1 of 24

What makes a

“Computer Science Person”?

Minoritized Students’ Sense of Identity in AP CSP Classrooms

Jean J. Ryoo & Kendrake Tsui

February 11, 2020

RESPECT Conference - Portland, OR

2 of 24

Focus on Students’ Perspectives:

ENGAGEMENT

IDENTITY

AGENCY

3 of 24

Research-Practice Partnership

4 of 24

Research Questions

To what degree do minoritized youth feel engaged with computing in AP CSP classrooms, and how does this compare to their sense of belonging in the field of CS?

Do minoritized youth coming from communities historically underrepresented in CS feel that they are, or can be “computer science people”?
What characteristics/features of “computer science people” do minoritized youth believe they can or cannot identify with?

5 of 24

Methods

Ethnography

Weekly observations of 4 high school classrooms (ECS and AP CSP)
In-depth interviews (teachers, students)
5 focal students per classroom

Pre/Post Surveys

Across ECS and AP CSP classrooms in entire district

All Post-Survey responses = NOT linked pre-to-post

861 AP CSP Students
Demographics:

50.5% Latinx
14.3% White
6.6% Black
28.6% Asian
59.4% Male
40.6% Female

Pre-/Post-Surveys (2018-19 school year) = Linked pre-to-post

631 AP CSP Students
Demographics:

59.4% Latinx
16.6% White
4.8% Black
16.2% Asian
57.2% Male
42.8% Female

6 of 24

Example Survey Questions

Open-Ended Questions:

Do you consider yourself a “computer science person”? Why or why not?
Describe a project you made in this class that you got the most excited about. What did you make and why were you excited about it?

Likert-Scale Questions:

I like computer science.
I am interested in learning more computer science either on my own or in school.
I have what it takes to become a computer scientist one day if I want to.
If I wanted to pursue a career in computer science, I would be readily accepted by people in the field.

7 of 24

Analysis

SPSS:

Within racial/ethnic groups: two-sample t-tests to determine statistical significance between male vs. female respondents
Calculated coefficient of variation between male vs. female responses to ensure that p-values were not impacted by variance around the mean; there was little to no difference in the variation from the mean for males vs. females in different racial/ethnic groups
Conducted analysis of variance (ANOVA) to compare racial/ethnic groups

Racial/ethnic groups = all-inclusive (mixed-race students counted in all groups)
MaxQDA to code open-ended responses

2 rounds of coding

Kendrake + Jean

For open-ended questions, we used MaxQDA to analyze students’ explanations for why or why not they considered themselves “computer science people.” This was conducted through two rounds of coding. The first round of coding surfaced specific codes that fell into larger themes. For example, codes such as “likes computer science,” “likes programming,” and “likes technology” fell under the larger theme of “likes CS/technology.” This first round of coding helped identify potential “outliers” among students’ responses that did not fall under clear categories. In the second round of coding, we looked more closely within categories to identify shared meaning across answers, as well as which responses carried nuance that required further investigation. In particular, close attention was paid to responses in which students used words such as “but” or “however” that signified ways youth wanted to express their engagement with CS despite not identifying as CS people.

8 of 24

Overall Positive Engagement with Computing

All racial/ethnic and gender groups had positive views of CS
No statistically significant differences between racial/ethnic groups
Some statistically significant differences between females/males within racial/ethnic groups

FINDINGS

9 of 24

Race / Ethnicity	“I like computer science.” (1-10 scale)
Race / Ethnicity	Gender Identity	Mean Likert Score	P-Value
Latinx	Male (n = 199) Female (n = 176)	7.82 6.64	< 0.001
Asian	Male (n = 63) Female (n = 39)	8.16 6.97	0.001
White	Male (n = 71) Female (n = 34)	7.69 6.79	0.066
Black	Male (n = 28) Female (n = 21)	7.14 7.38	0.719

10 of 24

Young men identify more as “CS People” than young women (post-survey)

Race / Ethnicity	Do you consider yourself a computer science person?
Race / Ethnicity	Gender Identity	Yes (%)	No (%)
Latinx	Male (n = 199) Female (n = 176)	132 (66%) 63 (36%)	67 (34%) 113 (79%)
Asian	Male (n = 142) Female (n = 70)	86 (61%) 29 (41%)	56 (39%) 41 (59%)
White	Male (n = 72) Female (n = 34)	41 (57%) 12 (35%)	31 (43%) 22 (65%)
Black	Male (n = 28) Female (n = 21)	10 (36%) 11 (52%)	18 (64%) 10 (48%)

FINDINGS

11 of 24

Why did students identify as “CS People”?

