Hidden in Plain View

Abstract

This article elaborates upon a model used to engage marginalized young people in a longitudinal study of youth transitions. The model PARTH elaborates upon a set of principles that were successfully used to engage marginalized young people in a 6-year project. PARTH principles focus researcher attention on the ways they think about and relate to young people and they support researchers to empowerment and the exercise of personal agency by young people. The model was instrumental in achieving high retention rates that exceeded 89% across the 6-year study.

What is already known?

The challenges involved in recruiting and retaining young people in research in general and the challenges of recruiting and retaining marginalized young people in research in particular are well recognized. Widely diverging retention rates have been reported with rates dropping to less than 30% when the focus is on marginalized youth. Such low rates have serious implications for data quality in both qualitative and quantitative research. Marginalized young people have been reported as difficult to locate, as comprising a high proportion of attrition rates, or constituting outliers and therefore present in insufficient numbers to support analysis that faithfully reflects their lived experiences. They have been reported to have levels of involvement in social and other service systems, to lack personal and financial resources enjoyed by other youth with more normative experiences, to need to manage atypically high levels of risks in the individual, familial, educational, neighborhood, and societal domains, and thus can be expected to face significant challenges in making safe transitions to adulthood. They are also more likely than other youth to face enduring challenges in educational attainment and maintaining secure and stable employment. Their absence from longitudinal accounts therefore leaves large gaps in our understanding of how young people navigate pathways to adulthood when they have been exposed to large amounts of risk and live fragile lives at the societal margins. Researchers are increasingly giving consideration to how to successfully recruit and retain these groups of youth and, in this regard, relational practices have been identified as helpful for forming meaningful research relationships with young people that encourage their active participation. The current article builds on this emerging interest. It elaborates on five principles that guided sampling, recruitment, and retention efforts and that maintained the active participation of young people in a national longitudinal, mixed methods study of marginalized young people’s transitions.

What this paper adds?

The five principles developed in the research and discussed in this article provide a new perspective on ways of structuring research projects so as to maximize the chances that sufficient numbers of marginalized young people will be recruited and retained in research. The principles focus researcher attention on how to create facilitative environments that encourage the active engagement of young people as research partners. They provide valuable guidance for the ethical management of such research and for working effectively with stakeholder groups in the recruitment and retention process. Importantly, this article identifies that “researcher factors” are more important than “young person factors” in successful recruitment and retention and it offers guidance regarding how to effectively manage these researcher factors.

Introduction

As the field of youth transitions research grows, it is increasingly recognized that research must take account of the heterogeneous nature of the population of young people and the diverse pathways through adolescence that young people take. This means that to be valid and meaningful research needs to be able to successfully identify, recruit, and retain diverse groups of youth in research projects (Abrams, 2010; Arnett, 2000; Berzin, 2010; Bonevski et al., 2014; Borek, Allison, & Cáceres, 2010; Hanna, Scott, & Schmidt, 2014; Hardgrove, McDowell, & Rootham, 2015; Hooven, Walsh, Willgerodt, & Salazar, 2011; Shahabi et al., 2011; Ward & Henderson, 2003; Wright, Burgos, Krygsman, & Brown, 2015). However, across the spectrum of academic disciplines, researchers have reported difficulties in recruiting and retaining subpopulations of young people variously described as “hidden” (Bonevski, et al., 2014; Vangeepuram, Townsend, Arniella, Goytia, & Horowitz, 2016), “hard to reach” (Abrams, 2010), “at risk” (Borek et al., 2010), “minority” (Shahabi et al., 2011), and “vulnerable” (Taylor, 2009; Ward & Henderson, 2003). While each of these terms references a slightly different subgroup of young people, what the labels share in common is the difficulty researchers have reported in trying to secure a sufficient number and range of these youth to create valid and trustworthy research. This has meant that, to date, it has been difficult for researchers to consistently ensure the meaningful participation of these youth and as a result, research has not been able to accurately reflect their diverse realities. In the current article, we refer to these young people as “marginalized” because they often exist at the margins of research, have been reported as difficult to locate, as comprising a high proportion of attrition rates, or constituting outliers and therefore present in insufficient numbers to support analysis that faithfully reflects their lived experiences (Abrams, 2010).

