"Not my Priority:" Ethics and the Boundaries of Computer Science Identities in Undergraduate CS Education

Researchers in the CSCW community have long problematized the separation of social and ethical considerations from design work. Despite increasing attention to tech ethics and ethics education, however, computer scientists' sense of ethical responsibility remains of concern. This paper offers insights on how this boundary between tech and ethics is maintained and reinforced for students as they develop their identities as computer scientists. Drawing on interviews with eight undergraduate computer science (CS) students at McGill University, we explore the role that ethics play in the legitimate peripheral participation of students inside and outside their formal education. We found that while individual opinions on the importance of ethics varied, students agreed that ethics are not valued or rewarded in their education, extracurriculars, or future work prospects. We describe how placing ethics outside the boundary of computing acts as a form of occupational closure, excluding both important multidisciplinary work and marginalized bodies. We argue that in order to promote ethical practice in the design of CSCW systems, we must make it in the interest of future designers to learn socially grounded ethics. This requires that designers, researchers, and future employers actively reshape the boundaries of computing by asserting social and ethical considerations as values of computing and design.


INTRODUCTION
Computing fields have historically crafted their claim to legitimacy by framing themselves as a natural science, separate from human and social sciences [38,41,91].Computing is typically presented as mathematical, logical, abstract, and disembodied as opposed to practical, physical, and contextual [21,121,130].Scholars in computer-supported cooperative work (CSCW), humancomputer interaction (HCI), and science and technology studies (STS) have long critiqued this separation, and advocated for the development of skills in social science and humanities alongside technical skills [38,41].Researchers in the CSCW community have also been pushing for this change via frameworks such as Value Sensitive Design (VSD), Critical Design, and Reflective Design Fig. 1.Graph depicting the change in student citizenship (defined in terms of attitudes on ethics and cultural inclusion) for different college and university majors after four years of college, as found by Nuñez et al [95].We have re-graphed their data; the source uses a vertical bar graph [95].More than any other major, the commitment to global citizenship in Computer Science decreased after four years of education.Engineering has the second most negative shift.Education, Social Science, the Humanities, and Math/Statistics have the most positive shifts.[5,28,56,92,113].However, the scientific and rationalist worldview remains dominant in many areas of CSCW and computing [6,21,121].
In order to address this, an increasing amount of work both in and adjacent to CSCW has focused on the integration of ethics in Computer Science (CS) education (see, for example [24,34,44,47,57,73,101,115]).Education is a valuable site for shifting what it means to be a computer scientist, as it is a space in which future professionals form their identities and ideas about the profession 1 .A majority of undergraduate CS programs in the United States now include a course dealing with the "social and ethical implications of computing" [101].There have also been increasing calls for the integration of ethics throughout the CS curricilum [89,101].Yet, undergraduate CS programs continue to perform poorly in instilling students with a sense of ethical responsibility.A 2021 study looking at 120 colleges and universities in the United States found that CS students' attitudes on ethics and cultural inclusivity worsened over the course of their degree [95].
If we are to succeed at building social and ethical considerations into computing culture, we must first understand why and how these considerations continue to be placed outside the boundaries of the field.Key to this is understanding why certain values and ideologies are adopted by future practitioners as they develop their CS identities, a question which remains underexplored.Prior research in CSCW has worked to uncover the histories and structures behind existing social norms, values, and ideologies in computing and engineering worlds [4,6,39].This paper contributes to this body of work by examining how the boundaries between ethics and computing are maintained through the process of CS identity formation in undergraduate CS education.
We draw on interviews with 8 undergraduate students at McGill University, which focused on their experiences as CS students, their interest in studying ethics, and their perceptions of what makes a "good" computer scientist and computer science student.Unlike previous work on ethics in CS education, our goal is not to evaluate students' level of ethical reasoning, which has been shown to be both poor [130], and a bad predictor of ethical behaviour [11].Instead, this paper discusses how the student participants viewed the role of ethics in building a credible computer science identity, and how they saw ethics valued both inside and outside their education.Our work also draws on prior work on CS identity formation which has found that CS students have a sense of what the ideal computer scientist is, and compare themselves to it [21].This idealized, normative conceptualization of what it is to be a CS (student) reproduces hegemonic norms about CS [21].
Our study is exploratory, and seeks to contribute new theory to address why ethics continue to be kept outside the boundary of computing fields.We offer insights on factors which inscribe ethics as something that is "not [a] priority" for CS students, keeping it outside the boundaries of CS.Given CSCW's long-standing interest in promoting ethical and equitable practices in design and design research (see e.g.[19,22,36,49,52,53,108,134]), this paper brings attention to current barriers to the adoption of ethical design frameworks in and beyond CSCW.Also at stake are best practices in research and design, approaches to ethics policy, and collective understandings of what it means to be a designer.

BACKGROUND
2.1 Values, Boundaries, and Occupational Closure A branch of CSCW research has focused on uncovering the ideologies, mythologies, and structures of power which underlie the values in design and computing worlds (e.g.[42,55,77,78,84,85,93,109,110,123]).Notably, Lucy Suchman has highlighted the political and situated nature of technology and design practice [120][121][122].Paul Dourish has explored how the mythology of ubiquitous computing exploits and reproduces power structures, and shapes how designers approach this research [39].Morgan Ames has problematized the gendered and Western assumptions behind constructionism, the dominant learning theory in educational technology research within CSCW and HCI, as well as in computer science education [4].
The gendered nature of values in design in particular has long been considered among feminist HCI scholars in the CSCW community [12].Daniela Rosner has produced a gendered history of the pillars of design tradition -individualism, universalism, objectivism, solutionism -and demonstrated how these limit who can be a designer [105].In other work, Ames and Rosner have examined power structures inherent in the ideologies behind these and other values upheld in CSCW, such as passionate interest in CS [4,6].
As these scholars illustrate, values upheld in computing work to create and maintain symbolic boundaries which determine who and what belongs within computing fields [80].In contrast to physical boundaries, symbolic boundaries are conceptual distinctions and systems of classification that are dynamically shaped as different social groups form, uphold, and contest different definitions of reality [80].The creation of these boundaries can be explained by the theory of occupational closure, as conceptualized by Anne Witz [132].Witz's theory of occupational closure builds on the theory of social closure by Frank Parkin, which describes the process by which social groups create boundaries which determine who and what belongs within respective groups [96].According to these theories, privileged groups maintain their position by adopting strategies which exclude subordinated groups (exclusionary closure).On the other hand, subjugated groups may attempt to enter these boundaries in order to access the resources of privileged groups (usurpationary closure) or increase their status by excluding other subordinated groups (dual closure) [97,132].
Occupational closure theory [96,132] has been found useful for explaining how scientific occupations [60,132] and computer science specifically [97,124] have walled themselves off from feminized topics and skills in order for an occupation to gain status and prestige.Elizabeth Patitsas has illustrated how the shift in focus towards math and engineering in CS education over the 1980's and 90's enacted exclusionary closure [97].By positioning computing in a way that would increase the field's status as a "hard" science, it closed the field off to women who were more likely to enter the field through subjects such as information systems, business, and new media [97].This trend was compounded by multiple boom-bust cycles in student enrollment [97].Periods of high enrollment were repeatedly followed by exclusionary policy aimed at reducing enrollment, leading the participation of women in CS programs in the United States to drop from 35 percent in the 1980s, to just 17 percent in the early 2000s [97].
In our paper, we are interested in the framework of norms which shape both the practices of designers and designers-to-be, as well as who is able to become a designer in the first place.We contribute to prior work in this area by looking specifically at how the boundary between ethics and CS is maintained in current undergraduate CS education.Drawing on the theory of occupational closure, we examine the relationship between this boundary work and processes of social exclusion in computing and design practice.By boundary work, we refer to the process in which individuals and groups mobilize to define the boundaries of a field -whether it be to subvert/alter them, or to maintain the status quo [81].

