Pew Research MethodsVerified account

@pewmethods

Nonpartisan, non-advocacy updates on polling, data science and other research methods. Go behind the scenes with our blog, Decoded:

Washington, D.C.
pewresearch.org/methods
Joined February 2015
353 Photos and videos Photos and videos

Tweets

You blocked @pewmethods

Are you sure you want to view these Tweets? Viewing Tweets won't unblock @pewmethods

  1. Pinned Tweet
    Mar 26

    NEW: ’s Methods team is proud to announce the public release of version 1.0 of pewmethods, an R package containing functions that we use in our day-to-day survey data work. You can find the package on GitHub here:

    3 replies 61 retweets 156 likes
    Show this thread
    Show this thread
    Undo
  2. 14 hours ago

    Our new report offers a detailed portrait of the demographic composition and vote choices of the 2018 electorate, and provides an update and comparison with findings from our study of the 2016 electorate. Check out the 2016 analysis here:

    7 retweets 6 likes
    Show this thread
    Show this thread
    Undo
  3. 14 hours ago

    NEW: To understand how Americans voted in 2018 and how their turnout + vote choices differed from 2016, we surveyed U.S. adults online and verified their turnout using commercial voter files that aggregate state turnout records:

    2 replies 7 retweets 10 likes
    Show this thread
    Show this thread
    Undo
  4. Sep 1

    Pollsters can use different scales to ask respondents about their political ideology. Our new Decoded post examines how a 7-point scale with only the endpoints labeled (“extreme left” vs. “extreme right”) compares with a fully labeled, 5-point scale.

    3 likes
    Undo
  5. Sep 1

    New on Decoded, we explore two different ways of asking telephone survey respondents in Europe about their political ideology. One approach uses a 7-point scale with only the endpoints labeled; the other uses a 5-point scale that is fully labeled. More:

    11 retweets 29 likes
    Undo
  6. Retweeted
    Aug 31

    Really excited about our brand new report that culminates a yearlong study on American's views of the news media.

    2 replies 24 retweets 27 likes
    Undo
  7. Aug 24

    Using a database of more than 1.5 million Facebook posts and over 3.4 million tweets produced by members of Congress since 2015, experts studied how lawmakers use the term Latinx in posts and how that has changed. Read more:

    2 retweets 9 likes
    Undo
  8. Aug 24

    NEW: ’s Data Labs team analyzed every tweet and Facebook post from members of Congress from Jan. 1, 2015 to July 6, 2020 to study how lawmakers’ use of the term "Latinx" on social media has changed over the years. Read the analysis here:

    3 retweets 5 likes
    Undo
  9. Aug 18

    In recent years, voters who have graduated from a four-year college have been more likely to answer surveys & also more likely to support a Democrat for president – meaning high quality polls must add an educational adjustment to their results. Read more:

    2 retweets 1 like
    Undo
  10. Aug 18

    . is searching for an accomplished professional with skills in research design, computational social science, communication/public speaking and people & project management to serve as new Associate Director for our Data Labs team. Apply here:

    1 reply 11 retweets 7 likes
    Undo
  11. Aug 18

    This report focuses on education levels, but other poll features matter too (transparency, sample source, etc). A plausible education profile should be on the checklist for trustworthiness in state polls – but there are other items on the list too:

    1 like
    Show this thread
    Show this thread
    Undo
  12. Aug 18

    Based on data from the U.S. Census Bureau's Current Population Study (CPS) Voting and Registration Supplement, these results can be used by state-level pollsters to inform their weighting adjustments for polls in the 2020 election cycle.

    1 reply 1 like
    Show this thread
    Show this thread
    Undo
  13. Aug 18

    Reviews of 2016 polling found that failing to adjust for overrepresentation of college grads was among the reasons many state-level polls underestimated Trump's support. Our new report provides data on the education profile of voters in all 50 states + DC:

    2 replies 9 retweets 6 likes
    Show this thread
    Show this thread
    Undo
  14. Retweeted
    Aug 12

    Students learning about surveys are generally taught that a very large sample size is a sign of quality because it means that the results are more precise. While that principle remains true in theory, the reality of modern polling is different.

    6 retweets 6 likes
    Undo
  15. Aug 6

    How can you analyze datasets too large to be reasonably examined by humans? In a new Methods 101 video, Director of Survey Research Courtney Kennedy and Senior Data Scientist explain how it’s possible with machine learning:

    1 reply 8 retweets 9 likes
    Undo
  16. Aug 6

    And for journalists, Director of Survey Research Courtney Kennedy recently joined a expert panel to share some advice for reporting on polls. Watch it here:

    2 retweets 3 likes
    Show this thread
    Show this thread
    Undo
  17. Aug 6

    Want to go even deeper? Check out our field guide to polling that provides a broad examination of the current state of the U.S. polling industry:

    1 reply 2 likes
    Show this thread
    Show this thread
    Undo
  18. Aug 6

    Transparency in how a poll was conducted is associated with better accuracy – This means disclosing essential information like the poll’s sponsor, where & how participants were selected, the mode of interview, field dates, sample size, question wording and weighting procedures.

    1 reply 1 like
    Show this thread
    Show this thread
    Undo
  19. Aug 6

    The margin of error only addresses the fact that random samples are likely to slightly differ from the population. But there are 3 other sources of error in polling: nonresponse, coverage error and mismeasurement. (The real margin of error is often about double the one reported.)

    1 reply 2 likes
    Show this thread
    Show this thread
    Undo
  20. Aug 6

    Different polling organizations conduct their surveys in different ways – While some use traditional phone polling methods, some use online opt-in panels. Others conduct polls online using a panel of respondents recruited offline (similarly to our American Trends Panel).

    1 reply 1 like
    Show this thread
    Show this thread
    Undo
  21. Aug 6

    Are you following this year's election polls closely? In a new post on our blog, we shared several key points the public should know about polling heading into the 2020 presidential election. Here are a few highlights:

    2 replies 6 retweets 8 likes
    Show this thread
    Show this thread
    Undo

@pewmethods hasn't Tweeted yet.

Loading seems to be taking a while.

Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.

    You may also like

    ·