Abstract
Behavioral sciences now routinely rely on digital data, supported by digital technologies and platforms. This has resulted in an abundance of new ethical challenges for researchers and ethical-review boards. Several contemporary high-profile cases emphasize that ethical issues often surface after the research is published, once harm has already occurred. Consequently, implementing safeguards in digital-behavioral research is often reactionary and fails to adequately prevent harm. In response, we propose the DECIDE (Describing Ethical Choices in Digital-Behavioural Data Explorations) framework, which encourages ethical reflections and discussions throughout all stages of the research process. The framework presents several questions designed to help researchers view their work from new perspectives and uncover ethical issues they might not have anticipated. We provide several resources to support researchers with their ethical reflections and discussions, including (a) the DECIDE framework spreadsheet, (b) the DECIDE desktop app, (c) information documents, and (d) flowcharts. In this article, we provide suggestions on how to use each resource to encourage proactive discussions of how ethical issues may apply to specific research contexts. By promoting continuous ethical considerations, safeguards can be put in place throughout the research project, even after research commencement. The DECIDE framework shifts ethical reflection away from being reactive toward a more proactive endeavor, reducing the risk of harm and the misuse of digital-behavioral data.
The digital universe is saturated with information about people, including their demographics, psychology, and behaviors. Harnessing these data for research provides abundant opportunities to investigate new topics and methodologies. Yet there remain conceptual gaps among researchers and ethical-review boards when assessing the ethical risks of such projects (Zimmer, 2018). This has been emphasized through several high-profile cases in which the publication of research was followed by significant ethical debates (Isaak & Hanna, 2018; Schneble et al., 2018; Shaw, 2016; Van Noorden, 2020; Verma, 2014; Zimmer, 2010, 2018). When using digital data generated by “human subjects” in research, concerns largely focus on consent, anonymization, the right to privacy, the protection of vulnerable populations, and whether online data should be considered “public” or “private” (Isaak & Hanna, 2018; Schneble et al., 2018; Shaw, 2016; Van Noorden, 2020; Verma, 2014; Zimmer, 2010, 2018). When consulting research participants, authors of one study found that people were more accepting of their digital data being used in academic research than by social media companies or journalists (Hemphill et al., 2022). However, what participants considered “acceptable” depended on what parts of their data would be reused, their understanding of who would use the data, and the reasons behind its use (Hemphill et al., 2022).
Consequently, there is an immediate need to transform research ethics into a proactive process via our DECIDE (Describing Ethical Choices in Digital-Behavioural Data Explorations) framework presented herein. In short, the framework includes a series of questions that prompt ethical discussions/reflections within research teams and individuals. The framework advocates that ethical reflections should not end on receipt of ethical approval from a review board but complements this by proposing ongoing reflection throughout a research project. We are not proposing continued governance (e.g., repeated ethical approvals); instead, DECIDE prompts informal discussions and reflections across all stages of a research project. Ethics is often misunderstood as a form-filling process, which serves to satisfy university governance, because this is often a requirement before any research can take place. We emphasize ethics as a process in which the reduction in risk emerges through ongoing discussions and reflections, which prompt a change in research practices. Although our guidance can inspire research governance (and thus might be useful to ethical-review boards), we are deliberately describing DECIDE as a framework to support culture change rather than a comprehensive research-governance framework.
The focus of the DECIDE framework is to provide contemporary considerations for individuals studying digital data that are generated by human activity (e.g., purchasing goods, posting on social media) or created through the surveillance/remote sensing of individuals (e.g., closed-circuit-television data). In other words, DECIDE illuminates emerging issues with the use of digital data that provide information on “human subjects” and their behaviors. This encompasses (a) digitally exclusive behaviors that take place natively and exclusively through a piece of technology (e.g., Facebook “like” data), (b) digitally mediated behaviors in which the behaviors can be conducted without using technology but can be mediated by technology (e.g., verbal conversations over video chat), and (c) digitally recorded behaviors that create an electronic log of nondigital actions (e.g., using fitness apps to measure the number of steps per day; Kaye et al., 2022). These data can be extracted in secondary data-analysis projects or generated in experimental work. Therefore, the scope of DECIDE encompasses both primary and secondary data initiatives while also considering the ethical risks associated with all stages of the research processes, such as during data collection, analysis, and sharing.
