PRELIMINARY EXAMINATION OF THE DATA

[Link Map]
begin comments text sound
line

Below are boxplot of the three predictors showing how the presence or absence of kyphosis depends on each predictor.

The black box on the boxplots represents the inter-quartile range of the data, with the white line representing the median value.
95% of the data lies within the dotted lines, and the lines outside this segment represent extreme values.

The boxplots show that for all the three variables, start, age and number, there is a noticeable difference between the cases in which kyphosis is present and in which it is absent, which suggests that all three variables might be useful in predicting the occurrence of kyphosis.

line
Generalized Additive Models 23.4.96 Page : 11a of 15