The Manhattan Institute has published a report by Duke professor Jacob Vigdor, “Comparing Immigrant Assimilation in North America and Europe” (Reihan takes a look at it over in his own corner of NRO). A composite index like this can be a useful exercise, and this one yields some unsurprising results: of course Mexican immigrants in the U.S., for instance, are more assimilated than Muslim immigrants in Europe.

But saying that some groups are “more assimilated” and others “less assimilated” begs the question of what “assimilated” means. Those inclined toward open borders (like the WSJ’s Jason Riley) will just accept the word at face value, but this is one of those times you actually have to read the methodology section of the report. This is especially true since the author’s Conclusions section makes explicit that he sees his findings as making the case for high levels of immigration, albeit along the lines of Canada — i.e. (though he’s not quite this explicit), more Asians and fewer Hispanics.

So, what are Vigdor’s metrics of assimilation?

Economic indicators used in the computation of the index consist of educational attainment, earnings, occupational prestige, employment status, and labor force participation rate. In the case of the last four indicators, the index performs separate comparisons of males and females, since differences between them in labor-market participation have been meaningful historically and remain so.

Civic indicators consist of citizenship and veteran status. In the case of the latter, males and females are considered separately, since military service is more common among males.

Cultural indicators consist of ability to speak English, marital status, number of children in an adult’s household, and whether a spouse is native- or foreign-born.

These generally make sense, but since they’re the easiest indicators to measure, there’s a certain element of looking for your keys where the light is better. While speaking English and having a job are clearly steps toward assimilation, what really matters is what John Fonte calls “patriotic assimilation“, or what Stanley Renshon describes as “emotional attachment” to the American national community. Vigdor’s “civic indicators” try to get at that, but his finding that Canada is a model shows how weak this index really is. Vigdor writes that Canada “consistently outranks the United States” in part because its “liberal attitude toward dual citizenship” results in higher levels of naturalization and thus a higher ranking in the civic indicators. But dual citizenship is the very antithesis of patriotic assimilation, something Teddy Roosevelt rightly called a “self-evident absurdity.” Any index that uses dual citizenship as a measure of the success of assimilation is inherently flawed.

There are ways of getting at this question, but they can’t easily be compared across countries and continents. Alejandro Portes and Ruben Rumbaut, for instance, conducted a longitudinal survey of young people from immigrant families in Miami and San Diego and found that going to high school significantly reduced the likelihood that they identified as Americans. Likewise, a 2008 Harris Poll (summary here, but I don’t think the full results have been published) found naturalized citizens to be significantly less patriotic than the native-born (support for international law over the Constitution, support for teaching ethnic pride vs. American patriotism in schools, opposition to renouncing prior loyalties as a condition of naturalization, etc., etc., etc.).

Attempting to measure assimilation is certainly worth doing, and Vigdor’s report gives us a partial look at the issue. But to get at patriotic assimilation — which is what really matters in a diverse, partly-propositional republic like ours — you can’t just use data from the Census Bureau’s American Community Survey because the light’s better there. You have to develop your own data, which is expensive, and ask questions that modern social scientists are often uncomfortable thinking about, let alone researching in depth.