Wide-form input logistic regression
The second and third approaches for fitting logistic regressions require "wide" format data:
x fail success Total successProportion
1 a 12 2 14 0.1428571
2 b 3 11 14 0.7857143
For the second approach, model a 2 column matrix of successes and failures (e.g. number of no's and yes's per group).
In this format, use the formula cbind(success, fail) ~ predictor
.
For the third approach, model the probability of success (e.g. group 1 had 75% yes and group 2 had 65% no) and the weight or number of observations per group (e.g. there were 40 individuals in group 1 and 100 individuals in group 2).
In this example, successProportion
= success
/ Total
.
In this format, the formula proportion of successes ~ response variable
is used with weights = number in treatment
.
This exercise is part of the course
Generalized Linear Models in R
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Fit a wide form logistic regression
lr_2 <- ___
# Print the output of lr_2
___