Notes · updated 2026-08-15
“Making It Your Own” Is Not Enough: The Ten Dimensions of Tojisha-sei, and How to Measure Them
No one disagrees that tojisha-sei — the standing and capacity of the person directly concerned — matters.
What, then, has to be observed for tojisha-sei to count as high? Here the conversation stops.
It stops because distinctions that everyday speech folds into one come apart in empirical research. Being related to a problem; feeling it as one’s own; deciding for oneself; being able to carry it out; owning the outcome; being able to influence institutions; holding the authority to speak about one’s own experience. These are all different.
This note sets those distinctions against the English-language research traditions.
The provenance ledger is in source/review/tojishasei-agency/papers.md.
Survey metadata
- Date: 2026-08-15
- Material: a report reviewing tojisha-sei in an integrated way (
source/review/tojishasei-agency/deep-research-report-tojishasei.md). The report carries a reference list, but not one entry had a DOI, so the bibliographic records were fixed through the Crossref, OpenAlex, arXiv, NDL Search, and J-STAGE APIs - Result of the bibliographic check: in six places the report’s description did not agree with the original (below). For the figures as well, those that agreed, those that could not be confirmed, and those that agreed only in part are kept apart
- Confidence: every bibliographic record was fixed by DOI, ISBN, or a stable URL. Items where neither the abstract nor the full text could be reached carry
[primary source unverified]
Six places where the source report and the originals came apart
| As the report has it | The original |
|---|---|
| Haggard, P. (2012), single author | co-authored by Haggard, P. & Chambon, V. |
| Sundblad, Y. (2010) | the Springer volume is 2011 (the conference was held in 2010) |
| Patzak, A. et al. (2025) | the authors are Patzak & Zhang, two people only |
| Tapal et al. 2017, “a 13-item structure” | began with 13 items and dropped two in Study 2; the final scale has 11 items |
| Liu et al. (2025), “using Traditional Low-Code Platforms and Large Language Models” | arXiv:2402.01156, v1 dated February 2024; the title correctly reads “using Traditional vs Large Language Model Support” |
| Pearce et al. (no year given) | 2022 IEEE Symposium on Security and Privacy |
Among the figures, too, some can be cited as they stand and some cannot. Each of the later sections states, case by case, whether a figure was confirmed.
No single English word covers it
Try to replace the Japanese tojisha-sei with one English word and something always drops out.
| Field | Term in the original language | What it centres on |
|---|---|---|
| Japanese sociology and social welfare | tojisha, tojisha shuken (sovereignty of the person concerned) | taking on one’s needs as one’s own and envisaging a different reality |
| Psychology | agency (Bandura) | acting through intention, forethought, self-regulation, and self-reflection |
| Sociology | agency (Emirbayer & Mische) | temporally embedded, relational social engagement spanning past, future, and present |
| Cognitive science | sense of agency | the experience of being the origin of one’s actions and their effects |
| Self-determination theory | autonomy | experiencing one’s actions as aligned with one’s own volition and values |
| Social cognitive theory | self-efficacy | the belief that one is capable of carrying out a particular action |
| Organizational psychology | psychological ownership | experiencing an object as “mine” or “ours” |
| Community psychology | empowerment | control, critical awareness, participation, power |
| Political science | political efficacy | the belief that one can take part, and the belief that institutions respond |
| Science and technology studies | lay expertise, situated knowledge | whose experience counts as knowledge |
The term that needs the most care is stakeholder.
It is rendered in Japanese as rigai kankeisha or kankei tojisha, but what stakeholder names is a position with respect to the problem, not what a person can envisage and do.
A patient being a stakeholder in a medical app is one thing; holding the right to decide its specification and to access its data is another. Stakeholder status can be a necessary condition for tojisha-sei, but it is not a sufficient one.
A tojisha is not a tojisha from the start
The Japanese-language discussion has a feature that English agency alone does not capture well.
Nakanishi Shoji and Ueno Chizuko’s Tojisha Shuken (Sovereignty of the Person Concerned; Iwanami Shinsho, 2003) is set against the arrangement in welfare where the needs assessments of family members and professionals take precedence over the person’s own. When Doba Manabu surveyed this lineage in Shakaigaku Hyoron (Japanese Sociological Review) in 2007, the argument he introduced runs as follows.
