Demographics are not Destiny: An Interview with David Allison
Editor’s Note: While Mr. Allison’s “Valuegraphics” is a product — and we aren’t in the business of promoting products — the Charged Affairs team believes this interview is of significant foreign policy value. Valuegraphics is part of an important, growing, and poorly understood trend in which big- and meta-data are used to better understand audience and messaging… and Allison has something important to say about what this means for voting patterns across democracies.
David Allison, the inventor of Valuegraphics, and a self-described data pioneer, remembers exactly where he was when he decided his career, and his worldview, needed to change. At that little table, in that little café, reading that seemingly innocuous article summarizing millennial preferences, desires, and hobbies, he realized how much of his work as a marketer and consultant had been based on sweeping flawed assumptions about large demographic groups neatly packaged into cookie-cutter corporate solutions. At that moment he decided to dedicate his life to testing his new controversial hypothesis by modernizing the lens through which we view our population, and digging deeper than outmoded categorizations based on age, sex, gender, race, or religion.
Allison believes traditional demographic analyses and polling strategies are flawed to the point of being useless to help us understand consumer, or voter behavior. Therefore, he says, we need a new social science data analysis model that digs deeper than outmoded demographic stereotypes, and rests on a 21st-Century understanding of a diverse, global, modern populace.
Enter Valuegraphics: the tool Allison has spent years developing to address this glaring need. To be clear, Allison is not the first social scientist to call for more transparent and effective methodological approaches to the collection of polling data. Since the 2016 US election took so many experts by surprise now more than ever prominent pollsters like Gallup, Rasmussen, and FiveThirtyEight make sure to highlight just how sound, statistically representative and thorough their methodologies are. Though Valuegraphics is not the first tool built to inform the analysis of electoral outcomes, or market trends, it is however the first to do so with as much commitment to teasing out the public’s deepest, lasting values, not merely their (stated) preferences at any given moment.
So what, from a methodological standpoint, makes Valuegraphics different? Allison cites its reliance on a random, stratified, statistically representative sample of respondents designed to optimally mimic the general population, and the over 250,000 survey responses he has collected worldwide, 100,000 of which have come from the US and Canada alone. These totals fall short of Gallup Daily Tracker’s 350,000 annual survey responses, for example, but are rendered more impressive when considering the number of questions contained in a Valuegraphics survey. Valuegraphics contains nearly 400 survey questions built to probe more deeply into a respondent’s psyche than traditional polls. Allison claims that by scoring respondents’ affinities to 40 core values, Valuegraphics combats survey bias more stringently than traditional analyses.
Valuegraphics’s most innovative ingredient lies within its algorithmic predetermination of a respondent’s interests to build a set of questions uniquely tailored to evaluate that respondent’s place along those 40 value spectra. By engaging with respondents through initial questions regarding their own unique preferences, no two people see the same survey, but all can be scored in the same categories. Most polls attempt to control for demographic covariates like age, education, race, and gender. While Valuegraphics collects this information, it groups respondents in terms of those values, not respondents’ external characteristics, and eschews sweeping demographic conclusions for inferences based on what respondents reveal their values to be.
Allison likens the experience to asking someone who has identified that they are a hockey fan questions about their favorite team. By analyzing responses to questions concerning components of their fandom, his thorough surveys burrow into deeper truths that expose our core values. How much one values tradition, family, trust, loyalty, and other values shine through in the results these surveys yield, and, Allison posits, correlate strongly with future behavior from the supermarket to the ballot box.
Regarding the latter, Valuegraphics analyses have revealed two major themes concerning the 2020 election, although Allison is careful to preface this proclamation with the disclaimer that his approach is not built to determine which candidate or party voters prefer. He fixates less on whom the American people will elect, and more on why the polity will behave the way it will.
Valuegraphics data suggest that economic anxieties will drive unprecedented numbers of infrequent voters to the polls, and that more frequent voters will be motivated to turn out because of their concerns over climate change, as opposed to, specifically, healthcare. Such sound, yet intuitive predictions are not unique to Allison’s model, which he believes only lends to its legitimacy. Arriving at conclusions similar to those resulting from more traditional model does not make his approach any less novel, and over time, he believes Valuegraphics will paint a clearer picture of the hearts and minds of voters and consumers around the world.
“We have created a tool that helps us understand each other better.” Allison’s voice softens under the emotions that arise when he thinks about his creation’s future. Listing off globally problematic assumptions that values-based data might combat helps him focus his next thought. He concludes with yet another maxim, a single truth he believes drives every flawed assumption in our culture, and perhaps serves as an ideal first line for Valuegraphics’ business case.
“We fear the things we don’t understand.”