DECIDE aims to help researchers recognize the broader risks associated with their research because it remains difficult to comprehend what issues will arise during a project. This is further complicated by various and long-standing conceptual gray areas regarding digital-behavioral research: “While many lessons from previous ethical breaches find their way into regulatory guidelines or law, unique ethical dilemmas arise as a natural part of any research phenomenon” (Markham, 2016).
One example gray area is whether data found on the internet is public or private and whether that data “should” be used in research. This was illuminated by a project that shared rich Facebook data from a cohort of Harvard students that led to some being reidentified (Lewis et al., 2008). In response to criticisms, the principal investigator said, “We have not accessed any information not otherwise available on Facebook,” making the argument that “the data is already public” (Zimmer, 2010). This example clearly shows conflicting viewpoints on whether certain types of social media data should be shared and analyzed.
Associated issues can become magnified when working with vulnerable populations. This is evidenced in a now retracted article that received public concern following the use of Reddit data created by individuals in the schizophrenia community (Lyons et al., 2024). Despite obtaining ethical approval from their host university and discussing issues pertaining to anonymity and consent in their original article, this did not mitigate the ethical issues that became apparent following publication. In their retraction statement, the authors reflected and discussed the difficulty in ascertaining the public or private nature of online forums. However, they also considered how some online communities are perceived as safe spaces, but for individuals with schizophrenia, gathering information from these communities can add to feelings of paranoia and transform these into “unsafe” places.
Researchers frequently conduct studies with the intention of minimizing harm based on their best understanding; however, the absence of clear guidelines makes them susceptible to unintentionally causing harm. This leaves researchers open to negative consequences, such as public shaming, the retraction of research, and reputational damage. Therefore, the unique nature of each research project, a diverse set of guidelines (see Table 1), and evolving gray areas mean that existing frameworks do not adequately support researchers when navigating this type of research and related issues.
Some Existing Ethical Guidelines for Psychological Science and Their Scope
Concerns specific types of digital data.
There are also large inconsistencies evident between research communities and institutional review boards in terms of both how decisions are made and what requires ethical approval. For example, not all review boards consider data scraped from the internet (e.g., public social media posts) to be “human-subjects research,” which therefore do not require ethical review (Metcalf & Crawford, 2016; Paxton, 2024; Proferes et al., 2021). This is then reflected back in how researchers view digitally scraped data as being seemingly decoupled from the individuals who produce it, thus neglecting to consider the impacts on people, communities, and society (Metcalf & Crawford, 2016).
However, even if researchers are not directly interacting with participants throughout the research process, the same ethical considerations applicable to in-person research should still be applied (Paxton, 2024). This is because digital data generated by individuals can reflect their thoughts, perspectives, identity expressions, and motivations. Therefore, most studies that routinely rely on digital data carry ethical implications, including an individual’s loss of control over their data and unauthorized processing of these data in the future, both of which can be considered a violation of someone’s privacy (Zimmer, 2010). To provide support for researchers and reviewers, the DECIDE framework treats all digital-behavioral data as human-subjects data (as recommended by Metcalf & Crawford, 2016), encompassing digitally exclusive, mediated, and recorded data.
In sum, without appropriate guidance, researchers may not have the expertise required to recognize the ethical issues that might arise from their research, leading to ethical questions being raised in the period after publication (described here as “reactionary ethics”; Favaretto et al. 2020; Verma, 2014). This is not a helpful route to maximize the collective good within and beyond psychological science (Tiokhin et al., 2021). A lack of comprehensive ethical guidance is further compounded by the static process of solely addressing ethical considerations before research. Thus, ethics are typically conceptualized in a “single event” (Head, 2020; Velardo & Elliott, 2018), and once approval has been granted, ethical implications are seldom thought of again. Indeed, although ethics can be seen as an initial precursor (or worse still, hurdle) to research, in practice, ethical dilemmas evolve and change with the progress and the opportunities of a given project. Consequently, researchers have questioned whether existing guidelines and processes are still “fit for purpose” (Favaretto et al., 2020) and have called for the implementation of dynamic guidelines that better address the challenges of working in the digital age (e.g., Anabo et al., 2019; Clark et al., 2019; Vitak et al., 2016).
