Online methodology may affect Rasmussen polls

Rasmussen Online Polls

            Rasmussen Reports is a polling company founded by Scott Rasmussen in 2003. Headquartered in Asbury Park, New Jersey, Rasmussen operates under the tagline, “If it’s in the News, it’s in our Polls.” In addition to polling, the site produces articles about the polls and current events. The company generates revenue from ad sales and subscription fees.

Methodology

            Rasmussen uses an automated polling methodology. Polls are mostly conducted by recorded questions fed to a calling program. According to Rasmussen, this is a positive because “every respondent hears exactly the same question, from the exact same voice, asked with the exact same inflection every single time.” While automated calls are more cost-effective for a self-funded website, critics would say that people are more likely to hang up on an automated call. This could result in a lower-than-average response rate that reduces the accuracy of the sample’s representation.

Bias

            With no outside sponsors, Rasmussen Reports appear to have no major influencers that would want to sway the polls a certain way. Although their methodology is not inherently biased, headlines on the site suggest reasons to question their impartiality.

 Here are some sample headlines from the home page on October 25, 2016:

·      Voters say Clinton has more to hide than Trump

·      Most fear voter fraud, say candidates shouldn’t accept early results

·      Should taxpayers help cushion spike in Obamacare rates?

·      Voters doubt more than ever that Obamacare will reduce health care costs

·      Most still say Clinton should have been indicted

·      Voters say a candidate’s policies more important than character

Fivethirtyeight polling expert Nate Silver says that historically, the polls “badly missed the margin in many states, and also exhibited a considerable bias toward Republican candidates.”

Rasmussen missed the final election margins by 5.8 points, worse than most pollsters, and 13 polls missed it by 10 or more. This included a poll of the 2010 Senate race in Hawaii where Rasmussen missed the final margin by 40 points. This, according to Silver, is the largest error ever recorded in a general election in FiveThirtyEight’s database. FiveThirtyEight gives Rasmussen a C+ grade with the largest bias toward Republicans in their top 11 poll providers.

            Rasmussen’s methodology, history and reputation inspire some skepticism in their polls.  

by Cal Mincer

Read More

YouGov's slightly left lean

Online polling site YouGov is considered to lean more to the left politically and this could be reflected in their polls. In the past, they partnered with The Economist and CBS News for polls. For the 2016 election, YouGov and The Economist collaborated on political polls. CBS News leans more towards the left while The Economist says their polling is bias-free. Since September, trends on YouGov poll stories have mostly positive headlines for Hillary Clinton and negative headlines for Donald Trump. 

The organization

YouGov is a public company that started in the United Kingdom and has operations in the United States, Middle East and Asia-Pacific. Their polling methodology is short online polls that a group of 1000 participants are invited to take. The results are weighted to reflect the general population, then published in an article that has the results and some insights to give context.

Strengths and weaknesses

Strengths of online polls methodology include convenience, they are quick and easy to take, and people feel a greater sense of privacy. Another strength is the fact YouGov uses close-ended questions, where users choose answers from a list of options instead of writing in a blank space. One good thing is their polls always feature an “I don’t know” option after every question. A weakness is that the poll is distributed online. The response rate for online polls are not as reliable as phone or in-person polls.

by Caroline Miller

Read More

CNN/ORC and phone survey methods

When conducting a poll of any kind, methodology is a vital question when considering if a poll is accurate. Methods of polling include phone, face-to-face, computerized and written surveys.

Different methodology means different people are being reached. It also determines the kinds of questions that are being asked and subsequently the respondent answers. For example, face-to-face surveys allow for more open-ended questions, as people are more likely to give longer, richer responses if the surveyor is sitting in front of the respondent.

Surveyors weigh the benefits and downsides of factors like response rates and cost.

Response rate can be a huge downside for computerized, phone and written surveys. It is rare that surveys release their response rate of face-to-face interviews.

A recent CNN / ORC International  poll surveyed 547 registered voters by telephone.

In this case, randomness is a key issue. Phone surveys can be more random than other methodologies since lists of numbers can be selected randomly. Issues can include people responding who aren’t actually the correct person or people declining to answer calls from strangers.. Phone surveys can also be quite expensive.

By Brandon White

Read More

Washington Post/ABC News collaborate on phone polls

Importance of Methodology

The Washington Post and ABC news collaborated to conduct polls on the election. According to its archives, the newspaper  collaborates with various groups in order to reach more diverse audiences and to have more resources to devote to polling. Organizations include universities, publications and other organizations.

For this poll, they wanted to know about voter attitudes towards the candidates. Therefore, they asked questions such as how enthusiastic voters were to be voting for their chosen candidate or if the tapes of Donald J. Trump’s comments about women influenced voters attitudes.

What Methodologies Were Used?

For this poll, 1,152 adults were interviewed. A telephone sample was created using random digit dialing. including both landlines and cell phones of both registered and likely voters in the continental US. Pollsters asked to speak with the youngest adult at home. Pollsters asked questions with pre-determined responses for the respondents to choose from. To make the sample more representative of the population, people with certain demographics had their responses weighted higher than others.

In theory, telephone surveys should be pretty accurate, since they allow for the randomness that makes surveys accurate. However, they are not always representative because of the trouble getting people to answer their phones and give answers.  The weighting the pollsters apply is an attempt to compensate for that. It isn’t clear why they didn’t choose to include Hawaii and Alaska since they are part of the US and will be voting. The plan of asking for the youngest adult could also introduce a bias, as younger voters have consistently polled as more supportive of Hillary Clinton.

by Alex Giglio

Read More