Hertzman et al. conducted a study to identify determinants of
    elevated blood lead levels in preschool chilcren in a Canadian community.
    Subjects were children between the ages of 2 and 5 years inclusive
    who lived in a Canadian town that is the site of one of North
    America's largest lead-zinc smelters.  Subjects were considered
    to have high blood levels of lead if they had 18 microg/ml or greater,
    and to have low blood lead levels if they had 10 microg/ml or less.

    The data file "soillead.txt" on the course web page contains the
    following variables:

    highbld --  coded as 1 if subject has high blood lead level,
                         0 if subject has low blood lead level
    soil    -- the level of lead in the soil in the subject's yard

    Subjects were NOT matched.  

    This example is taken from Daniel, Biostatistics:  A Foundation for
    Analysis in the Health Sciences