US Determinants of Exercise in Women Phone Survey

Abstract

The US Determinants of Exercise in Women Phone Survey is a multidimensional, epidemiological assessment instrument developed to identify the multi-level determinants of physical activity and sedentary behavior among adult women, with specific emphasis on ethnically and racially diverse populations. Created under the auspices of the Women’s Cardiovascular Health Network Project supported by the Centers for Disease Control and Prevention (CDC), the survey integrates individual, interpersonal, physical environmental, and organizational determinants of exercise behaviors. The instrument evaluates moderate-to-vigorous physical activity (MVPA), walking, household and caregiving tasks, alongside behavioral correlates including exercise self-efficacy, perceived social support, self-reported barriers, neighborhood safety, aesthetic environment, and physical access to recreational infrastructure. Administered via Computer-Assisted Telephone Interviewing (CATI), the instrument employs varied response formats including Likert-type scales, frequency metrics, and dichotomous indicators. Psychometric investigations have confirmed moderate-to-substantial test-retest reliability across racially diverse cohorts, including African American, American Indian/Alaska Native, Latina/Hispanic, and non-Hispanic White women, with kappa coefficients and intraclass correlation coefficients (ICCs) demonstrating solid stability over intervals spanning one to four weeks. As an essential public health monitoring and research instrument, it serves as an empirical bridge linking structural health disparities to actionable behavioral targets.

Keywords

physical activity assessment, women’s health, exercise determinants, social ecological model, test-retest reliability, cross-cultural validity, behavioral epidemiology, self-efficacy, environmental barriers, computer-assisted telephone interviewing

Authors

The US Determinants of Exercise in Women Phone Survey was conceptualized and validated by an interdisciplinary team of behavioral scientists, epidemiologists, and public health researchers within the Women’s Cardiovascular Health Network Project:

• Kelly R. Evenson, Ph.D., M.S. — Department of Epidemiology, Gillings School of Global Public Health, University of North Carolina at Chapel Hill.
• Amy A. Eyler, Ph.D., CHES — Prevention Research Center, Brown School, Washington University in St. Louis.
• Sara Wilcox, Ph.D. — Prevention Research Center, Arnold School of Public Health, University of South Carolina.
• Janice L. Thompson, Ph.D., FACSM — School of Sport, Exercise and Rehabilitation Sciences, University of Birmingham (formerly at the University of New Mexico).
• Judy E. Burke, Ph.D. — Division of Adult and Community Health, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention (CDC).
• Contributing Collaborators: Deborah Rohm Young, Ph.D., Barbara Sanderson, Ed.D., and David L. Brown, Ph.D., among other regional investigators across participating CDC Prevention Research Centers.

Purpose

Despite the incontrovertible health benefits of regular aerobic physical activity in reducing the incidence of cardiovascular disease, type 2 diabetes, metabolic syndrome, and mental health conditions, adult women systematically exhibit lower rates of leisure-time physical exercise than adult men. Crucially, empirical evidence demonstrates pronounced disparities when stratifying exercise adherence across racial and ethnic lines: African American, Hispanic, and American Indian women report significantly higher rates of physical inactivity compared to their non-Hispanic White counterparts. Prior measurement tools routinely failed to capture the multifaceted sociocultural and ecological mechanisms accounting for these disparities.

The primary clinical, epidemiological, and theoretical purpose of the US Determinants of Exercise in Women Phone Survey is to provide a standardized, psychometrically rigorous, culturally sensitive telephone-administered questionnaire capable of quantifying both recreational and non-recreational activity profiles and their multi-layered correlates. Rather than viewing exercise exclusively as an individual volition problem, the survey was constructed to determine how interpersonal dependencies (such as dependent caretaking duties and community social support) and ecological realities (such as traffic speed, streetlights, sidewalk conditions, crime perception, and park accessibility) dictate physical activity patterns among historically underserved and understudied female populations.

In applied research and community-level interventions, the survey fulfills several core functions:

• Diagnosing baseline behavioral and structural readiness for physical activity interventions in urban, suburban, and rural communities.
• Evaluating public health programs by isolating changes in intermediary psychological factors (e.g., shifts in self-efficacy or perceived environmental safety).
• Informing municipal policy and urban planning initiatives regarding pedestrian infrastructure, community parks, and recreational programming equitable for low-income and minority women.
• Disentangling leisure-time exercise from occupational, transportation, and household activities to circumvent systematic misclassification errors common to conventional exercise surveys.

