Posted:
Statistically, it has been demonstrated that once one passes a certain critical mass in a random sample size, the trend that then follows doesn't deviate much from that. If the numbers are taken from all over India, then the end result is more likely than not accurate. Same thing when you see opinion polls - when the sample size is above a certain number, and well distributed over variations of the population, properly representing the various segments, the final results is unlikely to deviate from the samples chosen.
Of course, there is the possibility that the distributions picked are not representative of the larger populations, which is when you're more likely to see variations.
