SOP Data Collection Project -- SSB Template 081012

Street Outreach Program Data Collection Project

OMB Information Collection Request

Formative Generic Clearance

0970 - 0356

Supporting Statement

Part B

August 2012

Submitted By:

Caryn Blitz, Ph.D.

Office of the Commissioner

Administration on Children, Youth and Families

Administration for Children and Families

U.S. Department of Health and Human Services

1250 Maryland Avenue, SW, 8^th Floor

Washington, DC 20024

B1. Respondent Universe and Sampling Methods

The respondents that will provide feedback will be street youth served by FYSB’s SOP grantees as well as street youth who do not currently utilize services from SOP grantees in each city. The eleven cities include: Austin, TX; Boston, MA; Chicago, IL; Washington, DC; Minneapolis, MN; New York City, NY; Omaha, NE; Port St. Lucie, FL; San Diego, CA; Seattle, WA; and Tucson, AZ. The youth served by these programs are between 14-21 years of age and are homeless. These homeless youth may use SOP drop-in centers to take a shower, eat a hot meal or obtain food coupons, received hygiene kits, and/or obtain referrals for medical, dental, mental health or social services. The youth who do not use SOP services are included as respondents to try and identify whether they have needs that the SOP grantees are not currently able to meet. This information will help SOP grantees to improve their services and better meet the needs of all homeless street youth in their city.

B1.1. Sampling Design and Recruitment

Data will be collected via focus groups and computer-assisted personal interviews. All data will be collected in person.

Personal Interview: An unbiased method for sampling the homeless has not been identified. Many studies have clearly documented the difficulty associated with any attempt to enumerate homeless populations (e.g., Bur & Taeuber, 1991; Dennis, 1991; Rossi, Wright, Fisher, & Willis, 1987; Wright & Devine, 1992). A promising method, the Respondent Driven Sampling (RDS) approach (Heckathorn, 2002; Heckathorn et al., 2002) has been used with adult and youth homeless populations (Coryn et al, 2007; Gwadz et al., 2010) and with small native communities in the Arctic (Dombrowski et al., in press).

Local research staff will recruit initial “seed” respondents by word of mouth and by handing out and posting the Interview Recruitment Brochure (Attachment A) in locations frequented by street youth. To assure the highest degree of coverage among youth, two initial seeds will be identified in each city. Once selected, a set of screening protocols that are asked before the bulk of the interview questions begin (see Attachment B: Interview Questions) will be used to verify that the initial seeds meet the study age range and definition for being homeless, and are screened out for the following: inebriated or high, cognitively impaired, or actively psychotic. To be eligible to participate in the study, individuals must be 14-21 years of age and meet the definition of “homeless” that best describes their age group. This study will employ the Stewart A. McKinney Act of 1987 definition if homeless for individuals over the age of 18: 1) An individual who lacks a fixed, regular, and adequate nighttime residence, or an individual who has a primary nighttimes resident that is a) a supervised publicly or privately owned shelter designed to provide temporary living accommodations (including welfare hotels, congregate shelters, and transitional housing; or b) a public or private place not designated for or ordinarily used as regular sleeping accommodations for human beings (HUD, 1995). For those participants 18 years of age or younger (to age 14), the study will employ the definition of “homeless” provided by the National Network of Runaway and Youth Services: defined as someone 18 years or younger who cannot return home or has chosen never to return home and has no permanent residence (GAO, 1989).

As mentioned above, initial seed respondents will be provided a $20 gift card for their interview and will be offered additional gift cards to recruit people from their social networks to participate in the interview. Each seed will be given three coupons (see Attachment C: Seed Coupon) to give to homeless youth they know. Coupons will have unique identifiers that will link them to their seed. Coupons will have an expiration date of 7-14 days to allow for tracking the rate of return and to reduce respondent burden. Initial seed respondents will be provided a $10 gift card for each returned coupon that results in a completed interview. Thus, a respondent who completes the interview and successfully recruits three peers would receive $50 total in gift cards. Each new participant will be offered the same number of coupons and incentives to recruit respondents from their social networks, up to the fourth seed.

Figure 1 (see below) is a visual representation of how the reimbursements will operate within the RDS design. Solid-colored circles represent homeless individuals who have completed a personal interview. Arrows indicate the recruitment coupons given to peers. Bold arrows signify that an individual was successfully recruited and completed a personal interview. Therefore, each bold arrow indicates a $10 gift card for the recruiter. Thin arrows that lead to circles marked with an “X” indicate that the coupon was not returned and the seed was unsuccessful.

Two seed respondents will be identified in each city. Previous work conducted by Dombrowski with indigenous populations in Canada has demonstration that about half of all initial seeds do not start network trees (see Figure 1, Seed 2) and for every three coupons given to a respondent, only 1.1-1.4 were returned for successful interviews.