No one is a tojisha from the start. Needs are formed when the present state is grasped as a shortfall against a desirable state and a different reality is envisaged. That is when a person becomes a tojisha.
Tojisha-sei is therefore not settled by objective harm or attributes alone. It includes a process of formation: from recognizing that something is off, to forming a need, to framing the problem, to envisaging a different reality.
This connects with Emirbayer and Mische’s framework. They divided agency into three aspects: repeating established habit, envisaging a future that does not yet exist, and judging within an uncertain present. “Grasping the present as a shortfall and envisaging a different reality” belongs to the second.
What is central to tojisha-sei is not “making it your own” but the capacity to envisage that a state other than the present one is possible.
An expanded new edition of Tojisha Shuken appeared in November 2024 (Iwanami Shinsho Shin-akaban 2042, adding a chronology of the tojisha movement). Citations need to specify the edition.
The ten dimensions
Separated out, the distinctions come to this.
| Dimension | What it asks |
|---|---|
| Affectedness | Is the person actually affected by the problem or the system? |
| Need self-definition | Have they put the problem into words as their own need? |
| Projective agency | Can they envisage a desirable state different from the present one? |
| Autonomy | Are they choosing on the basis of their own values and reasons? |
| Efficacy | Do they hold the belief that they can act, and the actual capability? |
| Epistemic agency | Can they explain, hypothesize, and test using their experience and evidence? |
| Psychological ownership | Do they experience the object as theirs? |
| Participatory power | Do they hold real power over agenda-setting, decisions, refusal, and the allocation of resources? |
| Collective agency | Can they act jointly with other tojisha? |
| Stewardship | Can they take responsibility for outcomes, maintenance, incidents, and externalities? |
One of these alone should not be picked out and called “tojisha-sei.” Research on it begins from stating which dimension is being handled.
Separate fields have treated the same problem under different names
The research bearing on tojisha-sei did not grow out of a single discipline.
Measurement of political efficacy began in the 1950s and later split into internal efficacy (I have the ability to understand and take part) and external efficacy (institutions respond to citizens’ demands). Arnstein classified citizen participation in 1969 as a ladder of eight rungs. Rappaport put empowerment forward in 1987 as a central concept of community psychology.
The origin of Scandinavian Participatory Design was not the usable screen. In the UTOPIA project, as Sundblad recalled it in 2011, what mattered was that trade unions and workers took part directly in the design and development of the technology. Rather than passing requirements to specialists, they negotiated over the change in work that the technology itself would bring. The starting point was democratic power over technological change.
What Haraway questioned in situated knowledges in 1988 was the notion of an objectivity detached from position and body. What Epstein described in 1995 was the process by which AIDS activists, rather than remaining recipients of information from scientists, acquired biomedical knowledge and intervened in the rules of clinical trials themselves. A model stronger than “listening to the voices of those with experience” was established here. It is the model in which tojisha take part in the rules of knowledge production.
In Japan, tojisha-kenkyu (self-directed studies by the people concerned) extended this further, into “the right to do research.” The official page of Bethel’s House in Urakawa records that it began around 2001 with the “studies” of a few people (confirmed against the primary source). It moves the person from an object diagnosed by professionals to a subject who studies their own difficulties. Supporters and peers become co-researchers.
What matters here is that tojisha-kenkyu does not hold up “understanding and controlling everything by oneself” as its ideal. Not-knowing is put outside as a hypothesis and opened to others. It amounts to an objection from within to any reading of tojisha-sei as “the ability to solve everything unaided.”
In HCI, end-user development connects to this lineage. When Fischer and colleagues defined meta-design in Communications of the ACM in 2004, what they placed at its core was returning the “owners of problems,” those who know the problem best, to the design process.
