Aligning vs moderating (1 Viewer)

[Lazarus]

Laz, isn't any capping determined year by year? If so do you know how they determine if a subject is to be capped and how much to cap it?

Yes.

A 'modifier' is calculated on the basis that the maximum possible mark in the combined 2 unit English course (which constitutes almost the entire HSC candidature) should be 50. I haven't had time to work out the calculations behind this yet.<!--:

max_E - the initial maximum scaled mark for the combined course
mean_E - the scaled mean for the combined course
stdev_E - the initial scaled standard deviation for the combined course

<blockquote>
(max_E - mean_E) / stdev_E * (stdev_E + modifier) + mean_E = 50

modifier = stdev_E(50 - max_E) / (max_E - mean_E)</blockquote>-->

In both 2002 and 2003, the modifier was determined to be 2.52.

The cap (or maximum scaled mark) for all other courses is then determined by:

max - the maximum scaled mark
mean - the scaled mean
modifier - the constant calculated above
stdev - the initial scaled standard deviation

<blockquote>
max = mean + (modifier * stdev)
</blockquote>
If the maximum mark produced by this formula is greater than 50, the maximum mark is set to be equal to 50.

What happens if a bright student enjoys subjects that are capped?
...
In other words, a great artist and a great physicist are being given different marks despite both their apparent abilities.

In most cases, the effect of imposing a maximum scaled mark seems to be negligible.

If there are enough great artists doing the course, the scaled mean and standard deviation should shift enough to allow the maximum scaled mark to rise. One student is generally not enough.

Beyond that, there isn't much I can say... write a letter to the Technical Committee on Scaling.

 

[~ ReNcH ~]

When the cap is something like 45, then it can have much more than a negligible effect though. But I guess there isn't much I can do. I doubt that one 16 year-old kid can influence the Technical Committee to change the whole process.

BTW. What exactly is the "standard deviation"? What exactly does it "show"?

 

[Lazarus]

They might not modify the process, but they would at least justify it for you.

The standard deviation of a distribution is the average distance of each score from the mean. For example, a standard deviation of 6.0 means that, on average, each score in the distribution was six points away from the mean.

It provides a measure of the 'spread' of the distribution. A large standard deviation indicates that the distribution stretches over a wide range of points. A very small standard deviation might indicate that scores were clumped around the mean.

Just google any of the statistical terms if you're not sure - there's plenty of info out there.

 

[~ ReNcH ~]

Hehe.
That's cool. Thx.

BTW. In the old HSC, would the standard deviation have been keep constant from year to year to maintain the "bell curve"? Or could it still vary by a lot?

 

[Lazarus]

The standard deviations of the distributions of raw marks would have fluctuated, just as the raw standard deviations would be doing so now (even though we don't know them).

But for the Board's reported marks - the old equivalent of our aligned marks - yes, the marks for 2U courses were transformed so that each distribution had a fixed mean and standard deviation. This meant that the same proportion of students would be scoring marks of 90+ in all 2U courses, etc.

That all changed with the new HSC.