Polls Measure Likely Voters. What About New and Disenfranchised Voters?

[Fawke Em]

From Reddit:

I've been thinking lately about the cycle we're in with "likely" voters. I understand why pollsters would only want to include people who are likely to vote, but it could definitely skew elections.

Say you have a candidate, candidate H, who appeals to people who voted in the past. Candidate H knows the demographics "likely" to vote, and does great within those demographics.

But there's another candidate, candidate B, who appeals to groups of people who are normally left out of the political process. Candidate B does great with these disenfranchised voters. But these people aren't considered "likely voters", so they're not included in the polls.

You might meet lots of people who support candidate B, and candidate B might do well in less formal polls that don't try to weed out "unlikely" (aka disenfranchised) voters, but candidate H would lead in polls of likely voters by a strong margin. And of course candidate H would try to say that the less formal polls are skewed or gamed or somehow flawed- that people should trust the polls of "likely" voters instead.

Candidate B's supporters, seeing their preferred candidate lagging in the polls, could give up. They could switch to candidate H or drop out of the process completely, deciding that it's unlikely that a candidate they support will ever win. This would perpetuate the cycle- they don't participate in the election, so in the next one they're still considered "unlikely" voters.

I think this is one of the largest challenges we face. Although polls are helpful to measure progress with likely voters, they don't tell us much about disenfranchised voters, who are precisely the kind of people Bernie appeals to most. This is why it's so important to keep spreading the word about how to register and get involved in the primaries. We need the people who don't normally turn out to show up in force- not just for Bernie, but to turn the tide of elections in the future.

https://www.reddit.com/r/SandersForPresident/comments/3w7x5l/cbsny_times_national_poll_clinton_52_sanders_32/cxu9ahj

[RoccoR5955]

1. Not to mention the sample size

Some 360 odd Democratic "likely" voters. How could a sample size that small render anything other than a flawed result?
Some say that it is in line with other "scientific" polls, but these other polls also use small samples from a limited number of voters. Once again, the polling outfits are doctoring their polls to get the results that they want.

Remember in 2012, they all said that Romney was the likely winner.

[LonePirate]

2. Measuring irregular or disenfranchised voters is the toughest part of polling, IMO.

There simply is no rock solid way of predicting turnout of this group. I think pollsters rely heavily on past behavior and so these types of voters are not counted much because chances are they will not vote. New voters, be it the 18-21 year olds or newly naturalized citizens tend to have high first time voting percentages; but the margins are small given the percentage of voters they represent, especially the latter group.

Also, national polls for the presidential general election are meaningless unless one candidate has a large lead and we're only a week or two away from the election. The only purpose they serve right now is to give the media something to report and maybe help the candidates to gauge messaging.

[hootinholler]

3. Ok, I'm gonna be the word Nazi here...

The word should be discouraged not disenfranchised. The disenfranchised are those who can not vote even if they want to.