Measuring it takes four layers
No standard scale of tojisha-sei has been established. Existing scales have to be combined according to what is being studied.
| Scale | What it measures | Psychometrics (state of confirmation) |
|---|---|---|
| Sense of Agency Scale (Tapal et al. 2017) | positive and negative sense of agency | began with 13 items and dropped two in Study 2, leaving a final 11 items. Study 1 (n=236) gave ω=.78/.76, with two-month test–retest r=.78/.74 (n=91). Study 2 (n=408) gave ω=.80/.75. Confirmed in the full text |
| Political Efficacy Short Scale (Groskurth et al. 2021) | internal and external political efficacy (4 items) | ω was .88 internal / .84 external for the UK sample and .86 internal / .86 external for the German sample; test–retest reliability after about three to four weeks was .83 / .68 for the UK and .87 / .75 for Germany. Not visually confirmed against the primary PDF; this rests on secondary confirmation only [primary source unverified] |
| Sociopolitical Control Scale (Zimmerman & Zahniser 1991; Peterson et al. 2006) | perceived leadership competence, perceived control over policy | the revised version detected method bias from reverse-worded items, rendered all items positively, and supported a two-factor structure |
| Psychological Ownership (Van Dyne & Pierce 2004) | the sense that something is “mine” | three field studies with more than 800 employees; Cronbach’s α of .87 / .90 / .93, with n=409 for Study 2. The individual n and test–retest reliability for Study 1 and Study 3 remain unreached. What the abstract confirms goes only as far as “showed no incremental value in predicting performance” [primary source unverified] |
| Psychological Empowerment (Spreitzer 1995) | meaning, competence, self-determination, impact (four dimensions, 12 items) | measured and validated in the original study |
Self-report reaches as far as sense of agency, ownership, and efficacy. Participatory power and structural opportunity cannot be measured by how the person feels.
Political efficacy research splitting internal from external, and empowerment research splitting the intrapersonal, interactional, and behavioural dimensions, are earlier solutions to this problem. Even where the person has agency, effective tojisha-sei is low if the organization does not hand over the data, the budget, and the authority to deploy.
Measurement therefore comes in four layers.
| Layer | Question | Examples |
|---|---|---|
| Subjective | Does it feel like my problem, mine to decide, within my power? | sense of agency, autonomy, self-efficacy, ownership |
| Behavioural | Did they in fact define the problem, choose, refuse, and change it? | self-initiated goal setting, rejecting AI proposals, editing, how help is sought |
| Outcome | Can they understand, fix, transfer, and maintain? | reproducing an explanation, finding bugs, writing tests, delayed transfer |
| Structural | Do they actually hold resources, decision rights, and a veto? | permissions on data and APIs, budget, deployment rights, right of refusal, right of appeal |
Prompt counts, lines of code, and the acceptance rate of AI suggestions should not be taken as measures of agency just as they stand. Someone issuing many prompts may be exploring on their own initiative, or may simply be stuck because the AI is not understanding them.
Cultivating it takes both freedom and scaffolding
Instructing people to “think of it more as your own problem” is a theoretically weak way to cultivate tojisha-sei.
The meta-analysis Patzak and Zhang published in Educational Psychology Review in 2025 synthesizes 94 studies and 110 effect sizes (both figures confirmed in the abstract, and they agree).
Autonomy support is set out as offering choice, giving rationales, and acknowledging students’ negative feelings and perspectives; structure as guidance, clear expectations, and informational feedback.
The individual correlation coefficients the report cited (r=.53 and others) remain unconfirmed [primary source unverified].
Beyond Springer’s authentication wall, the full text could not be reached through the author’s self-archive, institutional repositories, ResearchGate, or PsyArXiv either (retried 2026-08-16).
What can be said within what was confirmed is that autonomy support and structure are not treated as opposed. Letting the person decide, and giving clear expectations, scaffolding, and feedback, are used in combination.
In Youth Participatory Action Research, young people become co-researchers rather than research subjects. The systematic review by Anyon and colleagues in 2018 screened 3,724 records, included 63 studies across 67 reports, found agency and leadership the most frequently reported youth outcome at 75.0%, and counted only two randomized trials (all three figures confirmed in the abstract, and they agree).
There is theoretical soundness and a consistent run of positive reports, but the study designs are not enough to assert an average treatment effect with any force. The same holds for tojisha-kenkyu: its value as a history of practice and as qualitative research has to be read separately from confirmation of causal effects.
