"VCE School Leaders Briefing 2015" powerpoint- Statistical moderation

[ActiveSideofInfinity]

I just read a powerpoint which talks about statistical moderation indepth, its called "VCE School Leaders Briefing 2015"; you can find it ranked third in google if you search "VCE School Leaders Briefing 2015".

The last part is about how statistical moderation is performed, and it's pretty abrasive about it.
For example, If students are ranked equally, the top nth exam marks are averaged out, and that will be your SAC Score:

"In this example:
Three set of school results represented.
In all three schools the top three students have the same score and therefore the same rank in their Course Work Score (CRS)
In all three schools there is a variation in the exam scores.
As a result of the EXT being aligned against the Course Work score the students in school one all have confirmed moderated scores of 100
As a result of the EXT being aligned against the Course Work score all the students in school two have had their CRS adjusted positively.
As a result of the EXT being aligned against the Course Work score the students in school three have had their CRS
adjusted negatively .  In this case the EXT of 100+82+82 has then been divided by 3 to give each student a score of 88 on the internal assessment."


my schools' SACs are relatively easy and therefore the ranking system is volatile. So after exams, statistical moderation will do more harm than good- in this case.

"Students with anomalous results excluded   
Students with anomalous results are excluded from the moderation process so that they don’t adversely affect the moderation result.
Automatically exclusions - students with missing coursework scores, students with missing examination scores, students with DES applications, students whose results have been withheld & students who have transferred schools during the course of the year.
Statistical exclusions - student has an exam performance that is low when compared with the rest of the group (2.5 SD below the mean for the group). "

"Any adjustment to students’ coursework scores is made solely on the basis of the relationship between coursework scores and external scores for the group as a whole. It does not depend on the relationship between coursework scores and external scores for individual students"

For example, if the SAC average is 70% in a subject, but the mean is 50% in the exam, the only way that you would be saved from the nightmare scaling is if you were first in the cohort. The powerpoint also states that students with anomalous scores (low or high) will have their SACs scaled based on how their cohort did; it doesn't matter if you get full marks on the exam, you can, and will still probably get scaled down (harshly) if your cohort does worse than expected on the exam. The aim of statistical moderation is to make the mean of the SACs as close as possible to the mean of the exam marks, so if your cohort as a whole performs way worse on the exam... the mean of SACs will be scaled accordingly.

I love how VCAA states that it "Aims to eliminate marking differences between schools, not within" but it shows no mercy to students who do relatively well in the school SACs (no matter the difficulty) but are automatically scaled down if the cohort performs badly. A good example would be IT: applications; all sacs are computer based, and we learn little theory until all our sacs are complete. There is always a massive discrepancy between people who do well in SACs and people who do well in exams, but by the laws of statistical moderation, your sacs will be scaled according to how your cohort went. In the presentation, it was stated time and time again that individuals with anomalous scores will be scaled within the scope of the cohort, and so their SAC scores are heavily dependent on how their cohort went.


This should be a wake up call for people in a weak cohort, make every mark count, as bad as it sounds, its true that your cohort can have a massive effect on your study score. A drop in rankings (say 3) can be the difference between a scaled SAC mark of 85 and a scaled SAC mark of 70.

Anyone curious about statistical moderation should definitely read the powerpoint. It has strangely given me a stronger motivation to study; because I now know that my B+ and B in unit 3 will be most likely scaled down to a C and D+ respectively and there's nothing I can do about it except seek to redeem those lost points in UNIT 4, and the exam