Indexing Immigration and Integration Policy: Lessons from Europe

Abstract

Immigrant‐related policy indexes have become popular in both U.S. and European contexts, yet these projects boast distinct and divergent trajectories. European indices are characterized by rigorous conceptualization, specificity in elements of policy design (e.g., settings like fees or appeal process), and a variety of measurement strategies. By contrast, U.S. state‐level policy indices exhibit a lack of differentiation between immigration and integration policy and excessive generality in measurement and representation of policy instruments and settings, exacerbated in presenting a policy index as an aggregate count. This paper argues U.S. policy indexing can benefit from the European indexing experience. Following an overview of the state of each field, assessing concept differentiation, specification, and measurement in each, the paper illustrates how even the well‐conceived and specified European integration policy indices run into problems at the analysis stage. It presents a replication‐replacement study to illustrate divergent performance of highly correlated and conceptually agreeing indices, as well as methodological issues inherent to indices of low‐N, including using a policy index as a dependent variable and index selection absent a priori theorization. It concludes with suggestions for improving American immigration policy indices, as well as general observations on working with statistical power‐challenged indices and data limitations.

Is “Threat” in the Eye of the Researcher? Theory and Measurement in the Study of State‐Level Immigration Policymaking

“…However, because most studies of American immigration policy at the state level use simple counts, this study also uses simple counts. This does not mean that this methodological choice is without costs (Goodman, , ). As Monogan () and Goodman () show, the specification of the dependent variable has a major impact on the regression results.…”

“…This does not mean that this methodological choice is without costs (Goodman, , ). As Monogan () and Goodman () show, the specification of the dependent variable has a major impact on the regression results. For the purposes of this study, however, the dependent variables and all controls are held constant across models in order to focus on the demographic variables.…”

“…The same seems to be the case with models that disaggregate immigration policy into narrower domains such as education, welfare, and enforcement. Studies using the immigration policy indices that comparatives have developed to characterize the policies of many countries, and which have been the subject of rigorous design, produce similarly inconsistent results when used as dependent variables (Goodman, ). In all, the evidence suggests that aggregate policy models may not be the best vehicle for studying how racial threat or group political power influence policy outcomes.…”

One Model Does Not Fit All: The Varied Politics of State Immigrant Policies, 2005–16

“…This study speaks to methodological concerns central to this special issue, especially improving research designs so as to more clearly align dependent measures of immigrant policy with causal models. Given that state policies affecting immigrants cut across multiple policy domains, empirical findings are highly dependent on the manner in which measures of policy output are operationalized, a point underscored in this issue by Butz and Kehrberg (), Filindra (), Goodman (), and Berkman et al (). The results here particularly underscore Monogan’s () warning that the causal effects of variables of interest vary according to whether they are tested on specific policy areas versus aggregate indices.…”

“…The results here particularly underscore Monogan’s () warning that the causal effects of variables of interest vary according to whether they are tested on specific policy areas versus aggregate indices. In light of this, and as argued by Goodman (), researchers should take special care when using additive measures of immigrant policies, which are prone to excessive generality and lack of differentiation. This does not necessarily mean eschewing aggregation, but the findings here reiterate the need for clear theoretical and empirical justifications for aggregation choices.…”