Strategies for Longitudinal Research with Youth in Foster Care: A Demonstration of Methods, Barriers, and Innovations

Abstract

Longitudinal research on youth in foster care is important, but often challenging to accomplish. To assist the field, a thorough description of the development of the SPARK (Studying Pathways to Adjustment and Resilience in Kids) project, a longitudinal research project on the mechanisms of resilience for foster youth and their caregivers, is presented. Authors explain the difficult task for researchers in accessing youth in foster care and suggest strategies for success. Recruitment approaches for foster youth and their families are also provided along with examples of effective techniques. Data collection concerns are discussed, and the authors provide recommendations for researchers to consider when asking youth sensitive questions. Finally, data collection on academic information from teachers and how the SPARK project works with the academic community to gain information on school functioning for youth in the project is described. Suggestions for methodology utilized in future research along with examples of innovative adjustments to typical research procedures are provided as guidance for how research on maltreated youth can be conducted.

In the United States, the number of youth in foster care reached 408,424 in 2010 (US Department of Health and Human Services (DHHS), 2011). Unfortunately, the number of youth requiring removal from their biological parents and placement with alternative caregivers has not significantly declined over the past decade with just as many youth entering foster care as exiting in any given year (DHHS, 2011). Although clinical science can do little to effect change in the rate at which youth are placed in care, research can provide meaningful insights on the impact of foster placement on a child's well-being. Researchers have understood this priority as evidenced by the multitude of studies devoted to documenting and analyzing the mental health of the foster youth community (Oswald, Heil, & Goldbeck, 2010).

Placement in foster care is not always a temporary outcome (Barber & Delfabbro, 2003); therefore research must be able to assess the long-term impact of both the precursors to care and the additive effect of the out-of-home placement. To this end, the present discussion is intended to provide researchers with a roadmap; an example of how longitudinal research can be conducted on an important but hard to access population. Using the SPARK (Studying Pathways to Resilience and Adjustment in Kids) project as an example, suggestions for recruitment, data collection, retention, and ethical considerations germane to research on youth and families in the foster care community are presented

The SPARK Project

The SPARK project is a federally funded, five-year research project (from 2009-2014), devoted to documenting the process of psychological, educational, and physical outcome for youth ages eight and older who are placed in foster care. Because most measures are created for typically developing children, youth are excluded from the study if they demonstrate an IQ score in the mentally retarded range or if they have a diagnosis of Autism. Youth, who have been in state custody for at least 30 days, and their caregivers (i.e., foster parents, kinship care providers, and residential facility staff), are asked to complete a lengthy set of questions, including the child's history of maltreatment, mental health, social support, and family environment. All questions are collected via an audio computer-assisted self-interview during which the participants respond to questions read aloud by a laptop computer over headphones. Additionally, the child's school and teacher provide academic information and behavioral assessment of the child in the school setting via an online questionnaire. Participants are compensated for their time and asked to complete the study three times at intervals of three months. The project is currently in its third year of data collection.

Accessing the foster care population

Although researchers may be interested in conducting research on youth in foster care, most find accessing youth in state custody difficult (Gilbertson & Barber, 2002). Primarily due to their protected status in every state, the identity, location, and maltreatment history of youth in foster care is confidential. In addition to approval from the standard university review board, in most states, approval for research projects with foster youth requires a two-part process: 1) endorsement of the research proposal by the state social service agency, and 2) approval from the district circuit court. Allowing researchers to contact youth in protected legal custody of the state requires state authorities to weigh the importance of protection of youth and the benefit of the research to the youth and to the field. Researchers should, in addition to detailing the probable positive gains from the project, take particular care to consider and address all possible detrimental ramifications to potential participants. Although perhaps a simple point, consideration of possible negative effects is important as youth in foster care and their caregivers may not feel as able as non-state supervised families to say “no” to research requests, especially those targeted at foster youth.

For example, the SPARK project provided in writing that the identity of any participant in the project was kept confidential. For most research on youth, consent is given by the legal guardian with full awareness that the child will be in a research study. However, for research on foster youth, the legal guardian (i.e., the state social service agency) had to agree to comprehensive permission up front for all foster youth in the county as well as agree not to know which individuals actually participated. Because youth in foster care and their caregivers are monitored by the state, it is imperative to ensure that any participant in the SPARK project understands that agreeing to be in the study is not shared with state authorities. Without such a promise, some foster youth might believe their assent could impact decisions made by the state and could influence their ability to go home to biological parents sooner.