Love the subject, programming, technology, and/or computers (305 of 439 students; 70%)
Have the knowledge and experience, or skills and ability reflecting a CS identity (137 students)
Have plans to pursue computing in either college or their future careers (27 students)
Believe they can help others with computing or contribute to the larger world with their skills (12 students)

FINDINGS

12 of 24

Why did students NOT identify as “CS People”?

Dislike CS or technology (124 of the 422, ~30%)
However: 10% (n = 35) described really enjoying CS
Prefer other subjects (21% - 90)
CS is too stressful or frustrating (15% - 63)
Still have more to learn (n = 54)
No CS career plan (n = 37)
Think they’re not “good at” computer science (n = 28)

FINDINGS

13 of 24

Not a “CS Person” yet...

22% (91 students) used words such as “but,” “however,” or “although”

“I don't consider myself a computer science person YET…there are several things I have to and want to learn.”
“Although computer science is an interesting topic for me, I find myself lacking in this field…If I continue to study and engage in this field, I will definitely learn a lot more and someday consider myself a computer science person.”

FINDINGS

14 of 24

Sense of Belonging & Identifying as “CS People”

Correlation between engagement with CS and feeling welcomed in the field if they pursued it

Not as clear correlation between engagement with CS and feeling like people of one’s own race/ethnicity or gender do CS

Correlation between engagement and growth mindset:

"Even if a computer science problem is challenging and I may not have support from teachers and peers to solve it, I still want to work through it"
"No matter who you are you can improve your computer science abilities"

FINDINGS

Kendrake

NOTES:

correlation values reveal that lower engagement with CS for Latina, Asian, and White females is related to the fact that these same groups of young women did not feel that they would be accepted in the field if they chose to pursue CS; but not as clear correlation with feel as strongly that their race/ethnicity + gender do CS, or (rating near neutral among all groups). → people doing CS race/gender not as correlated with personal belief of acceptance

Latina, Asian, and White females were also less likely to consider themselves CS people than their male peers

Meanwhile, their male peers who rated a higher agreement to feeling like their race/ethnicity + gender do CS and that they would be welcomed into the field if they pursued it, had higher levels of engagement with CS.

While there was no statistically significant difference between Black male and female responses to statements about CS belonging, we wonder why Black males, on average, agreed less than Black females about being welcomed in the field of CS (rating the statement just above neutral, similar to Latina, Asian, and White females). Also, why were Black males less likely to consider themselves CS people than Black females?

"I have what it takes to become a computer scientist one day if I want to" is most correlated with "Computer science ability is something you either have or you don't have, and there is not much you can do to change it." This is partially correct (it was only true for Latinx groups). Upon further evaluation, the the initial question was also highly correlated with

15 of 24

Differences between young Black men and women

No statistical significance & numbers were small
HOWEVER:

Black males, on average, agreed less than Black females about being welcomed in the field of CS (rating the statement just above neutral, similar to Latina, Asian, and White females).
Black males less likely to consider themselves CS people than Black females

16 of 24

Key Implications

1) Discrimination in CS → Solutions to disrupt power differentials

2) Supports for students who enjoy CS but do not think they can yet be considered CS people

3) Stress, frustration, persistence, struggle

4) Connecting CS learning to personal passions and visions for the future

Jean + Kendrake

1) CS classrooms should openly discuss issues of discrimination in the field of computing, and ways that youth can disrupt learning and career contexts that do not welcome their intersecting identities of race/ethnicity, gender, etc. We must ensure that youth, who do not feel like they would be welcomed into CS fields, find ways to counter discrimination and build new communities. We must find ways to discuss difficult issues without discouraging students, while providing supports/resources/encouragement for youth who like CS but don’t feel they are welcome in the field.

