Calculating study scores

[dankfrank420]

I recently received my statement of marks from VCAA, and I'm unsure as to how they derive them.

For example, it states that the weighted average of my standardised scores for methods is 1.559. Using a st.dev of 7 and a mean of 30, this would give me a SS of 41. However, I received a 43. Similarly, in Further I got a 1.69 and theoretically should have received a SS of 42 yet I received a 46.

Not that I'm complaining, just curious as to how VCAA arrived at these numbers when my own calculations gave me different ones?

[nerdgasm]

I don't think it's expected that your "weighted average of standardised scores" will perfectly correlate with your Study Score. Perhaps a different way of looking at it is: ultimately, everyone in Victoria who took a particular 3/4 subject is *ranked* according to their "weighted average of standardised scores" (WASS). Then, you look at the number of students who took the subject, work out how many people are meant to get 50's, 49's etc. etc., and distribute the Study Scores accordingly. So for example, if the highest WASS happens to be 2.344, then that person is guaranteed a 50, by virtue of being the highest-ranked student in the state, even though 30+7(2.344) < 50. Suppose also, that a fair number of people are taking the subject, so ~15 people can get 50's. You would then take the next 14 highest ranked people by WASS and give them 50's as well.

Also, consider what the 'theoretical maximum' WASS is for any subject in any year. Note that 50 ~ 30 + 7(2.86), so if Study Scores were calculated in that way, you'd need to be 2.86 standard deviations above the state mean in all your GA results. Looking at past grading distributions, that simply does not happen. Especially for SACs (but for exams too), a person getting full marks will be less than 2.86 standard deviations above the mean, simply because that's how the data is distributed. It logically follows that the 'theoretical maximum' WASS is very unlikely to ever reach 2.86, so this isn't a completely feasible way to calculate Study Scores.

[dankfrank420]

Thanks guys.