Moreover, for foster parents, researchers had to make clear that declining to be in the study would not be communicated to their case-worker to prevent any misperceptions of participation as an expected part of their fostering responsibilities. In this way, participants in the SPARK project were informed fully and protected from any possible undue influence from the state or the court to either comply or not comply with the research request.

Beyond the protective actions designed to ensure the safety of the participants, additional measures were taken to keep partners of the SPARK project (i.e., social service agencies) abreast of how the research was progressing. For example, before the data collection process began, the director of the SPARK project met regularly, and in person, with the administrators of child welfare for the state to brainstorm possible concerns for data collection. Similarly, the project director met with county circuit court judges (who have final legal authority over the youth in state custody) to discuss how information would be stored, shared, and protected. Successful access to foster youth and their families requires ongoing partnerships between researchers and state custody agency authorities. To support collaboration, the director also established a consistent feedback loop for communication with the administrators of child welfare to provide information on the status of the project and how the agreed upon recruitment and data collection procedures were working. Research is not the mission of most child welfare agencies, which are often overburdened and under-resourced in providing services to youth.

Sensitivity to the needs of the agencies involved is paramount, and researchers are encouraged to learn the culture and values of social service agencies while ensuring that all aspects of the research process are as transparent as possible. In the SPARK project, along with regular update meetings about procedures, state agencies were provided with aggregate results of the project and encouraged to discuss the findings internally to determine how the project data and research questions could benefit the services the agency provides. As the project progressed, the research team communicated with state agencies about how procedures may be improved and how data security is maintained. Relationships with state agencies can be slow to grow, especially in states where research or open access to youth names and case files is unprecedented. Therefore, researchers are encouraged to include the agency's agenda into the project, while concurrently working with stakeholders to infuse the project into the agency's agenda. For example, child welfare agencies have a mandate to monitor placement stability for youth in their care; researchers also have interest in how placement stability may impact mental health. Thus, the SPARK project shares aggregate data findings regarding placement stability and mental health outcomes with the child welfare agency so that this information (albeit, a subsample of the entire foster youth population) may enhance services provided by the state.

Although it is not the only model, the SPARK project offers an example of how to successfully access the fostering community in research. Due to our efforts in building relationships, the SPARK project staff gained permission to contact over 2,700 youth in foster care custody in a specific county. Access, however, is only the first step in the research process. Once access to contact is granted, eligible youth must be recruited.

Recruitment of foster youth and families

Recruitment efforts can be hampered by concerns about the possible stressful nature of answering study questions. This is a particular concern for research on foster youth as some studies have found a non-response rate to requests for participation by foster families as high as 91% (Gilbertson & Barber, 2002). Even when researchers are given contact information for youths’ most recent placement, the transient nature of the foster care population makes tracking interested children difficult (Berrick, Frasch, & Fox, 2000). Moreover, the time-consuming licensing requirements placed on foster parents (e.g., attending trainings, meeting with case workers) make it difficult for those who may be interested in research to find the time to answer study questions. To maximize participation rates, the SPARK project utilized multiple methods to recruit participants.

Approaches to recruitment with foster families

Like efforts to gain approval for research, recruitment efforts are most effective when the research staff form relationships with foster parents and caregivers. Although the SPARK project had permission to access all foster youth in a given county, the project staff was sensitive to the wariness some foster families may have to requests for their time and their foster child's time, especially from research projects or agencies unaffiliated with social services. Moreover, current contact information that is obtained by the SPARK recruitment team on foster youth was and continues to be quickly outdated and incorrect, making exclusive use of cold calls or mailings inefficient. Recruitment to the project meant making personal contact with caregivers (i.e., foster parent, residential staff) when possible and providing comprehensive information about the project. SPARK project information was distributed through personal approaches for recruitment, which involved finding and attending events specific to the fostering community to make face-to-face contact. For example, when the social service agency held an event to educate and support older foster youth in their vocational goals, the research team was invited to talk to the youth about SPARK. Project staff also met potential caregiver participants at foster parent training events and support groups, and attended case-worker staff meetings. Whenever possible, the SPARK project staff explained the research in person, and if a potential participant indicated interest, a SPARK staff member called the family or residential staff administrator to determine if the youth was eligible.

