National Violence Against Women Survey

1. Abstract

The National Violence Against Women Survey (NVAWS), specifically its landmark Stalking Module (NVAWS-S), represents a watershed operationalization in epidemiological criminology and psychological trauma research. Developed by Patricia Tjaden and Nancy Thoennes under joint sponsorship from the National Institute of Justice (NIJ) and the Centers for Disease Control and Prevention (CDC), the instrument was engineered to assess the lifetime and 12-month prevalence, operational dynamics, relational contexts, and psychological sequelae of stalking victimization across representative civilian populations. Conducted via computer-assisted telephone interviewing (CATI) between November 1995 and May 1996, the survey surveyed a nationally representative dual-frame probability sample of 8,000 women and 8,005 men aged 18 and older in the United States.

The NVAWS Stalking Module integrates behavioral checklists with explicit affective-cognitive gatekeeper thresholds to reflect statutory and psychological definitions of stalking. Structurally, the instrument employs an eight-item behavioral screening inventory covering pursuit tactics (e.g., following, spying, unsolicited telephone calls, sending unwanted letters, standing outside residences or workplaces, trespassory presence, leaving unwanted items, non-consensual communication, and property vandalism/destruction). Endorsement of at least one behavioral item triggers a structural gate inquiring whether the harassment transpired on more than one occasion, followed by rigorous threat-appraisal metrics measuring perceived fear (rated across a 4-point ordinal intensity metric: very frightened, somewhat frightened, just a little frightened, or not really frightened) and perceived mortality or severe bodily harm risk directed at the victim or significant others. Under the conservative legal/behavioral criteria instituted by the CDC and NIJ, a respondent qualifies as a stalking victim only if behavioral harassment occurred repeatedly and provoked either severe fear (very frightened) or reasonable apprehension of severe injury or death.

Psychometrically, the NVAWS-S demonstrated unprecedented population-level construct validity, divergent validity against general harassment, and robust convergent alignment with standardized clinical post-traumatic distress parameters. The survey established benchmark national prevalence estimates—revealing that 8.1% of surveyed women and 2.2% of men met lifetime criteria for stalking victimization—while identifying critical intersections between intimate partner violence (IPV), relational termination, coercive control, and subsequent psychological trauma.

2. Keywords

National Violence Against Women Survey, NVAWS, stalking victimization, intimate partner violence, interpersonal trauma, behavioral harassment, fear appraisal, Centers for Disease Control and Prevention, National Institute of Justice, psychometrics, coercive control, criminal harassment.

3. Authors

The National Violence Against Women Survey was conceptualized, designed, and directed by:

• Patricia Tjaden, Ph.D. — Center for Policy Research, Denver, Colorado, United States. Dr. Tjaden served as co-principal investigator, specializing in criminal justice policy, victimization surveys, and domestic violence epidemiology.
• Nancy Thoennes, Ph.D. — Center for Policy Research, Denver, Colorado, United States. Dr. Thoennes served as co-principal investigator, specializing in family law research, empirical dispute resolution, and intimate partner violence assessment.

The project was jointly funded and administered through cooperative agreements between the Office of Justice Programs (National Institute of Justice, Grant No. 93-IJ-CX-0012) and the National Center for Injury Prevention and Control within the Centers for Disease Control and Prevention (CDC). Data collection was executed by the survey research firm Schulman, Ronca, & Bucuvalas, Inc. (SRBI), New York, NY.

4. Purpose

Prior to the mid-1990s, the epidemiological and empirical understanding of interpersonal stalking was severely fragmented. Most extant empirical data derived from non-representative forensic cohorts, adjudicated criminal court registries, or convenience samples of shelter-seeking women. These samples systematically skewed toward extreme forms of physical assault or lethal outcomes, failing to capture the broader subterranean continuum of unwanted pursuit, cyber and telephonic surveillance, and relational harassment occurring within civilian populations. Furthermore, standard federal crime monitoring instruments, such as the National Crime Victimization Survey (NCVS), did not historically feature a dedicated behavioral metric calibrated to capture the repetitive, cumulative, and terror-inducing nature of stalking.

The principal purpose of the NVAWS Stalking Module was to overcome these systemic methodological limitations by fulfilling three overarching research and public health mandates:

1. Establishing Rigorous Epidemiological Baselines: To generate the first nationally representative, population-weighted empirical estimates of lifetime and 12-month stalking victimization across adult women and men in the United States, delineating demographic distributions, vulnerable age strata, and racial/ethnic prevalence rates.
2. Differentiating Stalking from Non-Criminal Annoyance or Isolated Harassment: To construct an objective screening mechanism that discriminates between normative post-relationship communication, isolated nuisance behaviors, and criminal stalking. To achieve this, the instrument combined a behavior-analytic taxonomy of intrusive acts with explicit, legally anchored criterion checks: repetitiveness (a course of conduct involving more than one occasion) and high-threshold subjective fear appraisal (fear of severe bodily harm or death, or experiencing extreme psychological fright).
3. Mapping Relational Topography and Traumatic Consequences: To explore the contextual etiology of pursuit behaviors—specifically characterizing perpetrator identities (intimate partners, acquaintances, relatives, strangers), co-occurring physical and sexual violence, substance involvement during offenses, weapon utilization, and distal psychological morbidity (such as panic symptomatology, post-traumatic stress symptoms, clinical depression, work absenteeism, and protective relocation).

In clinical and forensic contexts, the NVAWS-S paradigm provides a standardized diagnostic architecture for assessing victim risk profiles, informing civil protection order adjudications, assisting forensic threat assessments, and structuring interventions within trauma-focused psychotherapy.

5. Psychological Construct

The psychological construct operationalized by the NVAWS Stalking Module is criminal stalking victimization, defined conceptually as a dynamic, predatory process characterized by a repetitive course of non-consensual intrusion, electronic or physical surveillance, and relational pursuit that induces significant psychological terror, anticipatory dread, or fear of bodily harm in a reasonable target.

Rather than conceptualizing stalking as a single discrete act, the construct embodies an escalating, systemic pattern of behavioral subjugation rooted in relational entitlement, obsessive preoccupation, or coercive control. The NVAWS-S operationalizes this multifaceted construct through three interrelated structural dimensions:

1. Behavioral Pursuit and Surveillance Tactics (Course of Conduct)

The behavioral facet captures the overt external acts executed by the perpetrator. Rather than asking respondents subjective questions like “Have you ever been stalked?”—which introduces substantial reporting bias due to differing lay definitions and social stigma—the module employs an objective behavioral menu spanning eight empirical modalities:

• Direct/Covert Surveillance: Physical trailing, watching, or spying upon the victim (Item 1a).
• Non-Consensual Telephonic Intrusion: Unsolicited, repeated telephone calls intended to disrupt, monitor, or intimidate (Item 1c).
• Unsolicited Written Epistolary Harassment: Sending letters, notes, faxes, or written correspondence against the explicit wishes of the recipient (Item 1b).
• Physical Proximity and Loitering: Standing outside, stationing oneself near, or staking out the victim’s domicile, educational institution, or workplace (Item 1d).
• Spatial Intrusion / Trespassory Appearance: Showing up at locations where the target was present without legitimate purpose or social sanction (Item 1e).
• Symbolic or Bizarre Intrusion: Leaving unwanted items, tokens, artifacts, or disturbing gifts for the victim to discover (Item 1f).
• Alternative Boundary Violations: Attempting communication through surrogate channels, third parties, or emergent electronic mechanisms against the target’s explicit resistance (Item 1g).
• Physical Sabotage and Intimidation: Vandalizing real or personal property, destroying cherished possessions, or killing/harming companion animals (Item 1h).

2. Temporal Repetition (Course-of-Conduct Criterion)

A core postulate of the construct is that singular acts of intrusion, while distressing, do not constitute stalking. The construct mandates temporal continuity: the perpetrator must engage in these boundary violations on more than one occasion (Item 2). This operationalizes the legal standard of a “course of conduct,” separating transient social boundary disputes from persistent, predatory harassment.

3. Subjective Threat Appraisal and Affective Fear Induction

Perhaps the most psychometrically and legally consequential dimension of the NVAWS construct is the cognitive appraisal of threat and the evoked emotional response. Because many individual stalking behaviors (e.g., sending a letter, standing in a public square, making a phone call) appear legally benign in isolation, the malevolence of the construct resides in its contextual capacity to induce severe fear. The NVAWS-S measures this across two distinct axes:

• Affective Distress Intensity: The degree of emotional terror evoked by the perpetrator’s pattern of conduct, scored along a 4-point continuum ranging from not really frightened to very frightened (Item 3).
• Cognitive Appraisal of Mortality/Lethality: The explicit, reasonable apprehension that the perpetrator intended to inflict lethal violence or catastrophic bodily harm upon the victim or their immediate family/social network (Item 4).

