Statistics

[origamist]

Dear lazyweb of people-who-are-real-scientists-and-thus-actually-know-statistics:

Suppose that I have two graders/three junior tutors/four lab TAs/etc. They give me some sort of number for each student relating to student performance. I want to know if there is a statistically significant difference between the graders. (Actually, I want to know

Comments

[camlost]

Isn't that a t-test?

http://en.wikipedia.org/wiki/Student%27s_t-test

[zwilichkl]

It is a t-test if you have 2 groups. If you have more than 2 groups, you should use an ANOVA.

[zathrus]

I don't know about an established procedure, but it seems to me that comparing the average scores they hand out and the standard deviations in their scores might provide you with some information. You should of course keep in mind the effect of sample size on the reliability of this - it's always possible that all the best students ended up in one section, for instance, although comparing their average scores with the average scores earned by people in their sections on things you graded (such as exams) might help you get a handle on the chances of this. (Having now suggested all these calculations, I really hope you already have all this data in an easily-sorted spreadsheet.)

Newt

[nemene]

Compute the mean and standard deviation for a sample population for each grader. The standard error of the MEAN is to then take that standard error and divide by the square root of the smaple size. (this shows the variation in the average instead of the variance of the measurements.) if the two means compared to each other (with ranges) do not overlap then you have a statistically significant differnce etween graders.