Intro

All of these exercises use the data in medTrialX.xlsx

Part I summary exercise

Create a new variable hdiff as the difference between hbefore and hafter.
Generate two new ranked variables: incgroup and postsize following the rules:

    If income >=  3rd Quartile , incgroup = “high”,
        if income < 1st Quartile  , incroup  = “low”,
            else incgroup = “mid”

and

    If hdiff > median(hdiff), postsize = "large",
        else postsize = "small"

What do the two new variables tell you?

Give appropriate labels to all variables.
Give appropriate labels to the values of gender, smoker and incgroup.
Generate detailed summary descriptive statistics for hbefore, hafter and hdiff.
Make box plots of hbefore, hafter and hdiff as a function of gender. Put appropriate titles on the plots. Save the plots.

Make a custom table showing the mean, sd and skewness for hafter tabulated by gender and smoker. Do not display totals. Save the table as a word document.
Test for any association between the variables gender, smoker, incgroup and postsize.
Test to determine whether gender, smoker or income group affect the mean level of hafter.
Create regression models to test for the effects on hafter of
- age;
- age and gender;
- gender and smoker;
- any interactions between gender and smoker. Describe your results.