How have discrete choice experiments changed over time? This is the question Soekhai et al. (2018) try to answer. They conduct a systematic literature review covering 27 years of data. Below I summarize some of their findings graphically.
First, you see a trend of an increasing number of DCEs.
We also see that whereas DCEs were previously concentrated in the UK, now DCEs are done all over the world.
The vast majority of DCEs (82%) have between 4-7 attributes, with 39% having 4-5 attributes.
Whereas only 9% of DCE were done online between 1990 and 2000, now 57% of all DCEs are administered through an online approach. Further, the vast majority of studies (83%) present individuals with two choices to consider for each question.
When we examine the type of people who are surveyed, patients are the most likely.
The mean sample size was 728, but the median was 401. Most studies (89%) used a fractional factorial design. Mixed logit (39%) and multinomial logit (39%) were the most common econometric approaches.
For more details, do read the whole article (it is open access!)
- Soekhai, V., de Bekker-Grob, E.W., Ellis, A.R. et al. Discrete Choice Experiments in Health Economics: Past, Present and Future. PharmacoEconomics (2018). https://doi.org/10.1007/s40273-018-0734-2