Another off-beat post and the reason I bring this post up is to show that a small sample size is sufficient for most needs. You might have heard during elections that there are tons of opinion polls about the election result. Also, notice that every poll has small print at the bottom giving the margin of error. Usually it is between 3-5%. So how do they arrive at this number?
It is simply the power of sampling, and by just taking a survey of around 1,000-2,000 people, a margin of error of ~3% can be reached. Without kicking around the bush, the general way to calculate margin of error in sampling is by this formula:
The above is NOT an exact formula for margin of error but an approximate one. In the above formula, n is the sample size and if we plug in 1000 for n, we get around 3.16% . Basically, I’m trying to explain that as n being just 1,000 , would not give a large margin of error.
Of course, opinion polls cannot always predict the outcome of an election because of several factors (even if you account for margin of error):
- Selection bias: Usually folks being surveyed may be of the same demographics and hence not representative of the whole population.
- Lying on the survey: People lie on the polls.
- Do not vote: Some do not vote on election day.
- Change mind: Some will change their mind on election day.
- Method of sampling: There are many ways to sample, and each method has its own set of pros and cons. You can check the Wikipedia page for more details.
So, how is this related to domains or .in domains in general? Well, it all boils down to the number of reported .IN sales over the years, as you are aware, fewer sales are made public and thus difficult to gauge the trend of value of the domain. If you look at the sales page I maintain here; there are about 1,000 sales in there. This might seem small at first compared to .COM and others ccTLDs, but these are good enough sample size to gauge a pattern/approximate value.