Why Offer Multiple Participation Options?

(Or, why not require Informed Consent for all experiment?)

We sometimes get this question from researchers, teachers, administrators, and students, and we understand that it’s coming from a desire to ensure that all experiments are conducted ethically. We share that desire.

But, first and foremost, this would not actually have the desired effect of enforcing teachers to use consent when conducting experiments. For instance, an instructor who wanted to violate institutional policies could run an experiment without consent in Canvas (and further, could publicly disclose students' educational records), without Terracotta. So, even if we created this "consent requirement" in Terracotta, it wouldn't prevent someone from doing research unethically. Terracotta is only usable by those with a "Teacher" role in Canvas, and an institution already places a great deal of trust in its instructional faculty to protect student data and to treat its students fairly and ethically. We believe this trust should extend to faculty compliance with research policies when using Terracotta. 

You may still be wondering, "Why would Terracotta even allow research to take place without consent in the first place?" Here it is important to note that consent is not legally required for experimental research on routine educational practice, which is "exempt" according to the 2018 Common Rule (45 CFR §46.104(d)(1)), and that a student's FERPA-protected education record may be disclosed without consent for research that aims to improve instruction (34 CFR §99.31(a)(6)(i)(C)). Thus, for most research in Terracotta, there is no legal requirement to administer consent, and for minimal-risk studies (such as surveys), some Institutional Review Boards (IRBs) have determined that there is more risk in administering consent than in allowing the research to proceed without consent.

Further, research has shown that students of color are significantly less likely to provide consent. Given that consent is not legally required by HHS or FERPA, a minimal-risk research study may have good justification to forego consent to avoid selection bias, particularly when the research goal is to measure and improve equity. 

Even if the concern isn't selection bias, we've learned that there are use cases where a teacher will want to run a Terracotta experiment without consent. For example, currently a teacher is running a study in Terracotta as part of a class activity. This is a college course specifically on learning strategies, and as part of the class curriculum (as outlined in the syllabus), students are being randomly assigned to different study supports for different portions of the semester, and then together the class will analyze their de-identified data. In this situation it would be problematic to require consent, because in effect, the institution would be requiring that students should be allowed to opt-out of a class activity that is described in the syllabus. In this specific case, the teacher is running the class activity in a Terracotta experiment without consent (for educational purposes). Separately, the teacher has a different experiment in Terracotta, in the same course, which invites students to consent to use their data in a future publication (for research purposes). Here, students do not opt in to the learning activity, but they do opt in for research utilization of their data. 

There are other cases where consent is inappropriate, such as research with minors, or situations where a school administrator provides blanket approval for a research study to be conducted. 

With Terracotta, we believe strongly in the virtue of research consent, and more broadly in respect for persons (as outlined in The Belmont Report). We've made it very easy to administer consent in education settings ethically. Indeed, in some cases, teachers are using Terracotta only as a mechanism to easily and privately administer consent (for research interactions conducted outside Terracotta). When a user attempts to create an experiment without consent, we show a red warning page suggesting that they might want to reconsider their decision. However, given the above, we do not feel it is appropriate to enforce consent in Terracotta.