Although the personal approach was effective in recruitment, less personal, but larger-scale recruitment efforts were also employed, including the dissemination of over 11,000 flyers in foster care organization newsletters and via direct mailings, foster parent listserv advertisements, over 75 cold-calls to all foster families in a given county, and local news and radio broadcasts about the project. The desire was to be as thorough as possible in informing the fostering community about the SPARK project. While the latter approaches were less personal, they were important for ensuring that all foster parents, especially those who might not attend trainings or belong to the listserv, were aware of the project. Information provided to potential participants via all recruitment efforts included introduction of the project, explanation of the rationale and expected benefits, description of eligibility requirements, description of the process and incentives, and answering questions or concerns. If eligible, the foster family or residential facility was then scheduled for data collection.

Approaches to recruitment with residential facilities

To recruit children in residential facilities, who thus far comprise at least half of the participants in the SPARK project, the principal investigator and other SPARK personnel arranged informational meetings with directors and staff regarding the potential benefits of participating in SPARK. Following a time of relationship building with staff at several residential centers, every residential center that was targeted agreed to allow data collection to occur on their campus, with staff reporting on children with whom they frequently worked. Following agreement from the facility, recruitment for specific youth was conducted primarily by an administrator at each location determining who at their agency was eligible for the study and the staff asking the youth if they were interested in participating.

Outcome of recruitment approaches

The project is currently in its third year of data collection and recruitment, and thus far, about 63% of eligible children contacted directly about the project have indicated an interest and enrolled in the project. To provide a sense of the result of the myriad of recruitment approaches, Table 1 offers a summary of the recruitment for a subsample of foster home participants (n=97). Because the project includes only youth who are age eight and older, many of the recruitment approaches reached interested, but ineligible foster families (e.g., parents of toddlers).

Table 1.

Outcomes of Foster Home Recruitment Strategies

Recruitment Strategy (number of times completed)	Interested Participants	Eligible Participants	Enrolled
Attending Trainings (9)	54	28	24
Mailings (5)	80	40	27
E-mail Alerts (7)	15	9	5
Attending Special Events (7)	17	11	7
Referrals from Social Workers	110	64	34

A contributor to the success in recruiting foster families to the study was the creation of an identity for the project. The SPARK project has its own logo placed on all recruitment flyers and give-aways, such as pens, magnets, cups, t-shirts, and water bottles. Also, the SPARK project has a website (www.sparkproject.ku.edu) where interested families could and continue to find out more about the project and indicate interest in participation via secured email.

Data collection

For longitudinal research, ensuring positive data collection experiences at one time point are central to prevention of attrition for the next time point (Stouthamer-Loeber, van Kammen, & Loeber, 1992). For the SPARK project, data collection was designed to be as convenient and pleasant as possible. For example, all appointments took place at a community location a few miles from the foster family (e.g., library or community recreation center) or on-site for those youth who lived in a residential facility. Reminder calls to parents and residential staff were made prior to the data collection date to reduce missed appointments. The staff also provided childcare for any other non-eligible children in the family to make it easier for families to attend data collection sessions.

For youth at each time point, research assistants read aloud from an assent document outlining the purpose of, the confidential and voluntary nature of, the risks and benefits of, and the incentive for participating in the survey. Any questions the child may have are encouraged and answered before starting the survey. Presenting the assent verbally to the youth is beneficial in that it gives the research assistant an opportunity to gauge whether or not the foster child understands the purposes of the project. For adult caregivers at each time point, prior to signing a written consent form, research assistants read aloud from the consent form explaining the purpose, confidentiality, risks, and benefits of being in the study. The conversational style of consenting also provides more opportunity for questions regarding the research process, ensuring that participants truly have consented and assented in an informed manner.

One possible barrier faced by the SPARK project was the utilization of a long survey (i.e., 1,000 questions) for participants to complete, taking approximately three hours. To make the data collection process as enjoyable as possible, the SPARK project staff provided the youth and families with snacks, games, and a flexible schedule (i.e., evenings, weekends) for data collection appointments. Participants also had the option to complete the data collection over multiple sessions making it easier for the youth to maintain focus on the survey questions. The SPARK project research assistants assessed participants’ mood prior to and immediately after data collection to screen for any possible distress or fatigue the participant might have felt after answering a lot of sensitive questions. Analysis of mood ratings suggested no significant changes from pre- to post-survey administration despite the long length of the survey (Jackson, Makanui, Beals-Erickson, Sughrue, Tang, & Queenan, 2011). Moreover, knowing how long the survey is at Time 1, out of 210 surveys completed thus far, only five youth refused to return for another data collection appointment. Although it is likely that some youth are eager to return to the study for the compensation, youth also report to project staff that they like sharing their story with the computer and enjoy the attention they receive during data collection sessions.

