File talk:General election polls 2016 Clinton v Trump.svg

Leads vs within margin of error

If a poll shows a candidate in the lead outside the margin of error, and a poll a few days later shows the same candidate leading, but within the margin of error - whether because the lead is slightly smaller or the sample size of the newer poll is just smaller - the current system seems to be to update the map to show that the difference is within the margin of error. This seems a bit oversimplified, as the polls collectively show that one candidate is very likely ahead. For example, a large sample poll showing a lead of 6 with a MOE of 2.5 shows a lead. A small-sample poll completed two days later may show the same candidate ahead by 7 but with a MOE of 4. The second poll should reinforce the lead, but would actually result in the map being changed to show the race as TCTC. I suggest that once a candidate has a lead outside the margin of error, the state is shown in their column until 1) three successive subsequent polls show the race shows the same candidate either a) leading within the margin of error or b) tied; or 2) any subsequent poll shows the other candidate ahead, whether or not it is within the margin of error. --Mmulroney (talk) 18:07, 27 June 2016 (UTC)