I've been handed a dataset where respondents were given a series of 12 discrete choice questions with four options each. "id" is the ID of the respondent, "choiceid" is the question number, and the remaining fields are the characteristics of each option (school offering the program, time to completion in 6-month increments, format, and tuition cost in 10k increments).

I'm trying to run it through the mlogit package in R to do a conjoint analysis and see how much they value each options. So far, I've come up with this:

test2 <- mlogit.data(test, shape = "wide", choice = "choice", varying = 4:19,
sep = "", id.var = "id", alt.var="choiceid")
ml.test <- mlogit(choice ~ cost + school + time + format | 0, test2)
summary(ml.test)
coef(ml.test)[-1]/coef(ml.test)[1]
When I run this on the full dataset there's a thousand respondents, yet none of the variables come up even close to significant so I'm worried something is awry. Thanks!