Taylor (2009) explains that while marginalized youth may represent between 5% and 7% of the general population of young people, they often have high levels of involvement in social and other service systems, typically lack personal and financial resources enjoyed by other youth with more normative experiences, manage atypically high levels of risks in the individual, familial, educational, neighborhood, and societal domains, and thus can be expected to face significant challenges in making safe transitions to adulthood. They are also more likely than other youth to face enduring challenges in educational attainment and maintaining secure and stable employment. All of these characteristics make this diverse subpopulation of youth of great interest to researchers and to the policy and service delivery communities. Their absence from longitudinal accounts therefore leaves large gaps in our understanding of how young people navigate pathways to adulthood when they have been exposed to large amounts of risk and live fragile lives at the margins. As a result, the importance of retaining the active engagement of these young people in longitudinal studies is increasingly occupying researcher attention.

The need to secure adequate representation of these subpopulations of youth raises a range of methodological issues, and increasingly, researchers have turned their attention to developing techniques for sampling and recruiting young people and then retaining their active participation in research. Some have suggested that random and probability sampling techniques are unlikely to generate sufficient numbers of marginalized youth to allow valid conclusions to be drawn and therefore that attention needs to shift beyond traditional research techniques to more diverse and responsive strategies that are specifically tailored to these subpopulations (Bonevski et al., 2014; Bull, Levine, Schmiege, & Santelli, 2013; Gwin Mitchell et al., 2015). Rather than undermining the validity of samples, it is argued that such approaches add value to representativeness enhancing the rigor and quality of research findings in the process (Bonevski et al., 2014; Borek et al., 2010). Common themes are emerging in the literature regarding how to successfully draw marginalized young people into longitudinal research including the central role of relational practices in forming meaningful research relationships with young people that encourage their active participation. The current article builds on this emerging interest. It elaborates on five principles that guided sampling, recruitment, and retention and maintained the active participation of young people in a national longitudinal, mixed-methods study of marginalized young people’s transitions.

The study

The purpose of this study was to identify which factors contributed to successful transitions to adulthood for young people who were exposed to high levels of chronic, enduring risks throughout their childhoods. As noted above, these young people are referred to in this article as marginalized youth. Traditionally, transitions research has focused examining the ways in which youth on relatively normative developmental pathways go about building an emerging adult identity and the factors that assist them to do this successfully (see, e.g., Blinn-Pike, Lokken Worthy, Jonkman, & Rush Smith, 2008; Riggs & Han, 2009). On the other hand, research on marginalized youth has focused upon the ways in which risks influence outcomes (see, e.g., Berzin, 2010; Mechanic & Tanner, 2007; Thomson & Holland, 2003). The current study brings these two research strands together by mapping over time the sense marginalized young people make of the processes of growing up as they balance the risks and resources around them. This mixed-methods study had retention rates that equaled or exceeded 89% across all six phases. It included structured surveys administered 3 times at approximately annual interviews (Time 1: n = 593; Time 2: n = 528; Time 3: n = 506) followed by semistructured qualitative interviews with a subsample at approximately annual intervals for a further 3 years (Time 4: n = 107; Time 5: n = 102; Time 6: n = 104).

The research was submitted to and approved by the University IRB prior to field work commencing. From the outset, it was clear that sampling, recruiting, and retaining this number of marginalized young people over the course of the study were likely to require careful attention and effort. Like all research, there were three key sets of methodological challenges that needed to be resolved to successfully complete this project (Abrams, 2010; Bonevski et al., 2014; Borek et al., 2010; Hanna et al., 2014): A sampling framework that gave confidence that a representative group of youth could be identified, recruitment strategies that minimized the likelihood of refusals while nonetheless respecting young people’s rights to decline, and tracking strategies that encouraged active engagement of young people to maximize retention rates and minimize the chances that any losses of participants would introduce biases into the data set.