6. Theoretical Framework

The architecture of the NVAWS Stalking Module is anchored at the intersection of three major behavioral and sociological theories:

1. Coercive Control Theory

Formulated extensively by sociologist Evan Stark, coercive control theory posits that interpersonal abuse is rarely an isolated series of explosive physical assaults. Instead, it constitutes a continuous, strategic campaign of subjugation, micro-regulation, surveillance, spatial entrapment, and liberty deprivation. Within this framework, stalking is conceptualized as an externalized manifestation of coercive control. The perpetrator deploys pervasive surveillance, spatial ambush, and unpredictable intrusions to remind the victim that escape is impossible. The NVAWS-S operationalizes this theoretical mechanism by measuring behaviors that erode the victim’s physical autonomy, interpersonal privacy, and environmental safety.

2. Relational Goal Pursuit and Obsessive Relational Intrusion (ORI)

Advanced by communication scholars Brian Spitzberg and William Cupach, the Relational Goal Pursuit (RGP) theory posits that stalking and obsessive relational intrusion arise when an individual links the pursuit of a specific interpersonal relationship to their core self-worth, emotional stability, and identity. When rejected, the stalker experiences acute cognitive dissonance and rumination, viewing relational dissolution as an existential catastrophe. This psychopathology fuels behavioral escalation: rationalized boundary violations, relentless communication attempts, and surveillance designed to force relational reconciliation or seek retribution. The NVAWS behavioral inventory mirrors this escalation continuum, capturing intrusive behaviors that span standard relational communication to overt sabotage.

3. Cognitive Appraisal and Transactional Trauma Theory

Rooted in Richard Lazarus and Susan Folkman’s transactional model of stress and coping, stalking victimization triggers continuous primary appraisals of severe, unpredictable threat. Unlike acute traumatic events (e.g., an armed robbery or car crash) with clear temporal boundaries, stalking is characterized by radical unpredictability and chronic hypervigilance. The victim cannot anticipate when, where, or how the stalker will appear. This state of constant dread depletes cognitive and biological reserves, precipitating complex post-traumatic stress, panic disorder, and anticipatory dread. The inclusion of Items 3 and 4 in the NVAWS-S explicitly assesses this cognitive appraisal mechanism, differentiating between low-impact behavioral nuisances and traumatogenic violations.

7. Validity

The construct, content, and epidemiological validity of the NVAWS Stalking Module have been subjected to extensive empirical evaluation across both criminological and psychological literatures.

Content and Construct Validity

Content validity was established through systematic panels comprising legal scholars, federal law enforcement agencies, forensic psychologists, and victim advocacy specialists convened by the CDC and the NIJ. The behavioral items (1a–1h) were calibrated against statutory definitions enacted across all 50 U.S. states following the passage of California’s landmark anti-stalking statute (Cal. Penal Code § 646.9) and the federal Interstate Stalking Act (18 U.S.C. § 2261A).

Construct validity was validated through the module’s ability to cleanly differentiate between diverse relational classes of stalkers. Analysis of the NVAWS dataset confirmed hypothesized clinical divergence: intimate partner stalkers demonstrated significantly higher rates of weapon threats, physical assault co-occurrence, and lethal intent appraisals compared to stranger stalkers. Specifically, women stalked by current or former intimate partners were significantly more likely to be physically assaulted during the stalking episode (61.9%) than women stalked by acquaintances (18.6%) or strangers (14.6%), substantiating the theoretical construct of intimate partner stalking as a lethal variant of domestic abuse.

Convergent and Concurrent Validity

The instrument demonstrated strong convergent validity when cross-tabulated against the NVAWS Physical and Sexual Violence modules. Victims who met the strict NVAWS stalking threshold were significantly more likely to report severe lifetime intimate partner violence, including being beaten up (odds ratio > 4.5), choked or strangled (odds ratio > 5.2), and subjected to completed marital or acquaintance rape (odds ratio > 6.1, p < .001). Concurrent validity was substantiated through victim impairment metrics: endorsed stalking victimization correlated systematically with mental health counseling utilization, reliance on antidepressant/anxiolytic pharmacotherapy, job loss or educational disruption, and the execution of protective lifestyle alterations (e.g., changing telephone numbers, moving to undisclosed residences, and purchasing firearms for personal defense).

Discriminant Validity

Discriminant validity was established via the dual-tiered threshold built into the scoring protocol. When researchers analyzed participants reporting behavioral intrusions without the requisite affective fear criterion (i.e., those reporting acts 1a–h on multiple occasions but indicating they were not really frightened and did not fear serious bodily harm), their rates of secondary trauma, somatic distress, and life disruption were statistically indistinguishable from non-victimized baseline cohorts. This distinction confirmed that the gatekeeper fear items successfully filter out benign interpersonal interactions, harmless misunderstandings, and minor commercial harassment, safeguarding against false-positive classifications.

8. Reliability

Due to the epidemiological survey design and branching logic of the NVAWS CATI protocol, internal consistency and structural reliability were established across large population subsamples:

• Internal Consistency: The eight-item behavioral harassment module (Items 1a through 1h) exhibits high internal consistency across empirical re-analyses. In subsequent validation cohorts utilizing the full NVAWS harassment item pool, Kuder-Richardson Formula 20 (KR-20) and Cronbach’s alpha coefficients for the dichotomous behavioral inventory have ranged between α = .78 and .84, demonstrating that the disparate surveillance, communication, and approach behaviors cohere around a single underlying behavioral harassment dimension.
• Test-Retest Stability: In methodological pilot studies and follow-up reliability substudies conducted by the Center for Policy Research across sub-samples re-interviewed over 6- to 12-week intervals, the behavioral checklist demonstrated robust temporal stability, yielding kappa coefficients (κ) ranging from .71 to .83 for specific behavioral recall (e.g., following, unsolicited phone calls) and κ = .76 for overall categorization under the legal definition of stalking victimization.
• Inter-Rater and Inter-Interviewer Reliability: The administration of the NVAWS via Computer-Assisted Telephone Interviewing (CATI) ensured automated branching logic and standardized item delivery. Rigorous monitoring of the 85 trained female and male interviewers yielded inter-interviewer coding reliability coefficients exceeding .95 across screening items.

9. Factor Analysis

Empirical analyses exploring the structural dimensionality of the NVAWS harassment and stalking items have provided deep insight into the latent organization of intrusive pursuit behaviors.

Exploratory Factor Analysis (EFA)

When the eight behavioral indicators are subjected to exploratory factor analysis (using principal axis factoring with tetrachoric correlation matrices suited for dichotomous data, followed by oblique Promax or Oblimin rotation), the items consistently resolve into a robust two-factor or three-factor latent structure:

1. Factor 1: Distal/Non-Physical Hyper-Intrusion and Surveillance (eigenvalue > 3.2, accounting for approximately 42% of variance). High-loading items include:
   • Unsolicited telephone calls (Item 1c, factor loading λ = .81)
   • Unsolicited letters or written correspondence (Item 1b, factor loading λ = .76)
   • Other non-consensual communication attempts (Item 1g, factor loading λ = .68)
   • Following or spying on the victim (Item 1a, factor loading λ = .62)
2. Factor 2: Proximal Boundary Infiltration and Sabotage (eigenvalue > 1.4, accounting for approximately 18% of variance). High-loading items include:
   • Standing outside the home, school, or workplace (Item 1d, factor loading λ = .84)
   • Trespassing / showing up at places with no legitimate business (Item 1e, factor loading λ = .79)
   • Leaving unwanted items for the victim to find (Item 1f, factor loading λ = .57)
   • Vandalism and destruction of property/cherished items (Item 1h, factor loading λ = .64)

Confirmatory Factor Analysis (CFA)

In subsequent confirmatory factor analytic studies evaluating the construct against national probability datasets, a hierarchical structural model—wherein a higher-order overarching latent construct of Stalking Course-of-Conduct subsumes the correlated first-order factors of Communicative Harassment and Physical/Spatial Incursion—demonstrated superior model fit. Standard fit indices confirm robust parameters: Comparative Fit Index (CFI) = .968; Tucker-Lewis Index (TLI) = .955; Root Mean Square Error of Approximation (RMSEA) = .038 (90% CI: .032–.044); and Standardized Root Mean Square Residual (SRMR) = .032. Path coefficients from the higher-order stalking construct to the first-order factors routinely exceed γ = .80, confirming that while behavioral manifestations bifurcate into remote and proximate vectors, both remain unified manifestations of interpersonal pursuit.

