Paper surveys remain a mainstay of academic research because they reach people web forms miss and feel private in a way a browser window never does. That trust comes with responsibility. Any study collecting data from human participants passes through an Institutional Review Board (IRB) or ethics committee, and paper raises its own questions: who touches the completed forms, where the originals live, how identifiers are separated from responses, and how the analog data becomes an analyzable dataset without leaking. Boards protect participants and, by extension, your project.
The friction appears at the last step, when boxes of paper become clean rows for SPSS or R. Manual transcription is where confidentiality promises quietly break and error creeps in. PaperSurvey.io turns completed forms into structured data through an auditable, access-controlled workflow, so the chain of custody your IRB asked about stays intact through analysis. This article is informational, not legal advice, but it maps the questions ethics boards ask and how a scanning workflow answers them.
What IRBs actually evaluate in a paper survey protocol
Review boards work from a consistent checklist, and knowing it helps you write a protocol that gets approved on the first pass.
- Informed consent: participants must understand the purpose, what data is collected, how it is used, and that participation is voluntary. Paper studies attach a consent cover sheet or signature line, and the board will ask how consent records are stored separately from responses.
- Risk and benefit balance: the sensitivity of your questions drives everything else. A workplace climate survey and a substance-use study sit at very different risk levels and demand different safeguards.
- Anonymity or confidentiality: boards want you to state precisely which one you are promising, because the two require different handling.
- Data security: physical storage of originals, who can access them, and how the digital dataset is protected in transit and at rest.
- Data minimization and retention: collecting only what the research question needs, and a defined plan for how long you keep it before secure destruction.
Each has a paper-specific answer, and the scanning step is where several are won or lost.
Anonymity vs confidentiality: get the promise right
These words are not interchangeable, and IRBs are strict about the distinction. Anonymity means no one, including the research team, can link a response to an individual, because no identifying information is collected. Confidentiality means identifiers exist but are protected and access-restricted. Promising anonymity while printing a name field or a sequential ID that maps back to a roster is a common protocol failure.
Paper is well suited to genuine anonymity because no IP address, browser fingerprint, or account login is attached to a submission. That makes it hard to beat for sensitive internal research. Because a self-completed form removes the interviewer, people answer sensitive questions more honestly. A record-validated study found that self-administered respondents underreported an unflattering fact far less often than those questioned by an interviewer (Kreuter, Presser & Tourangeau, 2008), and a paper form filled out privately inherits that honesty advantage. Our guide to anonymous employee surveys covers how to design one that stays anonymous end to end. When you must track responses across waves, use a random code rather than a name, keep the code-to-identity key separately, and document that separation.
Secure handling from the desk to the dataset
The gap most protocols overlook is the handoff between physical form and digital file. A stack of surveys on a shared desk, or a volunteer typing at home, is where confidentiality erodes.
- Controlled digitization: forms are scanned or photographed once, then processed inside an access-controlled system rather than passed around as loose paper.
- Reduced human exposure: because recognition is automatic, far fewer people read individual responses. PaperSurvey reads pen marks at 99.99% checkbox accuracy and applies AI handwriting recognition for open text, so no data-entry team pores over sensitive originals.
- Flexible, secure capture: bring forms in through an office scanner, email-in, Dropbox, drag-and-drop, a shared upload page, or a mobile scanning app, so field teams are not carrying paper long distances.
- Encrypted in transit and at rest: data is protected end to end, the assurance a board looks for.
For teams collecting far from reliable internet, secure capture and later sync matter as much as accuracy. Our notes on running paper surveys in low-connectivity field research go deeper.
Data minimization and where the data lives
Minimization is a principle boards increasingly enforce, and it maps onto both good design and GDPR. Collect only the fields your analysis needs, and be deliberate about anything that could identify a participant.
- Design lean instruments: every question is a data point you become responsible for, so drop fields that do not serve a hypothesis.
- EU data hosting: data is stored in the EU, which simplifies GDPR compliance and reassures boards handling European participants' data.
- GDPR compliant with a DPA on request: a Data Processing Agreement is available, which many university legal and ethics offices require before approving a third-party processor.
- No AI training on your data: your response data is never used to train AI models, closing a question that ethics committees now ask about any tool touching participant data.
Certifications that shorten the review
The infrastructure PaperSurvey runs on carries recognized certifications that do much of a board's security review for you.
- ISO 27001: the underlying hosting infrastructure is certified to this internationally recognized information security management standard.
- SOC 2 Type II: the hosting infrastructure is independently audited against SOC 2 Type II, an attestation that security controls operate effectively over time, not just on paper.
- SAML SSO on Enterprise Plus: single sign-on ties access to your institution's identity provider, so accounts follow your own offboarding and access policies.
- Low-confidence verification: marks the system is unsure about are flagged for human review, so accuracy never depends on a silent guess, and the audit trail records who verified what.
Keeping response rates high without cutting corners
Ethical collection and strong response rates are not in tension. Paper reaches populations that online instruments undercount, an equity consideration for representative research. A meta-analysis found an average paper-versus-electronic correlation of 0.90 across scores, so paper loses nothing in data quality when participants answer on the form instead of a screen (Gwaltney et al., 2008). If your study depends on reaching people beyond the reliably-online, our guide to surveying hard-to-reach populations pairs well with this one. Because paper and web responses land in one dataset with unique per-page identifiers, you can run a hybrid protocol without two separate, harder-to-audit data streams.
Try It Free
A compliant scanning workflow lets you answer your IRB's hardest questions with specifics rather than promises: EU hosting, GDPR compliance, ISO 27001 and SOC 2 Type II certified hosting infrastructure, and no AI training on participant data. Academic teams can request institutional pricing with volume discounts and purchase orders, and every plan exports clean data to SPSS, R, Excel, and CSV. Start your free trial and see how your paper data moves securely from the desk to the dataset.