Suppose it is of interest to use the data sets in Exercises 1.24 and 1.25 to derive a model that would predict staff salaries as a function of percent of families in a high income level for current school systems. Comment on any disadvantage in carrying out this type of analysis.