable
- 4Phi CoefficientThe association between two categorical variables.
- 5SpuriousAn association that is only attributable to the systematic mean differences on subgroups in the sample.
- 6Outliers(On-line or off-line.) Where a few cases stand out as much higher or lower than the sample.
- 7CurvelinearWhen the slope curves, resulting in an association without being detected by r.
- 8Directional ProblemWhere it is unclear which variable came first (temporal precedence).
- 9Third variable problemWhere there is a plausible alternative for the association between variables (internal validity).
- 10ModeratorA third variable that changes the relationship between two other levels.
- 11Spurious correlationWhen two variables do not have a direct relationship, but are instead related to a third variable.
- 12Partial correlationTaking out a third variable and seeing if there is still a relationship.
- 13Longitudinal DesignWhere the same variables are measures in the same people over time.
- 14Cross-sectional correlationsA correlation between variables that are measured at the same time.
- 15AutocorrelationCorrelation of one variable with itself at two different times.
- 16Cross-lag correlationsCorrelation between an earlier measure of one variable and a later measure with another variable.
- 17Structural equations modelingIf X -> Y has a higher correlation than X -> Z, then XYZ should be stronger than XZY.
- 18Path analysisA form of structural equations modeling that uses single measures of each construct.
- 19Latent variable modelingA more complex version of structural equations modeling where each construct has 2 or more variables.
- 20Multiple regressionTest the influence of the third variable.
- 21Predictor variableUsed to explain variance in the dependent or criterion variable. AKA the independent variable.
- 22Criterion variableThe variable researchers are trying to predict in multiple regression AKA the dependent variable.
- 23Standard multiple regressionEntering all of the predictor variables at once.
- 24Stepwise multiple regressionEnters the predictor variables one at a time, only those that are considered unique.
- 25Hierarchical multiple regressionEach predictor is entered in an ordered set.
- 26Multiple correlation coefficientHow well a variable predicts an outcome (R). Also coefficient of determination?
- 27Controlling for the third variableHolding a potential third variable constant while investigating an association. Checks for variability.
- 28BetaA value for each predictor variable. Shows direction and strength but without guidelines.
- 29Pattern and parsimonyA variety of correlational studies that point to one causal direction.
- 30MediatorA variable that helps explain the relationship between two other variables. Internal to to the causal variable, of direct interest.
- 31Factor analysisA statistical technique used to analyse interrelationships among a large number of variables. This could be one factor, or one factor per variable.