Title: Social Network Analysis
1Social Network Analysis
2UCINET--Introduction
- UCINETUCINET is produced by Analytic
Technologies. It offers a very user-friendly,
reasonably priced software system for network
analysis. - Throughout this discussion, well use the example
of the cosponsorship network of the 58
legislators in the lower house of the Arizona
legislature, 2001.
3Starting UCINET
- When you first open UCINET, set the default
directory to a directory of your choice, by
typing in the directory name (into the space at
the bottom edge of the UCINET window). Note that
the original default directory is just the c\
drive. - Note that UCINET produces many types of filesand
deleting any (before you are entirely done with
your analysis) may make it difficult to use some
of the others.
4How to Read Data into UCINET
- There are several ways to read network data into
UCINET. Ill review two basic methods?using
matrices, and using dl language. - UCINET can read in a matrix dataeither saved in
a text file, or saved in excel. - So, in the case of the Arizona cosponsorship data
that we will use as an example, there are 58
legislators and therefore 58 X 58 3,364
dyads.
5How to Read Data into UCINET
- If those data are saved in a text file (with 58
rows, and 58 cosponsorship frequencies listed
on each of those rows), they can be read in as
raw data into UCINET. - Arizona.txt is an example of such a file.
- Download this file. Then, in UCINET, click on
Data and Import and Raw. The text box
should allow you to search for the file youve
downloaded, and to read it into UCINET.
6How to Read Data into UCINET
- One way to double check that it is correct is
to look at the number of rows and columns that
are automatically filled into that dialogue box
it should be 58. - After youve read in the data, an output screen
will show you the data in matrix form. - You can also see the data in matrix form by going
to Data, clicking on Spreadsheets, then clicking
on Matrix. This will open up an emptry
spreadsheet window. Click File and then Open
(or, control-O) to open the file. The file is in
the default directory (as specified in the lower
row of the UCINET program), and it is called
Arizona.h.
7How to Read Data into UCINET
- An alternative is to read data in using a
standard excel file. An example of such an excel
file can be found here. - In UCINET, go to Data, then Import, then Excel
matrix. Be sure to tell UCINET whether the
matrix includes row / column labels. - (The UCINET program will immediately remind you
of the name of the programjust as a safety
double-check. Just click ok.)
8How to Read Data into UCINET
- Often, the easiest way to read data into UCINET
is to use DL language. A document with a set of
examples of DL programming can be found here. - When you use DL language, the default value of
the edge between each possible pair of nodes is
automatically set to 0 (that is, no connection).
- The DL language is then used to specify the value
of all edges. Labels for the rows and columns of
the matrix can also be specified through DL
language.
9How to Read Data into UCINET
- Some researchers find excel files easier to work
withsome find DL language easier. In the
Arizona case, the SAS program that I used
actually creates a matrix file that is relatively
easy to transfer into DL language. - DL language can also be useful if you need to
read in additional matrices. In the
cosponsorship example, I have an additional
matrix file that lists the number of shared
committees on which each pair of legislators
serve.
10Other Types of Data
- Network files can be thought of in terms of
matrices (whether or not they are actually read
in as matrices (versus DL language)). - A second type of file is an attribute file,
which lists attributes for each node. So, I have
a set of attributes for each legislator,
including race, sex, partisanship, seniority, etc.
11Reading in Attribute Files
- Reading in attribute files to UCINET is a bit
unwieldy. Well use two examples from the
Arizona legislative dataset?party (srepub,
coded 1 if Republican, 0 otherwise) and ideology
(measured by swnom01, a w-nominate score). - Download the two excel files srepub and
swnom01.
12Reading in Attribute Files
- Notice that the files are in matrix formthere
are 58 legislators in this network, so each of
these two files has 58 rows and 58 columns. It
is really only the first column that mattersthe
other columns are just filler columns (as UCINET
works only with matrix files.) - Read the two files into UCINET, as you did with
the cosponsorship file previously. In this case,
however, choose node attribute file rather than
network adjacency matrix. - At that point, you should have srepub.h (and
the companion file srepub.d) as well as
swnom01.h (and the companion file swnom01.d).
13Reading in Attribute Files
- In order to create an attribute measure, go to
Data, then click on Attribute. Browse to find
the srepub.h that you imported into UCINET
(after downloading the parallel excel file). - The attribute information for Republican is in
the first column (recall what the excel file
looked likethe first column was 0/1 (0Dem,
1Repub), and the other 57 columns were 99 to
just fill out the matrix.
14Reading in Attribute Files
- So, for Vector is Row or Column?, choose
column. For Which Row/Col, choose 1. - Now, you need to specify how you want the
attribute to be constructed. In this case, it
makes sense to create an attribute that stands
for exact matchestwo legislators who have the
same code will be of the same party (either both
Republican or both Democrat).
15Reading in Attribute Files
- So, under Method, select Exact Matches
- And the output data set defaults to
srepub-Column-1. It may be useful to rename it
to something that makes more intuitive sense,
such as sameparty
16Reading in Attribute Files
- Next, lets look at ideology as an attribute
(swnom01). - As before, enter in the name of the file (browse
to find wherever you saved the ideology file that
you imported into UCINET, after downloading the
relevant excel file from the web.) And, enter in
Column and 1 to specify that the attribute
info can be found in the first column of data in
the file.
17Reading in Attribute Files
- However, it no longer makes sense to choose
exact matches, since legislators will rarely be
exactly the same ideologywhat matters is how
close they are in ideology. So, in Method,
choose absolute differencewhich will represent
similarity between two legislators ideologies.
18Reading in Attribute Files
- And, it may make sense to name the output file
something that makes intuitive sense, such as
diffideol. - Okay, now what can we do with UCINET?
19UCINET
- As noted in previous discussions, UCINET has a
terrific tutorial. You can use the Arizona data
to practice reading in files, and calculating out
commonly used descriptives for networks and nodes
(such as the various centrality measures that
weve discussed.)
20QAP Procedure
- UCINET also allows the possibility of a
regression analysis, using a QAP procedure. - The QAP Procedure will be the focus of our next
discussion.