st: Probability distribution for a runs test of sequentially administered items

["Barth Riley" <BarthRiley@comcast.net>]

Dear Statalist

I am exploring the possibility of using a Runs test to assess whether a set
of sequentially administered item responses (i.e., as in a computerized
adaptive test) is "in process" (with alternating correct/incorrect responses
towards the end of the test) or out of process (reflecting aberrant
responding, such as guessing or specialized knowledge). To do this, I would
like to apply a probability distribution for the run length (number of
consecutive pass/yes or fail/no responses) one would expect at a given point
during the testing given that the respondent were answering the items as
expected (in process). I was wondering if such a probability distribution
already exists (and if so, what is it?) or if I would have to derive it via
simulation.

Barth


*
*   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/