Do smart people pull you down? (1 Viewer)

~ ReNcH ~

!<-- ?(°«°)? -->!
Joined
Sep 12, 2004
Messages
2,493
Location
/**North Shore**\
Gender
Male
HSC
2005
Supposing at school there is one person in a subject who generally performs substantially better then the rest of the class e.g. they get a mark of 90, while the person in 2nd gets around 70. Will that person hinder the rest of the class from also doing well? Will they actually "pull down" everyone because of the large gap between them and the rest of the class?
 

Lazarus

Retired
Joined
Jul 6, 2002
Messages
5,965
Location
CBD
Gender
Male
HSC
2001
If the relative differences between students for the exam are similar to the relative differences between students for assessment, there should be no such distortions.

However, if the difference between the top student and the second student were to substantially increase (relative to the other differences) on the exam... then the situation you describe becomes possible. The top half of the class would be pulled up with that top student, and the bottom half would be pushed down.

It is a product of using a quadratic (parabolic) transformation to moderate assessment marks.
 

~ ReNcH ~

!<-- ?(°«°)? -->!
Joined
Sep 12, 2004
Messages
2,493
Location
/**North Shore**\
Gender
Male
HSC
2005
Would such a situation occur if there was a large gap for school assessments, but if the whole class did well in the final exam, resulting in a small gap?
 

Lazarus

Retired
Joined
Jul 6, 2002
Messages
5,965
Location
CBD
Gender
Male
HSC
2001
It could, yes.

In that instance I think the whole class would be affected, and not just the bottom half.

It's hard to tell, because when the distribution is distorted that much (through a large disturbance of the relativities between students) it sometimes becomes impossible to determine a monotonic conversion curve. If the curve isn't monotonic (increasing or decreasing), and has a stationary point somewhere in the range being used, the rank order of students could be changed. The bottom moderated mark is artificially 'adjusted' to prevent this, but making that adjustment affects the rest of the distribution as well.

Makes it difficult to determine causation. But considering that's what will be done when the Board moderates the marks anyway, I guess you don't really have to worry about the true cause; just the effect.

Test it yourself - <a href="/moderate.php">/moderate.php</a>
 

Users Who Are Viewing This Thread (Users: 0, Guests: 1)

Top