Kati wrote:nandakumar wrote:Survey by definition is a sample and extrapolation of its findings to the universe. But the universe itself already stands measured. But if someone insists that because the extrapolated characteristics for the universe as derived from the sample are at variance with the observed characteristics of the universe, the observed characteristics of the universe must be deemed to be false is to turn the notion of truth on its head. I mean, all statisticians agree that even the most rigorously selected sample may end up not capturing the characteristics of the population. That is why there is something called a 'sampling error'. Of course if the point is that the methodology for measuring the aggregate GDP data (universe) is somehow flawed, then there must at least be an indication of what might be wrong with it. There was none in the story.
Though mostly I'm a passive BRFite, I'm going to stick my neck out here.
I would like to know how the surveys are done, who carries out the surveys, and using what method are the samples drawn. ....
The entire exercise is dubious at the best, and treacherous at worst.
May be I'm oversimplifying the whole exercise, but being a graduate of Bharat's premier Statistical Institute, of which NSSO was a part till 1980,
I have seen how the "surveyors" stayed home, played cards, and then cooked up numbers. On top of it, Bharat's economy is 50% unorganized, and hence always stays out of any so called "official surveys".
To put it simply, desi survey data are horribly unreliable. Whatever figures we see are from some organized sectors (like - power generation, consumer goods under the tax net, automotive, cinema ticket sales, trackable export-import, etc.), a vast, vast segment of the economy stays outside the official figures.
I would like to know answers to my above questions.
I don't know much about the operations of NSSO which is responsible for conducting these surveys. There are field units located in various parts of the country. They in turn outsource the data gathering work. A small sample of the sample is revisited for verification. On paper the process is rigorous. But that doesn't mean that the results always reflect the characteristics of the population. These surveys I think serve two purposes. It throws light on policy gaps . One such example is somewhat dated. In the 70s, the Indian Council of Medical Research (ICMR) set up a vector control centre in Pondicherry. Why? Because the place reported the highest incidence of filariasis (elephant foot) among all regions in the country. Another example I can think of is this. The aggregate data might show that GDP is rising. But does it translate into more equitable distributions of such income? If not what policy initiatives that need to be undertaken to address unequal distribution? The other reason such surveys are undertaken is to help identify how the population measurement needs to be refined. My principal objection about survey results is that the media uses it to question the accuracy of measurement of aggregate data without any evidence of where the measurement is erroneous.