BHStatistics

November 16, 1999

Predicting College Reputations

Define a derived variable that will give predicted reputation scores for colleges in the CollegesAll data set (that have the requisite information).

Produce a variable table, sorted by predicted reputation. Check your predicted ratings for colleges that are also in the Colleges 1-230 data set. A college’s predicted reputation score should be the same in both data sets. [Note: There may be issues here that Luis tried raising that you chose to ignore. Be attentive!]

Hand in this on Friday. You will have class on Thursday to continue working on this. However, if you wait until Thursday to begin working on it, you will not have enough time to finish.

1. The formula you used to predict reputation rankings for colleges in the CollegesAll data file.
2. A list of the schools in the CollegesAll data set that have the 20 highest predicted reputation scores (school name, predicted reputation score).
3. A list of the schools in the CollegesAll data set that have the 20 lowest predicted reputation scores (school name, predicted reputation score).
4. A list of schools in the top 20 among the CollegesAll schools who are not in the top 20 among Colleges1-230. Why do you suppose these schools are not in the top 230?
5. An explanation of the procedure by which you ended up with predicted reputation ratings for colleges that don’t actually have one. Explain this as if you were explaining it to another student who is not in this class. That is, assume you are explaining your method to someone who is reasonably bright but who has not been in this class. Your explanation should have as its goal that the person to whom you give it understands what you’ve done and why you did it.

Note: The next two pages explain how you sort variables and make a variable table.

1. Set the sort options to "descending" (sort from highest to lowest values)

2. Select the variables you want sorted. The "Y" variable is the one that will be sorted. The other variables will be carried along. Then sort the variables.

3. Make a variable table.