Hi.
Can anyone recommend example of "list experiment" (or vignette) or con-joint survey with data, survey sheet, and STATA (hopefully) code? I just want to figure out and see how the actual analysis process looks like. Most survey experimental papers do not come with actual data, survey sheet, and analysis process code, unfortunately.
Also, I would appreciate any information on helpful materials for actual practice of designing and analyzing survey experiment.
The article presenting the kict family of Stata commands to analyse list experiments is extremely detailed in this regard, including commands and outputs. The data is available from public sources, I believe: Article Statistical analysis of the item-count technique using Stata
Jilke & Tummers have an interesting recent paper in JPART with a conjoint experiment. You could ask one of the authors if they can share? I think some journals are now requiring data and code
https://academic.oup.com/jpart/article-abstract/28/2/226/4827854
Hi,
So, in conjoints, you first need to turn the data. Data are usually provided in wide format, and you need to transform them into long format. For iteration of the experiment you will have (usually) 2 profiles, so individual i. will have a number of observations equal the number of profiles.
So if you have 2 individuals and 4 profiles with 1 dimension, your original dataset will have 2 lines and 4 columns. To analyse the data in a conjoint section, you should think about transforming them in a sort of "panel": your transformed data will have 8 lines (4 per individual, each corresponding to a different profile) and only one column (because you had a single dimension, in this example).
- Badges
- Science topic
- Similar topics
- Papers