Where a Herzing IRB file has to agree with itself, document by document.

What is actually in the file?

More pieces than the word "application" suggests. Working from Herzing's own page-one manifests and the numbered steps of the two forms, a submission is assembled from these:

1. The application form, every step answered — identification, location and sponsor, faculty details, projected start and completion dates, and the substantive middle.
2. The substantive middle itself: on the improvement form, an aim statement, an overview setting out the purpose, the strategy, the measures and the outcomes you are after, and the improvement framework you will run; on the research form, a P.I.C.O.T. research question, an abstract naming purpose, variables and value, and a methodology section.
3. Every questionnaire and survey that will be used, attached in full.
4. The informed-consent documents, on the research route — everything that will be put in front of a participant before they take part.
5. A confidentiality and anonymity attachment, on the research route, describing how privacy will be maintained and confidentiality assured. This is a separate item from consent, and it is routinely forgotten because it feels like a repeat.
6. The data-security description, covering physical and electronic data alike.
7. The training certificate for those named, free to Herzing affiliates through Canvas.
8. Any external board's application and its approval letter, where another review body is involved.
9. The program chair's letter of approval from an affiliated institution, where that applies.
10. The signatures — three on the improvement form, including a program chair; on the research form, the principal investigator and, where the investigator is not faculty, a faculty sponsor who signs as co-principal investigator.

Ten items, and only the first two are things you sit down and write. The rest are gathered — which is why an inventory taken late is an inventory taken too late.

Which facts have to repeat, and where?

These are the cross-references a reviewer checks almost without deciding to. Read your own file down the middle column and then hunt every entry in the right-hand one.

Why does a paraphrase count against you?

Because the reviewer cannot tell a paraphrase from a change. If the overview says a practice change reaches two medical-surgical units and the consent form says it reaches "the unit", one of two things is true: either you wrote the same fact twice in different clothes, or the design shifted and one document did not hear about it. The reviewer has no way to know which, so they ask — and asking is what turns a single pass into a correspondence.

The discipline that removes this is unromantic: settle the plan first, then quote it. Not summarise it, not adapt it to each document's tone — quote it, word for word and number for number, into every field that touches it. Nobody has ever had an application returned for being repetitive.

The same rule governs versions. A superseded consent draft left in the folder is not a spare copy; it is a second, contradictory answer to a question the board asked once. If the application cites version four, version three does not travel with it.

What must the consent file carry on its own?

Herzing's research route asks for every consent document a participant will see beforehand, and federal law sets what those documents contain: that the activity is research and why it is being done, what will happen and for how long, foreseeable risks, any benefits, alternatives where relevant, how confidentiality will be handled, who to approach with questions about the study and who to approach about participants' rights, and a plain statement that taking part is voluntary and can be stopped at any moment without penalty.

Two features of the current rule catch people out. Consent has to lead with a short, sharply focused account of whatever would most help a person decide one way or the other — arranged so it can be understood, rather than piled up as isolated facts. And it may contain no exculpatory language: nothing that has a participant waive, or appear to waive, any legal right, or release anyone from liability for negligence.

Then there is the reading level, which is a legal requirement rather than a courtesy — the information has to be in language the person signing actually understands. Clinicians can be given clinical language. The people in their care usually cannot, and a consent form pitched at the person filing it rather than the person signing it is the single most-returned document in this file.

What do the data-security answers have to name?

Herzing asks the same four things on both forms, and asks them about physical and electronic data together. Vague answers read as no answer at all.

• Where the data will be stored — a specific place, not "securely".
• How that location or computer system is secured — the actual control, named.
• How long the data is retained — a period you could put on a calendar.
• How old data is disposed of — a method, not an intention.

Answer those as facts about your project and this section never comes back. Answer them as reassurances and it always does. Note too that the consent documents carry their own retention obligation, running three years past the project's completion date — a fact your storage answer should not contradict.

Which agreements are checked outside the file?

Three, and they are the ones you cannot fix on the evening before a deadline. A training certificate has to belong to a person actually named on the application, and cover work of the kind the plan describes. An external board's approval has to attach to the application it approved, which is why Herzing asks for both. And a chair's letter has to name the same project the form does, in the same words, signed by somebody with the standing to grant what it grants.

All three depend on other people's calendars. Request them before you polish anything you control. The reason that ordering matters is argued at length in why completeness is judged before quality, while the nine moves a Herzing file makes puts the whole sequence in order. Shorter answers, lifted whole, live on the FAQ page.

What to do next

Run one pass of your own before anybody else does. Take the middle column of the table above, find each fact in your file, then hunt every other place it should appear and read the two side by side. Where they differ by so much as a number, fix the document, not the memory of it.

If that pass turns up more mismatches than you can chase, send the folder over. Our free application review reads it the way a reviewer would — across the documents rather than down them — and reports in writing what disagrees, what is missing, and which pocket to open first. There is no charge for it and no obligation after it. How it works lays out the four tabs that follow for anyone who wants the file carried the rest of the way.

Sources

• Herzing University IRB, Application for Research Protocol Approval (PDF, October 2025) — checklist, P.I.C.O.T. question, methodology, participant section, data security, retention and signatures. IRB_Research_Protocol_Application_1025.pdf
• Herzing University IRB, Application for Improvement or Innovation Project Approval (PDF, November 2025) — checklist, aim statement, improvement framework, data security and signatures. Healthcare_Improvement_InnovationProject_Application.pdf
• Herzing University IRB, Getting Started: Read this First (PDF, October 2025) — which application each programme files. IRB_Getting_Started_1025.pdf
• Herzing University, Institutional Review Board (IRB). herzing.edu/institutional-review-board-irb
• 45 CFR 46.116 — what consent must say, how it must open, and what it may never contain. ecfr.gov — title 45, part 46
• CITI Program, the Human Subjects Research course and its biomedical and social-behavioural-educational tracks. about.citiprogram.org — HSR course

Field names, attachment lists and signature requirements move between editions of the forms; check yours against Herzing's current IRB handbook or portal. What has to agree with what does not move.