The strategy is to make the correlation between the Predictor and Reputation as high as possible at every step of the way. So, the variable that correlates the highest with Reputation is the one that will predict Reputation best. So start with it. Then look for the next highest, etc.
Here are the correlations among the variables that are common to Colleges230 and to CollegesAll.
Pearson ProductMoment Correlation 














zRepu 
zAcctRate 
zGradRate 
zTop10% 
zTest75 
zTest25 
zRepu 
1.000 





zAccRate 
0.656 
1.000 




zGradRate 
0.804 
0.600 
1.000 



zTop10% 
0.746 
0.708 
0.762 
1.000 


zTest75 
0.812 
0.708 
0.842 
0.845 
1.000 

zTest25 
0.825 
0.731 
0.853 
0.823 
0.961 
1.000 
Notice that the correlations are the same whether we correlate zscores or we correlate raw scores. Why? Because the formula for correlation converts every score to a zscore before summing the products, and the zscore of a zscore is itself. Click here for a more complete explanation.
Pearson ProductMoment Correlation 














Repu 
AcctRate 
GradRate 
Top10% 
Test75 
Test25 
Repu 
1.000 





AccRate 
0.656 
1.000 




GradRate 
0.804 
0.600 
1.000 



Top10% 
0.746 
0.708 
0.762 
1.000 


Test75 
0.812 
0.708 
0.842 
0.845 
1.000 

Test25 
0.825 
0.731 
0.853 
0.823 
0.961 
1.000 
The next table shows the steps I went through as I constructed my final Predictor formula, trying to make the highest possible correlation between Predictor and Reputation.