Source-linked AI summary
Invisible Labor in Open Source Software Ecosystems
John Meluso, Amanda Casari, Katie McLaughlin, Milo Z. Trujillo
TL;DR
Invisible labor in OSS is difficult to measure, despite its implications for workers and organizations. The study uses a cognitive-anchoring survey and mixed-methods analysis to assess labor visibility and attribution. It finds that roughly half of OSS labor is invisible and that attribution motivations may be at cross-purposes.
Problem
Invisible OSS labor can limit workers’ compensation and legitimacy and constrain organizational understanding, while its prevalence and contributing factors remain questions for this study.
Source: §1 · pp. 2–3Method
The study uses cognitive anchoring in a survey and explanatory mixed-methods analysis to assess OSS labor visibility, attribution, and factors shaping invisibility.
Source: §1 · p. 3, §3.2.1 · p. 9, §3.3.2 · p. 13Results
Roughly half of OSS labor is invisible, and attribution motivations and resource limitations may contribute to its prevalence.
Source: §6 · p. 23Takeaways & Limitations
Greater labor visibility matters for contributors’ compensation and legitimacy and for organizations’ understanding of OSS work.
Source: §1 · pp. 2–3The sample is too small for conclusive findings about all of OSS.
Source: §6.2 · p. 24
Abstract
from arXiv · showhide
Invisible labor is work that is either not fully visible or not appropriately compensated. In open source software (OSS) ecosystems, essential tasks that do not involve code (like content moderation) often become invisible to the detriment of individuals and organizations. However, invisible labor is sufficiently difficult to measure that we do not know how much of OSS activities are invisible. Our study addresses this challenge, demonstrating that roughly half of OSS work is invisible. We do this by developing a cognitive anchoring survey technique that measures OSS developer self-assessments of labor visibility and attribution. Survey respondents (n=142) reported that their work is more likely to be invisible (2 in 3 tasks) than visible, and that half (50.1%) is uncompensated. Priming participants with the idea of visibility caused participants to think their work was more visible, and that visibility was less important, than those primed with invisibility. We also found evidence that tensions between attribution motivations probably increase how common invisible labor is. This suggests that advertising OSS activities as "open" may lead contributors to overestimate how visible their labor actually is. Our findings suggest benefits to working with varied stakeholders to make select, collectively valued activities visible, and increasing compensation in valued forms (like attribution, opportunities, or pay) when possible. This could improve fairness in software development while providing greater transparency into work designs that help organizations and communities achieve their goals.
1 Introduction
OSS ecosystems rely on essential work beyond coding, yet labor can remain unseen or undercompensated, with costs for workers and organizations. This study measures the prevalence of invisible labor and factors shaping it.
1 Introduction
- OSS production depends on essential non-code tasks, including community management, finance, interaction design, and documentation.These tasks support production, repair, and safeguarding.Source: §1 · p. 1
- Lower task visibility can reduce contributors’ compensation and legitimacy, while limiting organizations’ transparency, coordination, and understanding of work.The passage identifies risks including burnout, precariousness, and disparities in representation.Source: §1 · pp. 2–3
- Fig. 1 reports roughly two thirds of OSS labor as not visible and about half of labor activities as receiving no credit.The figure counts labor shared with only one other person as not visible.Source: Figure 1 · p. 2
- Existing measurement often privileges code, leaving a need to establish how common invisible labor is and what shapes its prevalence in OSS.The study poses these as its two research questions.Source: §1 · p. 3
- Roughly half of OSS activities are invisible, based on measures of labor attribution and visibility across varied fields.The study uses an explanatory mixed-methods analysis.Source: §1 · p. 3
2 Background
The background frames invisibility as a problem of labor information and compensation, while noting that motivations, existing records, and governance can shape what becomes visible in OSS.
