purdueglobalirb

Home · Newsletter

The Purdue Global IRB application checklist: every document drawn to the same dimensions

A complete Purdue Global IRB packet carries a research plan, the recruitment materials exactly as they will run, written site permission, a participant-facing document, an ethics training certificate and the protocol form for your pathway. Assembling them is the easy half. The half that produces returns is agreement: the same sample, instrument, procedure and confidentiality promise reading identically wherever they appear. A reviewer need not decide which version is right to send the file back.

Nora Castellano, MSN, RN · 2026-08-23

Short answer

Purdue Global asks for a research plan summary, recruitment plan and materials, written site permission on letterhead, the participant document for your pathway, an assent form where minors are involved, and the training certificate — attached to the protocol form the pathway selects.

What goes in a complete Purdue Global IRB file?

The university enumerates it. Its submission guidance lists the required information as a research plan summary, a recruitment plan and any recruitment materials, agency approval in writing, an informed consent document where surveys or interviews are involved, an assent form or script where children take part, and a certificate of IRB training. Its Investigator Guide adds the pathway-specific layer on top: the protocol form and the participant template that your chosen pathway pulls.

What every packet carries, and what the pathway adds
DocumentApplies toThe dimension it is measured on
Protocol formPathway-specific: determination, exempt or expeditedWhether the study described matches the pathway claimed
Research plan summaryEvery packetQuestion, participants, setting, method, protections — stated, not implied
Recruitment plan and materialsWhere people are approachedThe artifact itself, word for word, as it will appear
Site permissionWhere a site, agency or employer is involvedWritten, on letterhead, signed by someone with authority, from an institutional address
Participant documentExempt information sheet or expedited consent formEvery required element present, and the version matching the protocol
Assent form or scriptWhere minors participateWritten for the age group it addresses
Training certificateEvery packetAn accepted course, current, held by the people who need it

Purdue Global also states the mechanical conditions around submission: use a Purdue Global email address, make sure the required items are attached because missing items delay the review, and expect the request to route to your thesis advisor, consultant or instructor, or supervisor for approval before the reviewers see it. And on the pathway templates, the instructional text printed in red is meant to be deleted and replaced with your own content in black.

What must the research plan summary actually answer?

Purdue Global frames the research question in a way worth taking literally: state what you intend to examine, not what you expect to find. A question written as a prediction has already told the reviewer that the analysis has a preferred outcome, which is a different conversation than the one you want to be having.

Beyond the question, the summary has to name people and places. Who are the participants, and what is your relationship to them — a distinction that matters enormously when the answer is "my own staff", "my own patients" or "my own classroom", because that relationship carries an influence question the board will want addressed. Where does the work take place. Then the methodology, described so that the protections are visible inside it: how information is collected and recorded, who administers the instrument, where the data is stored and who can open it, how long it is retained before destruction, and whether anything personally identifiable is collected at all. The guidance even reaches the detail of whether an online survey tool captures addresses that could identify a respondent.

Notice what kind of answers those questions want. Every one is answerable with a noun — a named platform, a named role, a named point at which identifiers come off. "Data will be stored securely" answers none of them, and it can be written by someone who has decided nothing.

What must a site permission letter say?

More than that the site is pleased about your project. Purdue Global's conditions are concrete: permission in writing, on letterhead, signed by an official, sent from an agency email address rather than a personal account. Where recruitment runs through an online community, its guidance points you at that platform's terms of use rather than treating a moderator's blessing as permission.

What the letter should also do — and this is where most letters we read fall short rather than fail outright — is name the access being granted in the same terms your protocol uses: which records will be reviewed, which people may be approached and how. A letter granting "support for this research" and a protocol describing eleven staff interviews are two readings of one dimension. The business and education sequences at Purdue Global make site permission a program condition as well as a board one, so the letter is load-bearing twice over.

What must the participant-facing document contain?

It depends which pathway you are on — an exempt research information sheet and an expedited informed consent form are different instruments, as the review paths article sets out. Where a consent form is required, the elements are enumerated in federal regulation at 45 CFR 46.116, and enumerable requirements are the easiest kind to fail, because a reviewer can check them off:

  1. A statement that the activity is research, with its purpose, the expected duration of the person's participation, and a description of the procedures.
  2. Reasonably foreseeable risks or discomforts.
  3. Benefits reasonably to be expected, to the participant or to others.
  4. Alternative procedures or treatments that might benefit the person, wherever such alternatives exist.
  5. How far the confidentiality of records is going to extend.
  6. Where risk is more than minimal, whether compensation or treatment is available for injury.
  7. Whom to contact about the research, about participants' rights, and in the event of harm.
  8. That taking part is voluntary, that declining forfeits no benefit the person is otherwise owed, and that they may stop whenever they choose.
  9. A statement about whether identifiable information or specimens might be used or distributed for future research.

