In a sense, this entire debate is a rather blatant example of how “political science” gets abused by being forcibly fit into conventional political thinking that it is not really compatible with–although Achen and Bartels did not exactly try to discourage this. What they were able to show is that, to a large degree, Sanders supporters are not consistently motivated by support for his “precise” policy program. This is hardly new in political science–this follows from a long tradition that showed that most voters, especially those who are not closely engaged with politics, lack a well-defined “ideology,” or awareness of what policy goals are usually bundled together, and are liable to change their mind on politics frequently.
What to make of Converse’s argument, of course, is not very obvious. While having a consistent ideology does indicate a closer engagement with politics, it is hardly an indication of a superior understanding. If anything, it is associated with pigheaded obscurantism and closed-mindedness that rejects with high frequency what does not fit their worldview. Ideology, after all, is not a logically consistently bundling that is inevitable, but a coincidental correlation that somehow emerges in course of politics. Awareness of ideology, then, is akin more to astrology than a real “science”: it is “true” because it is a repeated pattern–it’s true just because it is.
While Sanders voters may not be “ideological,” are they totally unaware of and unmotivated the issues? The lack of clear ideology does imply that “what they want” cannot be precisely measured and implemented along the DW-Nominate scale (if it meant anything in the first place–see my issues with its abuse). But symbolic politics ARE statistically associated with awareness of and desire for certain types of policy goals serving issues, EVEN if the linkage is defined with a very high variance, cognizant of the great deal of uncertainty on the part of the voters. Awareness of this linkage, too, is hardly a new discovery in political science–even if it may have been increasingly disregarded, unfortunately. In fact, this is the central argument of a classic in political science, Home Style by Richard Fenno. While Fenno noted that explicit mention of policy was rare and the vote choice was generally not proximately motivated by “ideology,” in favor of various “symbolic” concerns, it nevertheless formed the undercurrent that sustained the choice of symbols used in the politics: as Fenno states, the seemingly issueless symbolic politics “takes place within an issue context” (emphasis in the original) Symbols are powerful because they represent shared and strongly agreed upon values values on broad issue areas, even if not the specifics. Fenno’s entire argument is, indeed, a precaution against the urge to separate symbols and policy.
Amusingly, of course, separation of symbols and policy is precisely what has been taking place in American politics research since then. Symbols do not map neatly onto “measurable” policy (even if the measurements used, e.g. DW Nominate) are specious.) So the argument by many political psychologists, who subscribe to symbolic politics and “rationalization,” tends to overemphasize them as separate from policy, an alternate explanation, not something does map onto policy, but a bit more crudely, in non-spatial fashion–which, to be fair, would face a whole lot of additional objections as well as development of an entire body of alternate modeling framework (which, to be honest, is feasible, thanks to the advances in Fuzzy Logic and the like–but selling it is NOT easy given the present mindset of the field.)
I would not suggest “What is needed, I think, is a retreat from endless parsing of The Data and a little common sense” as The Week article suggests. Astrology, after you follow its logic for a while, might seem like common sense, especially when spiced up with “appropriate” data. What is needed, I think, is a dose of critical creativity: an interest in knowing what is being measured and why the data is showing up as it does, not just measuring data and making grand oracular pronouncements. This used to be known as “science,” but I suppose this is opposite of “data science” should be pushed aside in this new era, I suppose…