Ethics verdicts
Introduction to Social Research Methodology
Your brief
You are the junior research team at Meridian, a mid-sized Polish retail-and-services company: about 450 employees, 32 stores, and a growing e-commerce arm. Staff turnover has risen from 14% to 27% in two years, customer satisfaction is falling, and the board wants evidence, not hunches. Ideas for studies are now arriving from all over the company — and it has fallen to your team to act as Meridian’s ethics review panel.
Five research proposals are on your desk. None of them comes from a bad person; each comes from someone with a real problem and a plausible-sounding plan. Your job, working in groups of 3–4, is to deliver a verdict on each proposal:
| Verdict | Meaning |
|---|---|
| Permissible as it stands | The proposal can go ahead in its current form |
| Permissible with safeguards | The proposal is acceptable only if specific, named changes are made |
| Not permissible | No realistic safeguard fixes it; the study should not run in this form |
For every verdict you must do two things:
- Justify it against the principles from today’s session — voluntary participation, informed consent, no harm, privacy/anonymity/confidentiality, deception and covert observation, power asymmetry, purpose limitation, and data minimisation. Name the specific principles at stake; “it feels wrong” is not a justification.
- Name the safeguard(s) that would fix the problem, where one exists — and if you conclude that none exists, say why.
Be precise about vocabulary: if a proposal promises “anonymity”, ask whether it can actually deliver anonymity or only confidentiality. A verdict of “permissible with safeguards” is only as good as the safeguards you specify.
Task 1 — Read and assign (5 minutes)
Read all five cases below. In your group, briefly assign each case a first-impression verdict — one word, no discussion yet. You will revisit these at the end and see whether analysis changed your instincts.
Task 2 — Deliberate the cases (30 minutes)
Work through the five cases in order, spending about 6 minutes on each. For each case, record in writing:
- your verdict (one of the three options);
- the principles at stake (be specific — which principle, breached how);
- the safeguards that would fix it, if any — stated concretely enough that the proposer could act on them tomorrow;
- one sentence on what could go wrong for a specific person if the study ran unchanged.
Appoint a spokesperson: each group will present one case in the plenary.
Case A — The stockroom camera
Shrinkage (unexplained stock loss) at three Meridian stores has doubled. The regional operations manager proposes a study: install two additional cameras covering the tills and stockroom in each store, without telling staff, and have a research assistant code two weeks of footage for “handling irregularities” — noting which employee is on shift when each irregularity occurs. “If we announce it, behaviour will change and we’ll learn nothing,” he argues. “Besides, the employment contract already says the premises are monitored.”
Case B — The chat-log study
Meridian’s head office wants to understand why some store teams cooperate smoothly while others fracture. An internal analyst proposes to study team communication by exporting six months of message history from the company’s internal chat platform for eight store teams, and analysing tone, response times, and who talks to whom. Employees know the platform belongs to the company; they have not been told their messages will be studied. The analyst plans to read messages in full but promises the final report will “only discuss patterns, not individuals”.
Case C — The manager’s own survey
The manager of Meridian’s flagship Warsaw store is proud of her leadership style and wants data to prove it works. She proposes a CAWI survey, designed and sent by herself, asking her 14 subordinates to rate her own management — fairness, communication, support — on named accounts (“so I can follow up individually on any concerns”). Participation is described in the invitation as “voluntary, but I hope and expect everyone will contribute to making our store better”. She will read all responses herself.
Case D — The recycled exit interviews
An HR analyst studying the turnover crisis realises that Meridian already holds a goldmine: 74 exit interviews conducted over the past two years. Departing employees were told at the time that the interviews were “confidential and for HR purposes only”. The analyst proposes to re-analyse the full transcripts — which contain names of managers criticised, health disclosures, and personal reasons for leaving — for a new study of why staff quit, and to quote “anonymised” excerpts in a report to the board. Contacting former employees for fresh consent would be “impractical”.
Case E — The six-person team survey
Customer complaints about Meridian’s e-commerce support line are rising. The head of digital proposes a morale survey of the support team, run by an external agency, with a cover message promising that “responses are completely anonymous”. The survey asks for role, tenure, and satisfaction with “your direct supervisor”, plus open-text comments. The support team has six members. Results, including a breakdown by tenure band and the verbatim comments, will go to the head of digital — the direct supervisor in question.
Task 3 — The pattern check (5 minutes)
Look across your five verdicts as a set and answer these questions in your group:
- Which single principle was breached most often across the five cases?
- Which case was closest to the line — the one where your group hesitated longest or disagreed most?
- Compare your final verdicts with your first impressions from Task 1. Where did deliberation change your mind, and what changed it?
Task 4 — Plenary discussion (15 minutes)
Each group’s spokesperson presents one case (assigned by the instructor): verdict, principles, safeguards — two minutes maximum. Where groups disagree on a verdict, we will hear both justifications before moving on.
Then we discuss as a whole class. Be ready for these prompts:
- Several proposers argued that the data “already exists” or that monitoring is “already in the contract”. Why is lawful access not the same as ethical research use? What concept from the session captures the difference?
- Case A’s manager claims announcing the study would destroy its value. Is that ever a sufficient justification for covert observation of employees? What test would a covert design have to pass — and does his pass it?
- In Case C, suppose the invitation had said nothing about “hoping and expecting” participation. Would consent then be free? What is it about the situation, rather than the wording, that constrains refusal?
- Case E promised anonymity that a six-person team cannot support. Which is more damaging: making no confidentiality promise at all, or making one and breaking it? What should the cover message have said instead?
- If Meridian had an ethics committee, which of the five proposals do you think would have been submitted to it at all? What does that suggest about where ethical review is most needed — and who has to do it when no committee exists?