Identity and Recognition
Another related area of research is the work on credibility and identity formation in Science Education.In order to develop a "science identity," students must be recognized by others and themselves as a credible science person [9,10].Building a credible science identity requires that students have access to various forms of science capital -that is, economic, social, and cultural capital such as having family members in science, having access to scientific knowledge outside of school, or having people to talk to about science -which students can draw on for legitimacy and "resilience" against factors which work to exclude them from the field [7,62].Work in CSCW has similarly studied the processes of legitimate peripheral participation [82] in technological spaces, by which newcomers are introduced to community norms and practices in order to become productive members [25,26,48,106].It is through legitimate peripheral participation that people on the periphery can develop the knowledge to become credible members of a group.
STS scholar Samantha Breslin has noted how students measure their credibility in relation to what she terms the hegemonic computer science personhood -a normative, gendered, heteronormative, US-centric CS identity against which students are judged and judge themselves [21].According to Breslin, one value key to building this identity at Temasek University in Singapore was entrepreneurial passion [21].Similarly, in our previous work, we found that passion -in the form of both feelings of interest and demonstrations of passion (via extracurricular activities such as personal projects, hackathons, and independent learning) -acted as both a hegemonic norm, and a form of cultural capital which students could draw on for credibility in the field [32].
In this paper we draw on the same interviews that were part of the project on passion in computing education.Here we focus specifically on the role of ethics in developing a credible CS identity.

Ethics in CS Education and Design
The CSCW community has long called for ethical, accountable, and socially responsible design and design research (e.g.[19,22,36,45,49,52,53,74,108,131,134]).In order to scale up ethical practice in design, it is necessary to understand what is currently preventing social responsibility from being a widespread value in computing practice.This work seeks to inform efforts to bring socially and politically grounded ethical practice within the boundaries of computing and design.
This paper also contributes to the rapidly expanding body of work on computer science ethics education.Many HCI and CSCW scholars have been contributing to this area, such as Casey Fiesler who has analyzed ethics education in existing CS curricula and syllabi [47,57,101].Rua Williams and others have examined how CS students respond to ethical scenarios [70,128,130].Prior research has also shared approaches to CS ethics pedagogy [34,44,115], reported on pilot studies which integrate ethics into CS education [14,24,73,79], investigated the perspectives of instructors and tech practitioners on ethics education [8,116], and argued for the integration of ethics throughout CS education [46,89,101].We contribute to this work by examining undergraduate CS students' perspectives on ethics through the lenses of identity formation and occupational closure.

METHODS
Our research methods are based in feminist methodology, which has a long tradition in CSCW [3,42,51,54,66,71].As per Marjorie DeVault, feminist methodology has three goals: to make the lives and perspectives of women visible; to minimize harm in the research process; and to provide research which will lead to social change or action beneficial to women [35].In contrast to positivism, feminist methodology is rooted in standpoint theory, which appreciates the sociocultural context and limited perspective from which knowledge is produced [68,69,98].Building on this tradition, scholars like Suchman and Donna Haraway have advocated for located accountability in research and technology production; requiring researchers and designers to acknowledge and take responsibility for their partial perspectives [68,122].Following their practice, we write in first person, and will clarify our positions here [68,122].

Positionality
3.1.1Hana Darling-Wolf (First Author).I am a queer, non-disabled white woman who grew up in the United States, and moved to Canada for my post-secondary studies.At the time of writing, I am pursuing a master's degree in Computer Science at the University of Toronto.At the time of data collection, I was attending McGill University as an undergraduate student studying both computer science and gender, sexuality, feminist, and social justice studies.
My position as a fellow student affected the kinds of conversations I had with students during the interviews.I was familiar with the courses and extracurriculars available at McGill, as well as the course material.Having experienced the CS program myself, I could relate to the experiences of the students and had a feel for the general climate (in relation to attitudes towards ethics) of the program and student body.My experiences as a fellow student in computer science impacted both the questions I asked, and how I interpreted the data.
Other than a course taught by the second author (described in Section 3.2 below), I did not take (nor was I aware of) any courses that addressed the social or political implications of computing within my undergraduate CS education.Mention of social considerations or context in any of the CS courses I took was limited or non-existent.

Elizabeth Patitsas (Second Author
).I am a queer, disabled white Canadian.I am currently an assistant professor at McGill University, jointly appointed to CS and Education, and have only interacted with the CS programme as a professor.I teach two graduate seminar courses with significant socioethical content, and when teaching CS1 have incorporated some socioethical content.
When I was a CS/physics/math undergraduate at the University of British Columbia I took a fourth-year CS and Society course (CPSC 430).It was an elective but a popular one.My peers tended to see it as easy, as a joke.In contrast, the CS faculty that I worked as a TA for considered the course to be an important and vital part of the curriculum.(I am not sure how representative these faculty were given the course was an elective.) The curriculum of the course I took was highly procedural: we were taught different ethical theories and then taught to apply them like algorithms to a variety of scenarios.We were taught about laws that pertain to privacy and intellectual property, and relevant court cases, and asked to assess the legality of different scenarios in Canada vs the USA.The instructor was an enthusiastic Asian-Canadian sessional lecturer who had industry experience but seemed to lack background in philosophy or STS.The course materials contained memorable inaccuracies such as that John Stuart Mill was a social contract theorist.
The curriculum also failed to grapple with structural issues that cause malfeasance.In contrast, when I took a scientific ethics course from another department, we spent weeks on structural issues such as the limitations of peer review, the research grant system, and scientific journalism.McGill is a large research-intensive university, and a top ranking university in Canada.The School of Computer Science is located in the Faculty of Science, but offers bachelors of Arts, Science, and Arts and Science.The curriculum follows a standard path (CS1, CS2, etc.) similar to the curricula recommended by the Association for Computing Machinery (ACM).At the time of the study, there are no required ethics courses, and no CS faculty whose primary research area was in Human Computer Interaction (HCI).
The software engineering program (in the Faculty of Engineering) holds several overlapping courses with computer science, but includes a required course on "Software Engineering Practice" whose description includes mention of social, ethical, economic, and legal issues.The software engineering program also offers an optional HCI course.While there are several courses at McGill which explore socioethical issues in technology, nearly all of them are offered by departments other than CS (such as Communications, Sociology, and Information Studies), and most of them are graduate-level.At the time of this study, undergraduate students enrolled in the Faculty of Arts and Science (including students majoring in Cognitive Science) were required to take a course titled "Scientific and Technological Controversies" taught by Gabriella Coleman (who has since left McGill).The second author of this paper teaches two graduate seminars with significant engagement with CS ethics (a course on STS for CS graduate students, and a course on CS education that explores the sociopolitical aspects of CS education.)

