Researching your own workplace for a Purdue Global study: dual roles, coercion and the board's questions

Why does the board treat a workplace study differently?

Because voluntariness is harder to demonstrate when the researcher and the participant already have a relationship with consequences in it. The federal rule builds this in twice. Under 45 CFR 46.111(a)(3), selection of participants must be equitable, and the board is directed to be particularly cognizant of the special problems of research involving categories of people vulnerable to coercion or undue influence. Under 46.116(a)(2), consent may be sought only under circumstances that give a prospective participant sufficient opportunity to discuss and consider whether to take part, and that minimise the possibility of coercion or undue influence.

Purdue Global operationalises both. Its reviewer worksheet, IRB-314, carries a criterion demanding extra safeguards wherever coercion or undue influence is a live risk, and its protocol template for expedited research asks investigators to tick the participant groups involved — with Employees as one of the named boxes. Ticking it is not a confession. It is where the file starts answering a different set of questions.

What does Purdue Global say about convenience samples?

More plainly than most published guidance. The IRB-314 worksheet states that recruiting samples of convenience is generally not allowed, and names among such samples the use of departmental email lists and the recruitment of a known group of participants from a workplace or organisation familiar to the principal investigator. It goes on to say that recruitment strategies should align with the study's stated aims and research questions so that the recruited population gives the diverse participant pool needed to produce generalisable results — and that where answering specific research questions requires recruiting from specific pools, that must be justified.

That last clause is the door. The prohibition is not on studying your own setting; it is on studying it because it is yours. A file that says, in effect, "these are the people I can reach" has failed. A file that says "this question is about this population, in this kind of setting, and here is why the finding would not transfer from another" has done the work the worksheet asks for. Write that justification into the protocol's own population section, not a covering note; the protocol is what gets read against the criteria.

What questions does the protocol itself put to you?

Purdue Global's expedited protocol template asks a sequence a workplace study answers with unusual difficulty, which is why those answers are worth drafting first. It asks how potential participants will be identified — offering "individuals known to the investigators" as one of its own examples. It asks how the investigators will gain access to that population, and for evidence that the necessary number can be recruited. It asks how, when and by whom recruitment will happen, and in what setting. And it asks how the recruitment process respects privacy, offering examples such as limiting how many members of the team see identifiable information.

In a workplace study those questions interlock. Access is usually the relationship; identification is usually knowing people; the setting is usually somewhere participants can be seen being recruited. Each answer has to be given without leaning on the relationship as the reason. The template's own distinction is worth borrowing while you draft: privacy concerns the person and their information during collection, confidentiality concerns the management of data afterwards. A workplace study can pass on confidentiality and still fail on privacy — an interview room with a glass wall on the participant's own unit is a privacy finding no encryption plan repairs.

How is the dual role answered rather than assumed away?

By separating the roles in the record, in four places, so the same separation is visible wherever a reviewer looks.

• In the protocol. State your position at the site, state the relationship between that position and the participants, and state what you will not do in the researcher role that the working role would permit — accessing records you can already open, for instance.
• In recruitment. Name who invites, and make it someone whose invitation carries no employment weight. State how a person declines without anyone learning they declined.
• In consent. Say, in the participant's language, that participation is voluntary, that refusal and withdrawal carry no consequence for their work, and who to contact — and take consent somewhere and at a time that makes those sentences credible.
• In the data plan. Show what management receives, at what level of aggregation, and how small groups are handled so that a unit of three people cannot be reverse-read from a finding.

Where those four agree, the dual role has been drawn. Where they disagree — a protocol naming a neutral recruiter, a site letter with your manager circulating the invitation — the file has a tolerance failure in its most sensitive dimension.

Do you need site permission, and is quality improvement a way around this?

Site permission, yes, in the ordinary case. Purdue Global's initial-review procedure lists among required submissions a letter from an appropriate administrator permitting research activities on the premises, where research or recruitment happens in or through schools, businesses, care facilities or other organisations; program material for its applied doctoral work puts the same requirement in the candidate's own terms. The letter should describe the actual activity — which records, which people, what kind of contact — and it must agree with the protocol, because a reviewer reads both.

As for quality improvement: it is a real category, not a workaround. Purdue Global publishes a human research determination pathway, worked through its IRB-310 worksheet, for activities such as quality assurance, quality improvement or program evaluation intended for local improvement — work meant to benefit a single site or practice rather than to contribute generalisable knowledge. That description is also its own boundary. A comparison group, a validated instrument and an intention to publish generalisable findings all pull against a local-improvement claim, and the reviewer reads the design rather than the label. Where the project genuinely is local improvement, that pathway is the honest route and it produces a letter confirming non-research status.