120 Excel
STATPLUS TIPS
Once you break a scatter plot into categories, you cannot go back
to the original scatter plot (without the categories). If this is a prob-
lem, create a copy of the scatter plot before you break it down.
If your chart contains several data series, you can choose which
series to break down into categories by selecting the series name
from the Select a Series drop-down list box in the Display by
Series Category dialog box.With the data series now broken down into categories, we can compare
the universities’ graduation rates on the basis of the high school perfor-
mance of its incoming freshman students. On the basis of the chart, we
quickly see that schools in which 80% or more of the incoming freshmen
graduate in the upper quarter of their high school have higher graduation
percentages whereas those schools in which less than 80% of incoming
freshmen graduate in the upper quarter see a lower percentage of those stu-
dents graduate.
So we have two predictors of graduation percentage. One is the calculated
SAT score, and the other is percentage of incoming freshmen that graduate
in the upper 25% of their class. Notice, however, that within each category
(,80% and .580%), there does not seem to be a clear trend based on cal-
culated SAT values.
Essentially both measures are telling us the same general thing: uni-
versities that attract better student athletes will also have higher gradua-
tion percentages. That’s not surprising. What might be useful, however, is
determining which of the calculated SAT score or the high school gradu-
ation ranking is the better predictor. That is not something we can easily
determine from the chart. To answer that question, we have to perform a
statistical analysis of the data, which we’ll do in a future chapter.
There are other factors which we haven’t investigated. Does the size or
the location of the school matter? Does it make a difference if the university
is a public or private institution? And we have to be aware that we are only
looking at a sample of 11 universities; we don’t know if any of our conclu-
sions can apply to another sample of schools. All of these are questions for
future study.Plotting Several Variables
Before fi nishing, let’s explore one more question. The data include gradu-
ation rates for the athletes in the freshman class broken down by gender
and race. How do these graduation rates compare? A scatter plot displaying