Académique Documents
Professionnel Documents
Culture Documents
CITATION
Burge, S. K., Katerndahl, D. A., Wood, R. C., Becho, J., Ferrer, R. L., & Talamantes, M. (2015,
November 30). Using Complexity Science to Examine Three Dynamic Patterns of Intimate
Partner Violence. Families, Systems, & Health. Advance online publication. http://
dx.doi.org/10.1037/fsh0000170
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
BURGE ET AL.
the batterer knows hes gone too far and compensates with loving behavior, apologies, and
requests for forgiveness. Over time, the batterer
is unable to sustain attentiveness; tension
builds, and the cycle repeats.
In periodic dynamics, actions occur on a regular, predictable basis. Events and outcomes are
strongly correlated. An example might be a
regular Friday night battle: a man gets paid each
Friday, goes out drinking with his buddies, realizes he just spent half his paycheck, then
comes home drunk and angry and ready for a
fight. The paycheck, the drinking, the overspending, and the anger all lead predictably to
violence. Complexity science explains repeating patterns with a concept called an attractor,
described metaphorically as a magnet exerting a
pull on the system (Ward, 1995). Periodic systems are stable over time and respond predictably to interventions. Small interventions in the
system (such as advice) will create small
changes in patterns of behaviors; large changes
(such as alcohol rehabilitation) will have a big
impact.
Family systems theory views partner violence
as an ongoing interaction pattern resistant to
change. In most couples, the first event of violence is unexpected, and the offender is contrite
(Giles-Sims, 1983). The victim generally interprets it as an isolated incident, and her acceptance of him despite this offense can strengthen
their commitment to each other. However, the
act of staying can increase the likelihood of a
recurrence of abuse. Over time, violence itself
becomes grounds for resentment and conflict,
creating a higher risk for more violence. This
view of violence dynamics differs from the cycle of violence in the emphasis on actions and
reactions, each predicting the next. External influences can affect the interaction pattern by
threatening the stability of the relationship or
reinforcing the violent behaviors.
In chaotic dynamics, the overall pattern of
behavior recurs but the specific path is unpredictable. Chaotic patterns result when actions
and outcomes are separated in time, and when
feedback within the system varies in strength,
direction, and source. Like periodic systems,
chaotic systems have attractors shaping their
behavior; one can recognize repeating interaction patterns. Unlike periodic systems, chaotic
systems are sensitive to small changes which
may alter the path they follow. The metaphor
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
for this concept is the butterfly effect, referring to the idea that a butterfly flapping its wings
in India could affect weather patterns in Chicago. In couple dynamics, one wrong word
from the partner may send him into a fury, or
may go unnoticed. Chaotic systems do not respond predictably to interventions. Giles-Sims
emphasis on feedback, interdependent causal
factors, and stabilization of violence once established is consistent with chaotic dynamics
(Giles-Sims, 1983).
The Duluth model (power and control wheel)
posits that violence is one of several behaviors
used to control the partners actions. In contrast
to the cycle of violence, authors observe that
abuse is a constant force in battered womens
lives (Pence & Paymar, 1993). The power and
control wheel depicts eight nonphysical abusive
behaviors exhibited by men who batter: coercion, intimidation, emotional abuse, isolation,
denying, using children, male privilege, and
economic control. It illustrates that physical violence is part of a continuous pattern of controlling behaviors, not simply isolated incidents
or cyclical explosions. Batterers use of physical assaults may be infrequent and random,
but assaults reinforce the power of other controlling tactics, which are constant (Pence &
Paymar, 1993).
Random Dynamics
One type of random dynamics (criticality) is
common in complex systems. Criticality results
from constant stress on a system which is com-
Table 1
Predictions of Model Fit
Periodic
Chaotic
High predictability
High control
Causality by wife
Medium predictability
Low predictability
Low control
Causality by husband
More control strategies by husband
High stress
High hassles
Forgiveness
High stress
High hassles
Low closeness
High arguments
Mutual violence
Low closeness
High arguments
Mutual violence
Random
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
BURGE ET AL.
Sample
Investigators recruited 200 women from six
primary care clinics in San Antonio, Texas.
Eligible participants were women, married or
cohabiting with a man for one year or longer,
age 18 to 64, English- or Spanish-speaking,
with a recent history (past 30 days) of male-tofemale physical or verbal abuse. Women were
not approached for screening at all if a male
partner accompanied them to clinic that day.
