Dr Gary Allen
AHRECS Senior Consultant

Many national frameworks for human research ethics, such as the National Statement (2007 updated 2018) in Australia have respect as a core principle.  An essential component of respect is obtaining the prior consent of potential participants.

The role of consent in ethical research can be traced back through the Belmont Report and the Declaration of Helsinki, back to the Nuremberg Code and earlier to philosophical, bioethics and clinical texts.

Recent egregious ethical breaches such as the Cambridge Analytica, OKCupid and Emotional Contagion cases highlight that consent problems in research are not just an issue for biomedical research and not just a 20^th Century concern.

Where national and/or institutional policies discuss consent and questionnaire-based research, they will generally indicate that completion and return of a completed survey is a valid expression of consent.

Which is indeed reasonable, especially when it is important to conceal from the researcher who has participated in their research (e.g. where an academic is surveying their own students).

Most frameworks and guidance documents for the ethical conduct of human research will indicate participants should be able to freely withdraw from research without comment of penalty.

Indeed, this is again quite a reasonable position, given that genuine respect for our participants should include acknowledging that they must ordinarily be able to withdraw their consent without comment or penalty.

Solid ground thus far?  Good, because now we’re approaching the conundrum that prompted us writing this post dear readers.  For on-line questionnaires, does this mean stopping the completion of a survey whenever they want and not clicking the “submit” button?   Will this mean that the data already entered is not collected?  What if a participant changes her mind after submitting the data and wants to then withdraw her answers? If that survey is anonymous, consent and the submitted data cannot be withdrawn after submission, because the researchers won’t be able to tell which data was from which individual.

The advent of online questionnaires enabled the resolution of some problems that were largely only an online issue anyway and presented another practical ethics challenge.

A1        Using cookies to reduce the likelihood that an individual completes a survey more than once.

A2        Enabling an individual to save their progress through the survey and complete it over more than one session.

For A1 researchers should ensure the cookie does not enable them to identify respondents and ensure it will not compile any previous or future web activity.  This must be explained in the consent material and assurances provided with regard to these two matters.

For A2 the consent material should explain how password information is saved and the degree to which it can be used to identify respondents.

However, as our scenario indicates, the interesting question comes up if a respondent doesn’t finish the survey.

What happens to the information already entered into the incomplete survey?

What are the wishes of the respondents as to what should happen with these data?

The answers to these questions also depend on why the survey wasn’t completed.

Was it a combination of the participant forgetting, not having time, losing interest or struggling to log in?

Alternatively, was there a reason the individual no longer wished to participate?

Regardless of the reason, what does the individual want to happen with the use of their data?

Simple answers here are not necessarily helpful.  Assuming they forgot, etc may not be accurate and the use of their answers may be absolutely contrary to their wishes.  By the same token, losing potentially useful data merely because participants forgot might be a significant loss – especially if the number of participants is already low.

One approach in the National Statement could be used:

The relevant HREC could be asked to approve a waiver of the consent requirement, so as to enable the use of the partially completed surveys (as per NS Chapter 2.3) if they were accessible and this would need to be made clear in the survey instructions.

While the opt-out approach (also discussed in NS 2.3) might seem a promising strategy, the fact the researchers cannot correspond a set of answers with an individual means that an individual’s decision to opt-out could not be honoured.

Depending on why the individual didn’t complete the survey, the waiver of the consent requirement approach is not especially ideal.  It involves time and other resources that might be in short supply.

Proposals about the use of partially completed surveys should be discussed in the research ethics review application, the recruitment materials and consent materials.

This raises a related point: the reasons someone withdraws from a project might be of interest/importance to the researcher, their research centre, the research ethics review body or other researchers.

We propose the following strategy for a survey in which participants can be linked to their answers:

• The consent material should discuss what will be done with the answers if the survey is only partially completed;
• The revocation of consent process and form should be explained; and
• The resource material for researchers and research ethics reviewers should provide a matrix that explains the treatment of survey responses.

This mechanism uses an optional revocation of consent form.  It is essential that participants are told this form is optional, they can stop participating at any time, without explanation, but then using the form would be very informative to researchers.  In the subscribers’ area is both a suggestion for the questions in the revocation form and the associated text for the consent material.
.

We suggest the revocation form would provide some further clarity about the matters above.  In the case of situations where a revocation form is not provided it is perhaps prudent to conclude those individuals don’t want their partially completed survey to be used.
.

The same approach could be used for other kinds of research designs where data is collected at more than one sitting/point/session.
.
.

For surveys in which participants cannot be linked to their answers, we propose the following strategy:
.

• The consent material should discuss what will be done with the answers if the survey is only partially completed;
• The consent material could include an optional incomplete submission advice that provides researchers with information about why the survey was not completed, e.g. chosen from a dot-point list; and
• The consent material should also clearly state that once submitted, data in completed surveys cannot be withdrawn.

.

This post may be cited as:
Allen, G. (30 March 2020) Can I use your answers anyway? Research Ethics Monthly. Retrieved from: https://ahrecs.com/human-research-ethics/can-i-use-your-answers-anyway