[R list package] proportion of liars

[Graeme Blair]

Dear Ana,

The estimates of the proportions of liars from Table 7 in our paper
can be replicated using the summary() command following a list
experiment regression run via ictreg().

You can find the boundary proportions for our race example from the
US, for example, by using the commands:

data(race)

ceiling.results <- ictreg(y ~ age + college + male + south, treat = "treat",
		   	  J = 3, data = affirm, method = "ml", fit.start = "nls",
			  ceiling = TRUE, ceiling.fit = "bayesglm",
			  ceiling.formula = ~ age + college + male + south)

summary(ceiling.results, boundary.proportions = T)

This example will be included in an update to the package coming out
next week in help("summary.ictreg").

Best,

Graeme

On Fri, Mar 16, 2012 at 4:08 AM, Ana Nuno <ana.nuno08 at imperial.ac.uk> wrote:
> Dear Kosuke and Graeme,
>
> I am a PhD student at the Imperial College London, UK, and I have conducted
> a list experiment to quantify illegal use of natural resources in the
> Serengeti, Tanzania.
> I have been following your publications and I would like to analyse my own
> data using your list package. I find your developments in this field
> extremely useful!
>
> At the moment, I am been trying to replicate your tables using your help
> instructions. I have a very simple question: in practice, how do you get the
> estimated population probability of lying presented in the table 7 of Blair
> & Imai (2012)? For example, when I apply ictreg to fit standard design ML
> model with floor and ceiling effects and then  check the summary of results,
> where do the estimated proportions of liars presented in the table 7 come
> from?
>
> Thank you in advance!
>
> --
> Ana Nuno
> Munro Building N1.9
> Silwood Park
> Imperial College London
> Ascot, Berkshire
> SL5 7PY, UK
>
> +44 20 7594 2262
> http://www.iccs.org.uk/ana-nuno
> http://fp7hunt.net/