Dimensions of belonging: relationships between police identity checks and national identity in France

ABSTRACT

Researchers in many countries have identified racial differences in rates of civilian-police interactions such as identity checks or stop-and-frisk. Are frequent police identity checks associated with feelings of national identification and if so, in what ways? Using three dimensions of identity in a nationally representative French sample of immigrants and second-generation individuals, I find that experiencing frequent identity checks is associated with an increase in the likelihood of country of origin identification, a decrease in the likelihood of French identification, and a decrease in the likelihood of believing others see one as French. I also find that these relationships and their magnitude vary across immigrant generation and immigrant region of origin. These findings point to frequent police identity checks as an important mechanism through which street-level state practices impact immigrant integration.

KEYWORDS:

• Immigration
• France
• national identity
• policing
• discrimination

Acknowledgments

The author is grateful for the in-depth feedback and continued support she received on this paper from Robert Wuthnow, Patricia Fernández-Kelly, Paul DiMaggio, Oscar Torres-Reyna, Katherine McClelland, Samantha Jaroszewski, Amanda Cheong and Austin Huffman. This paper also benefitted greatly from the comments of participants of the Princeton Sociology Empirical Seminar and the CSSO Workshop on Social Organization.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1 Because male respondents are frequently stopped by police for identity checks, as reflected in the literature (Jobard et al. 2012) and this sample (among first, second and 2.5-generation respondents in the TeO sample, 85% of women had not experienced an identity check in the past year, compared to only 59% of men), it is possible that gender and frequency of police identity checks interact to make the experience of multiple identity checks inherently different for men compared to women when it comes to national identity. For this reason, I tested whether the interaction between gender and frequency of police identity checks significantly improved the fit of each model, which it did not (for country of origin identification: LR χ² = 1.62, p-value = .4443; for French identification: LR χ² = 0.14, p-value = .9327; for reflected appraisals: LR χ² = 0.54, p-value = .7619). In addition, when included in models, this interaction was not statistically significant. For these reasons, a gender and frequency of identity checks interaction was not included in regression analyses.

2 IRIS is a geographical categorization aiming to represent individual neighborhoods of homogenous composition with a population of approximately 2000 residents per IRIS. For more information of the IRIS as a geographical unit, see: https://www.insee.fr/en/metadonnees/definition/c1523

3 To examine this question, I performed a subset analysis of respondents aged from 20 to 29. Overall, this did reduce the difference in the percentage of respondents stopped two or more times between respondents of the second (including 2.5) and first generation for a given region of origin, but differences remain. Age thus may play a role in explaining differences in rates of frequent police identity checks across immigrant generations, but age differences are not the sole explanatory factor for variation in police check rates by immigrant generation. For more information please contact the author.

Additional information

Funding

The author thanks the Princeton Center for the Study of Social Organization for financial support.