Restructuring Dyadic Data
David A. Kenny
January 9, 2015
Background
• Dyadic Data Structures
– Individual
• One record for each person
• Own person’s variables
– Dyad
• One record for the dyad
• Both persons’ variables
– Pairwise
• One record for each person
• Both persons’ variables
• View:
http://davidakenny.net/webinars/Dyad/General/DDS/DDS.html
2
The Problem
• You have one data structure and you
want to convert to another.
individual to dyad
individual to pairwise
dyad to pairwise
• Other conversions are trivial and can be
accomplished either by deleting cases
or renaming variables.
3
Dyad ID
• For restructuring individual data, a
unique identification number for each
pair of persons is needed.
• For longitudinal standard design, the
“DyadID” is for each time point for each
dyad.
4
Usual Strategies to Restructure
• Restructuring by entering the
data the “right” way.
• Cut and paste
• Computer programs
–Built in routines to restructure
•SPSS:
davidakenny.net/webinars/powerpoints/Dyad/General/Restructuring.pdf
5
New Strategies to Restructure
– R programs
• Individual to pairwise
• Individual to dyad
• Dyad to pairwise
– Apps
• Individual to pairwise
• Individual to dyad
• Dyad to pairwise
– SPSS macro (no longer maintained)
• Individual to pairwise
• Individual to dyad
6
R Restructuring Programs
• Written in R
• Co-written with Thomas Ledermann of Utah
State University
• Information available at
–http://davidakenny.net/doc/RDDD.pdf
7
Details
• Installing R
–davidakenny.net/doc/InstallR.pdf
• Three programs
–ItoP.R: Individual to pairwise
• davidakenny.net/kkc/c1/ItoP.R
–ItoD.R: Individual to dyad
• http://davidakenny.net/kkc/c1/ItoD.R
–DtoP.R: Dyad to pairwise
• davidakenny.net/kkc/c1/DtoP.R
8
General R Program
• RDDD
– Description: davidakenny.net/doc/RDDD.pdf
– Program: davidakenny.net/progs/RDDD.R
9
Descriptive Information for Dyad Dataset
The dataset c:/ai.sav has been transformed from an
individual to a dyad dataset called c:/dyad.csv. The
distinguishing variable is Gender, and it has two
levels, Wives (-1) and Husbands (1). There are 148
dyads and 296 individuals, 148 Wives and 148 Husbands.
There are no missing data on any of the variables in
the dataset.
There are 7 variables, 1 between-dyad variable, 1
within-dyad variable, and 5 mixed variables. The one
between-dyad variable is Years Married, and the one
within-dyads variable is Gender. The within-dyads
variable, Gender, is a dichotomy and could be used as a
distinguishing variable. The descriptive statistics
for the variables as individuals are contained in Table
1 and the descriptive and inferential statistics as
dyads are contained in Table 2.
Table 1: Descriptive Statistics for Individuals (All Variables)
Variable
Years Married
Gender
Self Positivity
Other Positivity
Satisfaction
Tension
Similar Hobbies
Mean
-0.000
0.000
4.186
4.264
3.605
2.431
0.078
sd
7.707
1.002
0.412
0.498
0.496
0.687
0.646
Minimum
-11.214
-1.000
2.600
2.600
1.167
1.000
-1.000
Maximum
15.036
1.000
5.000
5.000
4.000
4.000
1.000
Intra. r
1.000
-1.000
0.087
0.235
0.618
0.319
0.281
Table 2: Inferential and Descriptive Statistics for Dyads (Mixed Variables)
Variable
Self Positivity
Other Positivity
Satisfaction
Tension
Similar Hobbies
Mean
Wives Husbands
p
4.291
4.082 <.001
4.246
4.281 .490
3.591
3.618 .451
2.520
2.341 .006
0.189
-0.034 <.001
sd
Wives Husbands
0.409
0.390
0.523
0.474
0.530
0.462
0.709
0.655
0.587
0.684
p
.568
.225
.034
.306
.052
r
p
.157 .056
.234 .004
.623 <.001
.340 <.001
.321 <.001
All calculations are based on 148 cases. Degrees of freedom for the test of
mean difference are 147 and for the test of standard deviation difference and
the test of the correlation are 146.
Restructuring Apps
• Uses the Ledermann & Kenny R programs
• Adaptation to apps done with the assistance of
Robert Ackerman
• Web-based, no need to download R or to install
R.
• Answer prompts
• Results
– Text on the screen
– Restructured data that can be downloaded
• Link: http://davidakenny.net/RDDD.htm
14
ItoP Illustration
SPSS Macros
• Steps
– Download the macro.
– Run the macro.
– Open the dataset.
– Create the call.
– Run the call.
• Macros
– pairwise.sps
• http://davidakenny.net/kkc/c1/pairwise.sps
– indtodyad.sps
• http://davidakenny.net/kkc/c1/indtodyad.sps
19
Calls (red is required)
• pairwise.sps
Pairwise dyadid = dyad i1 = 'A'
i2 = 'P' directory = 'c:\'.
• indtodyad.sps
IndToDyad dyadid = dyad
distvar = gender i1 = 'F' i2 = 'M'
directory = 'c:\'.
20
Issues
• 2 records per dyad
• No string variables for most methods
• The “Individual to Dyad” restructuring
programs always produce a new
variable called “partnum” (one
member is given a “1” and the other a
“2”) which can be useful in analyses.
21
Additional Readings
Kenny, D. A., Kashy, D. A., & Cook, W. L.
Dyadic data analysis. New York:
Guilford Press, Chapter 1.
Ledermann, T., & Kenny, D. A. (2014). A
toolbox with programs to restructure
and describe dyadic data. Journal of
Social and Personal Relationships,
online.
View as a webinar (small charge)
Special thanks to
Thomas Ledermann & Rob Ackerman!
22