The Power of Big Data and Psychographics [Fact Checking]

From the description:

In a 10 minute presentation at the 2016 Concordia Summit, Mr. Alexander Nix discusses the power of big data in global elections. Cambridge Analytica’s revolutionary approach to audience targeting, data modeling, and psychographic profiling has made them a leader in behavioral microtargeting for election processes around the world.

A highly entertaining but deceptive presentation on the state of the art for marketing political candidates.

Nix claims that most marketing companies base their advertising on demographics and geographics, sending the same message to all women, all African-Americans, etc.

Worse than a “straw man,” that’s simply false. If you know the work Selling Blue Elephants by Howard Moskowitz and Alex Gofman, then you know that marketers tweak their pitches to very small market slices.

But you don’t need to find a copy of Selling Blue Elephants or take my word for that. On your next visit to the grocery store see for yourself how many variations of a popular shampoo or spaghetti sauce are offered. Each one is calculated to attract a particular niche of the overall market.

Nix goes on to describe advertising in the 1960’s as “top down,” “hope messages resonant,” etc.

Not only is that another false claim, but the application described by Nix was pioneered for the 1960 presidential campaign.

Ithiel de Sola Pool, with others, developed the Simulmatics program for the computation of a great variety of factors thought to influence voting, for specific use in the 1960 presidential election. A multitude of influences can be introduced into the program, together with modifications of a strategic nature, and the results bear on both prediction and choice of strategy, much in the manner that elaborate market research influences business decision on manufacture and sale of a new product. The Simulmatics project assembled a basic matrix of voter types and “issue clusters” (480 of the former and 52 of the latter, making a total of 24,960 cells), consolidating as values the accumulated archives of polling on all kinds of questions. The records of the Roper Public Opinion Research Center at Williamstwon were used as source material. With no data later than 1958, the simulation achieved a correlation by states of .82 with the actual Kennedy vote.

(“The Mathematical Approach to Political Science” by Oliver Benson, in Contemporary Political Analysis, edited by James C. Charlesworth, The Free Press, 1967, at pp. 129-130)

I’ll grant that Nix has more data at his disposal and techniques have changed in the last fifty-seven (57) years, but there’s no legitimate reason to not credit prior researchers in the field.

PS: If you find a hard (or scanned) copy of The Simulmatics Project by Ithiel de Sola Pool, let me know.

Comments are closed.