I think the survey already cover too many aspects and I am dubious about what they will get out of it. At least if they make reasonable choices on the Sleeper hard product, leaving all else aside for the moment, that would in and of itself be a win IMHO.
I think they tried to do too many things at one time with the 16 sets of options. As I said before, there were way too many factors to weigh - frequency, cost, random amenities - and so I ignored a bunch of those factors. Frankly, the presence of a work space vs a panorama car vs a bar car doesn't impact my decision to take most trains - the presence of
a space is more important (though a kid's car could be a net
negative if it's basically a kids-only space). With something like that, you really only want to test a limited number of variables (with multiple values for said variables). They were testing something like seven variables (OTP, frequency, accommodation mix, shared space, WiFi, and price were the "main" ones; restroom/shower facilities seem to have also been tested a bit, while food options were only "paired" with accommodation typed [1]). Testing the eight train accommodations made sense, but "three train options plus four others" was a bit much.
Were I designing that survey, here's what I'd have tested in those slots:
-Accommodation type (see note below)
-Price
-Food service levels (namely, testing whether including "full" meals at a higher price point in coach/lie-flat space would work or making food options for-pay at the Slumbercoach/Roomette/Bedroom level would work).
-Competition options (see note below)
That's four variables, but that's more manageable.
Train frequency, OTP, type of shared space, etc. should have been asked about in separate questions, though I'd assume the presence of
some sort of shared space. One thing I
might slip in is testing the utility of a dedicated space (e.g. exclusive bar-lounge for the luxury spaces) would work.
I would also have adjusted the accommodation mixes (and alternatives) in line with respondent preferences - for a customer who
absolutely will not fly, they wasted a third of the space on the survey with non-starter options. Ditto one who will not take the bus (*
waves*) or who can't/won't drive the distance in question. Using a few earlier questions to weight out what else was going to be offered for some of the options. Dropping back to one airline option/adding first class air as an option, excluding the bus, or excluding the car on some sets and offering four train options would all make sense, as would adjusting the number of times a given option was included (weighing this by the number of travelers in the "expected" party would make sense - testing the Prestige space and the Luxury-for-4 options a combined eight times across the surveys for a solo traveler was probably a
bit much.
Edit: Doing this as a two-stage or three-stage study would probably also make sense, so as to more specifically work out some options.
[1] Coach/lie-flat only got a food trolley/cafe car. Slumbercoach/roomette/bedrooms got "meals in dining car". The "luxury" options got "all inclusive w/room service" (or some more-or-less meaningless variation in phrasing).