LLMs move the bottleneck of tojisha-sei
Once software can be built in natural language, the wall of implementation skill comes down.
In the empirical study of end-user development Weber published in 2026, 24 of 33 students (about 73%) in the mechanical engineering department at Kempten University of Applied Sciences in Germany completed a working questionnaire app with AI assistance (18 of 21 master’s students, 85%; 6 of 12 bachelor’s students, 50%). 28 of them (about 85%) recommended making it available to companies. At the same time, 18 raised the need for training and 10 the need for a dedicated platform. These figures were confirmed directly in the full text of the arXiv version, and all of them agree.
What follows from this is not that “anyone has become a software engineer.” It goes only this far: syntax, frameworks, and boilerplate are ceasing to be an absolute barrier to entry for non-specialists.
And the bottleneck moves. It moves to discovering needs, framing the problem, writing specifications, verification, security, data rights, maintenance, externalities, and joint governance.
What tells at the place it moves to is epistemic agency.
The preprint Shen and Tamkin released in January 2026 examines the relation between AI assistance and skill formation experimentally. In the pilot, the AI group completed the task faster (Cohen’s d=1.11, p=0.03) but scored significantly lower on a knowledge quiz (Cohen’s d=1.7, p=0.003). The abstract states that AI assistance impaired conceptual understanding, code reading, and debugging ability, that it brought no significant efficiency gain on average, and that participants who delegated wholesale raised their productivity while sacrificing mastery. These figures too were confirmed directly in the full text. It is an unrefereed preprint, however, and replication is needed.
When an instruction given in natural language yields an artefact, a sense of agency and psychological ownership arise easily. “I gave the instruction” and “I understand the causal structure” are different things.
The indicator to watch is therefore not the time it takes to build, but the time it takes to become able to control it unaided.
A transfer of responsibility also takes place. As Pearce and colleagues showed at the 2022 IEEE Symposium on Security and Privacy, AI-generated code can contain vulnerabilities. Handing non-programmers the ability to deploy is at the same time handing them responsibility for security.
No conclusion follows from this that “people with high tojisha-sei should be free to deploy.” The conditions of justification differ between a tool that affects only its maker and a system handling patient records, payment details, or employee data. Tojisha-sei does not override the rights of others.
Bandura’s framework is useful on this point. Besides personal agency, in which one acts directly oneself, people use proxy agency, working through others who hold resources or expertise they lack, and collective agency, producing results as a group. So relying on an LLM is not in itself a decline in tojisha-sei. What is at issue is rather what is delegated, where judgement is taken back, and which outcomes one holds responsibility for.
Where are the tojisha in gyaru-mind research?
This wiki records three studies that set out to measure gyaru-mind. Held against the ten dimensions, all three are empty in the same place.
Motode Rino 2023 (gyaru-mind-aspiration-scale) deliberately conducted no interviews with the people concerned. The reason is stated in the text. The judgement is that what Generation Z admires is not “Heisei gyaru now” but “Heisei gyaru then,” so the material should be taken from the fashions of the time rather than from anyone’s recollection. The source of the construct was placed in ten magazine issues.
That judgement has its reasons, and it is consistent with Motode’s question (how the fashions of the time make present-day admiration). As a result, though, need self-definition and epistemic agency were left outside the study. The line Haraway and Epstein established, in which those with the experience take part in the rules of knowledge production, is not used here.
In Ikegami Momoka and colleagues 2025–2026 (gyaru-mind-operationalization), the eight factors were produced by decomposing a single trade book through discussion among the authors, and the reference scores serving as ground truth were assigned by the first author alone. No inter-rater agreement is reported. Nor is there any built-in procedure by which people who identify as gyaru judge whether “this is our word.”
Epistemic agency is placed nowhere. Neither is participatory power.
The claim to novelty (gyaru-criterion-formation-novelty) is more direct still. The one line proposed as capable of rescuing theoretical novelty is a sequence in which tojisha deliberately amplify a difference, an evaluative word of theirs comes into being, that word becomes a rule of judgement, and the rule changes what is made next. Every step begins with the tojisha.
