OBJECTIVES: To describe survey methodology used to evaluate the burden of illness and racial differences in adult female acne (AFA). METHODS: A targeted, web-based survey was used to recruit a racially diverse sample of US adult females (25-45 years) with facial acne vulgaris from an existing pool of internet panelists.  Subjects who self-reported ≥25 visible facial pimples at screening were eligible.  Recruitment was stratified by age (50% 25-35; 50% 36-45 years) and race (50% White; 50% Non-White [25% Black/African American; 25% Hispanic/Asian/Other]). Survey outcomes included: sociodemographic and clinical characteristics; resource utilization; treatment satisfaction; quality of life; perceptions; coping behaviors; work environment/productivity; anxiety/depression symptoms; and skin-specific treatment preferences (non-White females only).  Validation rules were pre-programmed into the survey to improve data quality.  Descriptive statistics summarize results in the total sample; racial differences (White vs. non-White) were evaluated using descriptive statistics and t-test/chi-square analyses. RESULTS: The survey was fielded online from Oct-Nov 2011.  A total of 7245 panelists received survey invitations via email.  Of 4112 survey respondents, 208 (5.1%) were eligible and completed the survey.  Mean age of sample was 35±6 years; 48.6% non-White [Black (n=51); Hispanic (n=23); Asian (n=16); Other (n=11)].  Most females (80.3%) reported 25-49 visible pimples, followed by 50-75 (13.5%) and >75 (6.3%).  Median survey completion time was 24.5 minutes. CONCLUSIONS: Web-based survey methodology with an existing pool of panelists permitted focused recruitment of respondents to reach targeted sample sizes in each age and race/ethnicity stratum.  This method enabled recruitment of a targeted subset of patients, including non-White females and those with greater acne severity (≥50 visible pimples).  Web-based surveys are an effective method for collecting patient-reported data for stratified patient cohorts, while minimizing both time and cost.  This method was well suited for studying the burden of AFA in a real-world cohort.