Monday, July 30, 2007

The data is what it is...

Every now and then, you think something should be pretty obviously true, but it turns out to be false... or at least not entirely true. Consider the following supposition:

Students who take their remedial math class and first college-level math classes in consecutive semesters are more likely to pass the college-level class than those who take a break between the two.

On the face of it, it's seems arguably plausible and pretty simple. And when you look at the numbers, it turns out to be true. From the data available, we found that 52% of students pass when there is a break in between and 62% of students pass when there isn't. This is a 10% difference and a relative 19% increase.

However, when you take a closer look at the data, a surprising trend arises. For the weaker students (B- to C- students), the pass rate is approximately the same regardless of whether they took the classes sequentially! The data set got a bit small for this part of the analysis, with less than 100 students, but the signs point to something lurking underneath the surface that has not yet been identified.

The mystery to understand why this is the case begins...