Online polls vs online 'polls'

[dsc]

I think polls right now should only be studied for trends, if they are going to be studied or discussed at all, but if we are going to discuss them then we should at least be on the same page. There are legitimate online polls. They are conducted similarly to phone polls, you get an email asking you to participate, you click on the link with in, you answer questions. The link can only be used once. This is just a different method of doing what a phone pollster does. It is inherently no more, nor any less, scientific than a phone poll. If the sample is correctly designed it should be a proper proxy for the population within a given MOE. Thus we can use said poll to infer what the population thinks at a given point in time and given two or more polls by the same outfit on different dates we can make inferences about the trends of support for candidates in the population as a whole.

Now the kind of online 'poll' we get asked to DU is quite another matter. It is the epitome of a voluntary response scenario where people can answer multiple times and create chaos. It is a voodoo poll. No effort is made to match sample with population so literally no inference about the population can be drawn. No inference about the view of the population at a given point in time, nor any inference about trends in the population no matter how many such 'polls' we may have.

In short, simple being an online poll says nothing about a poll's scientific rigor or lack thereof.