Bad Polling Takes(@BadSurveyTakes) 's Twitter Profileg
Bad Polling Takes

@BadSurveyTakes

Methodology matters. Correlation is not causation. And "push polling" is not polling; it is headline click-bait.

ID:1272544608963252225

calendar_today15-06-2020 15:01:20

57 Tweets

2 Followers

33 Following

Bad Polling Takes(@BadSurveyTakes) 's Twitter Profile Photo

Breitbart News Never trust a survey that does not provide copies of all of the questions asked. Slide of 'results' do not the story.

Most organizations disclose the methodology used in accordance with the AAPOR Transparency Initiative

drive.google.com/file/d/1qUONKY…

account_circle
Bad Polling Takes(@BadSurveyTakes) 's Twitter Profile Photo

Sue Whitaker Not that I disagree with the sentiment, but it was a push poll. See Q3.1 ('Animals suffer and feel pain if they are slaughtered when conscious: To what extent, if at all, do you agree or disagree with the following statements?').

Fair, objective polling can be more persuasive.

account_circle
Bad Polling Takes(@BadSurveyTakes) 's Twitter Profile Photo

Thomas Paine A 20+ Q tracking poll is not informative on topics raised toward the end bc/ of fatigue. Also, the Q stem presumes there will be a vote ('When the Senate votes ...'), so respondents who think there SHOULD'NT be a vote are forced to make a choice.

assets.morningconsult.com/wp-uploads/202…

account_circle
Bad Polling Takes(@BadSurveyTakes) 's Twitter Profile Photo

Never trust a 'poll' that does not disclose the actual questionnaire used. What questions are asked before a given question can profoundly bias the results.

(The 'Just the News' site only discloses one question)

Oh, and btw, the Constitution [Art. I, Sec. 2] say otherwise.

account_circle
Bad Polling Takes(@BadSurveyTakes) 's Twitter Profile Photo

That is NOT what the survey shows

Only polled members. Thus, results not projectible to all 'U.S. Companies in HK.' Also, there is non-response bias

They even say 'This survey is not intended to be a scientific instrument.' It was an informa poll

amcham.org.hk/uploads/media/…

account_circle
Bad Polling Takes(@BadSurveyTakes) 's Twitter Profile Photo

Studies show that polling for the 2016 Election was quite accurate and consistent with historical results. Comey's statement on Oct. 28 has been shown to have had a significant impact not captured in the polling.

In other words, polls are accurate.

fivethirtyeight.com/features/the-p…

account_circle
Bad Polling Takes(@BadSurveyTakes) 's Twitter Profile Photo

The old WaPo question was poorly written. A respondent answering 'yes' to 'Doesn't it bother you?' does not necessarily mean 'it does not bother' the person.

In a recent study, roughly half of Native Americans considered the name problematic.

washingtonian.com/2020/02/21/a-n…

account_circle
Bad Polling Takes(@BadSurveyTakes) 's Twitter Profile Photo

High-quality polls (like those conducted by most news organizations) scientifically sample a defined population and report information about / opinions of the group of interest within a defined margin of error

Anecdotal evidence (attendance, etc.) is not science

account_circle
Bad Polling Takes(@BadSurveyTakes) 's Twitter Profile Photo

High-quality polls scientifically sample a population and report their opinions within a defined margin of error.

Further, most organizations disclose their methodology in accordance with the AAPOR Transparency Initiative

Anecdotal evidence (e.g., 'enthusiasm') is not science

account_circle
Bad Polling Takes(@BadSurveyTakes) 's Twitter Profile Photo

High-quality polls scientifically sample a population and report their opinions within a defined margin of error

Properly surveying ~1,000 people accurately reflects the views of the general pop. (+/-3%)

The chance of any one person being sampled is only about 1 out of 200,000

account_circle
Bad Polling Takes(@BadSurveyTakes) 's Twitter Profile Photo

High-quality polls scientifically sample a defined population and report their opinions within a margin of error.

Further, most organizations disclose their methodology in accordance with the AAPOR Transparency Initiative.

So, yes, they can be trusted.

account_circle
Bad Polling Takes(@BadSurveyTakes) 's Twitter Profile Photo

@NorthernLefty High-quality polls (like those conducted by most news organizations) scientifically sample a defined population and report the opinions of the group within a defined margin of error

Informal 'polls' are neither predictive nor informative b/c they are highly prone to bias

account_circle
Bad Polling Takes(@BadSurveyTakes) 's Twitter Profile Photo

Ed Rollins Donald J. Trump High-quality polls (like those conducted by most news organizations) scientifically sample a defined population and report information about / opinions of the group of interest within a defined margin of error.

Anecdotal evidence (e.g., donations, tickets) is not science.

account_circle
Bad Polling Takes(@BadSurveyTakes) 's Twitter Profile Photo

Justin Hampton Source?

It is unlikely this was a scientific study that could be projected across the supposed population of interest (here., 'professionals'). And if not, your statement is misleading.

account_circle
Bad Polling Takes(@BadSurveyTakes) 's Twitter Profile Photo

Bill Mitchell Polls are not 'models.' High-quality polls (like those conducted by most news organizations) scientifically sample a defined population and report information about / opinion of the group within a defined margin of error.

Rallies, yard signs, etc. are merely anecdotal.

account_circle
Bad Polling Takes(@BadSurveyTakes) 's Twitter Profile Photo

@RonnyJackson4TX Rep. Ilhan Omar Alexandria Ocasio-Cortez In 2009, during President Obama's first term, 78% of Republicans (responding to the same survey) said they were 'extremely proud.' Today, under Trump, it is only 67%.

(I'll let others distinguish 'proud' vs. 'love/hate'. You obviously can be disappointed yet still patriotic).

account_circle
Bad Polling Takes(@BadSurveyTakes) 's Twitter Profile Photo

Jordan Sather High-quality polls (like those conducted by most news organizations) scientifically sample a defined population and report information about / opinions of the group of interest within a defined margin of error.

Anecdotal evidence (e.g., lawn signs, tickets) is not science.

account_circle
Bad Polling Takes(@BadSurveyTakes) 's Twitter Profile Photo

Katie Hopkins A Twitter poll is not representative of any population (not even users of Twitter), no matter how many people respond.

To say it reflects the opinion 'of Brits' is misleading.

Also, the question was very poorly written. Many likely interpreted it as 'free' to the student.

account_circle