Assessing State Polls
*June 29 Update* Hillary Clinton has narrowly pulled ahead in North Carolina. Her leads in other states have slightly expanded in the past week, largely following the trend in national polls. The gender gap is currently favoring Clinton — though Donald Trump tends to do well with men, Clinton does even better with females. Trump is still struggling to consolidate Republican support. He’s polling in the high 70s to low 80s with Republicans throughout the states, bleeding some support to Gary Johnson (LP) and Clinton. To win, he’ll need to earn their support.
The PoliticalEdu state polls model uses polling data to analyze individual states in the 2016 presidential race. It accompanies the structural model and is combined with it in the aggregated 2016 electoral map model.
This model is developed through a simple process: Take the cross-tabs of each state poll and look at response by race, gender, and party identification. Those results are multiplied by inferred electoral composition of each group (determined by a linear extension of the trends displayed in 2004, 2008, and 2012). Demographic breakdown (race, gender, and party ID) is averaged and then multiplied by pollster rating (numeric values assigned based on the 538 assessment of polling outlets) and 1 divided by the days until the election from the poll’s end (this means that recent polls are weighted more than older polls). Results are then multiplied so the numbers are sensible (ie, so that when added together, the numbers are equal to the sum of poll values in the RealClearPolitics average).
Naturally, this model only applies to states that have been polled. Many have not, leading to a number of grey “undecided” states. Those will hopefully be filled in as the election approaches and more states are polled.
The model’s results, shown below, are favorable to Hillary Clinton. Thus far, she is faring well in state polls; however, it is still early and much can change between now and November.
Click the map to create your own at 270toWin.com
Poll-based states’ predicted results:
(You’ll notice these numbers do not add up to 100% — the polls released have options for “don’t know/other” and “wouldn’t vote,” thus preventing the candidates from adding to 1. The number of “don’t know” respondents should decrease as election day approaches.)
This post will be updated as more polls are released!