13. Assessing Relationships 3 Flashcards
What does goodness of fit compare?
- regression model
- mean
What does goodness of fit use?
SS_T, SS_R, SS_M
goodness of fit denoted by
R^2
What is R^2 a good gauge of?
size of relationship/prediction
What is also provided with a regression analysis?
f-stat
significant f-stat in a regression
means regression model is a good fit
f-stat and goodness of fit
measures how much the model has improved the prediction of the outcome compared to inaccuracy of the model
b_1 (slope) represents
change in outcome resulting from a unit of change in the predictor
if a regression coefficient is 0:
- unit of change in the predictor variable results in no change in predicted outcome
- regression line is flat
What is used to test the null hypothesis that the value of b is 0
t-stat
When is a multiple regression used?
when there are more than 1 continuous predictor variables (IVs)
steps to a multiple regression model
same as simple regression
methods of entering IVs for multiple regression
- hierarchical
- forced entry
- stepwise
- examiner decides order of entering predictors into model based on previous research
- most important predictor entered first
hierarchical
all predictor variables entered simultaneously into the model
forced entry
divided into forward and backward
stepwise
- SPSS determines order of entry based on correlation
- highest correlation with the DV is first and only retained if it improves the model
forward stepwise
- all predictor variables entered into the model and the SPSS removes them one by one based on whether or not they improve the model
backward stepwise
logistic regression
same as multiple regression except the outcome variable is categorical, not continuous
types of logistic regression
- binary
- multinomial
binary logistic regression
only two categorical outcomes (DV is dichotomous)
multinomial logistic regression
2+ categorical outcomes
