Home > Uncategorized > Let’s Detox From Polling

Let’s Detox From Polling

November 4, 2020

I cannot believe I fell, once again, for the polling that gave me the information I wanted to hear. It is indeed an emotional addiction, rather than a scientific curiosity, and I think we’d all be better off shedding our addiction to political polling. My latest Bloomberg Opinion column:

Polling Failed. It’s Time to Kick the Addiction

Doubling down won’t help Americans understand themselves.

For more of my Bloomberg columns, go here.

Categories: Uncategorized
  1. Lloyd Lofthouse
    November 4, 2020 at 6:09 pm

    Explain how polling failed. Poling did NOT fail except for Americans with a bad case of instant gratification wanting a clear winner by midnight on November 3rd.

    But this is not our average presidential election, and unless Trump pulls off a miracle in court to have mail-in ballots tossed out by the millions, he will lose lost the election.

    As I was writing this comment, it was announced that Biden took MI adding 16 Electoral Votes to his total for 264. He is also going to win NV and PA once all the mail-in ballots are counted.

    Once the mail-in ballots are counted in NV, and Biden is ahead in that state, he will have 270 Electoral College Votes. How do I know this fact? I fact-checked it and NV reports that Democrats turned in twice as my mail-in ballots as Trumplicans.

    IN PA, Democrats turned in more than three times as many mail-in ballots as Republicans, and by Friday at the earliest, those mail-in ballots will be counted and Biden will take that state.

    Then his Electoral College Vote will be 290.

    The same thing may happen in NC and GA once all the mail-in ballots is counted. When all the ballots are counted, it is possible that Biden will have a total of 321 Electoral College Votes, but he doesn’t need that many. He will win once Biden is announced the winner of NV.

    Meanwhile. Trump campaign minions, under orders from their mafia godfather, are filing lawsuits one after another in NV, PA, MI, NC, and GA in a desperate attempt to have those mail-ballots tossed out, and so far judges (even some appointed by Trump) keep tossing his cases out or ruling against him.

    Then there is the popular vote. Once all of the mail-in ballots are counted, the 2020 election will have had the biggest turnout in U.S. history, and Biden will also win the popular vote by a wider margin than Hillary Clinton did in 2016. Right now as I write this comment, Biden is already ahead of Trump by 3 million votes and there are a lot of mail-in ballots yet to count.

    Like

  2. M. Lewis
    November 8, 2020 at 5:58 pm

    I think the media should talk more about differential non-response. This is when folks who see their candidate down in the polls are less likely to respond to them. If this happens, it can make one candidate appear to be much further ahead than they really are. Weighting data can address this, at least in theory. In practice, it isn’t always so easy to do.

    Like

  3. Renee Wurth
    November 11, 2020 at 1:56 pm

    I see the failure stemming from hubris of quantitative research (I say this as a quantitative researcher). It took many years of building statistical models in the field of health inequities to realize that in order to understand what a population is thinking you NEED to talk to them. Big Data will more often misdirect than get to truth — especially when its done without proper qualitative grounding.

    Thanks, Cathy — for reminding us of this hubris we cannot quit.

    Like

  4. rockethire
    November 20, 2020 at 5:38 pm

    Polls = garbage in garbage out.

    Like

  5. Mark Schaeffer aka Albanius
    November 26, 2020 at 5:33 am

    Per the last data I checked yesterday, Biden led Trump by over 6.14 million votes nationally, or 4.0% of the votes cast between them, with more mail votes and affidavit votes, heavily Democratic, to be counted. 538’s final weighted poll average had Biden leading by 8.4% for all polls, regardless of quality. 538 did not show a standard deviation of their poll average, but I have the impression it is in the vicinity of +/- 3%.
    Nate Silver wrote on November 11 “a high percentage of states (likely 48 out of 50) were “called” correctly…that’s usually how polls are judged: Did they identify the right winner?
    …And yet, the margins by which the polls missed — underestimating President Trump by what will likely end up being 3 to 4 percentage points in national and swing state polls — is actually pretty normal by historical standards.”
    https://fivethirtyeight.com/features/the-polls-werent-great-but-thats-pretty-normal/

    Like

  6. Mark Schaeffer aka Albanius
    November 26, 2020 at 5:36 am

    BTW Cathy, I heard you on BBC radio this morning discussion implicit racial bias of algorithms, brilliantly. Please post a link to that on this blog.

    Like

  1. No trackbacks yet.
Comments are closed.
%d bloggers like this: