ADVHOCaT_SS_B 8-27-15 - ACF-OPRE Response - 11-20-15_clean

Accomplishments of the Domestic Violence Hotline, Online Connections and Text (ADVHOCaT) Study

OMB Information Collection Request

New Collection

Supporting Statement

Part B

August 27, 2015

Submitted By:

Office of Planning, Research and Evaluation and

Family and Youth Services Bureau

Administration for Children and Families

U.S. Department of Health and Human Services

7^th Floor, West Aerospace Building

370 L’Enfant Promenade, SW

Washington, D.C. 20447

Project Officers:

Seth Chamberlain, OPRE/ACF

Samantha Illangasekare, OPRE/ACF

Aleta Meyer, OPRE/ACF

Rebecca Odor, FYSB/ACF

Angela Yannelli, FYSB/ACF

B1. Respondent Universe and Sampling Methods

The possible universe of respondents to the web-based Preference of Use Survey includes all visitors to the NDVH website (www.NDVH.org or www.thehotline.org) and LIR website (www.loveisrespect.org). All visitors to these websites will be eligible to complete the survey. Those who will be included in the sample are individuals who click on the button on the websites to access the survey. Participation is voluntary and anonymous.

Survey Analysis

The analyses of the ADVHOCaT survey results will follow a systematic process of first examining the distribution of the responses to the categorical survey questions, qualitatively coding the open text box responses, and conducting sub group analyses to determine if differences exist in the categorical responses by the reported demographic characteristics (for specific survey questions, see Attachment A: NDVH/LIR Preference of Use Survey). The initial frequency analyses will determine the number of respondents and the percentage of total respondents for each question with a categorical response category. The survey will aid in the evaluation of the NDVH/LIR services to determine if there is a preferred mode of contact to the NDVH/LIR, and if so what mode of contact is more or less preferred by the overall sample. Open ended responses will be qualitatively coded. These codes will be analyzed quantitatively using numeric labels for thematically similar qualitative responses, to determine if reasons for a preferred mode of contact differ by participant in the overall sample.

In order to complete these analyses, a final sample of 5,000 participants has been projected: a sample of 5,000 will enable us to detect small to medium differences between subpopulations. For example, the evaluation will seek to determine if there are differences between mode of contact between young adult male and young adult female respondents. This sample size will allow for bivariate analyses to determine if there are medium to small effect size differences between young adult male and young adult female respondents in their reported preferred mode of contact, which would require a sample of at least 176 per group (see Table B below).¹

Table B: Sample size needed to detect small and medium differences between two groups   (based on two-tailed hypothesis, 80% power, group1 centered at 0.50, and significance level of 0.05 or less)
Effect size	Equals a difference of	Minimum number per group
Effect size of 0.2 (small)	10%	384
Effect size of 0.3	15%	176
Effect size of 0.4 (medium)	20%	91

B2. Procedures for Collection of Information

To collect information from individuals who visit the NDVH and LIR websites, the study team, with support from the NDVH and LIR, will create a link on the NDVH and LIR web home pages in the form of a button inviting website visitors to take the survey. This button will be placed in a prominent but unobtrusive location on each of the two home pages. Upon clicking on this link, individuals will be directed to a separate screen that includes information about the NDVH/LIR Preference of Use Survey, including its purpose, summary of content, the time it will take to complete, privacy information, and participants’ rights to refuse to answer any questions (see Attachment B: ADVHOCaT Consent Form for Survey). After reading the consent information, individuals will be prompted to click a link if they consent to participating in the survey. This link will take the participant to a separate screen with the NDVH/LIR Preference of Use Survey questions (see Attachment A: NDVH/LIR Preference of Use Survey).

The survey button, consent form, and survey will remain active on the NDVH and LIR websites for up to 12 weeks, or until the maximum number of respondents included in this ICR is reached (5,000 respondents).

The survey will take each participant up to 2.5 minutes (150 seconds) to complete.

B3. Methods to Maximize Response Rates and Deal with Nonresponse

Expected Response Rates

The NDVH and LIR have conducted surveys through their websites in the past on a range of topics related to domestic violence and dating abuse for their own organizations’ purposes. The number of responses to these surveys have varied based on the specific topic. The most recent survey conducted on the NDVH and LIR websites over 5.5 weeks on the topic of intimate partner violence and HIV/AIDS received 903 respondents through the LIR website and 232 respondents through the NDVH website. Based on these response rates, the study team would expect to receive at least 1970 respondents through the LIR website and 506 respondents through the NDVH website over a period of 12 weeks (for a total of 2,476 respondents). However, the study team anticipates that more individuals will choose to respond to the NDVH/LIR Preference of Use Survey, given that the information being requested is less sensitive than in past surveys. Therefore, we expect 5,000 respondents between the NDVH and LIR websites.

Dealing with Nonresponse

Only those individuals who visit the NDVH or LIR websites and are interested in participating in the survey will be included in this data collection. If website visitors do not want to participate, they do not have to click on the button to take the survey, or if they do not consent after reading the consent information, they do not have to click on the link to access the survey. Based on past NDVH and LIR surveys, the study team expects to receive 5,000 responses. If fewer individuals chose to respond to the survey, then the study team will work with the NDVH and LIR to leave the survey active on the websites for an additional number of weeks necessary to achieve the sample size.

Maximizing Response Rates

In order to maximize response rates to the survey, the study team will work with the NDVH and LIR to place the survey button in a prominent, but unobtrusive, location on each of the websites’ home pages. The surveys will remain on the websites for the length of time necessary to receive the necessary sample size.

B4. Tests of Procedures or Methods to be Undertaken

The study staff will test the survey button and link to take the survey after consent before activating the survey for access by website visitors in order to determine that the survey links are operational and that the data collected through the survey is accurately captured in a database.

B5. Individual Consulted on Statistical Aspects and Individuals Collecting and/or Analyzing Data

The study design plan and data collection protocols were developed by study staff at the George Washington University Milken Institute of Public Health who have experience in the area of domestic violence research and statistical analyses of large data sets. These include Dr. Karen McDonnell (Principal Investigator), Dr. Bart Bingenheimer (Lead for Quantitative Methods), and Nitasha Nagaraj (Senior Research Associate).

Input was also provided by the ADVHOCaT project officers at OPRE and FYSB: Seth Chamberlain, Samantha Illangasekare, Aleta Meyer, Rebecca Odor, and Angela Yannelli.

1 A sample size of 5000 will enable us to detect differences between young men who visit the website and other subpopulations. With a final respondent participant pool of 5000, 3750 are projected to represent adults and approximately 1250 young adults (based on current chat service usage, we anticipate 25% of respondents will represent a young adult population). Analyses of current users of NDVH/LIR services find 14% of contactors are male, 86% are female, and <1% are transgender. Thus, among young adult survey participants (n=1250), 175 are projected to be male.

4