***** To join INSNA, visit http://www.insna.org *****
I have two specific how-to questions about doing LR-QAP in UCINET, and one more substantive q:
1. In the menu for LR-QAP, on the right side there’s a box that says Data are: Symmetric (undirected), Non-Symmetric (directed). I assume that it is possible to have an undirected dependent variable matrix, and have a mix of directed and undirected independent variable matrices. But to which matrix/matrices does that box asking about directed or undirected data refer?
2. One of my independent variable matrices is geographic distance between households, but I missed one household’s gps coords. It’s a 105 by 105 household matrix. I know LR-QAP is pretty forgiving, but what does UCINET need in terms of entering a value to indicate missing data for the LR-QAP? Or is it still best to impute some kind of distance, as I see advised for other procedures?
3. In that matrix of geographic distances, I can calculate simple Euclidean distance, or I could use something like the log of the Euclidean distance, as I’ve seen for an ERGM. Thoughts on what is better? Using a geographic relationship like this is new for me.
Stacey A. Giroux
Assistant Research Scientist, Center for Survey Research and Ostrom Workshop
Adjunct, Department of Anthropology
Indiana University Bloomington
1900 E 10th St, Eigenmann 210-S, Bloomington, IN 47406