I wonder if anyone has ever given consideration to waiting until the votes are counted--or enough of them to be statistically valid in key precincts.
Exit polls will always have errors because people can lie, groups can be over sampled, etc
I wonder if anyone has ever given consideration to waiting until the votes are counted--or enough of them to be statistically valid in key precincts.
Good observation... That is what the pollsters do in their final analysis and in fact don't make predictive conclusions until after the results are coming in and they can verify the information they obtained. If it appears wrong they make corrections to reflect the actual results rather trying to make the actual results fit their data collection. They talk about this in the report.
The problem is that the early data releases are used by the media etc as being actual fact and it's not. It's in error and sometimes as in this election in large error.