Advertisements

Our Public Opinion Polls are Still as Awful as They Were in 2016

Today’s Campaign Update

(Because The Campaign Never Ends)

Why we should all stop paying attention to the polls. – In Florida’s GOP primaries, numerous polling groups conducted polls of the governor’s race over the past few months. The RealClearPolitics average of all the most recent polls pegged the race at a 6.7% advantage for Ron DeSantis, who was endorsed by President Trump.

The final result? DeSantis wins by a whopping 20% of the vote.  None of the polls taken in the race had DeSantis pulling more than 41% of the overall vote.  He actually won 56%.

In addition to all of that, barely two months ago, a poll commissioned by Fox News showed DeSantis trailing the other major candidate in the race, Adam Putnam, by 15%.  So the people at Fox are trying to convince us that this race turned around by 37 points in a little over 60 days.  Sure it did.

Stop worrying about polls. They are no better this year than they were in 2016. Most of them are conducted using awful methodologies and absurd samples for the purposes of a) creating fake news stories; b) generating clicks to the web pages of the entities sponsoring or conducting the poll; and c) attempting to influence public opinion.

The race between incumbent Senator Ted Cruz and Democrat challenger Robert Francis “Beto” O’Rourke in Texas is another great example of why many of these polls really don’t give us a shred of reliable information.  A recent poll in that race, sponsored by the Texas Lyceum, showed the contest actually in almost a dead heat, with Cruz polling 41% support to O’Rourke’s 39%.

Wow.  Shocker, right?  I mean, Texas is a state in which no Democrat has actually won a statewide election in a quarter of a century, a state won pretty easily by Donald Trump in 2016, one in which Democrat statewide candidates had trouble finishing within 20% of their GOP opponents in the last statewide elections in 2014.  So a win by O’Rourke, a man of Irish descent who is attempting to convince Texas’s Hispanic voters that he is one of them with his hip nickname, and Millenials that he is one of them by doing lots of cursing in his campaign events, would be a huge upset.

The Lyceum poll’s snapshot of a dead heat race appeared to gain some support with last week’s release of a poll from Emerson College showing the Cruz lead down to just 1%, at 39-38.  Again, wow, if only it were true.

The problems with these polls are numerous, however.  The Lyceum poll’s sample is fairly large at 1,178, but 27% of those polled aren’t registered to vote, and just 27% of respondents said they were “extremely interested” in the campaign.  All of which means this poll tells us pretty much nothing at all about what might actually happen at the polling places in Texas in November.

But it gets worse.  It turns out that the people who conducted this poll chose to pretend that Texas – a state in which no Democrat has won a statewide race since 1994 – is a dead even state in terms of party identification, choosing to question 39% Democrats, 39% Republicans and 22% Independents.  Now, based on past voting patterns, these percentages should really look more like 44% Republican, 30% Democrat and 26% Independent.  Something close to that, anyway.

All other aspects of this poll being equal, the choosing of a sample truly representative of the Texas adult population would probably have ended up showing Cruz with a lead of 8-10%.  Focusing the sample on likely voters most likely would have increased Cruz’s polling margin, given the historic disparity in party voter turnout patterns in Texas in recent elections.

The Emerson College poll is equally problematic.  While it does focus on registered voters, the sample size consists of just 550 people in a state with a wildly diverse population of 25 million or so. The number of different ethnic groups living in the city of Houston alone probably exceeds the total sample surveyed by this poll.

To make matters worse, party identification breakdown for the poll is impossible to discover.  It may be out there somewhere, but if so, it is extremely well-hidden.  One distressing result of the hilariously inaccurate polling outcomes of the 2016 elections has been a trend of these polling entities being reluctant to publish full, accurate details about the nature of their samples.

Increasingly, polling entities like Emerson do their best to avoid revealing key factors like party ID breakouts to the public.  This is the case with another recent poll in the Cruz/O’Rourke race – the NBC/Marist poll – which shows Cruz’s lead at 4%, but doesn’t really inform us how it got there.

What does that tell us?  It tells us that these polling companies, universities and sponsoring media outlets have little, if any, confidence in their polling methods in today’s rapidly changing society.  Rather than simply be honest and clearly inform the public of that reality, they hide aspects of their methodologies in order to minimize criticism.

Is the race between Cruz and O’Rourke fairly close at this point, with Cruz’s lead at least within single digits?  Yes, probably.  Irish Beto has thus far benefitted from the love affair he’s engaged in with the Texas and national news media, and the fact that most of the Texas population has little idea who he is or knowledge of his radical leftist policy views.  A fear of ending that knowledge gap is probably why Mr. Beto suddenly backed out of his scheduled debate with Sen. Cruz this Friday.  No use spilling those troublesome beans to the voters before he absolutely has to.

But is this a 1 or 2 point race, as these two highly problematic polls attempt to indicate?  Not bloody likely.  What these two polls really tell us is the same thing the disparity between the polls and the actual outcome of the race in Florida tells us:  We should all stop paying attention to these polls.

Just another day in the public opinion polls are a sad joke America.

That is all.

Follow me on Twitter at @GDBlackmon

Today’s news moves at a faster pace than ever. Whatfinger.com is my go-to source for keeping up with all the latest events in real time.

Advertisements

4 thoughts on “Our Public Opinion Polls are Still as Awful as They Were in 2016

  1. Robert Gaines - August 29, 2018

    Check out the vote totals dem vs Republican in Florida in the house districts where D and R challengers. HUGE turnout by R vs D. I saw no blue wave… Wont be reported.

    Still… Vote vote vote this Nov. #MAGA

    1. David Blackmon - August 29, 2018

      No, there will be no “Blue Wave” in November. That is purely an effort by the media to influence voters – never had been a reality and won’t be.

  2. Errant Yat - August 29, 2018

    Good analysis. But there is also another factor to consider. All the Trump supporters I know are refusing to answer any polling questions about their preferences in the upcoming election, either by phone or by mail. I think that’s how Trump sneaked up on HRC in 2016 and it’s a pretty good tactic in my opinion. But, it’s not good for the accuracy of the polls. MAGA

  3. Kay Sadeeya - August 29, 2018

    Here’s how the media had the 2016 Republican Candidates: (humor me, it could have happened this way)

    Republican Presidential Candidate Race

    ODDS CANDIDATE Note:
    1 10-1 John Kasich Goes out fast, nothing left at the end.
    2 2-1 Ted Cruz Favorite today, not sure of tomorrow
    3 4-1 Marco Rubio Ended up getting bumped in his last race
    4 7-2 Ben Carson Got everyone talking upset
    5 5-2 Jeb Bush If you’re into blood lines, suspect this one
    6 10-1 Jim Gilmore Not a lot of show in last outing
    7 12-1 Chris Christie Carries extra weight
    8 35-1 Carly Fiorina California claimer. Can run with the boys
    9 8-1 Rick Santorum Likes to get out front
    10 17-1 Rand Paul Tends to get boxed in by the field
    11 3-1 Rick Perry Looked good in training
    12 6-1 Scott Walker Breezed last time out
    12 76-1 Donald Trump Not a chance to even finish with this field

Comments are closed.

Scroll to top
%d bloggers like this: