LINEAR AND NONLINEAR DYNAMICS IN INTIMATE PARTNER VIOLENCE

LINEAR AND NONLINEAR DYNAMICS IN INTIMATE PARTNER VIOLENCE David Katerndahl, MD Sandra Burge, Ph. D Robert Ferrer, MD Johanna Becho Robert Wood, Dr. PH Reuben Mc. Daniel, Ed. D Acknowledgment: This study was supported by a grant from the National Science Foundation.

CYCLIC THEORY OF VIOLENCE 1. Tension-Building Phase 3. Calm, Loving Respite 2. Explosion!

POWER AND CONTROL WHEEL Using Coercion Intimidation & Threats Using Emotional Economic Power Abuse Control and Using Male Control Using Privilege Isolation Using Denying, Children Blaming, Minimizing

FAMILY SYSTEMS THEORY reaction Wife Interaction reaction Husband

SUMMARIZING DYNAMICS n Degree Of Nonlinearity – Meaning § Output Not Proportional To Input § Unpredictability – Types § § § n Information-Based Algorithmic Complexity Chaos-Based Dynamic Pattern – – – Periodic Dynamics Random Dynamics Chaotic Dynamics

Framework For Understanding Dynamics n Periodic Dynamics

Framework For Understanding Dynamics n Random Dynamics

Framework For Understanding Dynamics n Chaotic Dynamics

FRAMEWORK FOR UNDERSTANDING DYNAMICS n Periodic, Predictable Recurrences (Cyclic Theory) → Periodic Dynamics n Constant Stressors With Variable Violence (Power Wheel Theory) → Random Dynamics n Variable Strength & Direction Of Feedback (Systems Theory) → Chaotic Dynamics

METHODS Adult primary care women screened for IPV in prior month Severity of risk assessment & enrollment Completion of baseline interview (n=20) (Relationship & Violence History) Completion of 60 daily Interactive Verbal Response telephone assessments (Violence, Hassles, Arguments, Stress, Alcohol Intake, Closeness) Completion of end-of-study qualitative interview

Subject Protections No History Of Severe Violence 2. Community Resources Available 3. Hide Study Materials 4. Tell Partner About Involvement 1. -Not Study Topic Coded Interactive Verbal Response 6. Weekly Contact At Safe Phone 7. Taken to FJC at End of Study 5.

ANALYSIS n n Imputation Of Missing Data: TISEAN nstep procedure Assessment Of Nonlinearity n n Identification Of Dynamic Patterns: Lyapunov vs CD Saturation Identification Of Correlates Of Violence n Combination Of Measures Across Subjects – – – Information-Based: Algorithmic Complexity: Chaos-Based: Approximate Entropy LZ Complexity Lyapunov Exponent – Same-Day Correlates: – Prior-Day Correlates: Cross-Correlations Vector Autoregression Causality & Coefficients – – – Squaring rs (Glass, 1978) Z-statistics (Souffer et al, 1966) Weighted By SEs (Greenland, 1987) Cross-Correlations: p-Values: Coefficients:

STUDY RESULTS n 20 Women Enrolled – Ethnicity: – Age: – Education: – Employment: – Income: 80% Hispanic 20 - 62 Years Old <7 Years – College Grad 45% Full-Time $5 -10, 000 - $60 -70, 000

STUDY RESULTS n Household – – – Environment Husband’s Alcohol Intake Hassles Stress Arguments Closeness - 75% Husbands ≥ 9 Drinks Daily - Mean 1. 3 - 27. 2 (0 -30) - 38% Days “Extremely Stressful” - 57% Days With Arguments - 44% Days Felt “Very Distant”

STUDY RESULTS n Violence Reported – Frequent Violence § § § Threats Throwing Things Pushed/Slapped - 225/586 (38%) Days - 183 Days - 104 Days - 43 Days – Severe Violence § § § Kicked/Hit Beaten Weapons - 6 Women, 21 Days - 3 Women, 5 Days - 1 Woman, 31 Days

Patterns (Example) VIOLENCE • Constant Alcohol Intake • Same-Day Violence Relationships • Inverse - Closeness • Positive • Arguments • Stress • Period Of Sustained Violence • Arguments • Hassles

POTENTIAL MANAGEMENT IMPLICATIONS 1. Nonlinearity 1. Treatment Response Unpredictable 2. Dynamics-Specific Approaches 1. Periodic – Simple Interventions 2. Chaotic – Control / Anti-Control 3. Random – Multifaceted Interventions 3. Daily Diaries – Avoid Violence 1. Triggers 2. Prior-Day Correlates

POTENTIAL RESEARCH IMPLICATIONS Each Situation Unique 2. NO ONE THEORY CORRECT 1. Current Theories Only Part Of Whole 2. Transitions Between Theories PERIODIC RANDOM CHAOTIC CYCLIC POWER SYSTEMS WHEEL