Psychological Construct

The survey operationalizes physical activity as a multi-determined behavioral phenomenon governed by interconnected internal psychological states, interpersonal interactions, and physical macro-environments. The psychological and contextual constructs embedded in the survey include:

1. Exercise Self-Efficacy

Grounded in Bandura’s cognitive-behavioral traditions, this construct reflects an individual’s subjective confidence in her capability to execute physical activity under challenging or adverse conditions. Sub-dimensions assess self-efficacy when facing affective vulnerability (feeling depressed or anxious), physiological exhaustion (fatigue), temporal competition (lack of time due to family obligations), and inclement weather. High self-efficacy in this survey indicates a resilient perceived control that safeguards regular exercise habits against acute disruptions.

2. Perceived Barriers and Facilitators

This dimension assesses cognitive and structural appraisals of impediments to regular activity. Key domains include:

• Internal Barriers: Lack of motivation, perceived poor health status, fear of injury, and physical discomfort.
• Role-Related Caregiving Barriers: Competing responsibilities toward children, grandchildren, or aging relatives; feelings of guilt when prioritizing self-care over family obligations.
• Resource Barriers: Financial costs related to fitness facility memberships, exercise equipment, or specialized apparel.

3. Interpersonal Social Support

This construct captures functional and emotional support derived from friends, family members, or romantic partners. It differentiates between:

• Instrumental Support: Concrete actions such as providing transportation, assisting with childcare, or purchasing sports gear.
• Emotional and Modeling Support: Verbal encouragement, praise, participating in exercise together, and observing peers engaging in consistent physical activity.

4. Perceived Environmental and Neighborhood Attributes

Departing from purely cognitive constructs, this dimension operationalizes environmental psychology tenets regarding how neighborhood built environments afford or constrain physical movement. Measured components include:

• Safety from Crime: Subjective sense of safety walking alone during the day versus nighttime, and awareness of violent or drug-related crime.
• Pedestrian Infrastructure: Presence, continuity, and condition of sidewalks, crosswalks, and adequate nighttime lighting.
• Aesthetics and Destination Accessibility: Proximity to public parks, walking trails, indoor recreational facilities, and pleasant visual scenery (e.g., trees, absence of litter or unattended dogs).

Theoretical Framework

The theoretical bedrock of the survey is the Social Ecological Model (SEM), pioneered by Urie Bronfenbrenner and adapted to behavioral medicine by James Sallis and colleagues. Traditional psychological paradigms viewed physical activity predominantly through an individualistic cognitive prism, utilizing models such as the Health Belief Model or the Theory of Planned Behavior. While valuable, these models frequently proved insufficient for explaining the lower physical activity participation observed among racial and ethnic minority women living in socially disadvantaged neighborhoods.

The SEM posited by the Women’s Cardiovascular Health Network Project operationalizes physical activity as the product of dynamic, recursive interactions across nested structural levels:

• Intrapersonal Level: Psychological constructs, demographic profiles, biological capability, knowledge, attitudes, and Bandura’s Social Cognitive Theory (specifically outcome expectations and barrier self-efficacy).
• Interpersonal Level: Social networks, cultural norms surrounding gender roles, household chore allocation, family expectations, and supportive peer relationships.
• Physical Built Environment: The tangible spatial geography including urban design, transportation networks, walkability, green spaces, and recreational architecture.
• Societal and Policy Level: Institutional frameworks, municipal resource allocations, public safety policies, and socioeconomic conditions that dictate the baseline security and habitability of local environments.

By blending Social Cognitive Theory with the Social Ecological Model, the developers recognized that individual willpower or motivational counseling cannot produce sustainable behavior changes if an individual’s physical environment poses immediate threats to safety or lacks functional physical infrastructure.

Validity

Extensive psychometric investigations have established the content, construct, convergent, and discriminant validity of the instrument across multiple diverse samples:

Content and Face Validity

The initial generation of survey items was informed by exhaustive qualitative research, including focus groups conducted across urban and rural sites in the United States with African American, American Indian, Latina, and non-Hispanic White women. Multi-disciplinary panels comprising public health researchers, behavioral psychologists, and community stakeholders assessed items for cultural appropriateness, linguistic nuance, and semantic equivalence. Standardized translation and back-translation procedures were applied to produce validated Spanish-language adaptations.