2.1 Labor Visibility
- Full labor visibility requires information to be available, widely shared, and accessible; partial visibility can arise when any of these conditions is limited.Specialized work may be especially less visible because experts shield others from needing to know its steps.Source: §2.1 · p. 4
2.2 Labor Invisibility
- Labor is invisible when information about it is absent, unshared, unreachable, or unintelligible, or when work is undercompensated.Compensation includes credit, pay, and opportunities; visibility can also support worker autonomy, flexibility, and privacy.Source: §2.2 · pp. 4–5
2.3.1 Motivations.
- OSS participants pursue varied motivations, which can conflict because software, norms, and reward systems advance them differently.Attribution may support kinship or career goals but matter less to someone motivated by enjoyment alone.Source: §2.3.1 · pp. 5–6
2.3.2 Evidence of Invisibility.
- OSS records often omit or under-document activities beyond code, while even broad labels such as “code contributions” can obscure differences in work.Less-visible activities may include maintenance, moderation, financial accounting, and user support.Source: §2.3.2 · p. 6
2.3.3 Factors Shaping Invisibility.
- Conflicting motivations, platform features, and governance structures may shape whether labor is recorded, shared, accessible, or compensated.Governance can incentivize volunteer choices and influence recordkeeping and compensation practices.Source: §2.3.3 · pp. 6–7
3 Methods
The study combines a participatory feminist approach with an explanatory sequential mixed-methods design, using qualitative analysis to explain quantitative findings.
3 Methods
- The explanatory sequential design answers RQ1 quantitatively, develops RQ2 from those results, and then uses qualitative analysis to address it.The participatory feminist approach involved identifying community needs, co-designing the study, and discussing results with community members.Source: §3 · p. 7
3.1 Participatory Process
The participatory process drew on community collaboration to identify diverse OSS work and co-design a survey intended to assess its visibility and compensation.
3.1 Participatory Process
- The authors’ varied disciplinary and personal experiences both broadened and constrained how they understood and studied OSS work.The authors identify as English-speaking white academics and practitioners based in the West and Global North.Source: §3.1 · p. 7
- Over about 26 months, the team used mapping interviews, pilot surveys, and workshops to explore OSS activities and attribution across communities.The work included eight interviews and six workshops with 33 community members from five continents.Source: §3.1 · p. 7
- Community members identified a need for better labor taxonomies and metrics because existing tools inconsistently capture diverse activities.The team’s motivation was to make labor more visible when existing tools did not meet community needs.Source: §3.1 · p. 8
- The team co-designed an anonymous survey with community organizers and developer advocates, and avoided collecting identifying demographic data.Anonymity followed IRB advice for ethical compliance across jurisdictions and small demographic groups.Source: §3.1 · p. 8
3.2 Survey Instrument Design
The survey was designed to capture diverse OSS tasks while reducing social-signaling and recall biases. It measured credit as one form of compensation and estimated visibility by asking how many people knew about participants’ work.
3.2.1 Guiding Participant Attention.
- To limit ideological self-selection and social-desirability signaling, recruitment and survey wording avoided “invisible labor” and instead emphasized the familiar OSS concept of receiving credit.The authors acknowledge that no terminology is fully neutral.Source: §3.2.1 · p. 9
- Participants first considered recent collaborative OSS projects, then a broader range of ecosystem tasks, using “task” rather than “contribution” to elicit work beyond code.The two-year recall period was chosen to reduce underretrieval from excessive cognitive load.Source: §3.2.1 · p. 9
3.2.2 Compensation Questions.
- The survey asked about participants’ artifacts and use cases, then measured credit received from projects and for tasks, its medium, and satisfaction with it.Likert responses were numbered to reduce differences in how participants interpret response terms.Source: §3.2.2 · pp. 9–10
3.2.3 Labor Visibility Questions.
- Because ecosystem-wide activity is difficult to track, visibility was estimated from how often nobody, one other person, or at least two people knew about participants’ tasks.Question order was randomized between high-visibility-first (≥2, 1, 0) and low-visibility-first (0, 1, ≥2) to assess anchoring; self-reports remain vulnerable to frequency-estimation error.Source: §3.2.3 · pp. 10–11
3.3 Survey Procedures
The authors recruited 142 OSS participants for an explanatory mixed-methods study. Quantitative survey analyses were followed by thematic analysis of free responses about credit practices.