Given the contractor’s previous success in recruiting and interviewing homeless adolescents and women, they anticipate that for every three coupons given, two successful interviews will be achieved. An example of a successful seed can be seen in Figure 1, Seed 1. Based on Heckathorn’s analyses, “previous applications of RDS showed that the number of waves required for the sample to reach equilibrium is not large, generally not more than four to six” (Heckathorn et al., 2002, p. 58). Equilibrium indicates that further iterations are unlikely to change the demographics of the samples (Salganik & Heckathorn, 2004; Salganik, 2006). Based on these assumptions, it is estimated that each seed that is successfully initiated will result in 31 interviews by the 4th recruitment wave for a sample of 62 homeless street youth in each city. To account for seeds that never initiate, a third seed will be selected in cities for which half the sample has not been identified seven weeks after the data collection begins. Based on the work by Dombrowski, coupons should only be redeemed in person at the SOP agency. Prior experience suggests that some level of inconvenience is needed for respondents to recruit “closer” network members they believe will actually participate, increasing their odds for reimbursement. Without his level of involvement, it is difficult to saturate personal networks and draw conclusions about the representativeness of the sample. Recruitment and interviews will take place over a 5 month period.

Focus groups: A convenience sample will be used for the focus groups. SOP agency staff will recruit focus group participants (see Attachment D: Focus Group Recruitment Brochure, Attachment E: Focus Group Consent, and Attachment F: Focus Group Questions). Three focus groups of six participants each will be conducted at each of the 11 sites.

B2. Procedures for Collection of Information

This is a one time data collection. Statistical analysis will be limited to descriptive statistics for the personal interview and qualitative summaries of the focus groups.

Personal Interview. Local research staff will recruit initial “seed” respondents by word of mouth and by handing out and posting the Interview Recruitment Brochure (Attachment A) in locations frequented by street youth. To assure the highest degree of coverage among youth, two initial seeds will be identified in each city. Once selected, a set of screening protocols that are asked before the bulk of the interview questions begin (see Attachment B: Interview Questions) will be used to verify that the initial seeds meet the study age range and definition for being homeless, and are screened out for the following: inebriated or high, cognitively impaired, or actively psychotic. Interviews will most likely take place in private interview rooms at the SOP agency, but some interviews may take place in another agency, parks, cafes, or libraries. The interview location is largely determined by the respondent. Some respondents may feel uncomfortable completing the interview in the specified SOP agency; in that case, a more suitable location will be negotiated between the respondent and the interviewer. Computer Assisted Personal Interviewing (CAPI) techniques will be employed for the personal interview. Interviews will be conducted on netbooks programmed with Voxco software and set to upload directly to the contractor’s secure server. Data stored on the computers will be only identified by Random ID number and will be encrypted to protect participant's identities. The CAPI interviews will be audio recorded for transcription of open-ended data using audacity (a free open-source software program). The audio files will be password protected and will not contain any identifying information. These audio files will be programmed for upload to the contractor’s secure server along with the CAPI interview file. Respondents will be informed that their information will be kept private to the extent permitted by law and that their participation is voluntary. They may choose to not answer a question and can move on to the next one.

As mentioned above, initial seed respondents will be provided with a $20 gift card for their interview and will be offered additional gift cards to recruit people from their social networks to participate in the interview. Each seed will be given three coupons (see Attachment C: Seed Coupon) to give to homeless youth they know. Coupons will have unique identifiers that will link them to their seed. Coupons will have an expiration date of 7-14 days to allow for tracking the rate of return and to reduce respondent burden. Initial seed respondents will be provided a $10 gift card for each returned coupon that results in a completed interview. Thus, a respondent who completes the interview and successfully recruits three peers would receive $50 total in gift cards. Each new participant will be offered the same number of coupons and gift cards to recruit respondents from their social networks, up to the fourth seed.

Focus groups will most likely take place in private rooms at the SOP grantee agency in each city. This room will most likely be conference or group therapy rooms already utilized by SOP agency staff for these types of sessions. If all the participating youth request meeting at a different location, the focus group may take place at a different agency within the city. Each agency will utilize one audio recorder (Olympus WS-500 Digital Voice Recorder) to record the focus group sessions (total of 11 audio recorders). Audio recordings will be uploaded directly to the contractor’s server using a secure website that will be developed for this project. Audio recorders will be returned to the contractor once SOP grantees complete the focus groups, and interviews will be deleted from the recorders upon receipt. Respondents will be informed that their information will be kept private to the extent permitted by law and that their participation is voluntary. They may choose to not answer a question and can move on to the next one.

B3. Methods to Maximize Response Rates and Deal with Nonresponse

This is a one time data collection. Response Driven Sampling (RDS; see section B2, above) is being employed and has demonstrated effectiveness in maximizing response rates for homeless youth and adult populations. To account for seeds that never initiate, a third seed will be selected in cities for which half the sample has not been identified seven weeks after the data collection begins.

B4. Tests of Procedures or Methods to be Undertaken

A pilot test will be conducted during the first month of the data collection period at one SOP grantee site in order to test the RDS and data collection methods, and the CAPI technology for the personal interview. The goal of the pilot is to conduct 15 personal interviews and refine the instrument and methods before full implementation of the data collection with all 11 sites.

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

Contractor:

Dr. Leslie Whitbeck, Ph.D.

Department of Sociology

University of Nebraska at Lincoln

Email: lwhitbeck2@unl.edu

Phone: 402-472-5562

6