Investigators excluded women if they were
pregnant, they planned to leave their partner
within the next 12 weeks, or study participation
would endanger their safety. The study was
approved by the Institutional Review Board of
the University of Texas Health Science Center
at San Antonio (for more details, see Burge et
al., 2014).
Procedure
Bilingual female researchers screened
women, obtained consent, enrolled participants,
and administered baseline surveys. Daily for 12
weeks, women telephoned an automated survey
and responded to study questions about the previous days experiences. Weekly, participants
To quantitatively assess dynamics, time series data must be complete. Women made daily
reports on a mean of 63.2 days (15.9 SD) out
of 84. Missing data were imputed using the
nstep procedure in the TISEAN software package to maintain nonlinear characteristics (Heggler, Kantz, & Schreiber, 1999).
Analytic Assessment of Dynamic Pattern
Using 84 days of data points, we estimated
the dynamic pattern (i.e., periodic, chaotic, random/criticality) in male partner violence for
each participant. Women were classified into
three groups: the periodic group demonstrated
linear (predictable) dynamics of violence with
evidence of an attractor (repeating behaviors),
the chaotic group reported nonlinear (unpredictable) dynamics of violence with an attractor
(repeating behaviors), and the random group
had nonlinear (unpredictable) dynamics of violence with no attractor. We used the time series
of daily reports of man-to-woman violence severity to determine group assignments. Lyapunov exponents assessed linearity/nonlinearity
of violence dynamics (Wolf, Swift, Swinney, &
Vastano, 1985), while saturated correlation di-
Table 2
Schedule of Study Measurements
Baseline
survey
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Variables
Background
Demographics
Relationship and violence history (Pence & Paymar, 1993)
Childhood abuse (Dube, Williamson, Thompson, Felitti, & Anda, 2004)
Relationship history
Relationship function (Siegel, 2006)
Psychiatric problems
Partners drug use (Bradley et al., 1998)
Wifes psychological problems (Kroenke, Spitzer, & Williams, 2001;
Spitzer, Kroenke, & Williams, 1999; Spitzer, Kroenke, Williams, &
Lwe, 2006)
Attitudinal, behavioral and health outcomes
Violence appraisal (Vitaliano, 1985)
Hope (Herth, 1992)
Readiness for change (Prochaska & DiClemente, 1986)
Support/stress (Parkerson, Broadhead, & Tse, 1991)
Coping strategies (Carver, Scheier, & Weintraub, 1989)
Health care utilization (Brown & Adams, 1992)
Functional status (McHorney, Ware, Rachel Lu, & Sherbourne, 1994)
Symptom burden (Katerndahl & Oyiriaru, 2007)
Daily violence and environment
Hassles (DeLongis, Folkman, & Lazarus, 1988)
Frequency of arguments (Coleman, Weinman, & Hsi, 1980)
Husbands/wifes alcohol intake (Bradley et al., 1998)
Stress
Marital harmony
Husband-to-wife violence (Straus, 1987)
Wife-to-husband violence (Straus, 1987)
Upset about violence
Forgiveness
Qualitative information
Important life events
Field notes
Qualitative interview (42 Ss)
Daily
survey
Weekly
contact
End-of-study
Model Fit
Results
How well do the dynamic groups (periodic,
chaotic, random) fit the models of violence dynamics (cycle of violence, family systems, Duluth model)? To answer this question, we analyzed group differences in the Violence
Appraisal (assessing control, predictability and
causality) and Abuse Behavior Inventory (assessing batterers controlling behaviors) using
analysis of variance. Second, we examined
daily surveys to determine: group differences in
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
BURGE ET AL.
N (%)
34 (17%)
152 (76%)
30 (15%)
49 (25%)
64 (32%)
41 (21%)
90 (45%)
38.2 11.7
105 (53%)
24 (12%)
54 (27%)
85 (43%)
(Mean SD)
Table 3
Sample Description (N 200)
Individual characteristics
Race/ethnicity
White
Hispanic
Employment
Part-time
Full-time
Educationa
High school graduate
High school graduate
Some college
Age (Mean SD years)
Relationship characteristics
Family income $20,000 per year
Type of marriagea
Church
Civil
Common law
family system dynamics, but they did not. Instead, they were more likely to appraise violent
behavior as caused by the male partner (not
significant).
1.5 1.4
9.6 8.9
7.3 9.0
5.5 6.5
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Table 4
Violence Appraisal, Abuse Behavior, and Household Environment by Dynamic Pattern
p .05.
p .01.
p .005.
