Intro and data

 

This analysis is organized into two parts: (1) BGS-wide analysis and (2) grad group-specific analysis. The BGS-wide analysis describes under-represented minority (URM) vs non-URM applicants’ progression through the admissions process, examines the role of the early admissions (EA) pathway, and compares the racial and ethnic composition of the BGS applicant pool to the U.S. population. The grad group- specific analysis examines whether there is disparate attrition for URM vs non-URM applicants for each step in the admissions process (interview, admissions, matriculation) at the grad group level.

The data used in this analysis come from three datasets sourced from BGS administration. For the BGS-wide analysis, we use admissions cycle data disaggregated at the URM vs non-URM level for years 2018-2020. We further use a more granular racial and ethnic breakdown for URM-identified applicants from those same years. Finally, we utilize dataset containing all participants in Penn Post-Baccalaureate Research Education Program (PREP) and Penn Summer Undergraduate Internship Program (SUIP) who applied to BGS PhD programs in the 2016-2021 admission cycles.

The definition of ‘URM’ used by Penn BGS is [XXXX quote census def XXXX]. Applicants self-identify as ‘URM’ by checking a box on application materials that asks ‘[XXXXXXXXXXXXXXXXX]’. Applicants further describe their racial and/or ethnic identities (options: “American Indian or Alaskan Native”, “Black or African-American”, “Hispanic/Latino”, and “Native Hawaiian or other Pacific Islander”) on application materials in a section [XXXXXXXXXXXXXXXXXXX].

 

BGS

Figure 1: URM vs. non-URM application flow, 2018-2020

Figure 1 shows the count of URM vs non-URM applicants who progress through each stage of the application cycle (application, interview, admission, matriculation) for the most recent three years (2018-2020).

 

 

Figure 2: URM applicants flow by racial and ethnic identities, 2018-2020

Figure 2 shows the count of URM applicants, disaggregated by racial and ethnic identities, who progress through three stages of the application cycle (application, admission, matriculation) for the most recent three years of data (2018-2020). Interview data was not available for this sample. Importantly, many applicants identified with more than one racial or ethnic identity. For clarity of presentation, applicants were disaggregated inclusive of each identity - for example, an applicant who identified as both ‘Black or African-American’ and ‘Hispanic/Latino’ is represented in each of those flows in the above figure. Future analyses will explore alternative means of representing these intersecting identities.

 

 

Table 1: EA applicants (2015-2020)

Blah blah

 

Table 2: EA applicants by year versus all URM applicants

 

BGS applicant pool versus the U.S. population

 

We used U.S. Census 2019 national estimates here to calculate the proportion of the US population that is considered ‘URM’ (people who identify as Black or African American, Hispanic/Latinx, American Indian or Alaska Native/Indigneous and/or Native Hawaiians and other Pacific Islanders). We then compared it to the proportion of applicants categorized as such, via a one-sample test of proportions.

The proportion of the overall BGS applicant pool that is URM is significantly lower than the US population proportion (95% CI of the difference (0.17, 0.21) p = 0.000).

 

 

Graduate groups

Figure 3: BMB applicants 2018-2020

 

 

Figure 3: CAMB applicants 2018-2020

 

 

Figure 3: GCB applicants 2018-2020

 

 

Figure 3: GGEB-EPID applicants 2018-2020

 

 

Figure 3: GGEB-BSTA applicants 2018-2020

 

 

Figure 3: IGG applicants 2018-2020

 

 

Figure 3: PGG applicants 2018-2020

 

 

Figure 3: NGG applicants 2018-2020