Completed a Choice Modeling Survey

This week, I was a respondent to a choice modeling survey for the first time. The purpose of the survey was to find out about reliability of supply of water, electricity, gas etc. So this was something that I as a consumer have a lot of knowledge about and a lot of interest in and it is a clear private market. So all the criticisms about environmental valuation don't apply here. Choice modeling is often used to obtain environmental valuations and I think it is superior to other approaches in that application.

The online survey gave us eight different scenarios of possible packages with varying prices and utility reliability that we could pick from. Each scenario had three different packages - our current service plus two alternatives. Each alternative had around 10 different characteristics. I've always wondered whether survey respondents could handle that and thought that you just need to give each respondent one choice set and survey more people. I was just overwhelmed with information.



I chose the current package in every case. Partly because of the information overload but also because saving $50-$400 a year on my utility bill is just not worth it to me to suffer from multiple power cuts each year and the like, which is what the trade-offs were about. Even when I lived in East Jerusalem in the 1980s as an undergrad student we didn't have that many power cuts (Troy, NY was almost as bad :)). So I'm thinking the results of this survey will be pretty insensitive to price except for the more extreme scenarios at low incomes. $400 is a big deal to someone on $30,000 a year (Australian full time minimum wage), but $50 isn't for someone on $60,000 (average wages).

Being a social science researcher myself I always feel sympathetic to people carrying out surveys and often do them. But I get really annoyed when a surveyer has me on the phone for more than a quarter of an hour asking endless questions or when an online survey turns out to have tens of pages of questions. Why do people design surveys like this that are going to either have less patient people drop out of or refuse to respond or just give random answers to either from impatience or choice fatigue?