The new bi-annual Businessweek MBA rankings are now out. Everyone associated with business can now jump to the rankings to assess the caliber of the schools they went to, or wanted to go to. The large notoriety this year was that one of the oldest and most popular management programs -Harvard- has now completely placed out of the top 5 schools. This is the first for the Boston school, in the entire history of Businessweek rankings. Meanwhile Duke, ranked #6 just two years ago, vaults to the lead spot. Does this indicate that the rankings are completely out of whack with historical views of business school caliber? Not necessarily, though the changes are noticeable.
Look at the rankings on the right of the illustration below. The 2014 rankings are along the vertical axis, and the 2012 rankings are on the horizontal axis. Note a small detail that slightly less than 30 schools are shown, as we used only those schools which were jointly present in the top 30 for both rankings. More importantly, the pattern is fairly ok, and we can solve this serial correlation with a couple related techniques. We see here that schools have generally performed about where they performed last time. That should be expected.
Now in the same chart look at where Yale (colored blue) lies, and we notice it is the only clear outlier. Though still it doesn't wind up changing the overall relationship if removed. A school even closer to the core trend is actually Harvard (colored crimson). Even if it now ranks slightly worse than Yale (now that is noteworthy!)
The only reason we might be shocked by these ranking statistics -initially- is that we have been accustomed to much more of a 1-to-1 relationship, time after time after time. Particularly at the top schools. This used to be the way The New Jersey statisticians who invented these rankings for colleges used to defend them initially (e.g., "of course they are correct, see here- Princeton is at the top.")
We can see this for ourselves by looking at the previous 2012 rankings versus the 2010 rankings, shown on the left chart above. Notice the differences in charts, and how in the left chart many schools have no change in rankings at all, between 2010 and 2012. For example, Harvard remained at #2. And Yale remained at #21. Five other schools shared this 1:1 relationship, between 2012 and 2010. In 2014, on the other hand, only one school did.
Look at the rankings on the right of the illustration below. The 2014 rankings are along the vertical axis, and the 2012 rankings are on the horizontal axis. Note a small detail that slightly less than 30 schools are shown, as we used only those schools which were jointly present in the top 30 for both rankings. More importantly, the pattern is fairly ok, and we can solve this serial correlation with a couple related techniques. We see here that schools have generally performed about where they performed last time. That should be expected.
Now in the same chart look at where Yale (colored blue) lies, and we notice it is the only clear outlier. Though still it doesn't wind up changing the overall relationship if removed. A school even closer to the core trend is actually Harvard (colored crimson). Even if it now ranks slightly worse than Yale (now that is noteworthy!)
The only reason we might be shocked by these ranking statistics -initially- is that we have been accustomed to much more of a 1-to-1 relationship, time after time after time. Particularly at the top schools. This used to be the way The New Jersey statisticians who invented these rankings for colleges used to defend them initially (e.g., "of course they are correct, see here- Princeton is at the top.")
We can see this for ourselves by looking at the previous 2012 rankings versus the 2010 rankings, shown on the left chart above. Notice the differences in charts, and how in the left chart many schools have no change in rankings at all, between 2010 and 2012. For example, Harvard remained at #2. And Yale remained at #21. Five other schools shared this 1:1 relationship, between 2012 and 2010. In 2014, on the other hand, only one school did.
No comments:
Post a Comment