T-Score? Is that for Golf?
Back in the days when I was doing my PSLE, we simply took the total of our 4 subjects and add them together, for a total maximum part of 420 (English 110, Chinese 110, Mathematics 100 and Science 100). I scored 389 back in those days, putting me squaring among the top 10% of the national cohort.
In Singapore, where the PSLE is used to rank a student for his choice of secondary schools, such a simple method soon had its issues, especially if the cohort is very strong or very weak, or the paper very simple or very hard. When students cluster around certain grades, it makes it difficult for the school to sieve them out (which, again, is the real reason for the PSLE).
In order to avoid the clustering, and allow the entire cohort to resemble more of a bell curve, a transformed score is used instead of the raw score for each subject. Explaining exactly what the transformed score is would belong to a statistics lecture (which I am not qualified to give, even though I do a lot of data analyses for my MBA!). It is good enough to just say that a T-score comes in this format.
T-Score = P+[10(R - M)]/SD
Where:
P = Passing Grade of the paper (50, in the case of the PSLE)
R = Raw Score
M = Mean Score (Average) of Cohort
SD = Standard Deviation of Cohort
All 4 T-scores of each grade are added up for the final Aggregate score in the PSLE.
Using the T-Score means that the makeup of the cohort (and as such the difficulty of the paper) plays a big part in determining the child's final score.
Paul scored 90 in his paper in a cohort with a mean of 80 (easy paper!) and a SD of 10 (meaning about 16% of students scored 90-100 and another 16% scored 80-90). His T-Score is 60. Paul decided to retake his PSLE the next year (staying back a year to do so) and scored 90 again in a cohort with a mean of 60 (tough paper, weak cohort) and a SD of 20 (meaning about 16% of students scored 60-80 and another 16% scored 40-60, huge spread). His T-Score this time is a remarkable 105!
The example given above should also deal with two misconceptions many Singaporeans have about the PSLE T-Score.
Hopefully, this article can clear the air regarding that important national examination that all Singaporeans (okay, nearly all - my firstborn did the SPERS instead)!
In Singapore, where the PSLE is used to rank a student for his choice of secondary schools, such a simple method soon had its issues, especially if the cohort is very strong or very weak, or the paper very simple or very hard. When students cluster around certain grades, it makes it difficult for the school to sieve them out (which, again, is the real reason for the PSLE).
In order to avoid the clustering, and allow the entire cohort to resemble more of a bell curve, a transformed score is used instead of the raw score for each subject. Explaining exactly what the transformed score is would belong to a statistics lecture (which I am not qualified to give, even though I do a lot of data analyses for my MBA!). It is good enough to just say that a T-score comes in this format.
T-Score = P+[10(R - M)]/SD
Where:
P = Passing Grade of the paper (50, in the case of the PSLE)
R = Raw Score
M = Mean Score (Average) of Cohort
SD = Standard Deviation of Cohort
All 4 T-scores of each grade are added up for the final Aggregate score in the PSLE.
Using the T-Score means that the makeup of the cohort (and as such the difficulty of the paper) plays a big part in determining the child's final score.
Paul scored 90 in his paper in a cohort with a mean of 80 (easy paper!) and a SD of 10 (meaning about 16% of students scored 90-100 and another 16% scored 80-90). His T-Score is 60. Paul decided to retake his PSLE the next year (staying back a year to do so) and scored 90 again in a cohort with a mean of 60 (tough paper, weak cohort) and a SD of 20 (meaning about 16% of students scored 60-80 and another 16% scored 40-60, huge spread). His T-Score this time is a remarkable 105!
The example given above should also deal with two misconceptions many Singaporeans have about the PSLE T-Score.
- Does it mean that the elder brother who scored 220 marks three years ago is not as good as his younger brother who scores 270 marks now? This is simply not true. The T-Score is so much affected by the cohort that it is not meaningful to compare scores across the years. Unfortunately, this is still being done every year, even by government officials.
- Since the PSLE scores in Singapore are always reaching towards 300 (the highest ever recorded was 294, in 2007), does it mean that 300 is the top score? This is again simply not true. As my readers can tell from the above example, Paul would have scored 420, gaining national fame and at least getting the bang for his staying a year to compete with a weaker cohort!
Hopefully, this article can clear the air regarding that important national examination that all Singaporeans (okay, nearly all - my firstborn did the SPERS instead)!
Comments
Post a Comment