« Saudi Arabia is our most worthless "ally." | Main | Diazepam? »

08 August 2016


Feed You can follow this conversation by subscribing to the comment feed for this post.

Edward Amame


It was a response to Reagan's amnesty? Thanks, I didn't recall that. I DO remember that Pete Wilson was running a tough re-election campaign and he embraced it in a big way to try to save his ass.

Edward Amame

Col Lang

Will do.

Edward Amame


How the %$#@ do you know that Nate Silver is anti-gentile, anti-western and anti-American? Because he's a Jew?

Erich Newhill

Shepard, the LA Times poll - which performs the weighting I insist in necessary - has the poll results too close to call.


That is a poll I am more inclined believe.

BTW - I never said or implied you are "dangerous". Are you trying to be? I guess if someone is making important decisions based on your "thinking", you could indeed be dangerous.



Wilson embraced it as a Johnny come lately when he originally opposed it. Much like how Gov. Brewer was against HB17 in Arizona until her poll numbers tanked and then she became a national darling after supporting it with a full heart.



Or we can just stop importing them all.

This isn't a contest of where you go "OH YEAH BUT THESE CHINESE". They're ALL foreigners.



I think it’s time we looked at some data, because we're talking about different things (and my apologies, I tend to write sloppily on blogs). So far, we’ve established (and correct me if I'm wrong) that you believe polls do not ask about feelings or what someone is thinking at the time. You also believe that current pollsters are oversampling Dems. You believe this because of an additional set of polls that show a different breakdown for party affiliation.

Well, let’s look at the Gallup poll that now shows an even split between Democrats and Republicans: http://www.gallup.com/poll/15370/party-affiliation.aspx.

For what it’s worth, this is a trend line poll. Its purpose is not to establish exact numbers for affiliation at any one point in time; rather, it is intended to demonstrate over time whether the country is becoming more of one or the other (that’s why it presents its results in historical context). The data it gathers is both attitudinal (how you feel) and, as I predicted, highly volatile. Both of these things are plainly obvious from the survey page itself.

First, here’s the question it asks: “In politics, as of today, do you consider yourself a Republican, a Democrat, or an Independent?”

In other words, it is asking about how you feel about your affiliation today (not tomorrow, traditionally, or anything else). It does not ask about registration, membership, voter intention, or anything remotely approaching that. It specifically asks you to think about what you consider yourself right now.

In your thinking, political affiliation is a highly stable concept. I have stated the opposite, and this is the context in which I meant that. To see why, let’s look at the data.

The fluctuations from week to week in this survey may seem to a layman like a couple of points here and there. But you’re an actuary, so you have to recognize that they are wide swings. To use this data set to normalize other polls, you’d have to believe that it would be possible not only for Republican party membership across the entire country to change as much as 10% in a two-week period, but for this poll to accurately reflect that fluctuation. Maybe you liked the last iteration of the poll, which showed a 50/50 split. But two weeks earlier, there seemed to be substantially more Democrat-leaning voters than Republicans. Did the world really change that much in such a short period of time? And two polling periods before that, it’s the Republicans who had a slight edge.

Although it may not seem logical, if you ask about current attitudes towards a party, this is what you typically get. This survey should be looked at over time, with deep suspicions as to accuracy of any one data point. It cannot be used in the moment to support a claim of fraud.

Eric Newhill

shepard, too many confounds and too many opportunities for lousy sampling technique or just plain old random error. I would believe you if they called the same people at different times and then got variation in party affiliation. Mrs. Kowalski polled in January say she's *feels* like a democrat and favors Clinton. Asked in March, she feels like a republican and favors Trump....in June a democrat again, etc.

You are putting way too much faith, IMO, in these pollsters having good technique and their samples being both truly random and truly representative. You are asserting if the polls end up with a 36%/26% D to R ratio, then that is reflecting the mood (attitude) of the country. Prove it! I am saying it reflects bad sampling; either on purpose or due to sloppiness and a desire to get the poll out to the media and get paid. I do not believe that party affiliation is so subject to how you are feeling.

Anecdotally, I identify as Republican. I do not think Trump is a great candidate. I will vote for him, though. I did not like McCain at all and I voted for Obama in that election. Yet I identified as a republican voting for Obama. I did not identify as a democrat. So that goes against your theory (and a theory is all it is).

I think you are dancing past the fact that most of these polls are targeting registered voters and asking what they are registered as. That question eliminates your attitudinal theory; unless you want to tell me that they're running down to the county building and changing their registration every couple weeks or so.

The comments to this entry are closed.

My Photo

February 2021

Sun Mon Tue Wed Thu Fri Sat
  1 2 3 4 5 6
7 8 9 10 11 12 13
14 15 16 17 18 19 20
21 22 23 24 25 26 27
Blog powered by Typepad