Yet within the same examination it is pointed out that Matsui Takeshi analysed “kogal” in 2015 as an example of a social sign made by the media. That is the possibility that the naming was not done by the tojisha.
The core of the novelty claim rests on tojisha-sei being real, and tojisha-sei itself has not been verified.
What the three have in common is that in none of them is it designed who holds the legitimacy to define what gyaru-mind is. What Nakanishi and Ueno wrote about welfare was a criticism of the arrangement in which the needs assessments of family and professionals take precedence over the person’s own. In gyaru-mind research, magazines, a trade book, and discussion among the authors take precedence over the person.
There is something to be said in the other direction as well.
The theory of tojisha-sei grew up in the contexts of welfare, medicine, and disability. It is a tojisha-sei that starts from harm and disadvantage. What gyaru-mind starts from is aesthetic difference.
The theory of needs Doba introduces and the novelty claim about gyaru practice are similar in structure. One line has needs arise where the present state is grasped as a shortfall against a desirable state and a different reality is envisaged; the other has a difference rated negatively by existing criteria deliberately amplified and turned into a different value. Both turn a state placed on the negative side into a resource for envisaging.
The first pass through this search wrote that no study setting the two side by side had been seen. Change the route and search again, and they turn up. The route crossing crip theory with fashion studies had not been tried in the first pass.
Ben Barry and colleagues treat disability justice and the alteration of clothing within the same body of research. Wardrobe interviews with 40 disabled and non-binary gender people report that practices of dress dismantle the dominant norms of gender and disability, while also carrying tension between intersecting privilege and marginalization. In response to exclusion from the fashion industry, the sharing of clothing swaps and styling techniques arose, and the process by which an alternative fashion world is built from that has also been theorized. The same kind of juxtaposition appears in work on body-size stigma and fashion expression, and on the politics of Black hair and self-definition.
Robertson and colleagues, in a systematic review of research on Black women’s fashion, point out that the field has split into two: research centring resistance and oppression, and research centring a more holistic self-expression. That accounts starting from disadvantage and accounts starting from aesthetic self-presentation do not mesh is recognized as a problem within the field itself.
This is as far as the gap has been filled. What these establish is a framework that places the two side by side, not a demonstration of the causal chain gyaru-criterion-formation-novelty requires — the sequence running from the amplification of difference, through the generation of an evaluative word, to the generalization of a rule of evaluation.
What touches this gap from the measurement-design side is stage 0 of gyaru-mind-measurement-design (collecting situations from the tojisha).
References
URLs retrieved 2026-08-15. Every bibliographic record was fixed through one of Crossref, OpenAlex, arXiv, NDL Search, or J-STAGE.
Theories of agency and autonomy
- Bandura, A. 2001. Social cognitive theory: An agentic perspective. Annual Review of Psychology 52(1), 1–26. https://doi.org/10.1146/annurev.psych.52.1.1
- Emirbayer, M. & Mische, A. 1998. What is agency? American Journal of Sociology 103(4), 962–1023. https://doi.org/10.1086/231294
- Haggard, P. & Chambon, V. 2012. Sense of agency. Current Biology 22(10), R390–R392. https://doi.org/10.1016/j.cub.2012.02.040
- Ryan, R. M. & Deci, E. L. 2000. Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist 55(1), 68–78. https://doi.org/10.1037/0003-066X.55.1.68
Measurement scales
