[Ads-l] dark poll

[Laurence Horn]

> On Oct 26, 2016, at 10:15 PM, Ben Zimmer <bgzimmer at GMAIL.COM> wrote:
> 
> And Gallup did indeed have Romney slightly ahead in its final poll before
> Election Day.

Well, "a statistical tie"...  

> 
> https://urldefense.proofpoint.com/v2/url?u=http-3A__www.gallup.com_poll_158519_romney-2Dobama-2Dgallup-2Dfinal-2Delection-2Dsurvey.aspx&d=CwIBaQ&c=-dg2m7zWuuDZ0MUcV7Sdqw&r=wFp3X4Mu39hB2bf13gtz0ZpW1TsSxPIWYiZRsMFFaLQ&m=A9Eb61AlgzP0Kxpr6W554C6X5oscPwsB0XeNd1-mPM0&s=Xq2kHiJK8Y19eBv8EQeYMfORfiC9sIee2pCmrLYOKNk&e= 
> 
> On Wed, Oct 26, 2016 at 8:44 PM, Jim Parish <jparish at siue.edu> wrote:
> 
>> Laurence Horn wrote:
>> 
>>> I know there were some Republican-friendly polls that were apparently
>>> widely believed at the time that predicted Romney, but I don't recall
>>> whether Gallup was among them. There was the famous example (I assume
>>> actual and not apocryphal) of polls that predicted a landslide defeat of
>>> the Democratic candidate (Truman vs. Dewey? FDR vs. somebody?) that was
>>> flawed by having been conducted by telephone at a time when many voters
>>> were too poor to have one. But that wasn't a dark poll, just a stupid one.
>>> 
>> 
>> The poll you're referring to was by Liberty Magazine in 1936 - FDR vs
>> Landon. Truman vs Dewey was 1948, and Dewey led by such a margin in early
>> polling that the pollsters stopped polling in, I think, September, and
>> missed Truman's big comeback.
>> 
>> 
>> 
> 
> ------------------------------------------------------------
> The American Dialect Society - https://urldefense.proofpoint.com/v2/url?u=http-3A__www.americandialect.org&d=CwIBaQ&c=-dg2m7zWuuDZ0MUcV7Sdqw&r=wFp3X4Mu39hB2bf13gtz0ZpW1TsSxPIWYiZRsMFFaLQ&m=A9Eb61AlgzP0Kxpr6W554C6X5oscPwsB0XeNd1-mPM0&s=mLRDBupcEGvwh2SDZeo-aOCUCeqbt8XRRxD3VEuObOk&e= 

------------------------------------------------------------
The American Dialect Society - http://www.americandialect.org