Asians 17% more likely to make Dean’s Honors List at UW Math

Are east asians good at math? Common sense says so, but let’s see if the data backs it up.

First, we scrape Learn for a list of all undergraduate math students.

All UW math students are automatically enrolled in this course, and you can get a list of all 7755 students and their emails here. The interface only shows you 200 students at a time, so I wrote a scraper to automatically go through 39 pages of students.

How to determine which students are east asian? For this analysis, I consider east asian to be people of Chinese, Korean, and Vietnamese surnames. Here’s the 100 most common last names in UW math:

307 Wang
279 Li
244 Zhang
231 Chen
203 Liu
111 Yang
111 Wu
93 Xu
91 Huang
84 Zhou
78 Zhu
76 Lee
74 Zhao
73 Lin
62 Yu
61 Sun
56 Lu
48 Hu
47 Gao
45 Guo
44 Ma
44 Kim
43 Zheng
43 Luo
43 Jiang
41 Chan
39 Tang
39 Shi
39 He
34 Yuan
34 Jin
32 Pan
31 Song
30 Cheng
28 Zeng
28 Yan
27 Liang
26 Park
26 Cai
25 Ye
25 Xiao
25 Wong
24 Xie
24 Shen
24 Lai
24 Gu
24 Feng
24 Cao
23 Ng
22 Tan
22 Han
22 Deng
21 Ren
21 Qin
20 Nguyen
19 Zhong
18 Yao
18 Patel
18 Fu
17 Wei
17 Qiu
17 Leung
17 Lam
17 Dong
16 Xue
16 Tian
16 Khan
16 Ji
16 Du
16 Dai
16 Cui
16 Choi
15 Smith
15 Cho
14 Wen
14 Tran
14 Tao
14 Su
14 Meng
13 Zou
13 Singh
13 Peng
13 Ding
12 Yin
12 Ni
12 Kang
12 Jia
12 Guan
12 Fan
12 Chow
12 Chang
11 Yeung
11 Hong
11 Ho
10 Xing
10 Xia
10 Weng
10 Wan
10 Tsang
10 Shan

Most of these names are asian, with the first non-asian names being “Patel” with 18 occurrences, “Smith” with 15, and “Singh” with 13. For this analysis, I considered a student to be asian if their last name has ≥ 20 occurrences (so from Wang -> Nguyen).

This underestimates the number of asians, but it’s close enough for our purposes. Since we’re comparing the honor roll rate between two groups, the split doesn’t have to be exact. We split the population into 3415 asians and 4340 non-asians.

Next, I scrape the list of students on the UW Math Dean’s Honors List. Each term, all students with term average above 87% qualify for Term Dean’s Honors List, and their names are available on the web page.

For a sampling of students on the honor roll, I concatenated the list of all students on the honors list for the most recent 3 terms: Winter 2016, Fall 2015, and Spring 2015. This gives 1459 students (non-unique).

Of these 1459 students, 699 are asian (as defined above) and 760 are non-asian.

With this data, we can tabulate the ratio of honor roll counts to total math undergraduate counts for asians and non-asians:

Note: the ratio is NOT the proportion of asians that make the honors list — it is not the case that 19% of students every term enter the honors list. We don’t know how many students are in school each term (vs on co-op). However, as long as the proportion of students on co-op is the same for asians and non-asians, this ratio still reflects the difference in honor roll rates for the two groups.

From the ratios 0.205 for asians and 0.175 for non-asians, we see that asians are 17% more likely to make the honors list.

Is this statistically significant? Using R, I performed Fisher’s test of significance, which yields a p-value of 0.0064. So yes, the result is significant.