- Tapal, A., Oren, E., Dar, R. & Eitam, B. 2017. The Sense of Agency Scale. Frontiers in Psychology 8, 1552. https://doi.org/10.3389/fpsyg.2017.01552
- Zimmerman, M. A. & Zahniser, J. H. 1991. Refinements of sphere-specific measures of perceived control: Development of a Sociopolitical Control Scale. Journal of Community Psychology 19(2), 189–204. https://doi.org/10.1002/1520-6629(199104)19:2%3C189::AID-JCOP2290190210%3E3.0.CO;2-6
- Peterson, N. A., Lowe, J. B., Hughey, J., Reid, R. J., Zimmerman, M. A. & Speer, P. W. 2006. Measuring the intrapersonal component of psychological empowerment: Confirmatory factor analysis of the Sociopolitical Control Scale. American Journal of Community Psychology 38(3–4), 263–274. https://doi.org/10.1007/s10464-006-9070-3
- Van Dyne, L. & Pierce, J. L. 2004. Psychological ownership and feelings of possession: Three field studies predicting employee attitudes and organizational citizenship behavior. Journal of Organizational Behavior 25(4), 439–459. https://doi.org/10.1002/job.249
[primary source unverified](individual sample sizes, α, and test–retest reliability not reached in the full text) - Groskurth, K., Nießen, D., Rammstedt, B. & Lechner, C. M. 2021. An English-language adaptation and validation of the Political Efficacy Short Scale. Measurement Instruments for the Social Sciences 3, Art. 1. https://doi.org/10.1186/s42409-020-00018-z
[primary source unverified](ω and test–retest reliability not reached in the full text) - Spreitzer, G. M. 1995. Psychological empowerment in the workplace: Dimensions, measurement, and validation. Academy of Management Journal 38(5), 1442–1465. https://doi.org/10.2307/256865
Empowerment, participation, and lay knowledge
- Rappaport, J. 1987. Terms of empowerment/exemplars of prevention: Toward a theory for community psychology. American Journal of Community Psychology 15(2), 121–148. https://doi.org/10.1007/BF00919275
- Arnstein, S. R. 1969. A ladder of citizen participation. Journal of the American Institute of Planners 35(4), 216–224. https://doi.org/10.1080/01944366908977225
- Haraway, D. 1988. Situated knowledges: The science question in feminism and the privilege of partial perspective. Feminist Studies 14(3), 575–599. https://doi.org/10.2307/3178066
- Epstein, S. 1995. The construction of lay expertise: AIDS activism and the forging of credibility in the reform of clinical trials. Science, Technology, & Human Values 20(4), 408–437. https://doi.org/10.1177/016224399502000402
- Jasanoff, S. 2004. The idiom of co-production. In States of Knowledge, Routledge, 12–23. https://doi.org/10.4324/9780203413845-6
- Sundblad, Y. 2011. UTOPIA: Participatory design from Scandinavia to the world. History of Nordic Computing 3 (IFIP AICT), 176–186. https://doi.org/10.1007/978-3-642-23315-9_20
- Anyon, Y., Bender, K., Kennedy, H. & Dechants, J. 2018. A systematic review of youth participatory action research in the United States. Health Education & Behavior 45(6), 865–878. https://doi.org/10.1177/1090198118769357
- Patzak, A. & Zhang, X. 2025. Blending teacher autonomy support and provision of structure in the classroom for optimal motivation: A systematic review and meta-analysis. Educational Psychology Review 37(1), Art. 17. https://doi.org/10.1007/s10648-025-09994-2
[primary source unverified](individual correlation coefficients not reached in the full text)
Japanese-language work on tojisha
- Nakanishi, Shoji & Ueno, Chizuko. 2003. Tojisha Shuken (Sovereignty of the Person Concerned). Iwanami Shoten (Iwanami Shinsho), in Japanese. ISBN 4-00-430860-7. Expanded new edition November 2024 (Iwanami Shinsho Shin-akaban 2042, ISBN 978-4-00-432042-5)
- Doba, Manabu. 2007. Research trends by theme (tojisha-sei). Shakaigaku Hyoron (Japanese Sociological Review) 58(2), 221–230 (in Japanese). https://doi.org/10.4057/jsr.58.221
- Sekimizu, Teppei. 2011. The problem of “hikikomori” and the “tojisha.” Nenpo Shakaigaku Ronshu (Annual Review of Sociology) 24, 109–120 (in Japanese).
- Kumagaya, Shin’ichiro. 2019. What is tojisha-kenkyu? Joho Shori (IPSJ Magazine) 60(10), 955–958 (in Japanese).