Interviews
We conducted in-depth, semi-structured interviews [112] with eight undergraduate students at McGill University.Participants were recruited via an email advertisement directed to students in all undergraduate CS programs (in both Arts and Science).These interviews were part of a larger project investigating discourses of passion and belonging in CS education.Therefore, the advertisement included references to passion and issues of gender, race, and class but no reference to ethics.25 students responded to the advertisement.Participants were selected based on when they responded to the email (with priority to students who responded first).Slight adjustments were made to include more diverse gender representation (the vast majority of respondents were women).Due to constraints on time an resources, we originally asked 10 students to participate.However, two did not respond.
I (first author) conducted the interviews via the virtual meeting platform of the participant's choice.Interviews lasted approximately one hour each.The interview protocol is available in the appendix.In order to assess saturation, I performed a preliminary thematic analysis after each interview.This was important, as many experts on qualitative methodology have argued that it is not possible to determine the appropriate sample size a priori [20,29,43,64,72,102,114].We determined that sufficient saturation had been reached after eight interviews, and therefore did not recruit new participants.By saturation, we mean that new interviews exhibited no new major themes, and responses were consistent with both existing theory and each other.
All participants were currently enrolled as students at time of interview (2020/2021).The interview questions were open-ended, and explored students' understandings of what it means to be a "good computer scientist" or a "good computer science student." Only the following subset of questions mentioned ethics explicitly: • Are ethics an important part of being a good computer scientist?
• Is it important to take ethics courses to be a good computer science student?
• Are you interested in studying ethics?Why?Why not?
• Have you taken ethics courses?Why?Why not?
In the interviews, participants were not provided with a definition of the term "ethics." Although "ethics" has multiple meanings across different contexts and traditions (e.g.values in design, business ethics, professional codes), we use it here to encapsulate concepts such as responsibility, reflexivity, social impact, and social justice [47,107].Given increasing talk of "tech ethics" and ethical scandals among tech companies, we expected "ethics" to evoke the social implications of technology and design in a way that would be most familiar to CS students [47,107].While we personally advocate for considering ethics from a justice lens, participants were not limited in their definitions.(A justice lens evaluates systems with attention to systemic power imbalances, as opposed to a fairness lens which fails to consider how unfair situations came about in the first place, or how notions of "fairness" are socially and politically constructed [16,130].) Each interview was digitally video-recorded and/or audio-recorded according to the student's request.I (first author) also took hand-written notes during the interviews, and conducted an initial analysis of the general themes appearing in these notes after each interview.The interviews were then transcribed verbatim using the automatic transcription feature on Microsoft Stream, followed by a manual correction of each transcript.The transcripts were de-identified, and each interviewee was assigned an alphanumeric code (P1-P8).Video-recordings were used solely for convenience in recording and transcribing, and were not analyzed.

Qualitative Analysis
We then performed a qualitative content analysis [50,90] on the transcript data in order to draw out commonalities across interviews.We manually coded each transcript, collaboratively with the help of members from the Social Studies of Computing research group.The categories were developed inductively, with our codes organized under broad themes after the first round of coding.They were then placed on a virtual whiteboard and thematically analyzed again through affinity mapping with the help of members of the lab.The final themes were determined via several additional rounds of coding and discussion, until a consensus was reached.
Given our methods and sample size, our goal in this paper is to achieve analytic generalization and reader generalization.Reader generalization (or transferability) involves providing detailed descriptions that allow readers to extrapolate findings to other settings, or compare them to their own experiences [100].We provide raw data and thick description in order to allow the reader to draw their own conclusions about the data, and how these findings may apply to other settings [58].Quotes were lightly edited for clarity (e.g.removing filler words).Analytic generalization is when findings are generalized from particulars to broader constructs or theory [100].In order to achieve this, we place our findings in the context of prior work and broader theoretical constructs.We do not make any claims to statistical generalization, which is the extrapolation from a sample to a larger population.
This study is exploratory and aims to do initial theory building about how students view ethics in relation to their identity as computer scientists and their role in the field.Follow up work will expand on this study in order to verify whether these findings apply more widely to the student body.

Participant Demographics
To collect demographic information, we asked participants to self-identify any identity factors they would be comfortable sharing (e.g.gender, race, sexuality, etc.).Five participants were women, two were men, and one identified as queer (and later male).With regard to racial identity, our participants self-described as white (4), Black (1), Latinx (1), Asian (1) and mixed (1).The four nonwhite2 participants were all women, and the remaining participants (one woman, two men, and one queer person) were white.Half of the participants were international students.One participant self-identified as disabled.
Year numbering at McGill University is complicated because some students attend a CÉGEP (collège d'enseignement général et professionnel) before attending university, which includes the first year of post-secondary education.Years are therefore labelled between U1 and U4+, where students who have attended at CÉGEP start in U2, and other students start in U1.One student was in U1, two students were in U2, three students were in U3, and one student was in U4+.Only one student was majoring only in computer science.Double majoring and adding minors are institutionally straightforward and are popular amongst the student body.Three participants were CS majors with minors in other subjects.Another three participants were double majoring CS with another subject.The final participant was majoring in software engineering rather than CS.

Anonymization
Deidentified transcripts were coded collaboratively with the help of the Social Studies of Computing lab.Given only the identity factors and year of study of a participant, one lab member was able to confidently guess one of the deidentified participants.Without confirming or denying the guess, the lab member was recused from coding this interview.We remain cautious in how we communicate participants' races and genders in this paper.Therefore, we have refrained from including a table detailing their specific intersectional identities.
One participant mentioned taking a course taught by the second author.The second author recused herself from the coding of this interview transcript and saw only the upper-level themes that emerged from it.
"Not my Priority:" Ethics and the Boundaries of CS 174:9

Broad-Ranging Conceptualizations of Ethics
Before we discuss how our participants viewed the role of ethics in developing a computer science identity, we want to briefly provide some background on how our participants interpreted what "ethics" means.
4.1.1Levels of Exposure to Ethics.Although no students had taken ethics courses specific to computer science, two students were reminded of "general ethics" philosophy courses they had taken in their secondary education (P1, P4).P6 had taken the Arts and Science course mentioned in section 3.2.A couple students expressed feeling that they did not know much (P8) or enough (P6) about ethics.One international student, P7, looked up a translation of "ethics" during the interview.

