Data Reading Assignment


The objective is to familiarize yourself with the UCINET 5 and Krackplot software packages. NOTE:  Although extremely simple, Krackplot is a dos program and may be difficult to work with, depending on how much experience you have had with DOS.

Please read through all steps in each Part before beginning.

Part 1. Some made-up data

Step 0. Read the manuals! (Especially the UCINET user's guide) 

Step 1. Import the following dataset into ucinet:

dl n 10
format = nodelist1
labels embedded
data:
bs cj sb jb
cj rn sb jb jc dc
rn dc cj
sb bs cj jb jg rn jl
jb jg bs sb cj jc
bt df rn jc
jc dc cj bt bs
dc rn cj
df jl bt
jl jg df dc
jg jb sb jl

Step 2. Symmetrize the dataset (on the menu, choose Transform | Symmetrize) three ways: using the minimum option, the maximum option, and the average option.

Step 3. Determine whether the network is connected by running the networks | regions | components procedure (using both strong and weak options)

Step 4. Compute the density of (a) the network as it was imported and (b) the network as symmetrized by the maximum method.

Step 5. Compute the geodesic distance matrix for both the symmetrized and unsymmetrized matrix.

Step 6. Export the data (Data | Export | Krackplot) in krackplot format.

Step 7. Start krackplot, load the exported file and draw using the annealing method.

Step 8. Save the drawing as a gif file and exit Krackplot. If you have a postscript printer, you can save the drawing as a an encapsulated postscript file by first choosing options and turning encapsulation on, then choosing postscript. This generates publication-quality pictures.

Step 9. Insert the drawing in a Word file. Cut and paste all the outputs in the previous steps into this Word file as well. Intersperse your comments interpreting the output. Email the document to me.

Part 2. Some Real Data

Step 1. Import the following dataset into ucinet (Data | Import | DL), and call the output dataset "nsf92":

DL N=18
FORMAT = FULLMATRIX DIAGONAL PRESENT
LABELS:
HOLLY
BRAZEY
CAROL
PAM
PAT
JENNIE
PAULINE
ANN
MICHAEL
BILL
LEE
DON
JOHN
HARRY
GERY
STEVE
BERT
RUSS
DATA:
0 7 17 1 3 13 8 6 4 12 15 2 14 5 16 11 9 10
4 0 8 5 10 9 7 6 15 17 1 11 13 16 14 2 3 12
16 10 0 1 2 4 3 11 7 17 13 14 5 15 12 6 8 9
4 7 5 0 6 3 1 2 9 16 13 10 8 11 15 14 12 17
2 6 3 13 0 1 5 4 14 17 9 15 12 16 10 8 7 11
9 8 5 3 1 0 4 2 14 17 12 11 16 15 13 6 7 10
10 6 3 1 2 5 0 4 13 15 11 16 8 17 14 9 7 12
5 8 4 2 7 1 3 0 9 15 14 12 6 10 16 11 13 17
3 15 5 6 7 13 11 10 0 14 17 1 12 2 4 9 8 16
8 9 14 15 11 17 13 16 1 0 10 2 12 3 7 5 6 4
10 3 6 13 4 14 15 17 11 16 0 9 8 7 12 1 2 5
2 13 15 5 6 7 8 12 1 17 10 0 14 3 4 11 9 16
16 5 4 8 12 14 1 9 10 13 17 11 0 7 2 6 15 3
2 16 15 10 11 17 14 8 1 5 6 3 4 0 9 7 12 13
13 6 9 14 11 15 12 16 3 17 8 7 5 10 0 2 4 1
14 5 6 15 9 11 8 13 12 17 2 7 10 16 4 0 1 3
5 4 10 7 15 11 8 9 14 13 2 12 17 16 6 1 0 3
4 8 9 15 13 7 17 16 12 11 5 10 6 14 1 3 2 0

The data consist of rank-orders of interaction: the 1 in row 1, column 4 indicates that person 1 ranked person 4 first, which means person 1 spent the most time with person 4. They spent the least time with person 3 (rank = 17).

Step 2. Dichotomize the dataset (on the menu, choose Transform | Dichotomize) so as to obtain just each person's top 3 choices. In other words, we want a matrix that has a 1 in a given column if the row person ranked the column person in their top 3, otherwise there should be a 0. You can do this in the dichotomize procedure by setting cutoff = 3 and operator = LE.

Step 3. Determine whether the network is connected by running the networks | regions | components procedure (using both strong and weak options)

Step 4. Compute the density of the network.

Step 5. Compute the geodesic distance matrix.

Step 6. Export the data (Data | Export | Krackplot) in krackplot format.

Step 7. Start krackplot, load the exported file and draw using the annealing method.

Step 8. Save the drawing as a gif file and exit Krackplot. If you have a postscript printer, you can save the drawing as a an encapsulated postscript file by first choosing options and turning encapsulation on, then choosing postscript. This generates publication-quality pictures.

Step 9. Insert the drawing in a Word file.

Step 10. Import the following dataset, and call it "nsfsex"

DL NR =18, NC = 1
FORMAT = FULLMATRIX
ROW LABELS:
HOLLY
BRAZEY
CAROL
PAM
PAT
JENNIE
PAULINE
ANN
MICHAEL
BILL
LEE
DON
JOHN
HARRY
GERY
STEVE
BERT
RUSS
COLUMN LABELS:
SEX
DATA:
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2

Step 11. Display nsf92 (matrix from Step 1) using nsfsex (the vector from Step 10) as a blocking vector. (Choose Data | Display. For input dataset, enter nsf92. For "row partition" enter "nsfsex column 1". For column partition, enter "nsfsex column 1".)

Step 12. Cut and paste all the outputs in the previous steps into this Word file as well. Intersperse your comments interpreting the output. Email the document to me.

After this, you should try playing with various menu choices to see what they do.