[primary source unverified](whether this article extends to historical analysis was not confirmed in the full text; it may be a different piece) - Social Welfare Corporation Urakawa Bethel’s House, “Tojisha-kenkyu.” https://urakawa.bethel-net.jp/aboutus (the start around 2001 confirmed against the primary source; a self-description by a tojisha organization, treated as a description of practice)
End-user development and LLMs
- Fischer, G., Giaccardi, E., Ye, Y., Sutcliffe, A. G. & Mehandjiev, N. 2004. Meta-design: A manifesto for end-user development. Communications of the ACM 47(9), 33–37. https://doi.org/10.1145/1015864.1015884
- Weber, I. 2026. Feasibility of AI-assisted programming for end-user development. Advances in Natural Language Processing and Information Retrieval (NLPIR 2025), 1043–1053. https://doi.org/10.1007/978-3-032-20897-2_73 (preprint: https://arxiv.org/abs/2512.05666)
- Shen, J. H. & Tamkin, A. 2026. How AI impacts skill formation. arXiv:2601.20245 (not peer reviewed). https://arxiv.org/abs/2601.20245
- Pearce, H., Ahmad, B., Tan, B., Dolan-Gavitt, B. & Karri, R. 2022. Asleep at the keyboard? Assessing the security of GitHub Copilot’s code contributions. 2022 IEEE Symposium on Security and Privacy (SP). https://doi.org/10.1109/SP46214.2022.9833571
Unverified items
- The six places where the source report and the originals did not agree are recorded in the table in the body
- Figures that could not be confirmed: all of the individual correlation coefficients in Patzak & Zhang 2025 (r=.53, .59, .45, .36, .47, .35). Beyond Springer’s authentication wall, the full text could not be reached through the author’s self-archive, institutional repositories, ResearchGate, or PsyArXiv either
[primary source unverified] - Figures confirmed only to a secondary source (2026-08-16): for Van Dyne & Pierce 2004, Cronbach’s α across the three studies of .87 / .90 / .93 and n=409 for Study 2 — the individual n and test–retest reliability for Study 1 and Study 3 remain unconfirmed. For Groskurth et al. 2021, ω of .88 internal / .84 external for the UK and .86 internal / .86 external for Germany, and test–retest reliability of .83 / .68 for the UK and .87 / .75 for Germany. None of these has been visually confirmed against the primary PDF; they rest on secondary confirmation only
[primary source unverified] - Figures that agreed only in part: the “13 items” of Tapal et al. 2017 (the final version has 11), and “ω=.78/.76,” which is a Study 1 value only
- Whether Kumagaya Shin’ichiro 2019 contains the “historical analysis” the source report refers to was not confirmed in the full text. It may point to a different piece
[primary source unverified] - Hirose Shiina and Kishi Taichi, “A literature review on shutaisei” (2025, Proceedings of the 89th Annual Convention of the Japanese Psychological Association 979, DOI 10.4992/pacjpa.89.0_979) is a conference presentation abstract, and since it could not be confirmed to be the same thing the source report points to, it was not used in the body
- The meta-design paper by Fischer & Scharff 2000 (DIS ‘00) could not be reached; the ACM DL returns a 404
[primary source unverified] - Whether references 24 and 25 duplicate the existing corpus in
source/review/end-user-development/has not been checked - Studies setting a tojisha-sei that starts from disadvantage alongside a subjectification that starts from aesthetic difference were found in number in the renewed search of 2026-08-16 (see the relevant section in the body; the corpus is in section F of
source/review/tojishasei-agency/papers.md). The gap implied by the first pass’s “none seen” does not exist - OpenAlex and Semantic Scholar returned HTTP 429 (Retry-After roughly 8 hours), however, so citation-network tracing has still never been carried out. The lists of works citing Nakanishi and Ueno, Doba, Costanza-Chock, Thornton, Hebdige, Collins, Fraser, Siebers, and Agha remain unsearched
- Of the 15 references added, the abstracts of 5 (the book chapter by Barry & Nesbitt 2023, Barry 2022, Holliday 2001, Reddy-Best & Pedersen 2013, and the Hair Sankofa chapter) could not be reached; relevance was judged from title and bibliographic record alone
[primary source unverified] - Whether large-scale randomized controlled trials of the effects of tojisha-kenkyu exist was not searched