PPP (D) announced some big changes to its methodology. POS (R) discusses its experimental survey by smartphone app. And Charlie Cook wishes we'd stop rolling our eyes at the lessons of electoral history. This is HuffPollster for Tuesday, January 7, 2014.
PPP MAKES SOME CHANGES: The firm Public Policy Polling (PPP) which conducts automated telephone polls for public release and, separately, for Democratic campaigns announced some changes in its methodology on Tuesday: "Starting this week all of our public polls will have 20% of their interviews conducted over the internet, through Qualtrics. This will help us to reach the increasing share of voters who don’t have landline telephones. Our testing has found that the respondents we reach in the online component are much more age representative of the population than our telephone interviews, which should reduce the amount of weighting we have to do." [PPP]
Why conduct some interviews online? - According to the most recent report from the CDC's National Center for Health Statistics, more than third of Americans (38 percent) live in households with only mobile telephone service and no landline. Federal law prohibits pollsters that use an automated, recorded voice methodology, like PPP, from calling mobile phones, so their surveys are blocked from reaching as many as a third of potential voters. PPP's change -- which is similar to the approach already used by automated pollsters Rasmussen Reports, SurveyUSA and Purple Strategies -- is to try to reach cell-phone-only households via so-called "opt-in" internet panels. These panels consist of pre-recruited respondents that have volunteered to complete surveys in exchange for reward points or other financial compensation. They are typically not recruited using random samples, but rather via web advertising. (HuffPost partner YouGov draws samples from an opt-in panel). [CDC]
How will they select online respondents? - PPP director Tom Jensen confirms to HuffPollster that their approach will involve contacting volunteers drawn from opt-in panels, "screening respondents for whether they have a landline and conducting interviews only with those who don’t."
What does "through Qualtrics" mean? - Qualtrics is a Utah based market research company that describes itself as a "supplier of enterprise data collection and analysis." Put another way, their primary product is software that lets other companies conduct surveys online, but they are not widely known as an opt-in panel provider. Jensen explains: Qualtrics will "work with multiple sample providers, depending on the project. For our stuff they’ve been using SSI." [Survey Sampling International]." [See Qualtrics, SSI]
Statewide only - Reacting to the announcement on Twitter, Logan Dobson (R) wondered if PPP could do the supplementary online interviews in local polls, since few panels have enough potential respondents to survey geographic below the state level: "If someone running for State Senator or something hires PPP, they won't be able to get 20% online interviews in a district." PPP's Jensen responded via Twitter: "Can only do statewide/national. Clients can pay for live call supplement to CD polls if they want. We do few public House polls." [@LoganDobson, @ppppolls]
PPP also drops 'random deletion' - PPP also announced a change in its weighting procedure: "With a newly available upgrade in our statistical software we’re also adjusting our weighting procedures with the New Year so that we’ll weight for gender, race, and age simultaneously. The ‘random deletion’ procedure we’ve used since our founding in 2001 has served us well, but you can always do things in a more efficient manner and we hope this change will help make for more robust crosstabs as well." [PPP]
Qualified props from a PPP critic - Nate Cohn, who wrote a series of articles for the New Republic criticizing PPP's methods with particular focus on their procedures for weighting and random deletion, had this reaction via Twitter: "We'll see how it works, but that should allow a more representative sample and reduce fluctuation in demographic composition of @ppppolls...Even so, 20 percent cell-only is less than most of the live interview polls, and it doesn't look like they're weighting for cell use." [@Nate_Cohnhere and [here]
SURVEY BY APP - Republican pollsters Glen Bolger and Trip Mullen summarize an experimental survey conducted entirely via mobile phone app in October: "In conjunction with MFour Mobile Research, we recently conducted the first ever national political survey hosted entirely on mobile phones and tablets using MFour’s Surveys on the Go application. Surveys on the Go is a mobile app that allows users on iOS, Droid, or Mobile Web devices to respond to surveys sent directly to their devices. Using this platform, Public Opinion Strategies interviewed 800 adults over the span of two days on a wide range of political topics... [W]e asked respondents to take a photo of a nearby item that most reminded them of the Republican Party, and then one that most reminded them of the Democratic Party. As one might expect, the majority of the images for each party had negative connotations—there were toilets and trash cans. But respondents in this section also demonstrated a high level of creativity. One photographed a Star Wars DVD cover. The message: Republicans represent the Dark Side. Another respondent submitted a picture of a vise and commented that Democrats are 'squeezing the middle class.'" [Campaigns and Elections]
WHY THE 'SIX YEAR ITCH' SHOULD HAVE DEMOCRATS WORRIED - Charlie Cook: "Looking ahead to the 2014 midterm elections, the pattern of second-term, midterm elections shows significant losses for the party in the White House. These losses have occurred in five out of six such elections since the end of World War II, averaging six Senate and 29 House seats…Obviously, American voters do not have the date of each second-term, midterm election circled on their calendars to kick the party in the White House. But the novelty, energy, and excitement of newly elected presidents tends to dissipate in their second terms. We normally see a scarcity of new (good) ideas, and, to put it bluntly, a level of fatigue starts to plague the relationship between a president and the electorate. Statements, decisions, and policies from the first term can come back to haunt the administration during second terms….This pattern certainly doesn’t indicate an inevitable outcome, but it certainly isn’t accidental or coincidental. It is just the manifestation of the laws—or at minimum, strong tendencies—of human nature and politics. It doesn’t always happen. It doesn’t have to happen. But it usually does." [National Journal]
HUFFPOLLSTER VIA EMAIL! - You can receive this daily update every weekday via email! Just enter your email address in the box on the upper right corner of this page, and click "sign up." That's all there is to it (and you can unsubscribe anytime).
TUESDAY'S 'OUTLIERS' - Links to the best of news at the intersection of polling, politics and political data that we missed during our holiday break (starting with a few new items from today:
-Republican support for reducing unemployment benefits rose dramatically between 2009 and 2013. [Washington Post]
-Economic confidence improved in December. [Gallup]
-A slim majority of Americans are looking forward to the midterm elections. [Pew Research]
-Andrew Kohut thinks Barack Obama could drag down Hillary Clinton in 2016. [Politico]
-Larry Sabato predicts 2014 will be a good year for the Republicans. [Politico]
-Andrew Gelman questions whether Republicans are growing more partisan about evolution. [WaPost's Monkey Cage]
-Watching As Good As it Gets and The Rainmaker might make you temporarily more liberal on health care. [Mother Jones]
-There are no time travelers on Twitter, a study finds. [ABC]