3.3 Survey Procedures.
- The study recruited 142 participants through Twitter between January and June 2022, reaching OSS communities on at least five continents, and used a sequential quantitative-then-qualitative design.The authors report an estimated 5.5% margin of error under their stated assumptions and high internal consistency for numeric measures (ωTotal = 0.84, α = 0.78).Source: §3.3 · p. 11, §3.3.1 · p. 11
3.3.1 Quantitative Analyses.
- Quantitative analyses examined participant work across artifacts and use cases, credit and visibility frequencies, variable relationships, and causal relationships where possible.The analyses used descriptive statistics, correlations, and linear regressions.Source: §3.3.1 · p. 11
3.3.2 Qualitative Analysis.
- The qualitative analysis began with 162 free responses and used six-step latent inductive “Big Q” thematic analysis to develop codes and themes.The authors refined initial codes into 38 codes and six initial themes, then used analytic memos and comparisons to develop final themes.Source: §3.3.2 · pp. 11–13
- Participants distinguished personal recognition from organizational recognition, and one contrasted credit for code commits with the less-recognized work of reviewing and maintaining code.One respondent described healthy projects as recognizing contributions while larger organizations often failed to appreciate them.Source: §3.3.2 · p. 12
- Responses described credit as inconsistent, opaque, or misattributed, and participants contrasted manual recognition with automated attribution.Examples include frustration with misspelled names and arbitrary automated recognition, alongside accounts that large-project credit can lose value over time.Source: §3.3.2 · p. 12
4 Quantitative Evidence of Invisible Labor
Across the sample, participants estimated that two-thirds of tasks were invisible and about half received credit. The results also show that visibility-question order shifted reported visibility and the importance participants placed on credit.
4.1 Sample Captures Diverse OSS Labor Activities
- The sample spanned diverse artifacts: five categories were selected by more than 30% of participants, and six use cases besides infrastructure were each selected by at least 20%.Software infrastructure was the most common use case (68%).Source: §4.1 · p. 13, Figure 2 · p. 13
4.2 Half of OSS Labor May Not Receive Attribution
- Participants reported receiving credit for half of their tasks (mean = 2.96 on a 1 to 5 scale) and from about half of their projects (mean = 3.07 on a 1 to 5 scale).Responses varied substantially; ten percent or fewer said they always or never received credit.Source: §4.2 · p. 14, Figure 3 · p. 14
- Participants reported slight satisfaction with credit frequency (mean = 3.36) and credit medium (mean = 3.58) on the linearized response scales.Automated metrics, project documentation, social media posts, and project memberships were the most reported credit mediums.Source: §4.2 · p. 14, Figure 3 · p. 14
4 Quantitative Evidence of Invisible Labor
- 66.7% of OSS tasks were estimated to be invisible, while 50.1% of labor received credit on average.The visibility estimate combines the two question-order groups; the authors report that labor was only partially visible on average (51.9%).Source: §4.3 · p. 15, §4.4 · p. 17
4.3 Anchoring Shifts Perceptions of Labor Visibility
- Participants anchored to high visibility perceived invisible work as less common than those anchored to low visibility, whose responses suggested more frequent nonvisibility.Question order also shifted reported partial visibility; the figure compares group means and response distributions.Source: Figure 4 · p. 15, §4.3 · p. 16, §4.4 · p. 18
4.4 Visibility Anchoring May Shift Beliefs About the Importance of Credit
- 30.5% greater credit importance was reported by participants anchored to low visibility than by those anchored to high visibility (95% confidence interval ±19.5%, p < 0.002).The regression had adjusted R² = 0.061 with robust standard errors.Source: §4.4 · p. 17, Figure 6 · p. 17
5 Qualitative Explanation of the Commonness of Invisible Labor
Participants described invisible labor as work that is insufficiently visible or compensated; the analysis identifies conflicting attribution motivations and limited resources as factors affecting its prevalence.
5.1.1 Visibility.
- Participants said records and contribution tools often expose commits while overlooking non-coding work and misrepresenting effort, time, and skill.They also noted that labor information recorded in commit messages may remain inaccessible to people outside project-specific channels.Source: §5.1.1 · pp. 18–19
5.1.2 Compensation.
- Participants reported insufficient or unfair credit, including status- and location-related disparities, erasure, and harms to contributors and retention.Some described pressure or hostility around less-recognized roles, while project leaders connected fair credit with retaining and welcoming contributors.Source: §5.1.2 · pp. 19–20