Five principles collectively summarized as PARTH (see Figure 1) were used to resolve these methodological challenges. The PARTH principles focused researcher attention on creating facilitative environments which encouraged the active engagement of young people as research partners. Active engagement is particularly important in longitudinal research because of the significant ongoing commitment required of participants over time. The PARTH principles also played an important role in preparing researchers for the field in areas such as orientation to the research and ways of securing active participation across the study.

Figure 1.

The PARTH principles.

The following discussion explores the way these PARTH principles helped resolve the three methodological challenges: sampling, recruitment, and retention, drawing on both the extant literature and experiences in the study.

Sampling

Sampling frameworks that identify a sufficient pool of potential relevant participants are the cornerstone of successful research (Abrams, 2010). Schools are often used as convenience sampling frameworks in youth studies; however, given that many marginalized young people are not participating in mainstream education environments, schools will not always contain a sufficient number and range of these young people (Auerbach; 2010; Hooven et al., 2011).Traditional techniques such as random or probability sampling are also unlikely to generate sufficient numbers of marginalized young people creating issues for validity, bias in estimates, and loss of statistical power in quantitative research (Bonevski et al., 2014; Borek et al., 2010; Hanna et al., 2014; Hooven et al., 2011; Wright et al., 2015) and threatening the integrity and trustworthiness of qualitative research (Abrams, 2010; Lietz, Langer, & Furman, 2006; Taylor, 2009). For these reasons, it has been argued nontraditional sampling techniques are more likely to generate effective sampling frameworks for both qualitative and quantitative research (Abrams, 2010; Bonevski et al., 2014). Techniques such as snowball sampling and neighborhood-based recruitment have been reported as successful in some situations (Abrams, 2010; Bonevski et al., 2014; Hooven et al., 2011). In other situations, new technologies, such as social networking sites (SNS), have been used successfully to build effective sampling frames (Bull et al., 2013; Gwin Mitchell et al., 2015). Collaborations with service provider organizations have also been found to facilitate sampling (Bonevski et al., 2014; Leonard et al., 2003). This latter approach requires an investment of time at the outset to allow organizations to satisfy themselves about the relevance of the research and the trustworthiness of the researchers; and to negotiate access, ethics, and consent arrangements and recruitment protocols (Bonevski et al., 2014; Leonard et al., 2003; Ward & Henderson, 2003).

Using PARTH to Create a Sampling Framework

The study faced similar challenges to those reported elsewhere in creating an effective sampling framework. The relational, reciprocal, and adaptability PARTH principles were particularly important. Our focus was upon young people who were either already excluded or disengaged from school or who were at risk of not completing school. Accordingly, schools did not contain the mix of youth we were seeking; and although service involvement was one characteristic we sought in young people, there was no single database that allowed us to draw a national random sample. Rather, records were distributed across a range of administrative databases, some of which were electronic and some of which were paper based, and these databases were not harmonized with each other. In order to create a framework which we could be confident would generate a reasonably comprehensive list of potential participants, we adopted a community-saturation sampling strategy based around all the organizations that worked with young people in each of five geographic locations. The community-saturation sampling approach involved negotiating access to client lists and working sequentially through each list to identify the names of young people who met our criteria and using staff knowledge where necessary to confirm the presence of these criteria. We continued to search organizational lists in each area until no new names were generated. This locality-based saturation recruitment strategy generated a comprehensive list of all youth in each area who met the selection criteria.

The success of this strategy relied heavily on the strength of relationships between researchers and youth organizations, something others have also noted is critical to the success of research that seeks to meaningfully engage marginalized groups of youth (Leonard et al., 2003). Indeed, in the current study, active relationships with provider organizations became a central feature of the research. To achieve this, the lead researchers invested significant amounts of time at the beginning of the research, building a collaborative relationship with the target organizations. In addition to providing opportunities for them to contribute to the design of the study, researchers supported their activities by assisting with service evaluations, supporting preparation of funding applications, providing advice regarding the development of policies and practice guidelines, and offering professional development activities. This reciprocity paid dividends during the fieldwork when organizational staff responded by assisting with locating youth and providing space for interviews. As a result of this relationship building, youth organizations were strongly supportive of the research. This positive orientation was invaluable throughout the research from securing access to a comprehensive list of potential young people, during recruitment and tracking of youth and right through into the translation of research findings into practice. These organizations were also able to provide support to interviewers and opportunities for interviewers to reflect on research processes and the challenges faced by marginalized young people; these discussions provided an important context for the research.

