Data Inquiry Project:
Exploring AP Computer Science Data
Topic: Computer science is an increasingly important discipline, as computing becomes progressively more ubiquitous and continues to infiltrate the job market. In the coming years, there will be far more computing jobs than there are U.S. computer science graduates to fill those jobs. And sadly, there is great inequity in who is learning computer science and entering this field. In this data inquiry project, I will explore one metric that can illustrate this inequity: high school students' Advance Placement (AP) exam results in Computer Science. Through this inquiry, I will attempt to answer the following questions:
- Which demographic groups are underrepresented in computer science?
- How do these underrepresented groups comparatively perform on the AP Computer Science exam?
- How do trends in San Francisco compare to those in California and the United States as a whole?
View data spreadsheet and analyses here. (Note: some San Francisco data cannot be shared publicly and has been redacted.)
Data Sources:
- 2013 Advanced Placement data from the College Board:
2. California summary report
3. San Francisco Unified School District raw data (unpublished)
- California Quick Facts from the U.S. Census Bureau
- San Francisco Unified School District enrollment data (unpublished)
A Summary of Findings: (Note, all aforementioned statistics are from 2013 AP data.)
⧫ Only 3% of test takers were Latino, even though 26% of students are Latino.
⧫ There were no African American and no Native American test takers, even though these groups represent 9% of
the schools' population as a whole.
⧫ Only 12% and 10% of test takers were African American, Latino, or Native American (combined).
⧫ There were only 16 total African American females in the state of California that took AP Computer Science.
Science exam, while the national average for all students was 67%.
- A relatively small, but steadily increasing, number of students within San Francisco public schools take AP Computer Science.
- In the public schools of San Francisco, some demographic groups are greatly underrepresented in AP Computer Science: females, African Americans, Latinos, and Native Americans.
⧫ Only 3% of test takers were Latino, even though 26% of students are Latino.
⧫ There were no African American and no Native American test takers, even though these groups represent 9% of
the schools' population as a whole.
- Similar trends of underrepresentation in AP Computer Science exist in both California and the United States.
⧫ Only 12% and 10% of test takers were African American, Latino, or Native American (combined).
- Female minority students (i.e., female African Americans, Latinas, and female Native Americans) are doubly underrepresented in AP Computer Science, with the smallest proportion taking the exam.
⧫ There were only 16 total African American females in the state of California that took AP Computer Science.
- This underrepresentation is most extreme in San Francisco.
- African American and Latino students performed significantly worse on the AP Computer Science exam.
Science exam, while the national average for all students was 67%.
- While all female students did not perform significantly differently than the average, female African Americans and Latinas performed more poorly than all other measured subgroups.
- ⧫ Only 27% and 32% of female African Americans and Latinas, respectively, earned a 3, 4, or 5 on the AP Computer Science exam, while the national average for all students was 67%.
Reflection:
For this data inquiry, I began by analyzing AP Computer Science exam data from the San Francisco Unified School District in various subgroups and comparing to the proportion in the population of the schools as a whole, in order to determine which groups are underrepresented in computer science. I quickly learned that I needed to compare this city or district data to the state and national data, so I then analyzed U.S. and California College Board data, as well as U.S. Census Bureau data to determine the proportion of the populations these groups represent. While I suspected that females and students of color would be underrepresented in AP Computer Science, I did not anticipate such great disparities.
The numbers truly come to life when zooming in one particular group. Even though African American females comprise about 7% of the U.S. population, they only made up 1% of the AP Computer Science test takers in 2013. And, in the entire state of California, there were only 16 total African American females who took the AP Computer Science exam, and only half (or 8 total) earned scores of 3, 4, or 5 to qualify for advanced placement. Unfortunately, there zero African Americans -- male or female -- in San Francisco public schools to take the exam in 2013.
In this study, I compared AP data to current high school enrollment in San Francisco public schools and the California and U.S. populations as a whole. Future studies could examine differences between public and private schools and among states. It would also be beneficial to compare AP Computer Science test takers to all AP test takers and identify additional metrics for tracking access to computer science education in earlier grades.
While I focused my analysis on San Francisco, it is clear that this is not an isolated problem; rather, it affects the entire state and nation. I hope these grave statistics begin to surface this inequality of both access and achievement in computer science, and that as a result, legislators, educators, parents, and community members actively work to engage these underrepresented groups and improve the pipeline to some of the most lucrative and widely available tech industry jobs.
For this data inquiry, I began by analyzing AP Computer Science exam data from the San Francisco Unified School District in various subgroups and comparing to the proportion in the population of the schools as a whole, in order to determine which groups are underrepresented in computer science. I quickly learned that I needed to compare this city or district data to the state and national data, so I then analyzed U.S. and California College Board data, as well as U.S. Census Bureau data to determine the proportion of the populations these groups represent. While I suspected that females and students of color would be underrepresented in AP Computer Science, I did not anticipate such great disparities.
The numbers truly come to life when zooming in one particular group. Even though African American females comprise about 7% of the U.S. population, they only made up 1% of the AP Computer Science test takers in 2013. And, in the entire state of California, there were only 16 total African American females who took the AP Computer Science exam, and only half (or 8 total) earned scores of 3, 4, or 5 to qualify for advanced placement. Unfortunately, there zero African Americans -- male or female -- in San Francisco public schools to take the exam in 2013.
In this study, I compared AP data to current high school enrollment in San Francisco public schools and the California and U.S. populations as a whole. Future studies could examine differences between public and private schools and among states. It would also be beneficial to compare AP Computer Science test takers to all AP test takers and identify additional metrics for tracking access to computer science education in earlier grades.
While I focused my analysis on San Francisco, it is clear that this is not an isolated problem; rather, it affects the entire state and nation. I hope these grave statistics begin to surface this inequality of both access and achievement in computer science, and that as a result, legislators, educators, parents, and community members actively work to engage these underrepresented groups and improve the pipeline to some of the most lucrative and widely available tech industry jobs.
Further Reading: Exploring Computer Science (ECS) is a curriculum that was designed to engage a more diverse group of students in Los Angeles Unified School District. They have done their own analysis of AP Computer Science data.