## A lesson in citing irrelevant statistics

I refer to the articles “School system ‘still best way to move up’”, “MPs speak for kids from poorer families” and “New chapter in the Singapore story” (ST, Mar 8).

These articles were page after page (four pages) of statistics cited by the Education Minister to more or less dismiss MPs’ concerns that kids from poorer families were disadvantaged. I do not think I have ever seen so many statistics given to support a position in a Parliamentary debate!

Unfortunately, I could not find a single statistic which in my view, is “statistically” relevant to the debate.

Irrelevant Statistics

Statistics that are quoted, by themselves, may be quite meaningless, unless they are on a comparative basis.

To illustrate this, if we want to say that Group A (poorer kids) is not significantly worse off than Group B (richer kids), then it may be pointless to just cite the statistics for Group A, without Group B’s.

Let me now get into the specifics of the statistics cited:

“How children from the bottom one-third by socio-economic background fare: One in two scores in the top two-thirds at PSLE”

“One in six scores in the top one-third at PSLE”

What we need to know for comparative purposes, is the percentage of richer kids who scores in the top two-thirds too.

“How children from 1 – to 3-room HDB flats fare: One in five scores in the top one-third at PSLE”

We need this data for different time periods, as the proportion of those living in such flats had changed over the years.  What we need to know is has this proportion who score well, changed in the last 5, 10, 20, 30 years, etc.

“… one in five scores in the top 30% at O and A levels… One in five goes to university and polys”

What’s the data for richer kids?

Since the proportion of the entire population going to university and polys has increased substantially, this clearly shows that poorer kids are worse off!

“These figures have remained constant even though the number living in 1 to 3-room HDB flats has fallen sharply over the years”

This statement may be “statistically” irrelevant, as all it may indicate is that the lower-income’s chances of performing better, on a relative basis, has remained stagnant.

“Top PSLE pupils- The top 5% come from 95% of schools… Every primary school has at least 10 pupils in the top third of the cohort”

This may be “statistically” of no relevance to the debate, as logically every primary school is made up of both poorer and richer kids.

Citing individual examples?

According to the articles:

“Education Minister Ng Eng Hen calls Hong Siang Huat “a living example of social mobility”.  He came from a poor family but is off to Britain on a government scholarship.”

The Minister was quoted as saying:

“My  parents had six children.  My first home as a young boy was a rental flat in Zion Road.  We shared it as tenants with other families”

Citing individuals who made it, may be of no “statistical” relevance, as what we need are the statistics as to the proportion of poorer kids to richer kids, who get scholarships, proportional to their representation in the population.

“More spent on primary and secondary/JC schools.  This means having significantly more and better teachers, and having more programmes to meet children’s specific needs”

What has spending more money, which what most countries do, got to do with the argument whether poorer kids are disadvantaged?

I think Straits Times journalist, Li XueYing put the crux of the debate in the right perspective:

“Dr Ng had noted that ensuring social mobility “cannot mean equal outcomes, because students are inherently different”, But can it be that those from low-income families are consistently “inherently different” to such an extent?”

Relevant statistics

Perhaps the most damning statistics that poorer kids are disadvantaged was the chart from the Ministry of Education (provided by the Straits Times), which showed that the percentage of Primary 1 pupils who lived in 1 to 3-room HDB flats and subsequently progressed to University and/or Polytechnic, has been declining since around 1986.

The statistics cited by the Minister Mentor, that in top schools like Raffles Institution, more than half of the students had fathers who were university graduates, in neighbourhood schools the figure hovered around 10 per cent, etc, was perhaps clearer statistical evidence, that the odds may be stacked against poorer kids.

As to: “… now ITE students in the bottom 15th percentile income bracket (per capita household income of up to \$300) will receive \$1,000 a year, up from \$800 a year”, how significant is this extra help of about 55 cents to a total of \$2.74  a day, for a student whose family is clearly struggling on less than \$300 per person per month?

In summary, if not for the Straits Times’ reference to the MOE tertiary and Minister Mentor’s data, the entire debate may arguably be a good lesson on statistics for  Parliamentarians, on how to try to win a debate with entirely “statistically” irrelevant statistics!

1. leonggsh

Test