Beyond the Incident: Institutional Predictors of Student Collective Action

ABSTRACT

Since the original unveiling of the I, Too, Am Harvard campaign, which highlighted students’ experiences with racially based microaggressions on Harvard’s campus, more than 40 other student-led initiatives have developed their own similar campaigns. We used data on 4-year public and private, not-for-profit institutions during a 5-year period from the Integrated Postsecondary Education Data System to investigate the institutional characteristics that predict the birth of an I, Too, Am initiative. We did not find evidence that the racial diversity of higher education institutions is predictive of their propensity to have an I, Too, Am campaign. Instead, the general states of institutions’ selectivity, size, and percentage of Pell Grant recipients were more predictive. Also, while we found that the current state of institutions has some predictive power, we found less evidence of this relationship for changes in the institutional state. The data suggested that the tipping points that motivate student social movement mobilization may not be primarily related to any specific change in institutional characteristics, but rather that they existed in a context of standing institutional characteristics.

KEYWORDS:

• Diversity in higher education
• politics of education
• student collective action

Acknowledgments

The authors would like to thank Christopher Redding, Benjamin Skinner, and Sondra Barringer for their helpful comments. The authors bear sole responsibility for the content of this article.

Notes

1. By “students of color,” in this section, we are referring to Non-White students unless otherwise noted.

2. Approximately 18 months after the first search, we replicated the search using the first 20 pages of Google results. We found no additional movements, suggesting we have the population of movements fitting our criteria.

3. We excluded institutions that switched between not-for-profit and for-profit status during the analytical period. We also excluded institutions that did not report Title IV status in addition to the institutions that reported not having access to Title IV funds, U.S. services schools, and those not included in the Carnegie Classification to reduce extraneous differences between the I, Too, Am (ITA) institutions and the comparison institutions.

4. We use private to represent private not-for-profit institutions for the rest of the manuscript because no for-profit institutions were included in the analysis.

5. We refer to students who identified as Black or African American and Hispanic or Latino as Black and Latino, respectively, for the rest of the manuscript.

6. We used predictive mean matching due to the restricted range of acceptance rate (could not be negative; White, Royston, & Wood, 2011). The imputations included a burn-in of 500 and 25 additional data sets. The 25 imputed data sets produced sufficiently stable estimates based on the Monte Carlo error of the coefficients, t statistics, and p values across all the model specifications (White et al., 2011).

7. To test for robustness, we also estimated the state models using only this restricted sample and found qualitatively similar estimates in magnitude, direction, and significance. Estimates are available from the authors upon request.

8. These analyses only included the original data.

9. As suggested by Allison (2012).