Examples of Ethical Concerns.
Participants were not explicitly asked to define the term or provide examples of ethics.However, the interviews demonstrated a broad range of conceptualizations and levels of exposure to ethics.Examples of ethical concerns related to computing mentioned by students included: algorithmic bias (P1, P4), security (P1), inclusion (P1), social media algorithms and data privacy (P2, P6), AI ethics (P1, P4, P8), and black hat hacking (P7).According to P8, the "basic example for ethics" had to do with trolley problem-type decision making for self-driving cars: "The only contact I had with [ethics] was auto-driven cars and [...] if there was an accident [...] whether you had to sacrifice the one controlling the car or the one on the road.I think that's the basic example for ethics." (P8) No students mentioned the military history of computing, the environmental effects of computing, or the ethics of refusing to build harmful technology [1,105,130].
4.1.3Morality vs. Social Impacts.Some students' definitions of ethics centered individual intention and morality (P7, P8).Other students related ethics to questioning the benefits and effects of tech on others during the design process (P1, P2, P3).Several students also tied ethics to being exposed to perspectives from other fields (P3, P4, P5).
Conceptualizations of ethics were thus broad-ranging in both individual conceptualizations, and across participants.Our goal in this paper is not to evaluate students' comprehension of ethics, or ethical decision making (as others have elsewhere, see e.g.[11,70,128,130]).However, these examples do start to point to the differing ways students conceptualized the role of ethics in CSe.g. as moral and philosophical decisions during design, or already systematic issues in the field.The following sections address how students view the role of ethics in computer science more directly.
4.2 Are Ethics Needed to be a Good Computer Scientist?Two Views.When asked whether ethics are an important part of being a good computer scientist, students unanimously agreed that ethics are important to consider as a computer scientist.However, this question prompted different views on the boundaries of computer science.As the following subsections will address, two views prevailed: (1) While some students saw ethics as part of doing good computer science, (2) Others made a distinction between doing good computer science, and doing good or bad things.

4.2.1
Ethics are Important for Being a Good Computer Scientist.Most students interviewed agreed that ethics are an important part of being a good computer scientist.In fact, some students highlighted the importance of ethics (or doing computer science "for good") before they were explicitly prompted about ethics (P2, P3, P6).In their descriptions of a good computer scientist, P2, P3, and P6 noted that it is necessary to ensure that one's work is ethically good: "I would say another aspect of a good computer scientist is [...] somebody who's able to look at the ethics of their project." (P3) "if we're gonna make an app for someone one day we should be able to tell is that app, I don't know, ethically good." (P6) "I don't think you can be doing computer science solely for the purpose of computer science, but for the purpose of how it can be used in good ways." (P2) Whether or not they had a clear conceptualization of ethics or what it might mean to do CS for "good," P2, P3, and P6 already included ethics and "good" use in their imagination of a good computer scientist.P2 tied doing computer science for good not only to being a good practitioner (computer scientist), but to the practice itself (doing computer science).P2 further added that being a good computer scientist means considering how one's work affects others: "Being a good computer scientist to me, I think, involves how what you're doing and what you're learning will be used for, and how those affect the ones around you. [...] it's important to realize that there are so many people around you who don't understand computer science yet they're being highly affected by [...] how it's being used." (P2) When asked later whether ethics are important for being a good computer scientist, P1 and P3 similarly highlighted the importance of questioning one's work, and the potential harms and benefits it could produce: "to be a good computer scientist, definitely it would be good to at least give yourself the chance to think about what you're doing.Why are you doing it and how can it be used and how it will affect a population?[...] there's so many questions that you have to ask yourself before you create something." (P1) "I think especially now as technology is developing at such a rapid rate, there's the ability, obviously, to develop technology to harm people [...] I think it's important to ask yourself why you're developing this technology, who it's going to aid, who it could potentially hurt and weighing in all of these questions.Yeah, I think people who don't question the ethics of their project should not be developing at all." (P3) P4 also agreed that ethics are important in order to be a good computer scientist, but felt that this applied more for certain fields than others: "[First author: Are ethics an important part of being a good computer scientist?]Yes, 100%, yes.It's going to be, I guess, more or less important on the field, but people who go into anything involving algorithms ethics, fairness and understanding like bias in data sets and ways that things could go south, or just understanding what could cause the algorithm to make a biased decision and what is a biased decision, I think that's really super important." (P4) P5 also offered a more nuanced view of ethics.
"I think ethics are important in computer science, but it is so easy to lose your ethics for money or just for convenience or fitting in [...] like even for myself, when I was working at a team and I didn't agree with some of the things that were happening [...] there were moments where it was just like, there's nothing us people of color can do to convince them, and we kinda just have to give up [...].[M]aybe if I had had a higher paying salary [...] I'd just be like, guys let's just get it out [...].[T]here's always like a price to peoples' decisions, and a lot of people can be bought off, and a lot of people also just don't care about the ethics of computer science." (P5) Here, she highlights multiple structural factors which limit computer scientists' -including her own -ability or interest in doing ethical work.According to P5, both financial factors and social structures (in this case racial hierarchies in the workplace) influence the reality of ethical practice in the tech industry.
Nonetheless, a majority of students felt it was important to consider ethics in order to be a good computer scientist -with P2 even bringing into question the boundaries of computer science itself.

Differentiating CS Skills and Ethics.
The second view placed ethics outside of the boundaries of CS identity.Although P7 agreed initially that ethics are important to be a good computer scientist, P7 and P8 -unlike P2 -made a distinction between computer science skills and moral decisions.P7's ethical concerns lay in the fear of very skilled computer scientists doing bad things (like black hat hackers hacking a hospital to get bitcoin (P7)).P8 similarly drew a line between being a good programmer and doing bad things.For P8, this brought into question whether ethics should be considered part of being a good computer scientist at all: "How would you define a good computer scientist?Is it someone that can write good code?[sighs] This is complicated.I don't know if you should link having ethics with being a good computer scientist.I would not link the two.[...] let's say you have a really good computer scientist, but he's doing bad things, like for example [...] there's been like hacking of, I think Chicago city two years ago -and they ask for ransom [...].Can you consider those people good programmers?They're certainly skilled, but what is the meaning of good here?If it's skilled, then obviously yes, and if it's good like you're doing good things, then obviously no.So it's all about terminology, but I think I would link being a good programm[er] with being skilled and interested.Then obviously poorly intentioned people are skilled and interested in what they want to do, but it's not good things.So that's why this is a complicated question for me.I think I would qualify them as good programmers, even though they're doing bad things." Although P8 hesitates in his categorization of ethics' role in computer science, he ultimately concludes that being a good computer scientist meant being skilled at writing code.Writing code, in turn, was not itself an activity which required ethical consideration.Rather, ethics was something an individual "programmer" (used here interchangeably with "computer scientist") had -that is, a system of personal morals.Thus being a good computer scientist did not include considering ethics (i.e.having good intentions).
These views were reflected in P8 and P7's views on the role of ethics in CS education.In order to understand how students' views on ethics in good computer scientists reflected on their idea of good computer science students, we asked participants if they thought that it is important to take an ethics course in order to be a good computer science student.P7's response was informed by her understanding of ethics as personal morals, leading her to see ethics courses as unnecessary: "I don't think so.According to P7, ethics are intuitive, and a matter of conscious intention.P8's response was similarly guided by his understanding of ethics as morals separate from computer science: "I think it's important to take an ethics course.But I'm not sure that you have to take one in order to be a good programmer." (P8) 4.2.3Valuing Ethics is Uncommon.While other participants diverged from P8 and P7's view that ethics are separate from being a good computer scientist, a few noted that their own understanding of ethics as part of being a good computer scientist differed from that of other students and how they were being taught.P2 noted the "lack of moral compass" among students in CS: "Being a good computer scientist to me, I think, involves how what you're doing and what you're learning will be used for, and how those affect the ones around you. [...] I think a lot of computer science students don't really think about that.There's a lack of a moral compass involving technology." (P2) When we asked P6 whether ethics are important to be a good computer scientist, she answered that they should be, but are not: "I think it should.I don't think it currently is, but I think it should.Definitely, it's not at all actually right now.
[Well], the way we're taught.But it should." (P6) These comments suggest that viewing ethics as important for being a computer scientist was not a widely held view.P6 brings forth another important observation: that ethics were not valued in their education.
4.3 Ethics are Neither "Useful" Nor "Fundamental" to CS As we illustrate in the previous section, students had split views on the role of ethics in being a good computer scientist.One thing our participants agreed on, however, was that ethics were not presented as important or useful in their education.

