Multivariate statistical analysis Flashcards
(17 cards)
univariate analysis investigates one variable at a time using:
one sample t test
Bivariate analysis investigates the relationship between two variables using:
bivariate regression or chi-squared table analysis
Multivariate statistical analysis
- Investigates multiple variables at once
- Controls for third variables that may influence relationships.
Dependence methods
explain or predict one or more dependent variables on the basis of two or more independent variables
Interdependence methods:
techniques that are used to group things together and give them meaning
dependence techniques
- multiple regression analysis
- multiple duscrininant analysis
- logistic regression
- multivariate analysis of varience
- n way cross tabulation
interdependence techniques
- exploratory factor analys
- cluster analysis
- multi dimensional scaling
N way cross tabulation
where two non-metric scaled variables are compared after accounting for the effects of a third or more non metric variable
- Gender and brand trial post-Instagram ad.
Partial correltion analysis
Measures association between two linear variables after controlling for the effects of other variables
- exmple: Education vs. commission while controlling for experience.
n-way univraute analysis of variance (ANOVA)
simultaneously tests for the difference in the mean of a metric dependent variable among two or more non-metric independent variables
- example: Message type, appeal, and site entry affecting purchase.
multiple regression analysis
analysis that allows for simultaneous investigation of the effect of two or more independent variables on a single, interval, scaled dependent variable
- New brand intro with varying advertising, pricing, and distribution levels.
Analyzed to predict sales performance regionally.
Interpreting Regression Output
- R² shows how much DV variance is explained.
- Beta coefficients identify influence strength of each IV.
- Use unstandardised (B) for prediction; standardised (β) for influence.
Mulicollinearity - problem with multiple regression
It is difficult to completely separate the effects of one idependent variable from another, causing the parameter estimates to be unreliable
Binary logistic regression
establishes a rule for forecasting the value of a binary dependent variable from a combination of two or more metric independent variables
Detecting Multicollinearity
Use correlation matrix to identify moderate to high IV correlations.
If the analysis contains one dependent variable that is metric and several independent variables that are also metric, what is the correct statistical analysis?
Multiple regression analysis
What is the term for a problem in multiple regression where independent variables are correlated with each other, making parameter estimates unreliable?
Multi-collinearity