In response to the above, we present DECIDE—a dynamic framework that aims to drive ethical decision-making today and in the future. Specifically, DECIDE outlines key ethical questions for psychologists to consider when engaging in digital-behavioral-data explorations. DECIDE encourages action and reflection throughout all stages of the research process, reducing the risk of harm to individuals and the misuse of digital-behavioral data. It describes a process in which researchers can be reflexive and dialogical about their research practices. This reflective process is inspired by the Association of Internet Researchers (2019) guidelines, which state that ethics begins with “reflection on own research practices and associated risks.” Therefore, ethics is a method that should be embedded in research culture, and in the rest of this article, we aim to outline one such method of ethical reflection.
Using the DECIDE Framework
The DECIDE framework is hosted on OSF: https://osf.io/nsgxw/. The extensive resource set contains four main components: (a) the DECIDE framework spreadsheet, (b) the DECIDE desktop app, (c) information documents, and (d) flowcharts. A “frozen” version of all these documents at time of publication is available at https://osf.io/7qkre.
The DECIDE spreadsheet (see Fig. 1) can be downloaded from OSF and contains a list of questions for individuals or research teams to support reflection and/or discussions when embarking on a new research project (see column “Question”). The spreadsheet lists bespoke ethical questions for primary- or secondary-data considerations (see column “Type”). Every question has a related information document hosted on the OSF page to provide further support and discussion points (see column “Link”). These information documents can also be found in the files tab for the OSF project, with reference codes as titles (e.g., A1PD).
Screenshot of the DECIDE framework spreadsheet.
The DECIDE framework also encourages researchers to revisit each question at various points throughout a research project. To initiate this, the spreadsheet provides the research stage in which it would be beneficial to revisit each question (see column “Research Stage”). This is not intended to be prescriptive but instead a useful prompt to help encourage regular reflection. To incorporate feedback from the research community or to adapt as technology and data-collection methods continue to evolve, the spreadsheet and information files can be easily updated by any author of DECIDE (https://osf.io/nsgxw/).
However, researchers can also engage with the DECIDE framework using the desktop app (see Figs. 2 and 3). A link to the app can be found on the OSF page. The app pulls in “live” information from the spreadsheet and information documents, and therefore, it is automatically updated if any changes are made to these files. The app supports the easy navigation of the different ethical questions and displays information from the documents in an organized manner.
Screenshot of the DECIDE app allowing researchers to indicate whether they are embarking on a primary- or secondary-data research project.
Screenshot of the DECIDE app. A researcher can select a research stage and view relevant ethical considerations. The information provided should allow for reflection and ethical discussions whereby individuals or teams feel satisfied that they have considered each question in the context of their project.
When opening the app, the toolbar on the left side asks researchers to indicate whether they are planning a primary- or secondary-data research project. If a project involves both primary and secondary data, it is recommended that a researcher completes a sperate instance of DECIDE for each individual data set alongside a consideration of any data-linkage risks.
Once a data type has been selected, the DECIDE app displays two drop-down menus. The first lists several research stages. The second drop-down menu presents the ethical considerations relevant to each research stage. Ethical considerations are phrased as multiple-choice questions (e.g., “Have you considered how the findings of the research might be mis-used?”). A researcher can respond with the answers “yes,” “no,” or “need more information.”
If “no” or “need more information” are selected, the app presents the researchers with further information to help them contemplate the ethical consideration in relation to their own project. The Association of Internet Researchers (2019) guidelines stated that discussions should be had against the “accumulated experience and ethical reflection of researchers in the field,” and in providing additional information, we aim to mirror this guidance. The additional information provided by the app includes case examples from published studies and signposts for researchers to further resources to aid their ethical decision-making (e.g., British Psychological Society, 2021; Caliandro & Gandini, 2017)
We suggest that when researchers first use the DECIDE app that they review all research stages and questions sequentially by clicking the “next” button. This enables a research team to systematically examine each question, aiding the write-up of an ethically considered research proposal. If researchers refresh the page or revisit later, they can use the drop-down menus to return to the same place and then continue clicking “next” to see the remaining considerations. We suggest that researchers make notes for each consideration because these can be used when writing a research proposal, preregistration, or ethics application. Engaging in this process will provide researchers with some confidence that they have proactively identified and addressed ethical issues when using new forms of digital-behavioral data.