Periodic
(n 16)
Chaotic
(n 40)
Random
(n 79)
10.56
7.50
8.31
10.69
9.13
8.11
9.55
8.11
6.88
1.47
5.13
1.87
9.47
8.56
1.50
7.60
2.47
1.49
3.72
1.15
8.74
5.46
1.65
5.90
3.61
1.69
4.01
1.46
9.41
6.22
2.89
6.71
4.03
8.7
.9
.4
2.2
3.1
1.1
1.1
5.9
.5
.4
1.2
2.4
1.2
.5
7.7
.8
.5
1.5
3.0
1.1
1.0
p .001.
Discussion
This study used baseline surveys and daily reports of intimate partner violence to assess the
fit between time-series derived dynamic patterns and three models of violence dynamics
found in the literature. Women in the periodic
group, consistent with the cycle of violence, found
violence to be predictable and controllable. They
reported high stress and arguments on violent days
and high arguments and violence preceding violent days, representing a buildup of tension. In
contrast, women in the random group found violence to be unpredictable and out of their control,
consistent with the Duluth model. Partners violent behavior, especially verbal abuse, was most
frequent in this group, occurring on nearly half the
days. Women in the chaotic group, on the day of
their partners violence, reported high stress and
hassles, low marital closeness and high womans
violence. This may suggest action-reaction in the
couple dynamic, consistent with family systems
theory. However, all groups reported this dynamic; it was not unique to the chaotic group.
BURGE ET AL.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Table 5
Frequency and Severity of Violence in 84 Daily Reports
p .05.
p .01.
p .005.
Periodic
(n 16)
Chaotic
(n 40)
Random
(n 79)
17.4
15.4
5.0
5.8
4.1
3.7
3.5
3.6
14.0
12.7
1.9
3.4
1.5
1.2
1.0
.9
24.9
23.4
4.8
8.0
4.3
1.4
1.0
.5
35%
2.87
28%
2.35
48%
2.55
12.6
3.7
5.8
4.4
4.1
3.6
8.8
1.7
1.8
1.3
1.1
1.1
17.2
3.1
3.5
2.8
1.7
.9
22%
1.97
15%
1.78
29%
1.86
p .001.
transformative states (Btz, Carlson, & Carlson, 1998) adapting to recent changes in the
system. Btz et al. (1998) observed that during
times of change, families operate at the edge of
chaos, randomly grasping for solutions to their
distress. Forces continually crushing the family
system would create unpredictable, nonlinear
interactions. In our study, women described internal stressors such as personal histories of
victimization, trauma, family conflict, illness,
and isolation, as well as external stressors related to poverty, unemployment, and partners
addiction.
The chaotic group was strikingly different
than the other groups; frequency and severity of
violence was lowest in this group, and marital
satisfaction was highest. Stress, arguments, and
drinking alcohol were lower. In this sample of
high-conflict couples, the chaotic group appeared to be the healthiest. They represented
couples with repeating but unpredictable interaction patterns; perhaps they were displaying
the adaptive side of complex adaptive systems, transitioning toward a healthier system.
The circumplex model of family systems,
particularly the concept of family flexibility,
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Table 6
Time Series Predictors of Husbands Violence From Daily Surveys
Periodic
(n 16)
Chaotic
(n 40)
Random
(n 79)
.705
.416
.676
.279
.427
.553
.279
.529
.528
.662
.493
.703
.150
.409
.511
.382
.551
.400
.682
.434
.722
.215
.420
.585
.492
.449
.317
.40 [.09]
.19 [.09]
.02 [.02]
.16 [.07]
.12 [.14]
.05 [.04]
.07 [.05]
.21 [.08]
.07 [.48]
.54 [.39]
.02 [.04]
.06 [.04]
.02 [.01]
.06 [.05]
.03 [.03]
.02 [.02]
.01 [.02]
.04 [.03]
.01 [.12]
.02 [.12]
.02 [.04]
.01 [.05]
.02 [.01]
.02 [.06]
.09 [.03]
.02 [.02]
.05 [.03]
.01 [.05]
.20 [.14]
.23 [.15]
p .05.