Because the SPARK project involved self-report of previous or current maltreatment, particular care was taken to ensure that answering questions about potentially traumatic events was managed in the best way possible. Data collectors for the SPARK project are clinical child psychology graduate students, and all youth in the project were monitored closely when answering study questions. Based on standard practices for ethical research (APA, 2002) and the desires of the state agency, the SPARK project developed extensive HIPPA protected data storage procedures for both paper and digital data and a procedure for reporting any abuse events discovered during the process of data collection. Additionally, research with foster youth may require more than the typical debriefing procedure after data collection. For example, although all of the measures used in the SPARK project have minimal risk to youth, a three-part debriefing process was conducted for all participants in the study. Research assistants meet with the child alone, the caregiver alone, and the child and caregiver together to confirm all participants felt reasonably well after the data collection process. SPARK project staff also called every participant within 48 hours of the data collection session to inquire about their well-being. These extra steps were included to ensure participants in the study experienced little to no distress from being a participant, and that any distress that arose was addressed.

Accessing case file data

In addition to self-report information collected on laptops, the SPARK project also collects maltreatment history information from the child's official record with the state social service agency. Although other projects on youth in foster care also have included data from official case files (Knight et al., 2006), lessons learned in the SPARK project offer a few suggestions for future research.

Because some of the case file information required access to the state's secured computer system, a SPARK project case file liaison was hired to access and code the case files located at the social service agency. The SPARK project liaison also redacted identifying information from the file (e.g., social security numbers, identity of person making the hotline call) before it was given to SPARK project staff. Once the file was redacted, the unredacted copy was destroyed and the redacted version was coded. The liaison coded the relevant information for each hotline call (e.g. date of event, age of child) and sent the coding sheet with the redacted file to the SPARK project staff via secured courier for reliability coding. Each report of alleged child abuse and/or neglect received a severity code using the Modified Maltreatment Classification System (MMCS; English, Bangdiwala, & Runyan, 2005). The MMCS code reflects the frequency and the type of incident reported. The conclusion of the allegation (i.e., whether the allegation was substantiated or unsubstantiated) was also coded. Before sending the redacted paper file and coding information to the SPARK office, as a final step, the liaison entered the coded information into the Research Electronic Data Capture Program (REDCap; Harris et al., 2009). REDCap is a secure, web-based program used to create and manage databases. The file-coding liaison utilized an encrypted laptop, provided by the SPARK project, to enter the child's information into REDCap. Data entered into REDCap is stored on a HIPPA-protected server. Once codes from the file coding liaison were entered into REDCap, two graduate research assistants coded the same files and entered the information into REDCap to ensure reliable coding of case file information.

While file coding presents an opportunity for more comprehensive data on maltreatment, several barriers existed. For example, hotline calls were not always documented in the case file in a consistent way (e.g., some information in the narrative, some information in reports from the court). Hiring a social service agency staff member familiar with case file documentation was instrumental to developing a consistent method for organizing the information in the case files. The liaison also met with the SPARK staff responsible for reliability on a monthly basis to ensure that the information in the file was coded in a uniform manner across coders.

In addition to providing evidence on the relation between maltreatment and adjustment, the SPARK project also sought to understand how youth in foster care fare academically. Given the research on the associations between maltreatment and academic performance (Kaplan, Pelcovitz, & Labruna, 1999) as well as academic outcomes for children in foster care (Geenan & Powers, 2006; Trout, Hagaman, Casey, Reid, & Epstein, 2008), the academic functioning of children enrolled in the SPARK project is relevant to the proposed aims of the project. Teachers contribute unique information to the project by supplying data on the children's grades, school attrition, and peer interaction.

Teacher Data Collection

As families enrolled in the study, they were asked to identify a teacher who best knew the participants’ school performance. Because teachers are outside of the foster care system, the process of collecting data from teachers of youth in foster care created some unique challenges. One, foster youth across a large county often attend a number of different schools. Youth in the SPARK project, for example, attend 81 different schools, representing 14 different schools districts. Two, teachers are busy, and in general, do not have much time to complete survey questions. Three, school personnel are often sensitive to any research project that might take time away from the work day, especially projects whose primary mission is not to advance learning and education. Despite the fact that the research project is not technically an internal, school-based research project, some school districts still required that the project be reviewed and approved prior to teacher recruitment. To manage these and other concerns, the SPARK project implemented several approaches to teacher data collection.

