In order to know how many students have fared well in my subject, I want to find out the attainment level for various course objectives.
The method that I am following is:
- Calculate the class average for individual course objectives.
- Count the percentage of students who have scored above (or equal to) the average.
- Calculate attainment levels based on the following rules: (a) Attainment level 3 if more than (or equal) 70 percent of the students score more than (or equal) average marks. (b) Attainment level 2 if more than 60 percent (or equal) of the students score more than (or equal) average marks. (c) Attainment level 1 if more than 50 percent (or equal) of the students score more than (or equal) average marks.
The problems that I am facing are:
- I am not getting an attainment level of 3 in any of the 50 subjects. Even getting 2 is very difficult.
- If I take an example, if the marks are 2,3,3,3,3,3,3,5 the average will turn out to be 3.125 and number of students more than or equal to average will be only 1 whereas I can see that 7 students have scored more than or equal to average marks.
I am taking average instead of predefined number as there are different types of subjects and setting a predetermined attainment percentage is very difficult. Can any one can help me here with any insights/references?
barrycarter's rant is less off-topic than it seems. (And is definitely correct - penalizing high-performing students and rewarding low-performing students just because they happen to be in a class heavy in similar performers is abysmally stupid. And if that actually is the policy of your university, then they apparently were in classes filled with low performers.)
If your presumption that the class performance is normally distributed were correct (and ignoring the finite sample size issue), then your system would never give you anything other than attainment level 1, and just barely at that, as exactly half the students would be in the at-or-above-average category. This is what the "average" means for a distribution that is not skewed. Normal distributions are balanced about their average. Half the students will be above average, half will be below average.
In order to get more than half the students to be above average, you need several students to be major disasters, falling way below the rest, while few if any students are major achievers. The really bad students will drag the average down disproportionally to the size of that group, bringing those who would otherwise have been just below-average to just above average. On the other hand, the really bright kids are working against you. Their high achievements raise the average with the opposite effect. Your attainment system rewards you for suppressing both the brightest and dullest of students, while encouraging only those in the middle. Is that really what you want?
Comparing students to the average is a grotesquely bad idea. The reason teachers grade on a curve is because they are not sure of how to accurately assess their students capability to master the material and their own ability to teach that material. (I've been there.) They are covering over their failings, not the students. The way to do this is to set a fixed standard on what students are expected to learn in the course. Then measure against that standard, not against the class average.