- No assumption of normality, i.e., the residuals of the model are not assumed to be normal
- how to model non-normal distributions?
- count data - poisson distribution
- proportions - binomial distribution
- glm() with family = “gaussian” is the same a lm()
# Plot the data using jittered points and the the glm stat_smooth
ggplot(data = df_long, aes(x = dose, y = mortality)) +
geom_jitter(height = 0.05, width = 0.1) +
stat_smooth(method = "glm",
method.args = list(family = "binomial"))
- The plot is bounded by 0 and 1
- The plot is monotonic - only increasing
- this is an important assumption of binomial regression