The SPARK project staff coordinated with each school district represented by youth participants to determine the best way to identify the correct email address of the teacher. Information on school websites providing teacher email addresses were not always up to date, did not always provide email addresses, and, at times, published incorrect teacher email addresses. Because teachers were informed via an online survey the name of the child on which they were to report, it was important to verify email addresses with school personnel to ensure confidentiality of child participants. Although the procedures for the project were approved by the state, schools and teachers may also require proof of consent by the state for a teacher to answer academic-related questions on a child. For the SPARK project, an academic release provided by the state social service agency was forwarded to each district prior to teacher administration of the study measures. The SPARK project staff also contacted the principals of each school as well as school district administrators to share information about the project and to get approval for the study procedures. Although teachers and school administrators were generally willing, it was necessary to ask for permission from special services directors (i.e., those who coordinate special education) as well as each principal and teacher. The project had to employ a bottom-up (i.e., asking teachers first) and a top-down (i.e., meeting with and getting permission from administrators first) approach to request permission for recruitment of teacher reporters. Because teachers often would not complete the study measures without permission from their school principal, administrators were encouraged to share their approval for the project with the teachers in their school.

Researchers are encouraged to be persistent and diligent in their efforts to work with school leaders to make participation in research as convenient and mutually beneficial as possible. For example, the SPARK project staff shared the aggregate results of the academic findings with participating schools to provide them with information about how youth in foster care are benefitting from the academic environment. District administrators indicated interest in the utility of these findings for their schools and academic services, and this collaborative approach to the sharing of data opened doors for better communication with school personnel.

Once schools and teachers were comfortable that their information was approved to share with the SPARK project, teacher data collection proceeded through the following process. First, each child's teacher was invited via email to voluntarily participate in the study using a secure, online survey. By allowing teachers to complete the study questions online, the SPARK project enabled teachers, who rarely have spare time during the day, to participate in the study at any time that was convenient. If a teacher did not complete the survey within a week, SPARK staff called their school to leave them a reminder message about the survey. Once the teacher completed the survey, a gift card was mailed to the teacher.

Each teacher was contacted again about every three months to complete the survey for a second and third time. Given the longitudinal nature of the SPARK project methodology, good communication, successful maintenance of relationships, and tracking of participant movements remain paramount to follow up assessments. The following describes efforts made by the SPARK project to follow participants over time.

Re-recruitment efforts

By definition, youth in foster care are meant to be in their foster placement for as short a period of time as is possible. When research questions are cross-sectional, the transitional nature of foster youth is fairly unnoticeable; for longitudinal projects, however, the movement of youth to new foster homes or returning back to their biological home can be a challenge to retention. Success in retention depends on many factors, not the least of which may be the participants’ memory of what it was like to be in the study at the previous time point. For some of the follow up time two and time three surveys, it is sometimes the case, albeit rarely, when the child is no longer in state custody due to placement with a caretaker who has legal rights to the child. If the child is no longer in state custody during the course of the SPARK project, the recruitment staff contacts the legal guardian and describes the study to the potential participant. If the legal guardian indicates interest in participating (and for their child to continue participating), a data collection appointment is scheduled and the new legal guardian provides consent for participation. If the legal guardian indicates no interest in the study, or if the SPARK recruitment team is unable to make contact with an individual at the child's new placement, the child is no longer considered an active participant in the SPARK project.

Although attrition is common to all longitudinal research, studies that include youth in foster care are especially likely to be challenged to find and re-recruit participants for additional time points (Bulat, 2009). All families are informed at intake that the project requires meeting for data collection at three time points, and the SPARK project has been successful at retaining participants across time. Specifically, thus far, out of 210 participants at Time 1, 149 have completed a Time 2 survey, and 100 have completed a Time 3 survey. The SPARK project is currently ongoing; therefore it is not possible to know the true study attrition rate until recruitment and data collection has ended. However, an estimate of the attrition rate thus far (including the 25 active participants currently awaiting their second and/or third survey time points surveys lost) suggests the SPARK project attrition rate would be 68 participants or 31%. This is likely an overestimation of our attrition rate given the number of participants still active and the changes project staff have implemented and described in this paper to maintain contact with participants.

