Read this piece in yesterday’s IHE, and see if you can believe it.
Apparently, a D of Ed statistical study was conducted to examine the impact of race as a factor in student loan default rate, and results were published accordingly. One small problem emerged inadvertantly last December however when, as part of the discovery process in litigation involving the department, it turned out that the study had omitted any data with regard to black students.
Now, let’s note slowly, carefully and specifically what happened here: 1) the US Department of Education conducted a survey in which it sought to demonstrate the impact of race on repayment rates of student loans and 2) it reached its conclusions without including any figures about black students in the analysis. I think I’ve got it right.
One of the commenters noted that a private business enterprise might well face criminal charges for work like this, while several others suggested that the missing data were deliberately excluded because the results would have been politically inconvenient.
At the very least, the episode doesn’t leave you brimming with confidence about the D of Ed’s capacity to produce reliable and accurate quantitative work, especially if it’s going to figure in their rating of individual programs’ eligibility for student loans.
Looks as if Secretary Duncan had better mandate some crash remediation in statistical analysis for his data crunchers, eh?