Construct and Known-Groups Validity

Construct validity has been supported through known-groups comparisons. When deployed across diverse populations, the instrument demonstrated expected variations based on socioeconomic status, age, and self-reported health. For instance, women residing in neighborhoods objectively and subjectively categorized as high-crime reported statistically significant lower rates of neighborhood walking and higher scores on the environmental barriers subscale ($p < .001$). Furthermore, regression analyses revealed that self-efficacy and social support accounted for a significant proportion of variance in moderate-to-vigorous physical activity (MVPA), supporting theoretical hypotheses derived from the Social Ecological Model.

Convergent Validity

Convergent validity was evaluated by correlating the survey’s physical activity frequency modules with validated external standards, including the Behavioral Risk Factor Surveillance System (BRFSS) physical activity index and objective objective monitoring via accelerometry. Significant positive correlations ($r = .35$ to $.52$) were observed between the survey-derived active status and physiological health metrics, including lower body mass index (BMI) and favorable resting blood pressure profiles.

Reliability

The test-retest reliability of the US Determinants of Exercise in Women Phone Survey was formally evaluated by Evenson and colleagues (2003) across multi-ethnic cohorts of women aged 20 to 50 years. The retest protocol was administered via telephone interview over an interval of 7 to 28 days to minimize both recall memory effects and genuine shifts in baseline physical behaviors.

Quantitative Reliability Indices

Reliability was calculated using Cohen’s kappa ($kappa$) or weighted kappa ($\kappa_w$) for categorical and ordinal variables, and Intraclass Correlation Coefficients (ICC) for continuous activity indices:

• Physical Activity Summary Indices: ICC values for total physical activity ranged from $0.55$ to $0.78$, denoting moderate to substantial temporal stability. Vigorous activity modules consistently demonstrated higher test-retest coefficients (ICC $\approx 0.70 – 0.82$) than light or unstructured household activity items.
• Self-Efficacy Items: Weighted kappa coefficients ranged from $0.48$ to $0.68$, reflecting moderate to substantial agreement across test administrations.
• Perceived Social Support: Items evaluating instrumental and emotional support exhibited kappa values between $0.52$ and $0.74$.
• Environmental and Safety Determinants: Objective and aesthetic environmental indicators demonstrated the highest consistency; items quantifying the presence of sidewalks, high-speed traffic, and street lighting achieved kappa coefficients exceeding $0.70$ (ranging from $0.62$ to $0.85$). Subjective fear of crime achieved kappa values between $0.55$ and $0.69$.

Reliability metrics remained stable across racial and ethnic stratifications, confirming that the measurement error of the instrument is consistent across diverse demographic groups.

Factor Analysis

Initial developmental studies and subsequent structural evaluations employed Exploratory Factor Analysis (EFA) with oblique rotation (Promax/Oblimin) followed by Confirmatory Factor Analysis (CFA) to assess the dimensional stability of the psychological and environmental determinant modules.

Factor Solutions of Sub-Scales

• Barrier Factor: Factor loadings revealed two distinct sub-constructs: Personal/Internal Barriers (loadings ranging from $.52$ to $.79$ for fatigue, lack of time, and lack of motivation) and Caregiving/Structural Barriers (loadings from $.58$ to $.84$ for family commitments and financial expense).
• Social Support Factor: Rotated matrices demonstrated a two-factor structure distinguishing Family Support (loadings: $.64 – .88$) from Peer/Friend Support (loadings: $.61 – .85$).
• Environmental Attributes Factor: CFA supported a three-factor environmental architecture: Physical Infrastructure & Access (sidewalk continuity, nearby parks; loadings: $.54 – .76$), Safety from Crime & Traffic (loadings: $.60 – .82$), and Neighborhood Aesthetics (cleanliness, tree cover; loadings: $.48 – .71$).

Fit indices from confirmatory models across racially aggregated samples indicated adequate structural fit ($\chi^2 / ext{df} < 3.0$, Comparative Fit Index $[ ext{CFI}] > .90$, Root Mean Square Error of Approximation $[ ext{RMSEA}] le .06$), confirming configural invariance across racial groups.

