Logistic Regression Flashcards
What is centring?
Taking the mean value and subtracting the score from each value to centre the data, which is useful when the data set does not contain a 0
What are the conditions and use of a logistic regression?
when there is a binary dv, with continuous or binary ivs, used to answer diff questions than multiple regression e.g. group membership
What shape is a logistic regression distribution?
s-shaped, logarithmic, non-linear
What is a quadratic function?
u-shaped, parabella
What is a binomial logistic regression?
1 or 0 values, yes/no, is/is not
What is a multinomial logistic regression?
categorical, 1/2/3 options, you’re a dog, cat, or a frog, w/ no order
What is an ordinal logistic regression?
Categorical but with an order e.g. high/med/low
What are standard, sequential, or statistical logistic regressions?
standard is all entered at the same time, sequential is variables entered in sequence/blocks, statistical is forward or backwards entry
What is coding in logistic regression?
assigning a group, the higher value, as the target outcome, and the lower value as the reference group, ref group is group of not primary interest, if more than 2 cats, dummy coding is used
What are the assumptions of logistic regression?
independent errors, multicolinearity and singularity, linearity, outliers, sufficient data/5 cases at least per cell of contingency table, and data is not clustered/errors not related
What do you use to test model improvement for logistic regression?
chi-square test
What is good classifcation accuracy for logistic regression?
closer to 100%, the better, lower than 50%, not good
What is the odds ratio value for logistic regression?
change in the likelihood of dv in change of iv