logistic regression Flashcards
Logistic Regression
to investigate the relationship of a number of predictor variables (categorical or continuous) with one categorical outcome (criterion) variable
Types of Logistic Regression
Binary Logistic Regression/ multinomial logistic regression
Binary Logistic Regression
Used when the categorical criterion variable has two levels
Multinomial Logistic Regression
Used when the categorical criterion variable has three or more levels
Use of logistic regression
How well your set of predictor variables explain your categorical outcome variable/ How accurate your model is at classifying individuals into groups (in the sample)
binary logistic regression
Logistic regression is about assessing the probability of belonging to one group as opposed to another (based on scores for the predictor variables)
Goodness of Fit Tests
test compares the performance of your model (your combination of PVs) to a model with no PVs/ p < .05 is good fitting model
specificity
what proportion of the sample are correctly
classified as not belonging to the group of interest - True negatives
WALD test
WALD test is used to determine if a PV is a significant predictor of belonging to the group of interest.
Sensitivity
what proportion of the sample are correctly classified as belonging to the group of interest - True positives
Odds ratio
the odds of belonging to the group of interest based on a one unit change in the PV.
The directional effect of the PV
assessed through inspection of the Beta value (B)
what do statistically significant PVs indicate
indicate the variables that make a person more or less likely to belong to the group of interest.