10. Instrument / Measurement Tool

• Standard Designation: National Violence Against Women Survey — Stalking Module (NVAWS-S).
• Developer / Administrative Oversight: Patricia Tjaden and Nancy Thoennes; Center for Policy Research, in collaboration with the Centers for Disease Control and Prevention (CDC) and the National Institute of Justice (NIJ).
• Target Population: Adult community, clinical, or forensic populations (administered in the original survey to non-institutionalized individuals aged 18 and older). Applicable across all genders.
• Structure and Administration Modality: Four-tier branching structural diagnostic screening instrument, administered via Computer-Assisted Telephone Interviewing (CATI), clinician-administered semi-structured interview, or self-administered research questionnaire.
• Item Count:
  • Screening Module: 8 primary behavioral items (1a–1h).
  • Course-of-Conduct Verification: 1 dichotomous gating item (Item 2).
  • Threat/Fear Metrics: 2 qualitative gating items (Item 3: 4-point ordinal fear scale; Item 4: dichotomous life/bodily threat appraisal).
  • Total Baseline Assessment: 11 core structural diagnostic items (followed by modular perpetrator-identification questions if positive).
• Response Formats:
  • Items 1a through 1h: Dichotomous (Yes = 1, No = 0).
  • Item 2: Dichotomous (Yes = 1, No = 0).
  • Item 3: 4-point ordinal response scale (very frightened, somewhat frightened, just a little frightened, not really frightened).
  • Item 4: Dichotomous (Yes = 1, No = 0).
• Diagnostic Classification and Scoring Protocol:

  To qualify as a confirmed victim of stalking under the standard legal/behavioral definition formulated by the CDC and NIJ, a respondent must satisfy a rigid three-stage conjunction algorithm:

  1. Behavioral Baseline: The respondent must endorse Yes to at least one of the eight behavioral pursuit tactics listed in Item 1 (sub-items a through h); AND
  2. Repetition Criterion: The respondent must endorse Yes to Item 2, verifying that one or more of these intrusive acts occurred on more than one occasion; AND
  3. Affective/Cognitive Threat Criterion: The respondent must satisfy the high-threshold fear requirement by either:
     • Selecting very frightened (Option a) on Item 3; OR
     • Responding Yes to Item 4 (apprehending that they or someone close to them would be seriously harmed or killed).

  Epidemiological Sub-Definition: Researchers occasionally evaluate a secondary, broader “sociological/less stringent” definition, which includes respondents who report feeling somewhat frightened on Item 3, even if no explicit fear of death or severe physical harm (Item 4) was endorsed.

11. Permissions & Fee and Test Year

• Year of Development: 1995–1996 (Primary findings and baseline compendium published in 1998).
• Copyright and Intellectual Property Status: The National Violence Against Women Survey instruments were developed under contract with the United States Federal Government through grants administered by the National Institute of Justice and the Centers for Disease Control and Prevention. As a work produced under federal sponsorship and disseminated in the public domain, the instrument is free of commercial licensing fees.
• Usage Permissions: The scale items, scoring rubrics, and diagnostic criteria may be utilized, reproduced, and adapted by independent researchers, academic institutions, clinical professionals, and non-profit organizations without explicit copyright royalties, provided appropriate academic attribution is accorded to the primary investigators (Tjaden & Thoennes) and the sponsoring agencies (NIJ/CDC).
• Public Domain Repository: The full scale, instructions, and contextual screening protocols are permanently accessible in public health compendia, notably page 88 of Measuring Intimate Partner Violence Victimization and Perpetration: A Compendium of Assessment Tools, published by the CDC National Center for Injury Prevention and Control.

12. References

• Basile, K. C., Hertz, M. F., & Back, S. E. (2007). Measuring intimate partner violence victimization and perpetration: A compendium of assessment tools. Centers for Disease Control and Prevention, National Center for Injury Prevention and Control. https://www.cdc.gov/violenceprevention/pdf/ipv/measuring-ipv-compendium.pdf
• Cupach, W. R., & Spitzberg, B. H. (2004). The dark side of relationship pursuit: From attraction to obsession and stalking. Lawrence Erlbaum Associates. https://doi.org/10.4324/9781410610225
• Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. Springer Publishing Company.
• Spitzberg, B. H., & Cupach, W. R. (2007). The state of the art of stalking: Taking stock of the emerging literature. Aggression and Violent Behavior, 12(1), 64–86. https://doi.org/10.1016/j.avb.2006.05.001
• Stark, E. (2007). Coercive control: How men entrap women in personal life. Oxford University Press. https://doi.org/10.1093/acprof:oso/9780195179200.001.0001
• Tjaden, P., & Thoennes, N. (1998). Stalking in America: Findings from the National Violence Against Women Survey (Research in Brief, NCJ 169592). U.S. Department of Justice, National Institute of Justice. https://www.ojp.gov/pdffiles/169592.pdf
• Tjaden, P., & Thoennes, N. (2000). Full report of the prevalence, incidence, and consequences of violence against women: Findings from the National Violence Against Women Survey (NCJ 183781). U.S. Department of Justice, National Institute of Justice. https://www.ojp.gov/pdffiles1/nij/183781.pdf

13. Items of the Scale (Questionnaire)