Applying Analytics

1-Dimensional Data Sets 95

tion. Some features of the fit stand out. First, the fit tracks the shape of the

data fairly well, and the major features of the falloff are captured. Second, the

high end of the fit falls so much that very high ages have a negative popula -

tion, which makes no sense. In fact, this is a result of choosing a str aight-line

functional form that doesn’t match the shape of the data, particularly the lev-

eling off in the curve at high ages. This could have been avoided by choosing

a more appropriate functiona l form or by fitting the data over a more limited

range. Without looking at the details o f the fit, there would be no way to

know that the procedure produced a nonsensical answer for parts of the fit’s

range, s o it is always