Questions, answered first.

Purdue University Global — the online university built for working adults — and the IRB process its doctoral studies and capstones pass through; not Purdue University in West Lafayette, Indiana, which is a separate institution with a separate board and separate procedures.

The two share a name and a system and almost nothing else that matters to a human-research file. Forms, portals, training requirements and review cycles are set separately, so guidance written for one does not transfer to the other. Everything on this site concerns Purdue Global. If you are preparing a file for West Lafayette, nothing here describes your board.

If the project collects data from or about living people — surveys, interviews, observations, records with identifiers — it needs a determination from the university's IRB before any of that data exists, and the university's current handbook or IRB portal sets exactly what must be filed and when.

The determination may be that the project is not human-subjects research, or that it is quality improvement, or exempt, or expedited, or full board — but that reading is the board's to make, and it is made before collection, not after. A project that turns out not to need review still usually needs the board to say so. The safe reading is to assume a file is required and to let the determination settle how much of one.

Most returns are tolerance failures: the same dimension — the sample number, an instrument's name, the site count, the recruitment wording — read in two documents and found different, or a required element the board needs stated plainly and the file does not state.

The board reads one study across several documents — proposal, application, consent, recruitment pieces, site letter — and agreement between them is invisible while drift is a finding. "Twelve" in the method chapter and "fifteen" in the consent is the classic return; so is a consent from a superseded protocol version, a site letter that describes different data from the application, a data-security plan that says "kept secure" and names nothing, and a category claim the design contradicts. None of these is a judgement on the study's quality. All of them are measurable before submission.

No. The approval letter is the datum: the day it exists, data may begin to exist, and anything collected before it is not data the study can use — no amendment reaches backwards to make it valid.

The same edge governs change after approval: a new site, a reworded flyer, a revised instrument moves through amendment before it runs. Recruiting "just to line people up", piloting an instrument on future participants, or pulling records "to see what is there" all sit on the wrong side of the line. It is the one dimension with no tolerance at all, and the cheapest to hold, because holding it costs only patience.

You need the human-research ethics training the university's current handbook names — CITI is the usual program — current on the day the file is read, and in the module set required for your role and your design; the certificate is read before the study is.

Two measurements, not one: currency and module set. A lapsed certificate stops the review at the cover sheet; so does the wrong module set — social-behavioral where biomedical was required, or the reverse — before a single idea of yours has been considered. Which modules, and how long a certificate is honoured, are the university's to set; check Purdue Global's current handbook or IRB portal and, if anything has changed since you trained, retake before you file.

If any data is collected at, from, or about an organisation you do not control — a clinic, a school district, a practice, an employer — you need a letter on that site's letterhead, signed by someone with the authority to grant access, that names the actual data: which records reviewed, which people interviewed, how many minutes observed.

"Happy to support this study" is goodwill, not permission, and a board cannot act on it. The letter must also agree with the application: if the letter and the form describe different data, that is one dimension measured two ways and the file comes back to have it reconciled. When the site has its own review board, that board's approval is a separate dimension; the university's handbook describes how the two are sequenced.

They are three levels of review for human-subjects research, set by the federal Common Rule and applied by the university's board: exempt for minimal-risk research in defined categories, expedited for minimal-risk research outside them, full board for anything above minimal risk or involving vulnerable populations — and the board, not the applicant, decides which applies.

Exempt does not mean "no file" — it means a shorter determination, and the board still makes it. Expedited means review by one or more designated members rather than the convened board. Full board means the convened committee reads it. Claim the level your design actually supports: a comparison group, a vulnerable population, sensitive identifiable data will contradict an exempt claim, and the reviewer reads the design, not the tick-box. The university's current handbook names its own categories and forms for each path.

Quality improvement measures a local process to improve it in place; research produces knowledge meant to travel — and the category you claim must match the design you drew, because the reviewer reads the design.

