Multiple Regression
Terms
undefined, object
copy deck
- Multiple regression is
- a data analysis technique that enables researcher to examine patterns of relationships between multiple independent variables and SINGLE dependent variable
- multiple variables formula for DV
- DV= (coefficent 1(effect 1) = coefficient 2(effect 2) +....+constant+residual
- multiple regress formula
- DV=(slope 1 (IV1)+ slope 2((IV2)+...Yintercept+ residual
- what is adjusted R(sq) for Mult Regression
- the squared multiple correlation between predicted and actual Y scores..the adjustment deflated bias based on sample size and #IV
- beta coefficients?
- the beta slope in standardized form, with scores converted into standard scores (so apples vs apples comparison)
- first step in Multiple Regression is to examine the interaction of the multiple IVs is that does not pass the F value, you ...
- stop there and don't go one, if good then examine each IV
- trustworthiness of results from any analysis can be affected by problems of
- sampling, measurement, the role of chance, and the technical assumptions of chosen analytic technique
- in Multiple regression, multicollinearity occurs when and with what effect?
- occurs when IV are highly correlated, effect of making it harder to reject null hypothesis around regression coefficients
- assumptions for Multple regression?
-
1. normality of distribution-check by doing histogram and look for fairly normal curves
2. homoscedasticity-homogeniety of variance
3.linearity-data cloud can be summarized with straight line besy - Bonferroni adjustment
- takes the traditional p value( alpha) .05 and divides it by # of tests, and that sets the new alpha threshold