JDP wrote:
Here’s how I look at the 58-63% MPS.
Part of the analysis is guessing because only three buckets of scores are disclosed, so the most critical component in coming up with an MPS is what assumptions you use for 0-50%, 50-70%, and >70% buckets.
In guessing the MPS, the most generous scenario would be using 0%, 50% and 70% for each of the categories, and the most pessimistic scenario would be using 49.9%, 69.9% and 99.9% for each of the three categories.
Now obviously the two scenarios are unlikely, so you have to come up with a “reasonable assumption” for the three categories. You can argue 40, 60 and 80 would be “reasonable,” but you don’t know for sure. I would say coming up with a confidence interval doesn’t add any value because you don’t even have an observed mean to begin with.
The difference you talked about can happen when you have a band 10 failed result applied with not so generous assumptions and a barely passed result applied with generous assumptions.
And I wrote this during my study break…I’m heading back to practice tests.