The new BusinessWeek rankings of business schools are out. Thousands of future MBAs will pour over the statistically-insignificant differences between similarly ranked schools to decide which will receive their quarter million dollars.

Of course, we can disagree over the finer points of methodology. But first, BusinessWeek needs to show us that they are capable of basic math.

For Vanderbilt, BusinessWeek reports that 170% of students have jobs. Given that they consider their formulaic hocus pocus proprietary, we can’t know exactly where BusinessWeek screwed up. What we do know is that BusinessWeek shouldn’t be trusted with important mathematical derivations that impact actual people’s lives.

**UPDATE (19 November):** BusinessWeek associate editor Louis Lavelle emailed me to tell me that the error was Vanderbilt’s, confirming my hunch in a comment below. Louis writes

It was a reporting issue from Vanderbilt. The questions in the survey ask for placement AT graduation, and AFTER graduation but within three months. In the profile those two numbers are combined for total placement within 3 months of graduation. I don’t have [Vanderbilt's] survey open, but instead of giving us the two numbers we asked for (e.g. 70% + 12% = 82%) [Vanderbilt] gave us a “three-month” number that’s cumulative.

In sum, Vanderbilt fails at reading comprehension, but BusinessWeek fails to put provided data through even the simplest validity checks. While these numbers do not enter into BW’s rankings, there is no public verification of the numbers that do.

Graduating MBA students, here’s how to get your school atop the next rankings. When asked to rank your school on a scale of 1 to 5, answer 500. That should raise your survey average a bit.

**UPDATE (21 November):**
BusinessWeek finally corrects the page, and editor Louis Lavelle’s tweets dismiss any responsibility on their part.

To be fair, placement does not actually enter the BusinessWeek formula for ranking schools, which are exclusively based on surveys and a small bonus for writing books that BusinessWeek reviews. However, the methodology that they do use (discussed here) is incredibly vague and devoid of rigor.

This is most likely a data error made at Vanderbilt. BusinessWeek just enters the data the schools give it.

I don’t see this mistake for any other school, so this does seem to be a Vanderbilt thing.

My question would be how they handle multiple job offers… does 170% mean an average of 1.7 job offers? That’s not exactly an error if the other schools are able to use multiple offers for some grads to get up to the average of 0.8 job offers per student.

Chase, unfortunately that’s not the case. The question is what percentage of

studentshave at least one offer. The mistake seems to be that schools are asked what percentage of students have at least one offer “by graduation” and what percentage “after graduation, but within 3 months.” Vanderbilt’s reported answer seems to be 78% for the first and 92% for the second, which are summed (because theyshouldbe mutually exclusive) to obtain the 170%.