Tuesday, May 24, 2011
Sticky Statistics..hard to say and type! But onto my point..
So the Exit Exam has a standard error of measure, hovering around 8 when looking at the lowest passing score (350) and neighboring scores.
So if a student received a 347 they did not pass. If they take it again and get a 350 they do pass as there was growth and that growth met the minimum requirements for passing (the 350 or above score). Okay. But I could not say "student X's score improved from the first to second test by 3 points" because it is below the Standard Error of Measure threshold of 8.
Well then couldn't I argue that student X with a score of 347 did indeed possibly, maybe pass? (And the antithesis, the just-passing student at 350 perhaps did not pass). Couldn't I argue (with the in-arguabiltiy of numbers) to the state or ETS or whoever that student with a score of 347 should receive a passing status? I could fight for all the barely-missed it kids.
See because this is real. One of my students did get a 347 and it's sad. Last year a student got 347. Then 348. Finally, 352. But she didn't need to take it three times in my opinion, as all scores hovered around 350 and statistically were identical. Her original score could have been passing. And given her hard life and diligent effort, heck, she deserved it.