Assignment 2
------------

The research question is: Does seeding (action) increase rainfall? (Look at the
first sentence of handout.)  

(a) Most relevant summary statistics:

          mean and sd of rainfall for action=0,
          mean and sd of rainfall for action=1.
   
    Also, show mean and sd for `suitability' and `percent'. In addition, show 
    the scatter plot matrix.

(b) > lm(rainfall~action*suitability*percent)

    See diagnostic plots. Probably 1 point is outlier. Remove that point and 
    run the above regression again. This maybe your final model (Well, there
    maybe other reasonable models.) COMMENT on diagnostic plots.

(c) Most important comment is that on seeding (action). Look at the p-value 
    corresponding to `action' in the summary, and comment accordingly.

    Some of you said `action' is significant even when your summary table 
    gave large p-value for `action'. You were referring to `model p-value'.
    Model p-value (from F statistic) is one thing (it tells you whether the 
    model as a whole is OK), individual p-value for different covariates is
    another thing (tells you about the role of that particular covariate).