While the community-saturation approach worked well, as others have noted, sampling through organizations is not without its challenges (Ward & Henderson, 2003). The importance of organizational buy in to research is well recognized (Aaltonen, 2016; Ward & Henderson, 2003) and promises of support will not always translate into the list of names required to generate a sample. Ward and Henderson (2003) reported on factors that militated against successful sampling through organizations such as heavy workloads, high staff turnover, and also the presence of deliberately unhelpful staff. In the current study, while the research was received very positively by both statutory and Non-governmental organizations (NGOs), this enthusiasm did not always translate well into assistance with creating the sampling framework. For instance, a large, national child welfare agency supported the research and allocated a liaison person, a senior manager, in each of the sampling sites to facilitate the sampling and recruitment process. However, despite many meetings to fine-tune logistics, presentations to local staff about the research, and ongoing expressions of support from all staff involved, this positive energy never translated into the needed lists of names, and attention moved on to other locally based organizations in each area. At the other end of the organizational spectrum, a small, locally based organization was similarly enthusiastic about the research and keen to have a role; however, internal management issues caused ongoing disruption among staff that repeatedly diverted their attention away from the research. Again, researchers moved their attention to other organizations, while still maintaining positive relationships with this organization.

The community-saturation approach was invaluable here, as at no point was the research wholly dependent upon any single organization to deliver the required names. This meant that researchers could step back from some organizations without compromising the overall integrity of the sampling process and without damaging relationships, because young people who met our criteria were likely to be the clients of more than one agency. Strong collaborative relationships with numerous youth organizations were a priority in this project. It was important to allow time for relationship building and for working patiently with each organization to adapt our general sampling methodology to their particular needs. This approach meant that the research would not be unduly disruptive of their daily workflows and allowed time for us to adapt the recruitment approach to their ways of working.

Recruitment and Ethical Protocols

Even an excellent sampling framework does not guarantee that sufficient youth will be successfully recruited (Vangeepuram et al., 2016). The manner in which initial recruitment approaches are made is very important. This calls for a respectful orientation on the part of researchers, willingness to adapt to changing circumstances, and the capacity to wait for the right time (Urry, Sanders & Munford, 2014). While it has been recently suggested that relatively little is known about how to successfully recruit youth (Vangeepuram et al., 2016), others report on successful methodologies and draw attention to practical issues such as the need to decide how initial approaches will be made to young people, who will make this approach, and ethical issues such as the process by which consent will be secured (Bonevski et al., 2014).

Researchers have suggested that successful recruitment strategies involve plans that are able to be adapted in order to respond to the characteristics of the target population (Bonevski et al., 2014). This means that uniformity of recruitment methods is less important than a recruitment plan that will ensure the participation of sufficient numbers of relevant individuals can be secured. In some situations, this will require formative investigation, so that recruitment can be structured around the daily realities of potential participants. It appears that direct approaches by researchers to young people are less likely to result in positive responses than mediated or brokered relationships where researchers engage a third party to make an initial introduction (Aaltonen, 2016; Abrams, 2010; Thomson & Holland, 2003; Vangeepuram et al., 2016; Ward & Henderson, 2003). For instance, Bonevski and colleagues (2014) suggest that organization-based recruitment is effective because potential participants are introduced to researchers by practitioners who already have trust-based relationships with the young people. They do warn, however, of risks to validity if gatekeeper selection bias is not controlled. It has been suggested that the involvement of community members in recruitment processes, preferably as paid members of research teams, is a critical facet of successful recruitment, and further that active development and maintenance of positive relationships with a range of stakeholders throughout the research will support effective recruitment (Resnicow et al., 2001). Gatekeepers have a powerful impact on the success of research (Aaltonen, 2016; Bonevski et al., 2014; Hanna et al., 2014; Leonard et al., 2003; Thomson & Holland, 2003; Ward & Henderson, 2003), and therefore, investing time and effort in building genuine, positive relationships and maintaining contact with key individuals throughout the research process is critical to success. Researchers therefore need good interpersonal skills and the ability to maintain respectful relationships with stakeholders even when this is difficult. Research leaders also need to ensure that they provide adequate support to interviewers and opportunities to reflect on research processes and solve any issues in a timely and respectful manner.

