All About T-Score

[MOE Hater]

Grades are based on the raw score but cut off marks for grades are moderated depending on the level of difficulty of the paper

[Fongg]

atutor2001:

The above method for estimating aggregate used to be quite accurate in the past as the \"means\" of all subjects were around 70 to 75. However, it is not so anymore. With the increase in the \"means\" for math and Chinese, any student who does not perform well enough in either of these subjects will find their aggregate T-score being pulled down disproportionately. My friends' kids suffered such fate last year. A low \"A\" in any one of those 2 subjects can bring the overall T-score down by as much as 5 to 10 points, compared to a student with the same results in the past. The drop will be greater when the standard deviation is bigger.

Hi atutor,

Can I find out more about this, i did not know that now the means for chinese and maths is higher so does that mean that they are more important than english and science? (their weightage higher?)

[verykiasu2010]

In every cohort the performance is different hence the mean fluctuates


If it has been higher it is due to change in exam format and / or how the bell curve is drawn

Every subject is still of equal weightage

Don’t forget the recent hoohah on the proposed lesser weightage on MT

[MOE Hater]

verykiasu2010:

I think it is not possible to have t-score of more than 75 for any subject.


Max t-score for each subject is 75. Max aggregate is 300.

http://road-to-psle.blogspot.com/2007/11/how-psle-aggregate-and-t-scores-are.html

Of course it is possible to get more than 75

[verykiasu2010]

MOE Hater:

verykiasu2010:

I think it is not possible to have t-score of more than 75 for any subject.


Max t-score for each subject is 75. Max aggregate is 300.

http://road-to-psle.blogspot.com/2007/11/how-psle-aggregate-and-t-scores-are.html

Of course it is possible to get more than 75

Please illustrate with an example

thank you

[atutor2001]

verykiasu2010:

Please illustrate with an example


thank you

The formula for T-score = 50 + 10 x (X-Y)/Z

X = Raw score of student
Y = Average Score of the whole cohort
Z = Standard Deviation* (SD)

So if Z is very very small i.e. near zero, then (X-Y)/Z approaches infinity.

Mathematically, there is no upper limit to the T-score.

[Sun_2010]

verykiasu2010:

Please illustrate with an example

thank you

To take a practical example
Say for tough Maths paper and a smart child, the
mean = 60,
SD= 12
raw score = 96

T- Score = 50 + (96-60)10/12
= 80

[Sun_2010]

My 2 cents-

A small increase in SD has a larger impact on the T-score than in a simialr increase in mean.

Say the same example as above with mean increase by 3
mean = 63
SD= 12
raw score = 96
T- Score = 50 + (96-63)10/12 = 77.5

now with SD increase by 1 point
mean = 60
SD= 13
raw score = 96
T- Score = 50 + (96-63)10/12 = 77.69

So the impact of mean increase by 3 is affects slightly more than SD increase by 1.

[verykiasu2010]

thank you all for the wonderful examples !


what has been the range of the SD and the mean for each subject in recent years ? any one ?

[Augmum]

MOE Hater:

verykiasu2010:

I think it is not possible to have t-score of more than 75 for any subject.


Max t-score for each subject is 75. Max aggregate is 300.

http://road-to-psle.blogspot.com/2007/11/how-psle-aggregate-and-t-scores-are.html

Of course it is possible to get more than 75

regarding this issue whether is it possible to get over T-score aggregate of 300 or beyond 75 T-score for each subject,
although theoretically it may seem possible from the equation but maybe practically in PSLE , not feasible
as i had seen it on a slide presented by a Primary school (elite sch)to Parents that the max T-score for each subject is 75
and the max aggregate T-score is 300
(i tried to google it but could not find now as it was a few years ago)