Yes, those techniques are good

These techniques can have an important role in a better understanding of sensitive behaviors. There are however large challenges in their application and UCT and RRT both suffer from challenges in implementation. But there are other techniques that are worth considering as well for example the ballot box method. See a recent paper here tat has some reflections on these challenges: https://www.cambridge.org/core/journals/american-political-science-review/article/when-to-worry-about-sensitivity-bias-a-social-reference-theory-and-evidence-from-30-years-of-list-experiments/3B922CC54881BA694E4D4B07FD286060

Thank you very much for your advice Diogo Verissimo Andrew Paul McKenzie Pegman ! I think the main challenge for studying consumers of illegal, luxury wildlife products is how to approach them for interviewing. We have recently conducted choice experiments on 1,000 high-income consumers of rhino horn and tiger bone glue in Vietnam and realise that no interview could have been conducted without a good relationship with the respondents. These consumers are in general not motivated by any indirect benefits (e.g., tiger conservation) or small gifts and do not want to be interviewed by younger investigators. Maybe the situation could be different in other contexts and UCT, RRT, or the ballot box can be applied.

You could try to sit in on interviews by border staff, where people are compelled to answer questions?

I agree that reaching those consumers is hard although if you have managed to get to 1000 people within that group it seems you have cracked the code! This is similar to other specialized audiences like Doctors. It does depend on the relationship you have. UCT and RRT are generally pretty hungry for data so you do need lots of data. Which is one reason I think it is worth exploring other techniques

All my interviews relied on personal relationships of the data collection team and it was extremely difficult to find suitable enumerators. Thus I am not confident that my methods could be applied to other contexts. This kind of survey is similar to interviewing business owners about their tax invasion Diogo Verissimo

Hoai Nam Dang Vu This is a great research question and one that I would be very interested to see empirically explored! I can say that we used UCT in our study one tiger bone glue in Vietnam... and it didn't perform terribly but had such enormous errors that it was challenging to draw concrete insights (Article An updated analysis of the consumption of tiger products in ...

I would be very happy to discuss this more, too!

Elizabeth Oneita Davis Thanks for sharing your experience! I think choosing a specific method depends very much on the targeted behavior and the sampling approach. If we aim at the general population, UCT or RRT can work and provide some insights but this runs the risk of not being able to focus on the actual consumers of these products. For example, in Vietnam the use of tiger bone glue is more popular than the use of rhino horn. And consumers in general are not hesitant to talk about their behaviors related to tiger bone glue. However surveying the general population is unlikely to provide much useful insight as consumers of these products have unique characteristics. In most of the surveys on consumers of tiger products, bear bile, or rhino horn, actual consumers only account for a small proportion. In our surveys on rhino horn consumers, the mean age of consumers is about 50 and the average monthly individual income more than US$2,000. The main challenge in studying actual consumers is how to approach them for interviews. If we can do that, it doesn't matter which methods we employ. We just finished a Choice experiment survey on actual consumers of tiger bone glue and those with an explicit intention to purchase this product. However, all the interviews were conducted thanks to personal relationships of our enumerators and snowballing. It also happened in our surveys on rhino horn consumers. It means that such techniques UCT or RRT are unlikely to reduce response biases in studying sensitive behaviors. If a consumer agrees to participate in an interview (i.e., the enumerator is trusted), we don't need UCT or RRT to elicit information about their consumption of these products.

Hoai Nam Dang Vu agreed! You don't need to bother with SQT trickery, unless you want those raw prevalence estimates! (ps- really looking forward to seeing your choice experiment study on tiger bone glue!).

For us, we think that clustring methods Can be applied at these studies. We used for exemple PCA analysis in order to grouping consumers.

Article Asking sensitive questions using the Unmatched Count Techniq...

Thank you Timothy Kuiper for this link

One of the ways i managed to get around the issue of a small sample size, when i used RRT, was to bootstrap my data. However, i did already have a sample size of ~450 participants...

Article Exploring the Africa-Asia Trade Nexus for Endangered Wildlif...

Another study using the RRT, but I would question all surveys using these fancy techniques (RRT, UCT, Bean technique, False Consensus Bias, etc.) on traders, retailers, and consumers of illegal and luxury wildlife products (e.g., rhino horn, tiger bone, pangolin). The first question is always "why respondents agree to participate in your surveys?" In highly patriarchal and hierarchical Asian society like China and Vietnam, there is no way to engage with high-income respondents except being a friend of them or introduced by their trusted persons. These respondents are in no way motivated by any abstract benefits (e.g., conservation) or small gifts and generally do not want to spend their precious time talking to young enumerators... But if they trust you, the mentioned methods do not make sense.