There are a number of ways to do it. A good one is qap correlation. You have a valued network:
A
B
B
B
A
B
B
B
B
B
B
A
10
0
0
1
5
3
7
3
0
1
B
10
0
0
9
0
11
8
10
0
0
B
0
0
0
0
0
0
0
0
0
0
B
0
0
0
0
0
1
1
0
0
0
A
1
9
0
0
0
6
4
5
1
1
B
5
0
0
0
0
3
4
1
0
0
B
3
11
0
1
6
3
9
14
0
0
B
7
8
0
1
4
4
9
8
0
0
B
3
10
0
0
5
1
14
8
0
1
B
0
0
0
0
1
0
0
0
0
0
B
1
0
0
0
1
0
0
0
1
0
And a categorical attribute:
Type
A
1
B
2
B
2
B
2
A
1
B
2
B
2
B
2
B
2
B
2
B
2
(only 2 As, which is not much to work with).
You can convert this to a ¡°is the same type as¡± matrix using Data|Affiliations):
A
B
B
B
A
B
B
B
B
B
B
A
1
0
0
0
1
0
0
0
0
0
0
B
0
1
1
1
0
1
1
1
1
1
1
B
0
1
1
1
0
1
1
1
1
1
1
B
0
1
1
1
0
1
1
1
1
1
1
A
1
0
0
0
1
0
0
0
0
0
0
B
0
1
1
1
0
1
1
1
1
1
1
B
0
1
1
1
0
1
1
1
1
1
1
B
0
1
1
1
0
1
1
1
1
1
1
B
0
1
1
1
0
1
1
1
1
1
1
B
0
1
1
1
0
1
1
1
1
1
1
B
0
1
1
1
0
1
1
1
1
1
1
And then run Qap correlation:
This yields something like this:
QAP results for paulnet * paulatt-sameType (5000 permutations)
1 2 3 4 5 6 7 8 9
Obs Value Significa Average Std Dev Minimum Maximum Prop >= O Prop <= O N Obs
--------- --------- --------- --------- --------- --------- --------- --------- ---------
Pearson Correlation -0.1443 0.2935 -0.0003 0.2176 -0.3377 0.4357 0.7259 0.2935 5000.0000
The results are non-significant.
Steve.
From: ucinet@... <ucinet@...>
Sent: Sunday, November 18, 2018 17:45
To: ucinet@...
Cc: Dr. Paul Beckman <pbeckman@...>
Subject: [UCINET] Question on the statistical analysis of SNA data
Hi all ¨C I have a question about statistically analyzing SNA results. The situation is this: there is a group of 11 investors who each fall into one two types, ¡°A¡± or ¡°B¡± and who can invest jointly with any other investor (the "invests_with" event is non-directional). I¡¯d like to know if either group tends to invest more with their own type or with the other type, that is: "A:A" joint investments vs. "A:B" joint investments vs. "B:B" joint investments.
I¡¯m OK at applying SNA but I¡¯m not a statistician, and this situation is more complicated because there are different numbers in each group type (there are 2 ¡°A¡± investors and 9 ¡°B¡± investors). Also, each individual investor has jointly invested a different number of times; the table below shows how many times each investor has invested with each other investor. Given this brief background, is it possible to determine if the ¡°A¡± and ¡°B¡± investors tend to invest more their own group or with the other group? If it¡¯s possible, how do I do it?
Thanks for any help you can give!
A
B
B
B
A
B
B
B
B
B
B
A
-
10
0
0
1
5
3
7
3
0
1
B
-
0
0
9
0
11
8
10
0
0
B
-
0
0
0
0
0
0
0
0
B
-
0
0
1
1
0
0
0
A
-
0
6
4
5
1
1
B
-
3
4
1
0
0
B
-
9
14
0
0
B
-
8
0
0
B
-
0
1
B
-
0
B
-