Ethics are Not
Important for Being a Good CS Student.When asked whether ethics were important to be a good computer science student, P6 responded that ethics are not important for students' success in the field: "No.You can be the best computer scientist and not give a shit about ethics and you'll be an asshole but you will do great." (P6) Likewise, for P1, ethics were not important for students' success in the university, since ethics were not rewarded by instructors: "I think, maybe not for students.The students are kind of also trying to be guided by the University or the teachers" (P1) This suggests that students did not see ethics as something that was valued in their education.

It is Neither Possible, Nor in Students'
Interest to Take Ethics Courses.Students highlighted several factors which contributed to the sense that ethics were not necessary for being a good computer science student.Most notably, the lack of ethics courses at McGill: "First author: And is it important to take an ethics course to be a good computer science student?P3: You know, it's funny, I would say yes, but I don't-McGill doesn't offer anything like that, which is funny.Kind of have to go looking elsewhere for that." (P3) "I wouldn't know where to take a class [...], but something that I do is to get informed as much as possible -talk with people, try to ask myself questions." (P1) "To be a good computer science student?No.I think it would be interesting though -I'm a big fan of mixing ideas and being exposed to different things, so I would support the idea of including an ethics class in computer science curriculum." (P4) Thus, in spite of their interest in taking ethics courses, and their openness to learning about the social consequences of technology, students are left to find resources outside of school to educate themselves.Rather than incorporate different forms of knowledge that students can draw on (as students advocated for, see section 4.1), McGill limits computer science to "hard" math and computing knowledge: "I'm surprised [ethics courses are] not a mandatory requirement for computer science students.In fact, the only requirements are math and computer science courses essentially, which just seems a little outlandish because sure, those classes teach you the necessary hard knowledge you need to know.I think there needs to be some translation or application into the real life consequences or moral standards of it all." (P2) This quote points to another important reason ethics were not viewed as important for computer science students: course requirements."First author: And have you taken any ethics courses?And why or why not?P2: No, I haven't.Because I don't find I have the room to.Or haven't thus far.[...] I would really love to." (P2) "I'm interested [but] since I'm really short on credits for my minor [...] I'll be short on credits overall, but no, it is really something I'm interested about.And when I see opportunities to learn about it, I do, but I don't, you know, read ethics books on my own.But if I had a mandatory ethics class to take, [...] I would look forward to it." (P4) Once again, students must find their own opportunities to learn about ethics outside of school, even when they are interested in taking formal ethics courses.For P8 and P7, these practical reasons combined with their understanding of ethics as separate from more "useful," "fundamental" CS knowledge (or as P2 said earlier, "necessary hard knowledge you need to know" (P2)): "First author: Are you interested in studying ethics?P7: [...] I'm not sure... [...] if there's an ethics course about computer science, I don't think I'm gonna take it.'Cause for international students the tuition is very high and I can spend my money on the more useful stuff." (P7) "I don't know what courses are linked with ethics.And second of all, it's not my priority for now, like my... right now is to learn the fundamentals of computer science.So, the ethics part of it will definitely come after." (P8) P8 and P7 note the lack of courses addressing ethics in CS at McGill.However, even if ethics courses were available, these quotes highlight how McGill's curriculum currently places ethics outside of computer science and presents ethics as not "useful" or not a priority (something that can "come after").

Ethics
Education is Lacking and Depoliticized.Despite a lack of ethics courses, some interviewees suggested that integrating more ethics into their education would not be so difficult.P1 and P2 in particular gave examples of courses which failed to (sufficiently) address ethics when relevant: "In everything related to technology, I think ethics are very important.[...] I think that's something that in McGill, it's been lacking in the academic side.There's other groups that have been trying to tackle this problem with their own resources, but in class I don't think I've heard any professor telling us about the ethics of what they were teaching [...].For example, I took a computer vision class, which is very controversial nowadays because [of the] bias there is in models and, the professor never even talk [ed] to us [about it]." (P1) "the closest I had [to ethics within a course] was in my robots and artificial intelligence class.[...] I think he just mentioned it in passing, so I've never been tested on [it].But I think he brought up how robots are affecting society.But not from a good point of view or bad point of view, mostly just objective." (P2) The most exposure P2 reported as having in a CS course was that "objective" comment "in passing" on the effect of robots on society.For P1, ethical considerations were entirely missing from their courses -even when dealing with models which are prone to bias.Thus, ethics were not integrated into existing CS courses at McGill.When the social impacts of technology were addressed, they were presented as "objective" facts, rather than opportunities for critical reflection.This suggests that even in discussions of ethics, CS courses at McGill problematically reinforce the view of CS as "objective" and unimplicated in the social impacts of design -leaving "other groups" to "tackle this problem with their own resources" (P1).

