In a tricot project, ranking data is collected on all participants (either farmers or consumers) with each one of them contributing one ‘partial ranking’. The ClimMob software uses the statistical Plackett-Luce model to combine all these partial rankings into a full picture.
The image below visualizes an example. Here, three growers each rank their individual set of three technology options (for example, crop varieties). Participant 1, for example, reported that var 1 was the best technology option, while var4 was the worst, and var3 took a middle position.
The partial rankings from all participants are then combined, giving the overall ranking. When there are disagreements (for example, some farmers find that var3 > var4, but others report that var4 > var3), the Plackett-Luce model determines the most likely overall ranking.
Further reading
- Plackett, R. L. 1975. The analysis of permutations. Applied Statistics 24: 193–202.
- Luce, R. Duncan. 1959. Individual choice behavior: A theoretical analysis. New York: Wiley.
- Turner, H., van Etten, J., Firth, D., Kosmidis, I., 2019. Modelling rankings in R: the PlackettLuce package.