A representative-represented matrix: exploring the symbolic effect of minority representation

ABSTRACT

This study explores how symbolic representation varies by citizens’ demographic identities and bureaucracies’ demographic composition. First, it presents a representative-represented matrix to provide a conceptual framework from the citizens’ perspective. Then, this is followed by the application to the area of policing in the experimental setting. Citizen perceptions are explored in terms of minority bureaucrats’ representative role and the potential benefits of a representative bureaucracy. The empirical evidence demonstrates that the symbolic effect of minority representation varies and may not accomplish expected outcomes for all minorities. The symbolic representation may be manifested for women and minorities but limited for White men.

KEYWORDS:

• Representative bureaucracy
• symbolic representation
• policing
• experiment

Supplemental Material

Supplemental data for this article can be accessed at https://doi.org/10.1080/14719037.2022.2124536

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. Recent studies report mixed findings on the individual and organizational effects of minority representation. Nicholson-Crotty et al. (2016) found that individual-level effects prevail over organizational-level effects, while other scholars reported both individual- and organizational-level effects in the education setting (Favero and Molina 2018; Vinopal 2018; Zwicky and Kübler 2018).

2. Scenario based on the over-represented minorities and/or women is demonstrated in Appendix A1. However, assessing and interpreting the scenarios is less likely and not the focus of this study within the context of policing.

3. Randomization inference was performed with ritest with 1000 replications. Heß (2017) introduces this versatile command that allows more complex resampling methods for individual data within strata. The p-value was 0.002, similar to that in the regression, which suggests that the individual-level randomization was successful.

4. The null hypothesis of Levene’s test of equality of error variances (i.e. variances are equal) was not rejected for both models (p = .067 and .073). The null hypothesis of the F test for heteroskedasticity (i.e. homoscedasticity) was rejected for both models (p = .005 and .000). In the estimated marginal means, covariates were evaluated at the following values: political preference = 5.99, fulltime employment = 0.437, education = 3.21, age = 2.90.

5. After the BLM movement, 76% of Americans consider racism and discrimination a “big problem”, which is 26% point higher compared to 2015 (Cohn and Quealy 2020).

Additional information

Funding

The author appreciates the SPS Research Committee at the School of Public Service, Boise State University for funding this work (RC0043).

Notes on contributors

Sanghee Park

Sanghee Park is an associate professor in the School of Public Service, Boise State University, ID, USA. Her research efforts focus on public management, public sector governance, and representative bureaucracy.