Does anyone know how to implement longitudinal network analysis using
the Quadratic Assignment Procedure in STATA, SAS or MATLAB? While it is
a second best alternative to now famous SIENA, I want to test it out and
compare the results.
UCINET's has a QAP routine, but it is used for a cross-sectional
research design, right?
I know that Prof. Bill Simpson has implemented the QAP routine in STATA
(http://www.hbs.edu/research/stata/). He told me that his algorithm is
effective in taking care of the dependence among the dyads as it
scrambles the rows and columns, however, it doesn't randomize the time
periods. Thus, if the dependent variable and one or more predictors are
co-varying over time, the QAP results won't correspond to the null
Thank you as always.