An analysis comparing participants who remained in the study to complete all time points to those who did not revealed no statistically significant differences in demographic variables (e.g., age, placement type, ethnicity) across groups. Table 2 provides information on the reasons that youth dropped out of the study. The most common reasons were a child moving to an unknown or distant location and a lack of timely response to schedule an appointment by the foster or residential placement in time for the next data collection session. Other reasons for attrition included the child running away from his or her placement and failure to schedule in time due to other circumstances. Despite the expected attrition of some participants, the SPARK project has been relatively successful at maintaining participants across time points; a review of the procedures used provides a few suggestions for successful re-recruitment.

Table 2.

Reasons for Attrition (N=95)

Reason	Percentage of participants
Case worker requested the child not continue	1.1
Child no longer eligible	5.3
Family no longer interested	7.4
Child ran away from home	9.5
Child not scheduled in time	15.8
Child's contact person unresponsive to requests	27.4
Child moved to unknown location or too far away	33.7

One, regular meetings with stakeholders, child welfare agencies, residential facilities, and foster parent support groups ensured that the project stayed in the recent memory of those who provide permission and access to foster youth. Two, SPARK project staff worked with social service agencies to track youth who move from their previous residence. Because the project staff maintained good relationships with foster parents, case-workers and agency staff, getting updated contact information was often successful. Thus far, 48% of youth in the study have moved placements, however, strategies of tracking participants in this project have proven effective in maintaining contact with participants despite regular placement changes. The SPARK project staff works to build relationships with families to keep them involved in the project for the long-term; efforts such as individual hand-written thank you notes and reminder cards about the project to current participants keep participation in a large-scale project personal.

Discussion

Understanding the mental health and adjustment of youth is important; perhaps most important is accessing youth in all the contexts in which they live, even when those contexts, like foster care, make the task challenging. The conduct of this type of inquiry in the context of foster care requires flexible and responsive methodology, and the SPARK research team has successfully created, administered, and conducted a program of research that investigates one of the hardest populations to study. It is our hope that other researchers will be emboldened by our example and the example of other successful projects and include youth in foster care into future research studies.

Because youth in foster care can be hard to access, researchers may find that adversity fosters innovation. This was certainly true for the SPARK project as we found many obstacles that resulted in creative solutions designed with the help of our social service agency partners. To ease the process of participants answering sensitive questions, we created a computerized data collection system that was designed to make answering questions as comfortable and as safe as possible for youth. One of the innovations perhaps particularly unique to data collection with youth in foster care is the protection of the data. All research must keep the identity and the responses of participants’ private; however, research with data from case files must also ensure that the storage of the data and internal access to the data by researchers is regularly monitored. Efforts to keep the data safe demonstrate both to the social service authorities and to the participants the great respect the team has for the opportunity to conduct research with foster youth. Because the SPARK project intended to capture maltreatment from case file data, regular meetings with social service agency members were initiated so that information could be easily shared and safety of the data could be assured.

The SPARK project staff also discovered innovative ways to include teacher data into the project. Meetings with district personnel and special services directors informed the process of teacher data collection as well as provided important contacts for garnering information such as teacher email addresses. Similar to methods of bridge building designed to foster collaboration with child protection agencies done for needed permissions early on in the project, strategies for working with school districts in a way that was mutually beneficial provided more opportunities for successful academic data collection.

Perhaps one of the most important improvements made to the project protocol involved our process for tracking participants. By forming relationships with agency staff, the SPARK project staff is able to access the most recent placement and contact information for each child in the study. Regular placement changes of study participants makes the need for re-recruitment strategies apparent and the importance of having a working partnership with the state social service agency cannot be overstated. The SPARK project methodology serves as one successful example, but the procedures continue to evolve with the constant demand for following a very transient and hard-to-access population.

The SPARK project is currently in its third year of data collection with over 200 youth in foster care enrolled in the study. The project has built a positive brand in the fostering community evidenced by the common reports from foster parents that they heard positive things about the project before participating. The SPARK project also finds that many of the foster parents who have completed their time in the project are eager to try it again and contact the project staff when a new foster child comes to their home. Word-of-mouth recruitment is powerful, and it provides an indication that methodologies designed to promote the project are working. The SPARK staff work to ensure that foster parents and the social service agencies involved in foster care have a positive experience, because for most, it is their first time participating in a research project. We certainly hope it is not their last.