Once each question has been considered, researchers should record the date when discussions took place using the flowcharts provided (https://osf.io/nsgxw/). This creates a transparent record when demonstrating to others (e.g., funders, institutions, colleagues) a commitment to ethics and accountability. The flowcharts also help plan when each question will be revisited. Because of the nature of research and planning fallacies, the flowcharts allow researchers to document when these follow-up discussions actually took place. By completing this, researchers or a team are showcasing their willingness to engage with a dynamic ethics process. It flags the research as being ethically vigilant through proactively addressing any issues that may arise during the research process. Pertinently, all the above occurs before the wider dissemination of the research project, shifting ethics from being reactive to proactive.
Developing the DECIDE Framework
The framework was developed by all authors. The team had already engaged in several conversations surrounding the ethics of digital-behavioral data, which led to the cocreation of a literature review, documenting several key debates in the field. This review can be found as part of an earlier version of this article (https://osf.io/4q8bt), which documents the numerous publications we consulted. It became apparent that many of the ethical issues outlined in this review could have been mitigated if researchers had simply asked themselves a set of questions. So as a team, we began documenting what these questions would be. All six authors met regularly to review and refine these questions, and we further discussed which questions would be worth revisiting at each research stage. Given that our goal was to promote the discussion of each consideration among researchers and teams, we phrased each question to have a simple “yes” or “no” answer to indicate whether it had been thought about or discussed. We avoided open-ended questions, such as “Describe how your project has considered X consideration,” because this may imply that a detailed, formal response is required rather than an informal discussion. The team equally cocreated the expansive educational material linked to each consideration, and we built on the insights and writing generated via our initial literature review. The end product was a collection of questions paired with resources that can serve as a springboard for ethical discussions at different research stages. Researchers can skip any considerations they deem inapplicable to their research, and therefore, it is not required to verify or enforce that each question has been considered in the DECIDE framework app. Via an online survey, we then collected initial feedback on the framework (see https://osf.io/nsgxw/). Results highlighted its ability to guide ethical reflections and prompt researchers to consider important ethical issues. For example, one participant wrote the following:
I like how the DECIDE framework reflects the process through which scientific research is undertaken and how it prompts researchers and their collaborators to engage in discussion and thinking around these issues. It is clear to me that the DECIDE framework has been designed by researchers, for researchers - to empower those in our academic community to engage with digital data in a way that is purposeful.
Any reader or researcher wanting to provide further feedback on the DECIDE framework can do so by visiting the OSF page (see https://osf.io/nsgxw/).
Discussion
The DECIDE framework can help researchers evaluate research in a dynamic fashion. However, we recognize that any suggestion to have work reviewed, even as part of an ethical-review process, has the potential to add “red tape” and bureaucracy to the research process. Consequently, DECIDE is not meant to replace existing research-governance processes but instead provide questions to guide conversations. These discussions will, in turn, help balance how digital data can address key societal challenges and deepen theoretical understanding while mitigating any potential harm, an important debate that should continue to remain at the forefront of researchers’ minds today and in the future. Therefore, reviewing DECIDE’s considerations are likely to result in better quality and more ethically considered research outputs (Lakens, 2023).
The DECIDE framework can also be used by ethics committees and researchers. Specifically, it can serve as a tool for identifying potential issues within an ethics application, allowing committees to selectively address relevant considerations to provide feedback to researchers. Because these considerations are annotated with examples and guidance, ethics boards can use them to identify and communicate why an issue poses a risk and suggest potential solutions. This approach ensures that feedback is both constructive and actionable—something that in our experience is essential for the smooth operation of ethical-review boards.
Our “call to action” is simple. We want to instigate culture change by asking researchers to engage in meaningful ethical discussions throughout a research project. Our aim is to change the perception that ethics is an onerous and bureaucratic obligation that entails cumbersome paperwork and hinders the initiation of research projects. Instead, ethics is the process of reducing risk by engaging in reflections and group discussions that are sparked by the cumulative knowledge produced by the community.