10
BURGE ET AL.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Conclusion
No single dynamic pattern describes couple
interactions in relationships with violence. Yet,
specific dynamic patterns correlate with different features of these relationships. These observations may explain why some efforts to intervene in violent relationships work, and some do
not. Complexity science suggests that periodic
dynamics respond in predictable ways to interventions, while random dynamics do not (Btz
et al., 1997; Katerndahl, 2009). If patients describe a repeating predictor of aggressive behaviorsuch as heavy drinking every Friday
night, this periodic pattern may be amenable to
an intervention directed toward the cause. On
the contrary, if patients describe partners aggression as frequent, controlling and random,
simple causes may be difficult to ascertain, and
simple interventions will not likely be effective.
In a random dynamic, the most successful remedy may be termination of the relationship. Alternatively, multifaceted interventions or the introduction of new agents (such as support
persons) into the patients life may be effective
in changing the dynamic and increasing her
safety.
References
Black, M. C., Basile, K. C., Breiding, M. J., Smith,
S. G., Walters, M. L., Merrick, M. T., . . . Stevens,
M. R. (2011). The National Intimate Partner and
Sexual Violence Survey (NISVS): 2010 summary
report. Centers for Disease Control and Prevention: Atlanta, GA.
Bradley, K. A., McDonell, M. B., Bush, K., Kivlahan, D. R., Diehr, P., & Fihn, S. D. (1998). AUDIT
alcohol consumption questions. Alcoholism: Clinical and Experimental Research, 22, 18421849.
Brown, J. B., & Adams, M. E. (1992). Patients as
reliable reporters of medical care process. Medical
Care, 30, 400 411. http://dx.doi.org/10.1097/
00005650-199205000-00003
Burge, S. K., Becho, J., Ferrer, R. L., Wood, R. C.,
Talamantes, M., & Katerndahl, D. A. (2014).
Safely examining complex dynamics of intimate
partner violence. Families, Systems, & Health, 32,
259 270. http://dx.doi.org/10.1037/fsh0000013
Btz, M. R., Carlson, J. M., & Carlson, J. (1998).
Chaos theory: Self-organization and symbolic
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
11
exponents from small data sets. Physica D. Nonlinear Phenomena, 65, 117134. http://dx.doi.org/
10.1016/0167-2789(93)90009-P
Siegel, J. P. (2006). Dyadic splitting in partner relational disorders. Journal of Family Psychology,
20, 418 422. http://dx.doi.org/10.1037/0893-3200
.20.3.418
Spitzer, R. L., Kroenke, K., & Williams, J. B. (1999).
Validation and utility of a self-report version of
PRIME-MD: The PHQ Primary Care Study.
JAMA: Journal of the American Medical Association, 282, 17371744. http://dx.doi.org/10.1001/
jama.282.18.1737
Spitzer, R. L., Kroenke, K., Williams, J. B., & Lwe,
B. (2006). A brief measure for assessing generalized anxiety disorder: The GAD-7. Archives of
Internal Medicine, 166, 10921097. http://dx.doi
.org/10.1001/archinte.166.10.1092
Straus, M. A. (1987). Conflict Tactics Scales. In N.
Fredman & R. Sherman (Eds.), Handbook of measurements for marriage and family therapy (pp.
159 163). New York, NY: Brunner/Mazel.
Vitaliano, P. P. (1985). Manual for Appraisal Dimension Scale and Revised Ways of Coping Checklist.
Seattle, WA: Department of Psychiatry, University
of Washington.
Walker, L. E. (1979). The battered woman. New
York, NY: Harper & Row.
Ward, M. (1995). Butterflies and bifurcations: Can
chaos theory contribute to our understanding of
family systems? Journal of Marriage and the
Family, 57, 629 638. http://dx.doi.org/10.2307/
353918
Wolf, A., Swift, J. B., Swinney, H. L., & Vastano,
J. A. (1985). Determining Lyapunov exponents
from a time series. Physica, 16D, 285317.
Yeragani, V. K., Pohl, R., Mallavarapu, M., & Balon,
R. (2003). Approximate entropy of symptoms of
mood: An effective technique to quantify regularity of mood. Bipolar Disorders, 5, 279 286. http://
dx.doi.org/10.1034/j.1399-5618.2003.00012.x
Zhang, X.-S., Zhu, Y.-S., Thakor, N. V., & Wang,
Z.-Z. (1999). Detecting ventricular tachycardia
and fibrillation by complexity measure. IEEE
Transactions on Bio-Medical Engineering, 46,
548 555. http://dx.doi.org/10.1109/10.759055
Received March 30, 2015
Revision received September 13, 2015
Accepted September 25, 2015