Linos Book Symposium: Comments by Ryan Goodman

[Ryan Goodman is the Anne and Joel Ehrenkranz Professor of Law at New York University School of Law. You can also find him on Twitter:@rgoodlaw]

Katerina Linos’s book, The Democratic Foundations of Policy Diffusion, bears the hallmarks of great scholarship. She tells us something new, important, and counterintuitive about international law. And she predicates her theories on multiple, rigorous and innovative empirical methods. I thus share the praise that has been heaped on Linos’s impressive book by the other OJ commentators from yesterday and today.  Indeed, I have drawn upon and cited drafts of the book manuscript in my own work; included excerpts of Linos’s earlier research in a textbook that I co-edit; and blurbed the back of the book with immense praise.

Against this backdrop, I want to interrogate a specific part of the book—the chapter on the United States—and, in particular, Linos’s theoretical account of her political opinion experiments. Indeed, this is the key part of the book that tries to get “inside the minds” of individuals and understand the mechanisms for influencing their policy preferences.

The experiments ask a representative sample of the US population whether they support a social policy. Linos then compares that baseline group with groups that were told either that Canada had adopted the policy, that most western states had adopted the policy, that the UN recommended the policy, or that US experts recommended the policy. She also compares whether particular subjects—Republicans versus Democrats or highly informed versus poorly informed individuals—responded differently to the different prompts. In one experiment the social policy is universal health care, and in another experiment the policy is paid maternity leave.

According to Linos, the experiments suggest that foreign and international models provide a source of information for members of the electorate to determine whether their political representatives are proposing good social policies (an information-deficit theory). Does the data fit this theoretical explanation best? Are the data equally consistent with alternative theories that emphasize social and psychological conformity as a mechanism that explains the influence of foreign and international models on individual preferences?

Let’s start with three of the most interesting and important empirical findings in the chapter:

1. US citizens respond much more favorably to governmental policies—including ones that explicitly require tax increases—if they are told that other western states have already adopted the social policy or told that the United Nations recommends it.

2. Republicans respond more favorably than Democrats when informed that the UN recommends a social policy or that most western states have adopted the policy.

Note: This finding appears to compare shifts in support among Republicans who disfavor a social policy with shifts among Democrats who disfavor the policy. This comparison may involve systematic bias. The type of individual who self-identifies as Republican and favors core parts of the Republican Party platform is very different from the type of individual who self-identifies as Democrat and opposes core parts of the Democratic Party platform. Indeed, the former might be considered conformists and the latter non-conformists. And Linos’s findings show that the former are more likely to follow global trends and the international “mainstream.” Accordingly, the key explanation may boil down to a social conformity mechanism.

3. In important cases, individuals who are poorly informed about a social policy respond more favorably than well-informed individuals when told that other western states have already adopted the social policy or that the United Nations has recommended it.

I want to focus in detail on the third finding, and contend that it should be significantly qualified. Consider the experiment involving health policy. That study found (as Linos admits) no statistically significant difference between respondents with high levels of information and respondents with low information when these individuals are subject to the various prompts (e.g., when told that all western states had already adopted the policy or that the UN had recommended it). That is, respondents overall expressed higher support for the proposed policy when told about the foreign or international models (finding #1), but poorly informed individuals were no more compelled than well-informed individuals to increase their support. This experiment thus provides strong evidence against the theory that information is a primary mechanism for foreign and international influence in shaping individual preferences.

Next, consider the experiment involving paid maternity leave. The findings were essentially the same as health policy with respect to the UN prompt. That is, respondents overall increased their support by large percentages if told that the UN recommended the policy. The findings, however, showed no statistically significant difference between well-informed individuals and poorly informed individuals. Again, these findings contradict, or at least fail to support, the information-deficit theory.

That said, the family leave policy experiment did find that poorly informed individuals responded more strongly than highly informed individuals when told that most western states had adopted the policy (30% increase for poorly informed and 16% increase for highly informed respondents). Similar results occurred when subjects were told that Canada had adopted the policy. Three factors, however, suggest caution in interpreting this data too favorably for the information-deficit theory. First, well-informed respondents also increased their support considerably when told of these foreign models. The information-deficit theory has no explanation for this finding. Indeed, it may be a more significant finding for the purpose of developing a theory of mechanisms of influence: even individuals who are highly informed about a subject are susceptible to large shifts in support for a policy if foreign countries have adopted or global institutions have recommended the policy. Second, the data are based on self-reported levels of information. And, we can easily imagine that self-understanding and self-reporting of one’s own awareness of social policy is systematically skewed (e.g., in favor of reporting overconfidence generally, over- or under-reporting confidence among particular types of individuals).

Third, cases in which low-information subjects responded more strongly may mask a spurious correlation. That is, another factor—e.g., lack of concern about the social policy—might lead both to individuals’ being poorly informed and being more amenable to favor a policy simply because other countries have approved it. Such political apathy contradicts the type of robust democratic politics that Linos’s thesis envisions. Her argument depends on the notion that individuals who are influenced by international models are also invested in the political process concerning the relevant social policies. There are other possible confounders. Consider for example social conformists–people who are more prone to emulate a reference group’s choices rather than conscientiously evaluate the merit of those choices. Social conformity—the hidden variable–may thus lead both to being poorly informed about the costs and benefits of social policies and to follow international trends.

In short, the experimental studies may provide equal or greater support for the social conformity model than the information-deficit theory of global norm diffusion—at least with respect to the US electorate.