5.1.3 Dissenting Perspectives.
- A minority prioritized enjoyment, learning, or own-use over recognition, while most participants described some of their work as invisible.Some respondents said credit was lower on their list of reasons for contributing or that they did not need recognition.Source: §5.1.3 · p. 20
5.2 Attributing at Cross-Purposes
- Participants' motives often coexisted, but expressive thanks, systematic instrumental attribution, and non-attribution served different aims and could conflict.Expressive attribution centers on sincere personalized thanks; instrumental attribution systematically links labor data to people and makes it accessible; some contributors prefer no attribution.Source: §5.2 · pp. 20–21
- Participants described attribution trade-offs: automation could feel impersonal, manual credit was hard to scale, and mixed systems could pair personalized thanks with accessible documentation.Some also warned that project knowledge and informal credit records held by individuals could disappear when those people leave.Source: §5.2 · pp. 21–22
- Limited individual and collective resources make attribution practices that support one motivation constrain others, helping explain why invisible labor persists.The authors frame this tension as one factor in their dataset, not the only influence on invisibility.Source: §5.2 · pp. 20, 22, 23
6 Discussion
Invisible labor is common in OSS, but measuring it completely is impossible and making work visible can conflict with privacy and flexibility. The discussion argues for stakeholder-shaped visibility and compensation practices while identifying limits in the study’s scope and evidence.
6 Discussion
- About half of OSS work is invisible, and visibility anchoring shapes perceptions of work visibility.Resource limitations seem partly responsible for invisibility, amid tensions between attribution motivations.Source: §6 · p. 23
6.1 Implications
- Two in three tasks were not fully visible, often leaving valuable expertise difficult to understand, compensate, or manage.Work was most likely partially visible to one other person, so that labor could become unknown if that person left or forgot.Source: §6.1 · p. 23
- Labor cannot be measured completely or permanently, and documenting some tasks may undermine privacy, autonomy, or flexibility.Whether emotion regulation should be recorded depends on context.Source: §6.1 · p. 23
- Participatory design can help stakeholders decide what labor information to show, how to show it, and what to keep invisible.Playbooks and flexible platform artifacts may ease ecosystem-specific decisions, though designing the process for every ecosystem may be time-consuming and ineffective at balancing voices.Source: §6.1 · pp. 23–24
- Assuming OSS information is invisible despite an “open” label may counter overestimates of visibility and prompt selective design decisions.Visibility priming also reduced perceived importance of credit.Source: §6.1 · p. 24
- About half of OSS labor is compensated, and compensation may combine attribution, non-attribution, pay, and opportunities.Participatory choices should account for motivations and constraints such as changing time commitments and available funding.Source: §6.1 · p. 24
6.2 Limitations & Future Work
- The exploratory survey and small sample limit conclusions about all OSS, while its social-media convenience sample may overrepresent socially engaged ecosystems.The authors suggest targeted studies of particular communities or ecosystem characteristics.Source: §6.2 · pp. 24–25
- Future work should examine how technology affordances shape visibility and triangulate findings with other methods and contexts.Attribution-medium and motivation findings are correlational, and other factors may influence satisfaction; proposed factors include gender, governance, and ecosystem size.Source: §6.2 · p. 25
7 Conclusion
The conclusion synthesizes evidence of invisible and uncompensated OSS work, attribution tensions, and visibility anchoring. It recommends designing attribution and compensation systems with diverse stakeholders while assuming much OSS work is not visible or compensated.
7 Conclusion
- Conflicting motivations in attribution practices may partly explain why OSS labor is often invisible or uncompensated.The study estimates visibility and compensation rates and finds approximately half of OSS labor does not receive credit.Source: §7 · p. 25
A Invisible Labor Survey
The appendix reproduces the survey instrument, including questions about respondents’ OSS projects, task visibility, and credit experiences. It shows the prompts and response formats used to collect these self-reports.
A Invisible Labor Survey
- The survey began with consent and asked about projects respondents worked on during the previous two years with at least one other person.It also asked respondents to identify applicable project categories and use cases.Source: A Invisible Labor Survey · p. 29
- Respondents were prompted to recall their project tasks before answering questions about credit.The survey defined credit as the project telling someone that the respondent performed a task; recipients could be project members or outsiders.Source: A Invisible Labor Survey · pp. 29–30
- Credit questions covered how many projects and tasks received credit, the mediums used, and satisfaction with those mediums.The instrument offered multiple-choice or multiple-select formats and listed examples including automated metrics, documentation, and standardized tools.Source: A Invisible Labor Survey · p. 30
- Visibility and attribution questions asked how many people knew about tasks and how important credit was to respondents.The survey also asked about satisfaction with the amount of credit received and invited free-response comments on credit practices.Source: A Invisible Labor Survey · pp. 30–31