May 27, 2008

Number blues

News item in the Hindu (May 9), written by T Ramakrishnan, is headlined "More cars than two-wheelers in the State", the said State being Tamil Nadu. First sentence in the report is: "Cars have overtaken two wheelers in the State."

Now this intrigued me. There's a table with the report, listing the absolute numbers and the yearly growth rates for cars and two-wheelers in TN over the last four years. It tells me that on April 1 2004, there were 5,547,755 two-wheelers and 564,949 cars in TN: a nearly 10-to-1 ratio. How did that turn around in four years to there being more cars today than two-wheelers?

Quick answer: it didn't. On April 1 2008, according to the same table, there were 8,260,019 two-wheelers and 829,789 cars in TN: still a nearly 10-to-1 ratio.

So why the headline, why that first sentence?

Because T Ramakrishnan has confused growth rates with actual numbers. Still from the same table, over the last year, the number of two-wheelers has increased by 10.08%. The number of cars, by 11.33%. This is the first time, still going by the table, that the growth rate of cars has been higher than the growth rate of two-wheelers.

And that first time, of course, occasions a news report to announce that there are "more cars than two-wheelers" in Tamil Nadu.

Update: The article in question is here, though note that the title has since been corrected to "Growth rate of cars higher than two-wheelers". (Should I take credit for that?)

The first sentence, though, remains the same: "Cars have overtaken two wheelers in the State."


Then there's "City, Thane and Konkan will get 102% rain: Study" in the Times of India (May 27).

What does "102% rain" mean? You might think, 2% above what the area usually gets, 2% above "normal". To begin with I would puzzle about a monsoon prediction that's so precise. How was it deduced?

But there's a still more intriguing sentence in the report: "Anything between 5% and 100% is classified as normal rainfall."

That's right. If in 2008, we get a twentieth as much rain as we have done in any given monsoon season gone by, that'll be "classified as normal rainfall".


Raj said...

There's a simpler explanation. T.Ramakrishnan just went by the first 3 digits and concluded rightly that 829 was more than 826.

Rahul Siddharthan said...

It seems many journalists, and their editors, are ignorant of the most basic statistical concepts. I keep seeing headlines that say "prices falling" when what they really mean is "inflation falling". I'm sure the government is happy with the confusion. If they bring down the inflation rate, the papers are happy, though prices are still getting higher.

R. said...

do two wheelers include cycles? (which btw is soon to be the preferred mode of transport for the discerning indian, right after petrol crosses Rs. 100 a barrel)

R. said...

geez! i meant a litre

Anonymous said...

It is a useful measure. For example, if last year, we had 95% rain, and this year we get 102% rain, it does tell us how much more we had this year as compared to last. It doesn't matter what you select as 'normal'. Perhaps it is an average over several years.

Hope this helps. Koi shak?

Anonymous said...

What happens is the Met people run their simulations based on data available and come up with some number.

The number is actually meaningless because to come up with one number for a large geographical area makes no sense. Perhaps what they are estimating is a measure of the force of the monsoon this year, rather than the rain. It is a fascinating concept. Please try and think about it instead of making fun of it. Thanks.

Sidhusaaheb said...

It is a direct result of 'dynamic', 'go-getters' with 'good communication skills' being employed, without actually giving much thought to the level of competence, methinks.