Marginalized Students Carry the Burden of Ethics Education
As noted in the previous sections, the participants' responses indicated that ethics are not currently prioritized in undergraduate CS education at McGill.However, if students do not learn about ethics in their formal education, where do they learn?Moreover, who is learning and teaching ethics?4.4.1 Who is Doing Ethics in their Free Time?Despite a lack of ethics in their formal education, several students expressed interest in taking an ethics course (P2, P4, P6).Some students tied their knowledge of ethics to current projects or work they did in their free time: "I'm already involved in some things on the side that involve ethics in machine learning.So I guess that's my way of keeping up to date with that and staying informed.But otherwise I don't have much plans to learn that on my own." (P4) "All my knowledge comes from documentaries, and just like, reading." (P6) P4 states that his work in machine learning involves ethics (although it was unclear in what ways).P6 credits her knowledge to documentaries and reading outside of class.Elsewhere in her interview, P6 mentions the importance of following the news.This indicates that some students build their understanding of ethics from popular and news media outside of their formal education.
However, two students -both women with minoritized racial identities -were particularly involved with ethics in their free time.P1 recounts how she organized a panel on ethics in AI, and was disappointed by the lack of turnout by computer scientists (although she notes that some law students joined the event): "I feel like people don't see the value of it.And, I don't know why, but people don't see-I guess people tend to go more to more technical stuff.It might be linked with this mentality of always adding up to the CV.So a workshop would be more valuable than going to discuss ethics in AI. " (P1) She notes her frustration with other students' focus on the "technical" aspects of computing which contribute to their CV.This suggests that students believe internships and industry jobs similarly prioritize "technical" knowledge and fail to value the ability to question the ethics of tech.
Similarly, P3 worked within an extracurricular club to organize a workshop which invited professionals to speak on ethics in their field: "the only thing that I've done is run a workshop on the ethics in tech where we had four specialists from different fields in technology who described the ethical issues behind their things, which was really, really, really, really eye opening.Although she found this experience valuable, she still brings up a desire to participate in an ethicsfocused course.P5, also a woman with a minoritized racial identity, described her involvement with several projects and organizations related to social justice and sustainability.
Although these were not the only students who expressed interest in or highlighted the importance of social and ethical considerations in tech, they do bring into question who is doing the work of educating themselves and others on ethics, and how this work is gendered and raced.The idea that students with socio-historically marginalized identities are more likely to care about ethics falls in line with prior work that has demonstrated that women in CS are more concerned and motivated by social impacts than men [88].This has implications not only for the distribution of labour tied to social justice and ethics, but also suggests that ethics education (or the lack thereof) has an impact on persistance and belonging for minoritized students in the field.If socio-historically marginalized students are more likely to be doing CS for a "purpose" [76,87,88], then rewarding and supporting this motivation must be part of CS education in order to make it more inclusive [61,127].

DISCUSSION
These findings suggest that ethics is placed outside the boundary of CS at several levels: in students' formal education, in their extracurricular activities, and in their bid for future employment.Students' understandings of ethics were typically shaped by information drawn from extracurricular activities or popular and news media.However, unlike other forms of passionate labour aimed at developing technical skills (e.g.hackathons, personal projects, FOSS, independent learning), time spent on ethics did not grant recognition in the field.While individual opinions on the importance of ethics for becoming a good computer scientist varied among participants, students agreed that knowledge of ethics was not valued by fellow students, the university, or future employers.Although some students expressed interest in taking ethics courses, it was neither in their capacity to do so (due to course restrictions), nor in their self-interest.In other words, ethics did not contribute to legitimate peripheral participation or the formation of a credible CS identity.

Comparison with Prior Literature
5.1.1Employability Shapes Students' Attitudes Towards Ethics.This paper has found that students prioritize technical knowledge over ethics because ethics do not bring them recognition from the university or employers.This finding is supported by Ella Sarder and Casey Fiesler's recent work, which finds that computer science students believe that employers do not value ethics, and as a consequence are less likely to value ethics themselves [107].
Moreover, they describe the "trade-off" between personal circumstances and unethical workas they note, many of the students they interviewed were willing to overlook ethical issues for a high enough salary [107].This view was reflected by participants in this study.For example, P5 expressed that it is "easy to lose your ethics for money."Financial considerations also influenced students' attitudes towards their education: P7 noted that the cost of courses as an international student impacted her perception of which courses were most important to take.The ability to find employment had a heavy influence on students' attitudes towards ethics.However, as Sarder and Fiesler note, this financial trade-off is particularly important for groups underrepresented in computing, who are more likely to feel pressured to secure a high paying job [107].This suggests that who is able to afford to learn and practice ethics is shaped by social factors such as race, class, and cultural background.

Unrewarded Labour: Passion for Ethics is not Passion for CS.
Given the examples of students pursuing ethics in their extracurriculars in subsubsection 4.4.1, some may question whether a formal CS ethics education is necessary.If students are already passionate about ethics and social justice, one may argue, they can simply learn on their own.However, it is important to note that students did not value passion for ethics and social justice in the same way as other extracurricular demonstrations of passion for computing.
Breslin's work has shown that passion for computer science -expressed in the form of independent learning and extracurriculars like hackathons and personal projects -acts as a hegemonic norm which is necessary for employability and credibility in computer science education [21].In contrast to other extracurricular work, free time spent on tech ethics did not contribute to students' sense of legitimacy and employability -as evidenced by the lack of engagement with the AI ethics workshop mentioned in the previous section, in favor of "more valuable" technical workshops (P1).In other words, interest in issues of ethics and social justice was seen as separate from interest in computer science.Choosing to spend time learning ethics meant forgoing time that could be spent building their CV.That is, a passion for ethics did not contribute to the performance of interest for computing that is required for credibility and employment in computer science.
Our findings support prior work which suggests that gender and race influence which students do spend time educating themselves on issues of ethics and social justice [42,87,88].Not only do these students do this labour in addition to their school work, but they must also do it on top of the passionate extracurricular labour perceived by other students as more valuable and necessary for building their CV (unlike work on ethics and social justice).Students with socio-historically marginalized identities are also more likely to have additional social obligations and responsibilities which limit the amount of time they have to spend on extracurricular activities in the first place [21,88].The deprioritization of ethics in both formal CS education at McGill and the labour market thus has serious implications for who can gain belonging and recognition in computer science.

Not
Valuing Ethics Fosters Epistemic Exploitation.We found that the work that is done by students to learn and teach ethics is gendered and raced.In this study, we found that all of the interviewees that provided specific examples of projects they participated in related to ethics or social justice were women with minoritized racial identities. 3In addition, we found that organizing and participating in activities related to ethics was not considered a form of legitimate peripheral participation, suggesting that the work minoritized students did to educate their peers was not rewarded.This falls in line with prior work on epistemic exploitation (that is, placing the responsibility of educating dominant groups about social issues on marginalized people) in computing fields, including CSCW and HCI [42,118].

Placing Ethics
Outside the Boundary of CS is a Form of Occupational Closure.Moreover, prior work has shown that women are more likely to be interested in careers which are perceived to have a broader social impact [23,61,88].If minoritized students are more likely to be interested in ethics, then failing to include ethics in the curriculum reflects a lack of interest in supporting their motivation and persistence in CS.Placing ethics outside the boundary of computer science therefore acts as a form of exclusionary closure.Past work on the history of computing has already documented how CS has repeatedly cordoned off feminized subjects and skills in order to gain or retain access to resources and prestige [31,37,41,124].5.1.5Understandings of Ethics are Vague and Depoliticized.One positive finding was that many participants did express their belief that ethics are important for being a good computer scientist.However, as Ben Green notes in his paper "Good" Isn't Good Enough, recent enthusiasm for doing "social good" in CS has not been followed by a clear sense of what "good" means [63].Rather, computer scientists typically uphold vague and de-politicized notions of "good" without any rigorous foundation in theories of social change [63].Similarly, Daniela Rosner demonstrates in her book, "Critical Fabulations," how designers tend to create tech solutions which reflect their own interests, and not the supposedly "universal" needs of users they often claim to address [105].Current understandings of "good" in CS also tend to promote techno-solutionism and Technological Benevolence -defining the world in terms of decontextualized "problems" and "solutions" to be charitably solved by computer scientists [21,121,122,130].These techno-solutionist 'fixes' typically address only the symptoms of larger social issues, and often lead to more harm than good [1,63,105,133].
These attitudes were reflected among the participants in this study, whose definitions of ethics and CS "for good" were often vague.Some students discussed their commitment to applying tech to social issues, but did not make any mention of questioning whether tech was the best way to address these problems (P3, P5).On the rare occasion ethics were mentioned in students' courses, the social implications of tech were presented as apolitical and "objective" (P2).
As other scholars have argued, computing disciplines must stop only valuing technical knowledge and solutions over knowledge in ethics and social science if computer science is to do actual good (and, crucially, to stop doing harm) [41,63].