2) CS classrooms should openly address the feeling of stress and frustration that youth face when solving challenging computing problems. We must encourage youth to persist and accept that they don’t have to know the answers immediately all of the time. How can computer science be demystified so that youth can see that even expert computer scientists struggle too?

3) We must encourage youth to see connections between CS learning and their personal passions and visions for the future. For youth who do not feel like there are connections between what they learn in AP CSP and their academic or career pathways, we must make learning more relevant so they can feel empowered to be computer science people.

4) We must build better supports for youth who are on the boundaries of computing—students who enjoy CS but do not think they can yet be considered CS people.

17 of 24

Check out Student Voices on Film: http://csequityproject.org /real-cs�

Jean J. Ryoo & Kendrake Tsui

(jeanryoo@ucla.edu) (tsui.kendrake@gmail.com)

18 of 24

TABLES - EXTRA SLIDES

19 of 24

Race / Ethnicity	People with my same racial/ethnic background do computer science.
Race / Ethnicity	Gender Identity	Mean Likert Score	P-Value
Latinx	Male (n = 190) Female (n = 166)	6.52 6.24	0.383
Asian	Male (n = 61) Female (n = 37)	7.70 6.76	0.045
White	Male (n = 71) Female (n = 32)	7.66 7.16	0.398
Black	Male (n = 26) Female (n = 20)	6.31 6.50	0.834

20 of 24

Race / Ethnicity	People with my same sex or gender do computer science.
Race / Ethnicity	Gender Identity	Mean Likert Score	P-Value
Latinx	Male (n = 190) Female (n = 164)	7.16 6.67	0.117
Asian	Male (n = 62) Female (n = 37)	8.29 5.84	< 0.001
White	Male (n = 72) Female (n = 32)	8.15 5.94	< 0.001
Black	Male (n = 26) Female (n = 20)	8.46 6.90	0.033

21 of 24

Race / Ethnicity	I have what it takes to become a computer scientist one day if I want to
Race / Ethnicity	Gender Identity	Mean Likert Score	P-Value
Latinx	Male (n = 191) Female (n = 163)	7.39 6.58	0.004
Asian	Male (n = 63) Female (n = 37)	8.37 6.30	0.001
White	Male (n = 72) Female (n = 32)	8.22 6.38	0.943
Black	Male (n = 26) Female (n = 20)	7.46 7.40	< 0.001

22 of 24

Race / Ethnicity	If I wanted to pursue a career in computer science, I would be readily accepted by people in the field.
Race / Ethnicity	Gender Identity	Mean Likert Score	P-Value
Latinx	Male (n = 190) Female (n = 163)	7.03 5.93	< 0.001
Asian	Male (n = 63) Female (n = 36)	7.35 5.00	< 0.001
White	Male (n = 72) Female (n = 31)	7.92 5.19	< 0.001
Black	Male (n = 26) Female (n = 19)	5.92 7.26	0.676

23 of 24

Race / Ethnicity	“I think computer science is interesting.”
Race / Ethnicity	Gender Identity	Mean Likert Score	P-Value
Latinx	Male (n = 199) Female (n = 175)	8.15 6.86	< 0.001
Asian	Male (n = 63) Female (n = 39)	8.43 7.15	0.001
White	Male (n = 71) Female (n = 34)	8.11 6.85	0.007
Black	Male (n = 28) Female (n = 21)	7.18 8.05	0.220

24 of 24

Race / Ethnicity	“I am interested in learning more computer science either on my own or in school.”
Race / Ethnicity	Gender Identity	Mean Likert Score	P-Value
Latinx	Male (n = 199) Female (n = 174)	7.16 5.22	< 0.001
Asian	Male (n = 63) Female (n = 39)	7.75 6.10	0.001
White	Male (n = 71) Female (n = 34)	6.54 5.62	0.123
Black	Male (n = 28) Female (n = 21)	6.37 6.76	0.650