Exploring birth-weight data

The Birth_weight data frame comes from a study of the risks factors for being underweight at birth. Let's explore the factors that might be related to birth weight.

One way to explore data is by building models with explanatory variables that you think are important, but in my view this is really confirmation rather than exploration. For instance, consider these models. The first involves explanatory variables that relate to social or lifestyle choices and the second involves biological variables. (Note: income is given as one of 8 levels, from poorest to richest. baby_wt is in ounces: 105 ounces is one kilogram.)

model_1 <- rpart(baby_wt ~ smoke + income, 
                 data = Birth_weight)
model_2 <- rpart(baby_wt ~ mother_age + mother_wt, 
                 data = Birth_weight)

Build these models and look at them, e.g.:

prp(model_1, type = 3)

The results might suggest to you that some of these explanatory variables are important and others aren't.

Now build a "bigger" model, combining all of those variables. Based on this "bigger" model, interpret the relationship among the explanatory variables as they relate to baby_wt. Select the single true statement from among these:

Possible Answers

Smoking is associated with baby_wt, but only for low-income mothers.
Smoking matters for higher-weight smokers.
Income matters only for non-smokers.
Income matters for higher-weight smokers.

Exercise

Exploring birth-weight data

Instructions

Possible Answers