The Exclusion of Both Multidisciplinary Work and Marginalized
Groups.The devaluing of ethics in computing education and industry is not surprising given the ways computer science has historically crafted its claim to legitimacy by presenting itself as a natural science, separate from human and social sciences [38,41,91].This falls in line with work from Wendy Cukier et al., who have noted how the demarcations created between ICT fields such as CS, CE, IS, and IT has served to narrow the definition of computer science in a way that excludes both women and multidisciplinary perspectives [31].HCI and CSCW are vulnerable to this exclusion, both for being inherently multidisciplinary and for having a reputation as being a more female-typed subfield of computing [119].In this study, some participants conflated computer science with programming, omitting the crucial non-technical aspects of design [12,30,83,94,105,111].
The promotion of male-typed values like abstract logic and entrepreneurial passion in CS also works as a form of occupational closure, shaping who is able or interested in participating in the field [21,27,75,88].This closure is reinforced by placing ethics outside the boundary of CS.The deprioritization of ethics in CS therefore has implications for both the harm done by technology, and the exclusion of minoritized students from the field.

The Exclusion of Ethics Places an Additional
Burden on Minoritized Groups.This boundary work also has social implications within the field.The students in this study that had the most experience with extracurricular activities centered around ethics were all women and racialized minorities.This finding suggests that marginalized students are more likely to do the extra labour of learning and teaching ethics on top of building the skills necessary for employment and credibility.In other words, placing ethics outside the boundaries of CS increases the likelihood of epistemic exploitation.Bringing ethics into the boundaries of CS is necessary to decrease this burden.

Redefining
The Boundaries of Computing.Making CS less harmful and exclusionary requires valuing knowledge of ethics at the same level as technical knowledge.In doing so, we must be careful to ground the value of ethics in social and political theory in order to prevent our efforts from becoming "nonperformative speech acts" [2] -that is, the substitution of real action with depoliticized verbal commitments, often referred to in tech as ethics washing [126,129].Otherwise, we risk perpetuating the myth that designers are objective and not responsible for the socio-political consequences of their work.Scholars in HCI and CSCW have already done extensive work to create practices which integrate these values (see, for example, work on speculative design, value sensitive design, critical design, and reflective design, among others) [33,40,56,83,105,113,117]. Centering and rewarding these interdisciplinary and critical practices in design is a useful starting place for creating a less exclusionary, and more ethical culture of computing.
Integrating interdisciplinary and socially grounded practices into computer science education is a first step in this direction.But if it is not in students' perceived interest to spend time learning ethics in their formal education or in their free time, it is likely that marginalized students will continue to carry the brunt of this work, while the boundaries of the field remain fixed.

Showing
Students that Ethics Knowledge is Employable.In line with other studies [13,18,59,65,99,125], we found in our own prior work [32] that students perceive their formal training in CS as disconnected from future employability, and hence lacking in utility.As such, students largely draw their understanding of what knowledge is useful for employment from outside the classroom: job interview processes, career fairs, reddit, etc.This information influences students' choices of extracurricular activities.
However, students' perceptions about what skills are important for employability are known to be inaccurate [13,104].In particular, CS students are known to underestimate the importance for employability of "soft skills" such as communication, leadership, design, project managements, requirements, and -yes -ethics [13,103].We recommend that CS educators directly address students' interest in employment, and to speak directly to it to counteract skewed information on social media.
Our findings support extensive prior work which advocates for the integration of ethics throughout computing education [46,89,101] (see also [14,24,73,79] for guidance on how to do so).
What we contribute to this literature is a reminder that students are motivated by perceived employability after graduation.Educators need to connect the their teaching of interdisciplinary and socially grounded computing practices to employability.As is, students are unconvinced that studying these practices are beneficial to their careers.Demonstrating to students how these practices may be beneficial for their careers may convince more students it is worth their time.

5.2.5
We Must All Work to Reshape Computing, Within and Beyond Education.Given students' perception that ethics are not useful for their careers, we must make it in the interest of computer scientists to gain knowledge of ethics from both inside formal education and outside.Changing the central values of a field is not an easy task.As Dourish argues, escaping a "legitimacy trap" requires asserting the goals and values we want to redefine legitimacy in the field [38].
It is up to us -as students, researchers, designers, and other tech practitioners -to assert these values, and assign credibility to other forms of knowledge within our jurisdictions of care and influence [67].CS educators and program directors can advocate for integrating ethics in CS education, as well as interdisciplinary pathways for CS students, and joint programs with fields in social science.Employers can assert these values in job advertisements, job requirements, and interview questions.Tech practitioners can highlight these values in career panels, Ask Me Anything (AMA) posts or sessions, or mentorship programs.Ethics knowledge can be encouraged on reddit or other social media platforms students use to seek career advice.Researchers can also read, cite, and hire scholars in STS, social science, and the parts of HCI which are more grounded in the social sciences.Professionals with sway over hiring decisions (whether in academia, industry, etc) can advocate for hiring more people with expertise in critical computing, STS, and interdisciplinary research.
Workshops can be organized across institutions to familiarize students, instructors, researchers, and designers with relevant background in socially grounded ethics and social science.Finally, creating scholarships for students to learn about ethics and socially grounded computing practices is a direct way to reward these studies.

