Getting IRB approval for a survey study is only the first step. The instrument you deploy to participants must match the instrument your ethics board approved : exactly. Every question, every answer option, every consent statement. Any deviation may require a protocol amendment, delay your data collection, or invalidate your results.
This guide covers how to design a survey that meets IRB requirements from the start, what the most common compliance issues are, how platform choice affects your obligations, and how to get your approved instrument into a platform without altering a word.
The Three Belmont Principles Applied to Surveys
IRB requirements in the United States are grounded in the Belmont Report (1979), which established three core principles for research involving human subjects. Understanding how each principle applies to survey design helps you build compliance into the instrument rather than retrofitting it after approval.
Respect for Persons
Participants must give informed, voluntary consent. They must be able to withdraw at any time without penalty. Survey design must make this autonomy visible : not buried in small print.
Beneficence
Research should maximise benefits and minimise harm. Question wording, sensitive topic handling, and survey length all affect participant wellbeing and must be considered in the instrument design.
Justice
Benefits and burdens of research should be distributed fairly. Subject selection criteria, language accessibility, and demographic scope all fall under this principle.
Designing the Informed Consent Section
Informed consent in a survey context is not a signature on a form : it is a section of the survey instrument itself. IRBs typically require the consent section to include specific elements, and the exact wording is usually approved as part of the protocol. Changing it after approval requires an amendment.
Required elements in most IRB-approved consent sections
- Study title and principal investigator : name, title, institutional affiliation, and contact information
- Purpose of the study : plain language description of what the research aims to learn
- What participation involves : time required, types of questions, how data will be used
- Voluntary participation statement : explicit statement that participation is voluntary and refusal has no consequences
- Right to withdraw : participants may stop at any time without penalty
- Confidentiality or anonymity statement : what data is collected and how it will be protected
- Risks and benefits : honest description of any potential discomfort and what participants or society may gain
- Contact for questions : researcher contact and IRB contact for concerns about rights
Confidentiality vs anonymity: a critical distinction
Most online surveys can only offer confidentiality, not anonymity. Anonymity means no one : including the researcher : can link a response to a specific participant. Confidentiality means the researcher knows who responded but protects that information.
Online platforms typically record IP addresses, completion timestamps, and sometimes device identifiers. Unless you have explicitly disabled all identifiable metadata collection, your survey is confidential : not anonymous. Your consent language must reflect this accurately, or your IRB may flag it.
Question Design Requirements for IRB Compliance
IRB reviewers look at question wording carefully. Several common design patterns raise compliance flags that can delay approval or require revision.
Leading questions
Questions that imply a preferred answer bias respondents and undermine research validity. IRBs flag these because they affect data quality and may cause participants to feel pressured toward particular responses. The AI Survey Advisor flags leading questions automatically, including capitalisation that directs respondents (e.g., "ONE area") and phrasing that implies the correct answer.
Sensitive question handling
Questions about mental health, trauma, substance use, sexuality, immigration status, or other sensitive topics require specific handling in IRB-approved protocols. The survey instrument must include a trigger warning before sensitive sections, optional skip logic that allows participants to bypass questions they are not comfortable answering, and signposting to support resources where relevant.
Scale design consistency
Inconsistent scale labels across sections : for example, using 1-5 in one section and 1-7 in another, or reversing the direction of a scale mid-survey : confuse respondents and affect data quality. IRBs increasingly flag scale inconsistency as a methodological concern, particularly for validated instruments where scale fidelity is part of the approval.
Age and demographic categories
Demographic questions, particularly age brackets, must be granular enough to support the analysis described in your protocol. An age range of "56+" when your analysis requires age-specific subgroup comparisons will be flagged. The approved demographic categories in your protocol must match exactly what appears in the deployed instrument.
Platform Choice and IRB Obligations
The platform you use to deploy your survey affects your IRB obligations, particularly around data storage, security, and participant privacy. This is an area many researchers underestimate.
| Platform | IRB Considerations | Best for |
|---|---|---|
| REDCap | Hosted on institutional servers, built-in audit trails, HIPAA-compliant infrastructure, data access controls | Clinical research, sensitive data, longitudinal studies, healthcare surveys |
| Qualtrics | Requires institutional licensing, data residency options vary, advanced anonymisation features available | Experience management, complex logic, large-scale academic surveys |
| Google Forms | Data stored on Google servers, GDPR compliance requires institutional Google Workspace agreement, limited anonymisation controls | Low-risk pilot studies, classroom research, non-sensitive data collection |
| Microsoft Forms | Tied to institutional Microsoft 365 agreements, data residency within institutional tenant | University-administered research using institutional Microsoft accounts |
GDPR and International Research Considerations
Researchers collecting data from participants in the European Union must comply with GDPR in addition to their institutional IRB requirements. GDPR introduces obligations that go beyond traditional IRB frameworks.
