Experimental Design
We are partnering with an Ayahuasca center in Brazil, where ayahuasca has been legal for ritualistic purposes since 2010. Our partner will recruit adults who have never taken ayahuasca before, screening people based on disease history, as is their standard practice. For safety, those with severe depression, or a family history of bipolar or schizophrenic disorders are not allowed to participate. Recruitment will happen on a rolling basis, as per the center’s status quo. Our goal is to recruit about ≈500 participants split across ≈15 sessions, averaging 33 participants per session. However, the final number of participants will hinge on how many first-timers the center can successfully recruit.
The randomization will be performed by our partner, and our only role is to administer the surveys. Participants will be randomly assigned to one of two groups upon arrival at the center: (i) take a regular dose of ayahuasca, offered in their usual ayahuasca sessions, with 3:1 concentration of ayahuasca:water, or (ii) take a less concentrated dose of ayahuasca, with 1:1 of concentration ayahuasca:water. To guarantee there is no selection into treatment, the study will be single-blinded (i.e. participants will not be informed about their treatment assignment until they fill out the last survey).
The sessions are preceded by a ≈15 minutes lecture given by the center’s owners. The lectures could be about either (i) happiness and pleasure or (ii) taming the mind. They will alternate which lectures are taught in each of the study sessions, allowing us to estimate suggestive effects of the lecture type on the nature of trips, and to explore treatment effect heterogeneity by lecture type.
We will also survey participants who booked a session but canceled or did not show up to the ceremony. For them, we will be interested in outcomes we measure at baseline, such as depression, (predicted) happiness, and work satisfaction. This group will be our “pure control group,” though given selection into the group, we will use a difference-in-difference approach for suggestive evidence of effects of the placebo relative to the pure control.
Randomization is stratified by sex, and it takes place when participants’ list for each study session has been closed (i.e. about one day before a study session). The center will perform the randomization by generating random numbers in a spreadsheet, allowing us to validate whether the randomization was implemented correctly.