Limitations, Reflexivity, and Future Work
Because this work was part of a larger project on passion and identity formation in computer science education, the number of questions related to ethics in the interviews was limited.Moreover, these questions appeared towards the end of the interview.The interview with P5 ran long, which meant we were not able to interview her about ethics in as much detail as other participants.
These interviews also took place in the particular context of McGill University.Although an increasing number of computer science programs require ethics course in their curriculum, CS programs at McGill currently have no ethics requirements.McGill, like many other universities, does not have any standard ethics courses for undergraduates.A handful of graduate-level seminars, mostly outside the department, address ethical issues in CS; these would all be electives if taken by undergraduates, would require the undergraduates to know how to enrol into graduate courses, and all of them are offered irregularly and without dedicated course codes (i.e.topics courses).
Although this means that our interviewees did not have any formal ethics training, they were also less likely to feel that their education had properly trained them in ethics, which can be equally problematic [15,101].Indeed, our second author was surprised that only one participant discussed trolley problems, as in her experience CS students conceived of ethics through trolley problems and rendering ethics into procedural processes, in line with Williams' findings on CS student perceptions of ethics [130].We hypothesize that this a result of whether an institution has a CS ethics course: the second author took such a class from her alma mater.In contrast, our participants' institution has no ethics class at all.
Since the time of the study, some new graduate-level topics courses which have ethical content have been offered.The department has hired more tenure-track faculty in areas directly relevant to ethics (e.g.bias in AI, inclusive design), and there has been some discussion about increasing the ethical content in graduate-level machine learning courses.But we know of no effort to reform the undergraduate CS curriculum to incorporate more ethical content, and it is unknown to what extent any changes have been available or evident to undergraduate students.
Future work could investigate how students' conceptualizations of ethics and the role it plays in CS education might change in CS programs with greater access to ethics education, or required ethics courses.
Our study is also shaped by our sample size.This study is exploratory, and seeks to offer preliminary findings in order to support the initial development of new theory.Given our research methods we make no claim to statistical generalizability [100].While we perceive that there are plenty universities like our own, we make no claim that this university is particularly representative of all CS departments.Similarly, while we feel our participant pool gave us a useful view of the student body we cannot assess what proportions of students agree with our participants.Nevertheless, interesting patterns emerged across our interviews.Future work will verify whether these findings apply more widely to the student body.In this study, we aim instead for analytic generalization -that our findings can be generalized to broader theoretical frameworks [100].
We believe our data to be sufficiently saturated for these generalizations.We define saturation here as no new major themes found in new transcripts, and consistency both across transcripts and with existing theory.Comparable projects [29,43], have found saturation after eight and seven analysed transcripts, respectively.Greg Guest et al. found that metathemes were present by the sixth interview, and full saturation occurred between seven and 12, with new themes emerging infrequently and asymptotically after 12 (they checked for saturation in increments of six) [64].However, extensive research on qualitative methodology has argued that it is not possible to determine sample size a priori, and resists the idea that there is a minimum quota for saturation that can be applied across qualitative research [20,64,72,102,114].Rather, saturation is dependent on factors such as sampling methods, population size, population homogeneity, and research question(s) at hand.Some work has even brought the usefulness of the term saturation into question [20].
Information power has been offered as a possible alternative to data saturation [20,86], which has researchers assess if the information they have collected is sufficient to answer their research question(s).Studies with larger information power hence have lower requirements for sample size.Factors which contribute to information power include study aim, sample specificity, theoretical backing, quality of dialogue, and analysis strategy [86].Our study is well supported by established theory, our interview dialogues were strong, and our aim was for an exploratory and in-depth analysis of narratives and discourses.Therefore, we believe our sample size to be appropriate relative to its information power and goals.
Because of our concerns surrounding the anonymity of our participants, however, we do not offer a detailed intersectional analysis in this paper.Given that our study was not designed for an intersectional analysis, we are concerned about overgeneralizing the experiences of our participants (e.g.our sole Black woman participant).While we identify trends across our participants along the lines of gender and race, future work is needed to provide a true intersectional analysis with appropriate sampling and interview protocol.We believe that studying how CS students understand the boundaries of the field would benefit from further work that attends to the intersections of gender, race, nationality, and disability.Only one participant in our study self-identified as disabled; studying disabled CS students warrants future work.
Other future work would include expanding the study to include other members of the university (e.g.instructors, graduate students), recent graduates, and industry professionals; investigating the role of ethics in CS identity formation across other institutions and cultural contexts; and organizing workshops which offer insights for instructors and designers on how to promote the incorporation of socially grounded ethics into CS education and practice.

CONCLUSION
Although there is extensive work in CSCW to create and foster interdisciplinary, critical, ethical, and responsible design practices for computing professionals (e.g.[33,40,56,83,105,113,117]), work remains on scaling up these practices within the IT sector.In this paper we have identified a barrier to this disciplinary change, which is that engaging with ethics in computing is not perceived by undergraduate students as useful or necessary for legitimate peripheral participation in the field.
Students placed ethics outside the boundary of useful skills required to complete their degree or get a job.Unlike extracurricular activities focused on developing technical skills (e.g.hackathons), time spent on ethics outside of school did not count towards their performance of passion for CS, and as such did not garner students recognition or employability in the field.Rather, ethics were presented as something that was neither "useful," nor a fundamental part of CS.This illustrates how current attitudes towards ethics inside and outside of formal CS education work to inscribe social and political considerations as outside the boundaries of computing and design.
Consistent with prior literature, we found that there were gendered and racialized patterns in who did do the work of crossing these disciplinary boundaries -in our study, three women with minoritized racial identities were the only students who intentionally sought out opportunities to learn about tech ethics.These students spent time and energy educating themselves and others about ethics on top of developing the skills that were necessary for gaining credibility.That marginalized students are currently taking on the additional burden of educating other students about ethics is sadly in line with research on marginalized groups within academia and computing fields [17,42,118].It also means that valuing both knowledge in social science and social motivations for doing computer science is important for making the field less exclusionary, which is also in line with prior work [61,88].
The exclusion of ethics from CS education and practice can be explained by the theory of occupational closure [60,96,97,132].In placing "soft" topics outside the boundaries of the field, computer science maintains legitimacy while excluding both important multidisciplinary work, and particular bodies.This has consequences for the wider adoption of multidisciplinary and socially grounded design practices espoused in CSCW such as participatory design, value-sensitive design, critical design, and reflective design [40,56,105,113,117].
This exclusion also leads computer scientists to develop vague and depoliticized notions of social good which are not informed by social or political theory [63].In order for CSCW systems to transcend social and disciplinary boundaries, we need future designers to be both representative of the communities we hope to affect and support, and well trained in multidisciplinary approaches grounded in social and political theory.Therefore, bringing ethical practice within the boundaries of computing must start at CS education: when students are developing their understanding of what it means to be a designer.
Approaches to integrating ethics must go beyond formal education, however.In order for it to be in the interest of students to learn ethics, we must convince them that these skills are valuable -and not detrimental -to their careers.It is up to us as researchers, designers, and tech practitioners (and also employers, administrators, and teachers) to assert these values in our research and practice, and recognize and reward others who have knowledge and expertise drawn from the humanities and social sciences.In order for future designers to feel ethical responsibility, we must make it in their interest to do so.

3. 2
Context: McGill University The interviews for this study were conducted in 2020 and 2021, at McGill University, in Montréal, Canada with approval from the Research Ethics Board 2 at McGill University (REB file#: 20-11-013).
[...] I think after 18 years old, most people already know what they should do and what they shouldn't do.So even [...] for some people who has bad ideas, even though they took the ethics course [...] the bad idea does not disappear.They still stay in their mind.'Cause for the ethics everyone knows what is the right answer, but [to] really change their ideas is very hard.[...] And also I think I know what should I do and what [...] I shouldn't do already so I don't think I need to take that course." (P7) But I would like to do a course at McGill or a course externally from McGill, just to, kind of, see what ethical questions arise that I haven't even thought of yet." (P3)