S-estimate definition

[markp]

Hi

I have 2 questions about the S-estimate

1) I just wanted to get confirmation/correction of the definition of the S-estimate reported by Ngene: "the minimum sample size needed to estimate statistically significant parameters values at the 95% level, assuming the prior values are correct."

2) Is the S-estimate the maximum of the parameter-specific minimum sample sizes?

Thank you

[Michiel Bliemer]

1) Yes that definition is correct. Note that the level of significance can be changed, e.g. ;eff = (mnl,s(3)) would require a t-ratio of 3 instead of the default 1.96 for 95% significance.

2) Yes, the S-error is the maximum of all S-estimates of the individual parameters.

Note that the sample size is assuming that all choice tasks are faced by each respondent. If you use blocking you need to multiply with the number of blocks.

More information in this journal article:
Rose, J.M. and M.C.J. Bliemer (2013) Sample size requirements for stated choice experiments. Transportation, Vol. 40, No. 5, pp. 1021-1041.

Michiel