Who this list is for
This list is for doctoral researchers in management, marketing, information systems or business ethics who want to study how organisations take responsibility for their digital practice: how they collect and use data, how they deploy automated systems, and how they answer for the consequences.
Gursoy et al. (2025) review the corporate digital responsibility (CDR) literature and set out future research directions across ethical, societal and environmental dimensions of digital business practice. Alongside that review, two features of the wider field shape what a doctorate can contribute. First, published ethics guidance converges on a small set of principles such as transparency, fairness and accountability (Jobin et al., 2019), while how firms act on them varies, and principles alone do not settle practice (Mittelstadt, 2019). Second, regulation is now explicit: the European Union’s artificial intelligence regulation imposes documentation, risk-management and governance duties on providers and deployers of higher-risk systems (European Parliament & Council of the European Union, 2024). Both make implementation, not intention, the interesting object of study.
How to use this list
Treat each row as a starting point. Read the review, then the two or three empirical papers closest to your row, and check what has appeared since. The tables give a possible research question, the design and data it would most likely need, and why the topic matters now. If you are still choosing a broad area, start with how to choose a PhD topic and how to select a PhD topic. Neighbouring lists cover business and management, governance in modern enterprises and cybersecurity in business; read them so your project does not overlap a neighbouring one.
Theme 1: How digital responsibility is built and governed
The first strand asks how responsibility practices come into existence: who proposes them, which committee owns them, what happens when they conflict with a product deadline. These are process questions, and process questions are answered by following organisations over time or by comparing cases, not by a single survey. Audit practice is one useful entry point, since ethics-based auditing now has a documented literature of its own and a stakeholder-facing purpose (Laine et al., 2024).
| Topic | Possible research question | Likely design and data | Why it matters now |
|---|---|---|---|
| How digital responsibility practices develop in one sector | How do firms in one sector develop digital responsibility practices across a defined period, and what triggers each change? | Longitudinal comparative case study of three to five firms: interviews at two or three points, policy documents, board and committee minutes where available (Eisenhardt, 1989) | Process evidence on how these practices emerge is scarce in the reviewed literature (Gursoy et al., 2025) |
| Governance structures and accountability for digital decisions | Which governance arrangements do firms use to assign accountability for data and algorithmic decisions, and which arrangements do staff report as effective in practice? | Comparative case study plus document analysis of published policies, audit reports and regulatory filings; interviews across compliance, product and engineering roles | Regulation now requires named governance and documentation duties, so arrangements can be compared against a reference point (European Parliament & Council of the European Union, 2024; Laine et al., 2024) |
A warning about access. Firms rarely open their ethics committee minutes to a doctoral researcher without an existing relationship. Build the project around what you can reach: published policies, audit reports, regulatory filings, and interviews with practitioners who will speak in a personal capacity. State in the proposal what you will do if one case withdraws.
Theme 2: Measuring digital responsibility
The second strand is measurement. Without a defensible measure, questions about outcomes cannot be answered, and the review notes the absence of validated instruments as a limit on empirical progress (Gursoy et al., 2025). Two routes are open to a doctorate: develop a self-report instrument properly, or build an observable measure from disclosures and documents. The second avoids the problem that firms and respondents have reason to overstate what they do.
| Topic | Possible research question | Likely design and data | Why it matters now |
|---|---|---|---|
| Developing and validating a corporate digital responsibility scale in one population | Which dimensions make up digital responsibility as managers in one defined population understand it, and does a scale built on them show acceptable reliability and construct validity? | Item generation from the literature and interviews, expert review, pilot, then two independent samples for exploratory and confirmatory analysis; test discriminant validity against existing responsibility scales (Podsakoff et al., 2003) | Empirical work is limited by the lack of a validated instrument (Gursoy et al., 2025) |
| A disclosure-based measure of digital responsibility practice | Can digital responsibility practices be scored reliably from public disclosures, and how does that score compare with managers’ self-reports in the same firms? | Content analysis of reports and policies with a published coding frame and two independent coders; inter-rater reliability; comparison with survey responses from a subsample | Disclosure-based measures are established in the environmental, social and governance literature and can be adapted here (Brooks & Oikonomou, 2018) |
Scale development is a legitimate doctorate, but it is a methodological project: plan for two data collections, not one, and be explicit that a new instrument needs validation beyond the sample that produced it. If your real interest is outcomes, borrow an instrument and spend the time on the outcome model instead.
