Item-count technique/List experiment

[Nicolas Orgeira]

Hi all,

I'm analyzing list experiment data using the kict commands to perform multivariate analysis. The first thing I'm checking is whether there are any design effects, that is to test whether respondents' answers to nonkey (not sensitive) items differ depending on their treatment status, with the null hypothesis stating that all estimated joint probabilities are positive (alternative is at least one joint probability is negative).

However, when running the command below, I get an error message, "__000001 has zero variance", and I don't why. I'm sharing a sample of the data (20 out of 2'070 observations) which displays the same error message when running the kict deff. See further below that the data does not seem to exhibit any floor or ceiling effects. Any help would be more than welcome!

Thanks

kict deff list_experiment_count, nnonkey(4) condition(list_experiment_group)

clear
input float(list_experiment_count list_experiment_group)
2 0
4 1
3 0
3 0
3 0
3 0
3 0
3 0
2 1
2 0
3 0
3 0
3 1
4 1
3 0
3 1
2 0
3 1
3 0
3 1
end

count / groups:	Control	Treatment (with sensitive item)
1	32	14
2	373	204
3	616	307
4	26	495
5	0	3