FIELD BLOG SUBSCRIBE TO RSS
New Platforms, New Public Opinion?Posted by: admin on Thu, 2012-05-24 16:24
By Tim Cooper, Research Director, InterMedia
Gayatri Murthy, Senior Research Associate, InterMedia
With the continued growth of new communication media and technologies, the public opinion and research sector is abuzz with equal doses of optimism and skepticism for its future. In a world of falling response rates and increasing costs for phone and face-to-face surveys, does this new frontier ask us to merely measure the chatter on Twitter and Facebook or does it reframe the definition of public opinion itself? This is among the many questions challenging the Digital Team here at InterMedia.
Let’s start with a confession. In 2006, the editor of Market Research magazine in the U.K., invited comment on the lasting power of Facebook. “It’s a fad, it will last about two years” was my sage response. And here we are six years on: Facebook is valued at close to $100 billion, and connects more than 901 million users (as of May 2012) and traverses the globe. Incredible!
We attended the American Association for Public Opinion Research (AAPOR) conference for the first time, last week, and were excited at the prospect of learning more about this year’s theme - Evaluating New Frontiers in Public Opinion and Social Research. We weren’t disappointed – attempts to systematically, even randomly, select samples from the Twitter fire hose, modeling pathways between news consumption on Facebook and political behaviors, second life user avatars carrying the values and opinions of their creators – were all real nuggets.
However, much emphasis is being placed on fitting tried and tested research methodologies to a context where much remains unknown. That is, trying ambitiously to fit sampling and statistical paradigms to new dimensions of observable behavior or trying to ask respondents questions of opinion in all too familiar ways. It’s an honest approach – moving the field on, it’s also very easy to shoot down. The cynics are out there and they won’t easily be dissuaded.
At the AAPOR Plenary Session, Examining the Value of Non-Probability Sampling in Social Research, it was refreshing to hear Doug Rivers of Stanford University and the Hoover Institution talk about his change of heart on the efficacy of non-probability sampling in the world of online panels. Other panelists presented cases where there was real value in non-probability sampling, outlining the strengths of the approach, when it can be used effectively and with confidence, and when it cannot. In particular, George Terhanian of Toluna eloquently presented an approach for predicting panel respondent characteristics that justified over sampling at the design stage in order to maximize representativeness after data collection. This is an approach inspired by John Tukey’s observation that 95 percent of statistical analysis is conducted post data collection, but only 5 percent informs design.
Together, the turn of faith and intelligent design were honest, innovative and necessary. Let’s face it – achieving true randomness is too expensive and non-response presents a mighty chasm.
From this view point we look again to the future and offer another prediction – hopefully better than the last, and then consider one or two implications on the future of research as it adapts to new digital information horizons.
By 2022, 80 percent of the world’s population will be using the internet. Estimates currently stand at around 30 percent. There likely will be greater censorship and state control, and market forces will demand returns. But, as with radio across Africa, television across much of Asia, the mobile phone nearly everywhere – internet use will be close to ubiquitous and access will largely be through mobile phones.
It is estimated that by 2016 close to 40 percent of the world’s population will be using smart phones and 25 percent will own one. This will, of course, be higher in developed nations, but the gap between developed and emerging economies is increasingly blurred. There will be more cross-cultural sharing, tweeting, linking, viewing, commenting, liking, etc., while on the go – whether in the office, the field, the factory or at home. Greater amounts of opinion shaping information and news will come to people via their friends and connections on social networks, presenting both opportunities and challenges to well-established media channels.
This increase in both internet use and smart phone access offers game-changing methodological opportunities, building from a solid base of empirical social science to ensure that data and insight carry credibility, retain trust and inform decision making. At the same time, there should not be a reluctance to innovate and share what works and what doesn’t as the research industry adapts in the digital age.
Much is being done already. The past few years have seen significant growth in online and mobile data collection. Jon Krosnick of Stanford University argued that data gathered online can be very accurate and can erase some of the shortcomings of phone and face-to-face surveys. The perceived advantages include lowering of costs, approaching sensitive issues more effectively and enhanced interaction through creative audio and video stimuli. There is also strong growth in the use of apps and games for conducting research and gathering data – from self-complete mobile headache diaries to crowd sourced polling apps – examples of real innovation. Up to now, gathering public opinion has generally meant adopting obtrusive methods and aggregating individual responses to surveys, and presenting it as the view of the collective. But increasingly, public opinion can be gathered using social media data and unobtrusive methods.
Looking forward, careful attention needs to be placed on ensuring that the identity of a digital respondent is really known. This is easier when constructing an online panel but much less so when analyzing Twitter conversations. No doubt, international panel builders are already looking to the future and building user profile data for multi-mode panel delivery at every opportunity. Social media owners, however, attract digital accounts from individuals, organizations and aggregators; it is easy to create both false and multiple identities, or even produce a spambot to do this for you, if you are moderately technically inclined.
It is also critical to understand what is being said and how this discourse shapes opinion one way or another. Sentiment analysis still has some way to go. A greater emphasis needs to be placed on listening to digital discourse, (whether local, national or global) to both shape the questions that are asked, and to develop analytical approaches that produce nuanced cultural insight.
As with all of the thinking shared by the AAPOR group, we are equally committed to sharing what we learn as we navigate the digital future.
If you’d like to read on, you can find some of the work of the InterMedia Digital Team below.
Networked Audiences - 10 Rules of Engagement
Obama in Brazil - InterMedia white paper
Information Horizon - InterMedia's digital lab
Interview: InterMedia's Ali Fisher on Changing Digital Media Landscape
Analysis: Social Media Advocacy as Challenge to Putin Regime