The regulation also asks that consent begin by presenting, briefly and in focus, whatever a person most needs in order to decide. That is easy to satisfy and easy to skip; a form opening with institutional throat-clearing satisfies the list and not this clause. Purdue Global's toolkit publishes a procedure for informed consent requirements and a reviewer worksheet on criteria for consent, so it is checked against a written standard.

Which dimensions have to agree across every document?

These are the cross-checks. Each is a single number, name or claim appearing in more than one document, and each is a place a file can quietly disagree with itself. Read them with the documents closed, then open, and note where the readings differ.

  1. The sample. How many participants, stated identically in the protocol, the recruitment material, the consent and anything the site was told.
  2. The population. Who is eligible and who is excluded, in the same words everywhere.
  3. The instrument. Named once and used with that exact name in every document — not "the survey" here and the tool's full title there.
  4. The time asked of a participant. If the consent says one length and the recruitment message another, the reviewer has found a defect without reading the science.
  5. The procedure. What actually happens to a participant, in the same sequence in the protocol and the consent.
  6. The data claim. Anonymous, de-identified, coded or identifiable — these are four different promises, and only one of them can be true.
  7. The site. The setting named in the protocol, named in the letter, and named in the consent.
  8. The version. The consent in the packet must be the consent the protocol cites, dated so that both are provably the same document.
  9. The category. The pathway claimed on the cover must match the design described in the methodology.

Nine readings. Returns come from here far more often than from an absent attachment: a missing document is obvious and gets fixed before submission, while a drifting dimension survives every proofread because each document is internally correct. That is why our first move reads every document against every other.

What does the board read the file against?

Its own worksheets. Purdue Global's toolkit publishes reviewer instruments — criteria for approval and criteria for consent — alongside standard operating procedures for initial review, exemption determinations and expedited reviews. Behind them sit the approval criteria at 45 CFR 46.111: sound design holding risk down, that risk weighed against the benefit anticipated, fairness in who is chosen, consent obtained and recorded or waived on proper grounds, monitoring of data where the study warrants it, privacy and confidentiality protected, and extra safeguards wherever participants are vulnerable. A packet drawn against those seven headings is drawn against the marking scheme.

What if the study crosses languages, sites or a transcription service?

Then Purdue Global has already published the instrument, and the mistake is improvising instead of finding it. Its toolkit carries guidance for international research and for non-English materials, a confidentiality agreement for transcription and translation services, parameters for letters of collaboration, an authorization agreement for use with another board, an amendment guidance sheet and an appendix for multi-site work. A study that sends audio to a transcriptionist without accounting for it in the confidentiality claim has a dimension reading two ways: the consent promises only the researcher hears the recording, and the workflow says otherwise.

After approval, changes move through the post-approval request the guide links, and the board maintains procedures for amendment requests, records management and protocol closure. The file does not end at approval; it ends at closure, and the documents keep having to agree the whole way. The questions page covers amendments and returns in short form, and the step-by-step sequence puts the route in order.

What to do next

Print nothing. Whatever exists — a finished packet, or a proposal and three half-drafts — goes through the intake sheet, and the first measure costs nothing. Every document gets read against every other, the way a reviewer will read them, and the reply is a written findings list: each drift with the places it appears named, each element the board requires that this file does not yet state, and the word clean against anything needing no work. Prefer to hand the plate over? Then category, documents drawn to one set of dimensions, filing and every reviewer reply become ours until a letter arrives. Throughout it, the question is yours, the data is yours, the conclusions are yours. The ruling is the board's alone.

Sources

Give us the entire IRB process. Keep the study.

The free application review is where it starts. From there we carry everything — the determination and the plan, every document drawn to the same dimensions, the submission itself, and every reply to the board until the approval letter exists. The study, the data and the findings remain yours; the board's decision is its own.

Request the free application review Send whatever exists — drafts count.
Has the board measured your file yet?