Challenging Friendship Network Matrix problem

[L T-3]

Hi,

I have a data set of participants (1st column <ID>) who have listed their
friends (2nd column - 1 to X number of friends, to a maximum of 20 friends).

These participants also indicate the interrelationships between their
friends (Lower half-triangular matrixes <1 - relationship present, 0 -
relationship absent>). Note: 20 columns are given to this friendship matrix
since we have a maximum of 20 friends.

Finally, participants also list down friends who are close to them (Last
column: 1 - close, 0 - not close).

1 1            1
1 2 1          0
1 3 01         1
1 4 100        1
1 5 1100       1
1 6 10111      1
1 7 011111     0
2 1            0
2 2 0          1
2 3 11         0
2 4 010        1
3 1            0
3 2 1          1
3 3 11         0
3 4 101        1
3 5 0011       0
3 6 10011      1

Question: How would I obtain the number of relationships each CLOSE friend
has with all other CLOSE friends in each network. So from the data above,
ID1 has 7 friends. 5 of them are close friends. Friend 1 has relationships
with 4 other friends in the network. If we count only close friends, Friend
1 (Close) has relationships with 3 other close friends.

This problem is compounded by needed to count both rows and columns for
friends 2, 3, 4 onwards...

It would be great if someone could help with a syntax on this! Thank you in
advance...

[Spousta Jan]

Hi,

IMHO the key task is to to prepare the data, then the rest is simple.
The solution can be something like:

1) Restructure the file (using the erstructure wizard) so that a case
corresponds with one interrelationships between two resopndent's friends
(e.g. For the first respondent you will have 1 + 2 + 3 + 4 + 5 + 6 = 21
cases (from the lower triangle submatrix); keep or add the info about ID
and status of friends (close/others).

2) Create the mirror file for symetric relationships (the mising upper
triangle submatrix)

3) Merge both files (so that the first respondent will have 21 x 2 = 42
cases in the file).

4) Use simple selects / crosstabelations / frequencies to compute all
you need.

Jan

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
L T
Sent: Thursday, January 04, 2007 1:53 AM
To: [hidden email]
Subject: Challenging Friendship Network Matrix problem

Hi,

I have a data set of participants (1st column <ID>) who have listed
their friends (2nd column - 1 to X number of friends, to a maximum of 20
friends).

These participants also indicate the interrelationships between their
friends (Lower half-triangular matrixes <1 - relationship present, 0 -
relationship absent>). Note: 20 columns are given to this friendship
matrix since we have a maximum of 20 friends.

Finally, participants also list down friends who are close to them (Last
column: 1 - close, 0 - not close).

1 1            1
1 2 1          0
1 3 01         1
1 4 100        1
1 5 1100       1
1 6 10111      1
1 7 011111     0
2 1            0
2 2 0          1
2 3 11         0
2 4 010        1
3 1            0
3 2 1          1
3 3 11         0
3 4 101        1
3 5 0011       0
3 6 10011      1

Question: How would I obtain the number of relationships each CLOSE
friend has with all other CLOSE friends in each network. So from the
data above,
ID1 has 7 friends. 5 of them are close friends. Friend 1 has
relationships with 4 other friends in the network. If we count only
close friends, Friend
1 (Close) has relationships with 3 other close friends.

This problem is compounded by needed to count both rows and columns for
friends 2, 3, 4 onwards...

It would be great if someone could help with a syntax on this! Thank you
in advance...