Author here. Wanted to common on the brand salience component since a few friends have mentioned it to me. I ran a follow-up experiment where I replaced the vendors with non-existent companies. With n=~500 runs, the base scenario (depersonalized) simply reduced into positional bias. In this case, the models picked the vendor listed first ~85% of the time. This is entirely precedented based on other work I've read.
Happy to answer any questions or run other variations/scenarios people are curious about.
When I originally wrote the solver, it ran completely inside the browser with JavaScript. The main downside of this approach was that users had to download the massive dictionary file (~3mb uncompressed in Trie format) before the application could deliver results.
Since then, I ported the solver logic to Node and set it running on one of my boxes. It's pretty quick at solving, but the scoring algorithm has a rare bug - which I'm still hunting - that causes the score to be off by a point.
Misleading title. Being someone who is "Very arrogant and expresses contempt against airport passenger procedures." is different than someone who complains.
Someone who "became visibly upset and arrogantly pointed his finger in the agent's face when asked why he did not have an airline ticket for a return flight." is different than someone who expresses discontent in a non-threatening manner.
Happy to answer any questions or run other variations/scenarios people are curious about.