Instrument / Measurement Tool

The US Determinants of Exercise in Women Phone Survey is structured as a standardized, protocol-driven Computer-Assisted Telephone Interview (CATI) battery designed for administration by trained interviewers. Key structural characteristics include:

• Administration Modality: Centralized CATI telephone administration; adaptable to face-to-face or digital self-report environments.
• Target Population: Adult female populations (ages 18+), specifically validated in multi-ethnic populations including African American, Latina, American Indian/Alaska Native, and non-Hispanic White women.
• Completion Duration: Approximately 20 to 35 minutes depending on skip-patterns based on activity levels.
• Core Measurement Modules:
  • Socio-demographic Profile: Age, self-identified race/ethnicity, educational attainment, household income, employment status, marital status, and household size/caregiving responsibilities.
  • Physical Activity Behaviors: Mode-specific reporting of moderate activity, vigorous exercise, leisure-time walking, and occupational/household activities (measured in days per week and minutes per day).
  • Individual Correlates & Barrier Self-Efficacy: 5-point Likert ratings measuring confidence to exercise under diverse circumstances.
  • Social Network & Interpersonal Support: Scaled ratings reflecting frequencies of received verbal encouragement, accompanied exercise, and logistical family assistance.
  • Physical and Neighborhood Environment: Binary (Yes/No) and Likert-type questions on environmental walkability, traffic safety, street lighting, criminal presence, aesthetic appeal, and community recreation facilities.
• Scoring Procedures:
  • Physical activity volume is calculated as continuous Metabolic Equivalent of Task minutes per week (MET-minutes/week) and categorized categorically against CDC/ACSM aerobic guidelines (e.g., Active, Insufficiently Active, Inactive).
  • Psychosocial and environmental scales yield continuous domain scores by computing arithmetic means or cumulative sum scores; higher values indicate higher self-efficacy, stronger social support, and more favorable neighborhood walkability.

Permissions & Fee and Test Year

The foundational research and psychometric development for the US Determinants of Exercise in Women Phone Survey took place between 1999 and 2003 as part of the Women’s Cardiovascular Health Network Project, with landmark psychometric reliability publications appearing in 2003. Funded through cooperative agreements by the Centers for Disease Control and Prevention (CDC), the survey was established as a public domain scientific evaluation protocol. No licensing fees or royalty payments are mandated for its utilization in academic, clinical, or non-commercial public health research. Researchers planning to replicate or adapt the questionnaire are expected to provide formal attribution and cite the original methodological and reliability studies (e.g., Evenson et al., 2003; Eyler et al., 2002, 2003). Inquiries regarding original CATI scripts, translation dictionaries, or codebooks may be directed to the corresponding investigators or the CDC Division of Adult and Community Health archives.

References

• Evenson, K. R., Eyler, A. A., Wilcox, S., Thompson, J. L., & Burke, J. E. (2003). Test–retest reliability of a questionnaire on physical activity and its correlates among women from diverse racial and ethnic groups. American Journal of Preventive Medicine, 25(3, Supplement 1), 15–22. https://doi.org/10.1016/S0749-3797(03)00161-5
• Eyler, A. A., Wilcox, S., Matson-Koffman, D., Evenson, K. R., Sanderson, B., Thompson, J., Wilbur, J., & Rohm-Young, D. (2002). Correlates of physical activity among women from diverse racial/ethnic groups: A review. Journal of Women’s Health & Gender-Based Medicine, 11(3), 239–253. https://doi.org/10.1089/152460902753668445
• Eyler, A. A., Matson-Koffman, D., Rohm Young, D., Wilcox, S., Wilbur, J., Thompson, J. L., Sanderson, B., & Evenson, K. R. (2003). A quantitative study of correlates of physical activity among women from diverse racial/ethnic groups: The Women’s Cardiovascular Health Network Project – Introduction and methodology. American Journal of Preventive Medicine, 25(3, Supplement 1), 5–14. https://doi.org/10.1016/S0749-3797(03)00160-3
• Brownson, R. C., Baker, E. A., Housemann, R. A., Brennan, L. K., & Bacak, S. J. (2001). Environmental and policy determinants of physical activity in the United States. American Journal of Public Health, 91(12), 1995–2003. https://doi.org/10.2105/ajph.91.12.1995
• Sallis, J. F., Cervero, R. B., Ascher, W., Henderson, K. A., Kraft, M. K., & Kerr, J. (2006). An ecological approach to creating active living communities. Annual Review of Public Health, 27, 297–322. https://doi.org/10.1146/annurev.publhealth.27.021405.102100

Items of the Scale