- Data minimisation : only collect the data you need for your stated research purpose. Every question must be justifiable in terms of your research objectives.
- Right to erasure : participants have the right to request deletion of their data. Your platform and data handling procedures must support this.
- Data residency : data from EU residents must not be transferred outside the EU without appropriate safeguards. Check your platform's data residency options.
- Explicit consent for sensitive categories : health data, genetic data, and other special categories require explicit consent under GDPR, which is more stringent than standard IRB consent requirements.
- Privacy notice : a separate privacy notice detailing how data will be processed must be provided to participants, in addition to the consent section within the survey.
Deploying Your IRB-Approved Instrument Without Altering It
Once your instrument has been approved, getting it into a platform accurately is critical. Any change to question wording, answer options, or survey structure between the approved document and the deployed survey may constitute an unapproved modification, requiring a protocol amendment before data collection can begin.
This is where manual platform programming introduces risk. A programmer who paraphrases a question, reorders answer options, or slightly rewrites consent language : even with good intentions : creates a compliance problem. The deployed instrument must match the approved document character for character.
Pirai AI converts the approved document : Word, PDF, or Excel : directly into the platform file without altering, paraphrasing, or reordering any content. The same text that was approved by your IRB appears in the deployed survey. For healthcare and clinical research using REDCap, Pirai AI generates a REDCap Data Dictionary CSV that preserves all field labels, choice values, and instrument structure exactly as in the source document.
Why exact text preservation matters for IRB compliance
- Your IRB approved specific wording : any change is technically an unapproved modification
- Validated instruments (PHQ-9, GAD-7, SF-36) have psychometric properties tied to exact question wording
- Protocol amendments take time and delay data collection
- Discrepancies between approved and deployed instruments can invalidate data during audit
The AI Survey Advisor is designed to be used before IRB submission : during the instrument design phase. Once your instrument is approved, Pirai AI converts it to the platform format without touching the approved content.
Multilingual Studies and IRB Requirements
Research involving non-English-speaking participants requires translated instruments, and those translations must typically be approved by the IRB as part of the protocol. A back-translation procedure : translating from English to the target language, then independently back to English : is the standard for demonstrating translation accuracy.
For multilingual survey deployment, Pirai AI extracts each language version from the approved bilingual document and deploys it to the target platform without altering the approved translations. See the multilingual survey guide for the full workflow. Pirai AI supports 16 languages and preserves all non-Latin scripts and Unicode characters during platform conversion.
IRB Survey Design Checklist
Use this checklist before submitting your survey instrument for IRB review:
- Informed consent section includes all required elements: purpose, procedures, voluntary participation, right to withdraw, confidentiality statement, risks, benefits, and contact information
- Consent language accurately describes the data collected : confidentiality if IP or metadata is recorded, not anonymity
- No leading questions, directive capitalisation, or implied preferred responses
- Sensitive topic sections include trigger warnings and optional skip logic
- Scale labels are consistent across all sections and the direction is not reversed mid-survey
- Demographic categories match the subgroup analysis described in your protocol
- Survey platform has been confirmed with your IRB or is on the institutional approved list
- Data residency and storage arrangements comply with GDPR if collecting from EU participants
- Translated instruments have been back-translated and the back-translation submitted with the protocol
- The deployed instrument matches the approved document exactly
Frequently Asked Questions
An IRB compliant survey protects participants through informed consent, voluntary participation, confidentiality or anonymity, minimal risk design, and equitable subject selection. The instrument wording, data handling procedures, and platform security must all meet the standards set by your institution's IRB or ethics committee.
Converting a survey from a Word document to a platform format like Qualtrics QSF or REDCap data dictionary does not change the IRB approval, provided the question wording and structure are preserved exactly as approved. Tools that alter, paraphrase, or reorder questions during conversion may require a protocol amendment.
REDCap is purpose-built for academic and clinical research with built-in audit trails, data access controls, and HIPAA-compliant infrastructure through institutional servers. Qualtrics offers more survey design flexibility and advanced logic but requires institutional licensing and security configuration for sensitive research. The choice depends on your research type, institutional access, and data sensitivity requirements.
Anonymity means no one, including the researcher, can link a response to a specific participant. Confidentiality means the researcher knows who responded but protects that information from disclosure. Most online surveys can only offer confidentiality because IP addresses and metadata are typically recorded by the platform. Your consent language must accurately describe which applies to your study.
Yes. While the core principles : respect for persons, beneficence, and justice : are shared across most frameworks, specific requirements differ. GDPR adds data minimisation, right to erasure, and data residency requirements for European research. UK Research Integrity and other national bodies have their own consent and data handling standards. Cross-cultural research often requires approval from multiple ethics bodies.
Convert Your IRB-Approved Instrument Without Altering a Word
Pirai AI converts Word, PDF, and Excel survey instruments into Qualtrics, REDCap, and 25+ platforms : preserving every approved question exactly as written.
Start Converting Free