Regents scores are not percents
It is that time of year again. Final grades are being estimated. Regents exams are scheduled. Graduation parties are being planned. But what about those students who don't get to graduate or who are scheduled for summer school because the scores on their Regents exams are inappropriately included in their final average for a course. (And the average is probably a mean instead of a median - another way to ensure higher summer school registration and lower graduation rates. More to be said in another blog post)
But aren't Regents exams measures of student achievement like finals are? And aren't the Regents on a 0-100 scale like many classroom grading systems? On the surface, one might answer yes to both, but upon further investigation, 1) we know that the Regents are not end of course exams as they are not all given at the end of the corresponding course, and they are not all completely aligned with NYS Learning Standards (the Regents English exam is a good example of both scheduling and content). And 2) as per the NYSED link attached, Regents scores are scale scores, not percent correct, in order to account for the varying levels of item difficulty (thanks to good 'ole item response theory).
So let's see an example of a similar type of inappropriate averaging with other data...
I ask the local weather person what the average temperature is for the coming week. For me, I will do yard work if the average is over 70; indoor work if under 70. Here are the readings:
Monday - 70
Tuesday - 80
Wednesday - 76
Thursday - 81
Friday - 72
Saturday - 24
Sunday - 75
So the average temperature is 69 (using the mean), right? I will work inside. Well, the math is correct, but the data are not on the same scale! Saturday's temp is in Celsius! The converted temp for Saturday is 75 degrees so the mean is actually 76! I should be gassing up the mower!
Working outdoors or indoors is not a big issue, but graduation or not, summer school or not, credit recovery or not - those are BIG issues. And the wrong calculation could make the difference for students in high school and beyond.
But aren't Regents exams measures of student achievement like finals are? And aren't the Regents on a 0-100 scale like many classroom grading systems? On the surface, one might answer yes to both, but upon further investigation, 1) we know that the Regents are not end of course exams as they are not all given at the end of the corresponding course, and they are not all completely aligned with NYS Learning Standards (the Regents English exam is a good example of both scheduling and content). And 2) as per the NYSED link attached, Regents scores are scale scores, not percent correct, in order to account for the varying levels of item difficulty (thanks to good 'ole item response theory).
So let's see an example of a similar type of inappropriate averaging with other data...
I ask the local weather person what the average temperature is for the coming week. For me, I will do yard work if the average is over 70; indoor work if under 70. Here are the readings:
Monday - 70
Tuesday - 80
Wednesday - 76
Thursday - 81
Friday - 72
Saturday - 24
Sunday - 75
So the average temperature is 69 (using the mean), right? I will work inside. Well, the math is correct, but the data are not on the same scale! Saturday's temp is in Celsius! The converted temp for Saturday is 75 degrees so the mean is actually 76! I should be gassing up the mower!
Working outdoors or indoors is not a big issue, but graduation or not, summer school or not, credit recovery or not - those are BIG issues. And the wrong calculation could make the difference for students in high school and beyond.
posted by DAS at
Monday, May 23, 2011
0 Comments:
Post a Comment
<< Home