it is purely laughable that someone claiming to have done a PhD thinks that 0.01% of a population can miraculously cover the whole 1.25 billion population....if that was the case every scientific mistake would have been accepted instead of being discredited as statistically insignificant!!
Seriously make use of your education dont behave as a troll!
Ok I see you do speak with the a self-professed conviction of some religious fundamentalist....just like people have said about you in this thread already.
No matter how much Science or rationale reasoning we give you, you will always resort to the bare line of faith trumping it all. The "nyaah nyaaaah you are wrong no matter what!"
Again I direct you to read at the very least:
Sampling Distributions
and then bring up specific points if there is something you don't understand.
But it seems you are too lazy, unwilling or just outright unintelligent to do so before opening your mouth and continuing your braying.
Any reasonably intelligent person reading this thread will easily see who is being the troll here.
Seriously some of you have foreign flags and know not even the basics....or pretend to be smarter than people who actually have a degree in such a field
LOL this gets even better. You are claiming you have a degree in this field? You are the one avoiding the basics at very elementary statistical explanations like the one provided earlier. Central limit theorem, normal distribution, confidence intervals and limits, standard error, bias, estimators....it is quite clear these are all very alien concepts since you have not once brought any of them up yourself, or established any debate on them....since it is clear you know nothing about them.
Instead you completely skirt the very fundamental definition of what a normal distribution is, what happens to all distributions at sample size>30 and the very basis of the central limit theorem....and instead jump right to sample ratio % and probability which have next to nothing to do with normal distributions in the context of sample sizes and confidence levels in the context we are talking about.
The very basic reason is similar to flipping a coin. Say you flip a coin 10 times and write down the number of heads.
Then you do this process 100 times ( a total of 1000 flips) and each time you write down the number of heads. So your sample size is 100. You will find this is normally distributed.
Now if you do this process 1000 times(total of 10,000 flips), your normal distribution does not become 10 times more statistically significant or better...since we have already established the randomness of an essentially binomial coin flip. It will get a somewhat better precision and confidence interval resolution, thats about it. In fact I can show you the exact calculation to prove this, but you first have to respond with some modicum of understanding and intelligence on this subject, which you have thus far failed to do.
Now lets say someone secretly weights the coin to an unknown degree and you dont know this weight (basically the coin is no longer 50/50 fair but you dont know the exact level and need to flip to test it)....dont you think it would be silly to flip it millions of times when you will probably see the result much earlier on? You will see it both in the sheer binomial distribution and normal distribution too (if you decide to structure it that way like described earlier).....because essentially it is still a perfectly random process (coin flipping) for all intents and purposes....even if the coin itself is weighted.
It is only that randomness that can thus come into question when we are talking about the relevance of a sample to its larger population as long as the sample itself is more than 30 (a rule of thumb since this is when many various distributions converge in various parameters - and their sample distributions show normal distribution)
If you know of some Pakistani member on this forum that actually has a background in statistics, please ask him/her to continue this debate on your behalf....they may actually bring some points that matter from a fundamentally solid position w.r.t statistical analysis. It is pointless to have you humiliated further. I will not waste any more time trying to explain basic things to you since you are quite obviously not from this background and are making no attempt to actually learn some basics to debate constructively. Anyone else that reads this is free to engage with me on whatever point of interest.
Yeah obviously! It is quite clear now. She definitely would not engage in this level of nonsense in a field she has no idea or inkling about.
and yes that would be interesting but where was the survey taken? Online...
Good now we are getting somewhere (just a little). You finally dont attack it on the basis of its sample size (which made you look really dumb...and dumber when you stuck to it - though it seems you still havent abandoned that line of reasoning).
Yes having the survey online will present a bias, since we cannot assume it is a perfect random sample anymore. The extent will of course need further analysis and study, but who is going to really care about doing that? The topic itself is quite an insignificant one to conduct such a time consuming and empirical analysis for.
Quick response online surveys are the best bet for these kinds of questions....those who want to believe in the results will do so....those who wont....will not. Increasing the randomness (and therefore data acquisition cost) is simply not a good return on investment for this case.
How many people of india have access to internet and spend time doing online survey?
You will have to do a representative study on that topic alone. For now it can only be labelled as a source of bias. We have no correction coefficient data to employ to account for it.
How many answer it properly and not to troll?
Its a yes or no answer. How does one troll a binomial selection?
How many say what they really feel or arent ashamed to feel like a black sheep?
Their identities are made known to everyone after they poll? Or this is an anonymous poll? Do you have any idea how dumb you sound right now?
Lastly, who processed the survey? were they politically motivated?
So even if the sample size was to your liking (lets say hundred of millions), you will still pull out this desperate last resort if the result is not to your liking. Glad no one cared enough to pre-emptively assuage a minority that cannot accept it is the minority....and their concerned brethren across the border
....because they will just say "it was politically motivated" and "controlled" in the end....no matter how much is done to try satisfy their "concerns".
