Iceman145 said:

How does SAM do it so accurately, given that Table A3 and Table A8 from the UAC are so vague?

It fills in the blanks.

Its accuracy will depend on the method by which it does this.

Currently, SAM combines Table A3 with Tables A4 and A5 to produce up to 12 reference points for each course instead of relying solely on the five points provided in Table A3. You can check this out by looking at the scaling profile for any course in SAM.

I'm actually revising the method which SAM uses to fill in the gaps, so it should be even more accurate in 2006.

Iceman145 said:

Is there a more detailed document that relates aggregates to the final UAI?

Occasionally, some additional equivalences will be published in the full scaling report (as occurred in 2004), but generally Table A8 is the only data source. The key is to have a proper understanding of the nature of the relationship between scaled aggregates and UAIs.

Whilst strictly speaking there is no single mathematical formulae that can be used to find all the aggregate/UAI equivalences (well, there are none that can be created without having those equivalences at hand to begin with), there are a number of non-linear equations that can be used to estimate the intermediate aggregates. I determine one of these each year for SAM on the basis of the data in Table A8.

The Technical Committee on Scaling determines the UAI corresponding to each aggregate by just counting down the aggregates - the top 0.0025% get 100, the next 0.05% get 99.95, the next 0.05% get 99.90, etc (rough analogy anyway).

Iceman145 said:

Oh, and.. is Table A3 completely accurate (for the figures shown), given that raw marks are used by the UAC in calculating aggregates, whereas Table A3 shows HSC marks?

We have no way of verifying the accuracy of Table A3.

If you have faith that the Technical Committee is publishing the proper statistics, then yes, Table A3 will be accurate. Raw marks are linked both to the scaled marks (by the Technical Committee's scaling processes) and the aligned marks (by the Board's standard-setting processes). Once the scaling has occurred, equivalences can be obtained by working backwards.