Keyboard Shortcuts
Likes
Search
some questions about QAP
|
Hi all, I have two questions relating to QAP in UCINET. 1. MR-QAP and LR-QAP application in binary network After reading the chapter8 of ASN, I am wondering that the relationship between MR-QAP and LR-QAP is the same as between their analogues in logit and ordinary regression; and MR-QAP is more suitable for valued dependent variable and LR-QAP is more suitable for binary dependent variable? However, I find that some binary cases like the high-tec example is analysed by MR-QAP. Or may MR-QAP works in both cases? ?? 2.coefficient comparison I have two different undirected collaborative networks: N1(35*35) and N2(41*41). And I want to run QAP to (1) examine factors predicting collaboration (homophily, sender effect, and transitivity; the identical model specification, but the outcome objects are different); (2) explore the relative contribution of above effects within network; (3) and compare the strength of certain effect across QAPs (e.g. is the homophily effect in N1 stronger than N2). Could I ues MR-QAP and compare the standardized coefficient of N1 model and N2 model and run T-test? Any information would be appreciated. Thank you! Best, Longxia Huo Ph.D. Student |
¿ªÔÆÌåÓýI¡¯m not a statistician, but I will give you my opinion on your first question. Hopefully, someone more qualified will be willing to comment as well. The relationship between mrqap and lrqap is just as you say, and you may prefer to use logistic regression to predict a binary variable. But using ordinary regression to predict binary variables is not unusual (google ¡®linear probability model¡¯). It does introduce heteroskedasticity, but since qap doesn¡¯t use standard errors in any direct sense, I don¡¯t see a problem. ? As for Q2, I don¡¯t see how one could do that. Again, it would be great if someone with greater statistical knowledge would comment. ? steve ? From: [email protected] <[email protected]>
On Behalf Of ??? via groups.io
Sent: Monday, June 20, 2022 22:21 To: [email protected] Subject: [ucinet] some questions about QAP ? CAUTION: External Sender ? Hi all, I have two questions relating to QAP in UCINET. 1. MR-QAP and LR-QAP application in binary network After reading the chapter8 of ASN, I am wondering that the relationship between MR-QAP and LR-QAP is the same as between their analogues in logit and ordinary regression; and MR-QAP is more suitable for valued dependent variable and LR-QAP is more suitable for binary dependent variable? However, I find that some binary cases like the high-tec example is analysed by MR-QAP. Or may MR-QAP works in both cases??? 2.coefficient comparison I have two different undirected collaborative networks: N1(35*35) and N2(41*41). And I want to run QAP to (1) examine factors predicting collaboration (homophily, sender effect, and transitivity; the identical model specification, but the outcome objects are different); (2) explore the relative contribution of above effects within network; (3) and compare the strength of certain effect across QAPs (e.g. is the homophily effect in N1 stronger than N2). Could I ues MR-QAP and compare the standardized coefficient of N1 model and N2 model and run T-test? Any information would be appreciated. Thank you! Best, Longxia Huo Ph.D. Student |