A comparison or control group, an instrument you are validating, a hypothesis about a population beyond the site, a plan to publish generalizable findings — each of these points toward research whatever the cover sheet says. A single-site cycle of measure, change, re-measure with no intent to generalize points toward QI. Many DNP projects sit honestly on the QI side; the board's own determination governs, and the university's handbook sets the form for requesting it. Claim the study you actually drew and the category question becomes the easiest one on the form.

The application in the university's current form; the consent (and any assent or information sheet) with every required element and the right version; the recruitment materials exactly as they will run; the instruments, named once and used everywhere with those words; the data-security plan with a name in every answer; site permission letters naming the actual data; and your current training certificate — each drawn to the same dimensions as the protocol of record.

The exact list is the university's to set and Purdue Global's IRB portal states it; what does not change is that the board reads the pieces together, so the file's first job is to agree with itself. Sample size, site count, instrument names, recruitment wording and the data-handling promises must read identically wherever they appear. That agreement is the whole of the complete file; the documents are only where it is measured.

The required elements, each present and findable: the purpose and what participation involves, the risks and the benefits, the limits of confidentiality, the right to decline or withdraw without consequence, whom to contact about the study and about their rights — and the version stamp that ties it to the current protocol.

A consent is enumerable, and a return names the element that is missing. The other return is version drift: an old consent stapled to a revised protocol, which boards are trained to measure first. The university's template, where it provides one, sets the order and wording of the elements; the numbers inside it — how many, how long, how many sites — must match the application and the proposal to the digit.

Four answers, each with a name in it: where the data is stored (which encrypted service, not "a laptop"), who can open it (named roles, not "the research team"), when identifiers come off and where the key linking codes to names is kept, and when and how the data is destroyed.

"Data will be kept secure" is the one sentence reviewers are trained to send back, because it measures nothing. Name the storage, name the roles, date the de-identification, date the destruction. If a site's own policy governs part of the plan, say which policy and where it applies. The plan must also agree with the consent's confidentiality promises — one dimension, two documents.

The return letter names findings; you answer each one on the record and fix the drawing everywhere the finding appears — not only the sentence the board quoted — then resubmit; a return is a question the file could have answered, not a verdict on the study.

The discipline is the same as the file's: if the board found "twelve" against "fifteen", the response changes every document that carries the number, and says so. A response that fixes one place and leaves another creates the next return. Read the letter for what it measured, reconcile the dimension across the whole file, answer point by point, and resubmit through the university's portal in the form it asks for. In Full Handover we draw the response and answer every reply after it until the approval letter exists.

An amendment is the board's approval of a change to an approved study, and you need one before the change runs — a new site, a reworded flyer, a revised instrument, a changed sample, a new team member — because the approval covers the study as drawn, not the study as it evolves.

The change is drawn first, filed through the university's portal in its amendment form, approved, and only then run; the first site need not stop while the second is added. Running a change the board has not seen is operating outside your own approval, and no amendment reaches backwards to cover it. Continuing review is the related obligation: when a study runs long enough, the board asks for a periodic report before approval lapses, and the university's handbook states its cycle.

No to both: the board's decision is its own and no review by this practice changes, predicts or guarantees it; and we do not quote review durations or our own turnaround times, because an honest number does not exist before the file is read and the board's timing is the board's.

What we do promise is the drawing — one study, every document to the same dimensions, every required element present, every reply answered — which removes the returns that are tolerance failures and leaves the board reading the study rather than the file. Purdue Global's current handbook or IRB portal describes the board's own cycle; a consultant who quotes you a faster one is selling, not measuring.

We do not conduct the research — the study, the data and the findings stay yours — but we handle the entire IRB process end to end if you want it handled, and the cost is quoted in writing after the free application review, as a fixed fee for one of three packages: Determination & Plan, The Complete File, or Full Handover, the premium, which carries every document, the submission and every reply until the approval letter exists.

No prices appear on the site because an honest figure does not exist before the file is read; nothing is billed by the hour and nothing is added afterwards. The free review may also return one line — submit it — with no quote at all. The pricing sheet sets out the three packages and what moves the quote; how it works sets out the three moves they carry.