The DECIDE framework is also a response to existing calls to actions made by the research community. Paxton (2024) stated, “researchers’ choices must grow out of personal accountability and a deep culture of scientific ethics, not mere adherence to legal and regulatory structures” (p. 499). This is because “the slow pace of governmental change means that legal and regulatory systems will always be a lagging indicator of permissible behaviours” (Paxton, 2024, p. 499). Consequently, the DECIDE framework can support a shift in culture by kickstarting discussions around emerging topics that are yet to be governed by formal rules and regulations. This can help avoid “ethical blind spots” whereby certain issues may be overlooked if they are not mentioned in existing research-governance policies. Note that “researchers have wanted to create a unified legal–ethical framework for online data for over a decade but have been stymied by a complex network of interacting technological, corporate, governmental and individual issues” (Paxton, 2024, p. 499). Therefore, we perceive grassroots initiatives, such as the DECIDE framework, to play an important role in mitigating emerging ethical issues that are difficult to regulate. DECIDE is designed and configured to be a dynamic resource that can be transparently amended and expanded to reflect the emergence of ethical conundrums and emerging practice.
The DECIDE framework also provides the means for researchers to adhere to three further calls to action that have been outlined by the research community. These are (a) recognizing the human involvement in data generation even without direct interaction, (b) documenting and sharing the ethical decision-making processes to promote thoughtful practices, and (c) for researchers to hold themselves to a higher standard than the minimum legal requirements (Paxton, 2024). The DECIDE framework meets these calls to action. It encourages researchers to consider the implications of their data collection, analysis, and sharing in projects that do not involve direct contact with individuals. Discussions prompted by the DECIDE framework can become an integral part of the ethical decision-making process, and these can be logged in the flowcharts provided or documented via other means. Many of the considerations in the DECIDE framework go beyond current legal requirements, so even if a project is technically permissible, it encourages researchers to also consider the moral implications.
Strengths and limitations
We developed the DECIDE framework with a focus on behavior data that are created when people interact with technology (e.g., digital interactions). This includes social media activity, app usage, sensor data, and other digital traces. However, the DECIDE framework remains flexible enough to address new and emerging forms of data. Any data that are laden with information about individuals have additional privacy concerns compared with data on nonhuman subjects (e.g., big data sets logging weather patterns). Therefore, it can support the mitigation of ethical issues in artificial-intelligence, Big Data, or human-subject projects more generally.
While the framework is designed to provide a robust foundation for ethical decision-making, we acknowledge that it may not align perfectly with all ethical viewpoints because these can vary widely across cultural and disciplinary contexts. Although grounded in existing research, the framework offers opportunities for further empirical validation to enhance its practical application in diverse real-world scenarios. To this end, we have provided a much-needed starting point when engaging researchers in constructive and actionable ethical discussions and reflections. Furthermore, we have provided the launch pad for future developments in this space. For example, as technology and data-collection methods continue to evolve, DECIDE may also require updating. However, the DECIDE framework has the flexibility and accompanying platform (the OSF) to be updated regularly, ensuring it remains relevant and effective over time. We encourage researchers to complement their use of DECIDE by seeking input from ethical experts, which can help tailor the framework to specific contexts and address any potential limitations effectively.
Conclusions
To conclude, the digital-data universe holds tremendous opportunity and potential to advance the scope and depth of current psychological understanding. However, alongside this progress comes risk—that research can have consequences that are often unforeseen and unrecognized (as illustrated by specific examples reflected on in the DECIDE framework information documents). The ubiquity of personal data, the power of computing algorithms, and the novelty of digital paradigms all contribute to uncharted waters when using emerging digital technologies. However, rather than just denying or recoiling from the scale of the challenges ahead, we posit that psychologists can be trained, enabled, and supported to make more ethical decisions. The DECIDE framework serves as a starting point, and it is our intention that psychologists are prepared to engage and lead research that uses new and emerging forms of digital-behavioral data.
Footnotes
Transparency
Action Editor: Yasemin Kisbu-Sakarya
Editor: David A. Sbarra
Author Contributions
