Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Re: single sample pre/post comparison of proportions


From   "Michael I. Lichter" <mlichter@buffalo.edu>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Re: single sample pre/post comparison of proportions
Date   Thu, 11 Jun 2009 13:10:55 -0400

Svend: The researcher who asked me about this likes the idea of reporting CIs and forgoing the explicit hypothesis test.

José: Unfortunately, it looks like the server ate the first line of your post so I'm not certain what you're suggesting. Is p(known) the pre-intervention proportion of adopters, is it an external estimate, or is it something else? If it's the first, I'm not sure I could justify that as a benchmark.

Thanks to both of you for your suggestions.

Michael

José Maria Pacheco de Souza wrote:
bound p(known), test a hypothesis Ha: p(new)>p(known) vs H0: p(new)=p(known), using the at risk of improving. Or as Svend presented, just estimate the proportion of new, among those at risk. In this case, aftward it will difficult to resist the temptation to compare this result with the p(known).
Cheers,
José Maria

Jose Maria Pacheco de Souza, Professor Titular (aposentado)
Departamento de Epidemiologia/Faculdade de Saude Publica, USP
Av. Dr. Arnaldo, 715
01246-904  -  S. Paulo/SP - Brasil
fones (11)3061-7747; (11)3768-8612;(11)3714-2403
www.fsp.usp.br/~jmpsouza
-----

Michael asked and Joseph responded (not shown) - and Michael then wrote:

The suggestion of a one-sample test restricted to pre-intervention ADOPT=NO crowd makes sense. I think you are also sneakily suggesting that the most obvious null hypothesis -- "H0: p = 0" is not a good choice; there would probably be some adoption even in the absence of the intervention, and the intervention probably cannot be called a success unless the proportion of adopters exceeds a minimum cost/benefit threshold. Instead, I could choose, e.g., "H0: p < .25" (a one-tailed test). That seems reasonable.

===============================================================

I wonder whether a P-value related to a somewhat arbitrary null hypothesis is useful. I think the following is more informative:

Assume that you had 90 participants, 40 of whom already had the good habit, leaving 50 "at risk" for improvement. 20 (40%) of these improved. The 95% CI for this estimate is 26%-55%:

. cii 50 20 , binomial
-- Binomial Exact -- Variable | Obs Mean Std. Err. [95% Conf. Interval] -------------+---------------------------------------------------------------
            |         50          .4     .069282        .2640784 .548206

Hope this helps
Svend


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/

--
Michael I. Lichter, Ph.D. <mlichter@buffalo.edu>
Research Assistant Professor & NRSA Fellow
UB Department of Family Medicine / Primary Care Research Institute
UB Clinical Center, 462 Grider Street, Buffalo, NY 14215
Office: CC 125 / Phone: 716-898-4751 / FAX: 716-898-3536

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index