Theme 3: Stakeholder response and firm outcomes
The third strand asks whether digital responsibility is noticed and whether it pays. Here the main risk is overclaiming. A correlation between a responsibility score and firm performance has many plausible explanations, and the literature on environmental, social and governance disclosure shows how sensitive such results are to measurement and model choice (Brooks & Oikonomou, 2018). Say what your design can and cannot support.
| Topic | Possible research question | Likely design and data | Why it matters now |
|---|---|---|---|
| How one stakeholder group judges corporate digital responsibility | How do customers (or employees, or investors) in one market interpret a firm’s digital responsibility claims, and which claims affect trust rather than suspicion? | Experimental vignettes with manipulated claim types, plus interviews to develop the stimuli; measure trust and legitimacy with published scales | Evidence on how stakeholders receive these claims is thin relative to the volume of firm-side writing (Gursoy et al., 2025) |
| Digital responsibility practice and firm performance over time | Is a disclosure-based measure of digital responsibility associated with later financial or reputational performance, once sector, size and prior performance are controlled? | Panel dataset of listed firms over five or more years; fixed-effects models with lagged predictors; report sensitivity analyses and the limits on causal inference (Wooldridge, 2010) | Longitudinal outcome evidence is scarce in the reviewed literature (Gursoy et al., 2025) |
For the second row, decide early whether the question is about firm value, reputation or employee retention, and build the sample around a measure you can obtain for every firm and year. Panels with gaps in the responsibility measure are common and will dominate the write-up if not planned for.
How to test a topic before you commit
Run each candidate through the FINER criteria: feasible, interesting, novel, ethical, relevant (Hulley et al., 2013). Four checks matter most here.
- Access: for case and interview work, name the firms and the person who can introduce you. For panel work, confirm the responsibility measure exists for the years you need.
- Measurement: decide whether your measure is self-report, disclosure-based or behavioural, and defend the choice in the proposal.
- Regulatory grounding: anchor normative claims in a verifiable source such as a statute, a standard or an audit framework rather than in a list of principles (European Parliament & Council of the European Union, 2024; Mittelstadt, 2019).
- Novelty: search for work published since the review; this field moves quickly, and a 2025 gap may be partly filled.
For turning the chosen row into a problem statement and proposal, see problem identification, gap identification and the research proposal.
Common mistakes
- Treating corporate digital responsibility as a slogan and studying attitudes towards the slogan rather than practices.
- Promising a validated global measure; a doctorate validates an instrument in a defined population.
- Using the same three research questions for three different titles.
- Reading published ethics principles as evidence of what firms do (Jobin et al., 2019; Mittelstadt, 2019).
- Inferring that responsibility causes performance from a cross-sectional correlation (Brooks & Oikonomou, 2018).
- Collecting the predictor and the outcome from the same respondent in one questionnaire (Podsakoff et al., 2003).
Key takeaways
- Corporate digital responsibility research divides usefully into governance, measurement and outcome questions (Gursoy et al., 2025).
- Published ethics principles converge while practice varies, so study implementation rather than intention (Jobin et al., 2019; Mittelstadt, 2019).
- Regulation now supplies verifiable duties against which governance arrangements can be compared (European Parliament & Council of the European Union, 2024).
- Disclosure-based measures borrowed from the environmental, social and governance literature are often more defensible than self-report (Brooks & Oikonomou, 2018).
- Narrow each row to one question with one coherent design, then test it for access, measurement and novelty (Hulley et al., 2013).