Based on the literature, successful recruitment approaches have two key characteristics (Aaltonen, 2016; Bonevski et al., 2014; Bull et al., 2013; Vangeepuram et al., 2016). First, a good introduction to the research and the researcher is important. A third party who can introduce researchers to young people helps deal with privacy and confidentiality issues and it also assists with establishing the credibility and trustworthiness of the researcher and the research. Second, face-to-face approaches work the best. As early as possible in the recruitment process, researchers need to be able to meet and directly talk to young people. At this point, researchers need to be able to communicate passion for the research and a positive, enthusiastic orientation to young people (Vangeepuram et al., 2016).

Early in their development, it was thought that new technologies such as SNS would become valuable recruitment tools because they offered a degree of separation between researcher and participant that was not mediated by other individuals. It was also argued that because young people have high rates of engagement with these technologies that they would respond positively to recruitment approaches by researchers (Bull et al., 2013; Gwin Mitchell et al., 2015). However, the latest evidence suggests that while SNS and related technologies may assist in establishing sampling frames by helping to identify pools of potential participants, these technologies are not always productive or efficient at recruiting young people (Bull et al., 2013; Gwin Mitchell et al., 2015). Therefore, such approaches need to be combined with a range of other strategies in order to ensure that researchers are reaching all potential research participants.

There are clear tensions in recruitment processes that pull researchers in different directions. Researchers need to balance the need to recruit sufficient participants to make their research viable while at the same time not coercing reluctant individuals to agree to become involved. Some suggest that an initial refusal should not be taken as a definitive rejection (see, e.g., the “two nos” rule, Farrall et al., 2015 cited in Aaltonen, 2016); however, others argue that researchers need to take care to avoid the appearance of “grooming” or otherwise placing undue pressure on individuals to participate (Aaltonen, 2016). The issue is more complicated than this however, because while participants may express a clear view regarding involvement in research at one point in time, this does not mean that their views will not change (Author, 2014b).

Consent issues are particularly acute in research with marginalized young people because age and power-related dimensions can distort the research relationship (Aaltonen, 2016). While age and stage of development is an important issue, equally it is important to take every opportunity to encourage active and direct involvement of young people in research. IRBs are likely to place clear conditions on the ways in which young people may be recruited, but consent, particularly in longitudinal research, is an ongoing negotiated process, not a single-point administrative event (Aaltonen, 2016). Thomson and Holland (2003) remind us that consent is always conditional, particularly in longitudinal studies; it can be withheld at one point and then given at some point in the future, as circumstances change. The tensions between procedural ethics and ethics in practice are particularly acute in longitudinal research, and ethics processes do not always successfully accommodate these fluctuating types of circumstances (Aaltonen, 2016). IRBs, however, are likely to want to be assured that full account will be taken of the issues youth vulnerability raises for the giving of informed consent and they will also want to be assured that researchers actively take account of the impact the age of youth has on their capacity to understand what they are agreeing to and to freely choose to be involved.

Using PARTH to Recruit Participants

In the current study, based on well-evidenced arguments presented by the researchers, the IRB waived its usual requirement for securing parent/guardian consent prior to approaching young people. The IRB recognized that this group of young people was a specific population whose families may have constituted risks for them and whose parents/guardians may also have had reasons of their own for denying the involvement of their teens in the research. The IRB recognized that such gatekeeping would potentially skew the sample and, further, that parental gatekeeping could deny these young people the opportunity for their voices to be heard by participating in research. When young people are able to make the initial decision to participate themselves, there are benefits for sampling and data quality, and being able to decide for themselves also supports the active agency of young people contributing in the process to their positive development (Vangeepuram et al., 2016). On the other hand, there is also the risk that young people will decline to be involved when gatekeepers (such as family members) might otherwise have persuaded them to participate (Aaltonen, 2016), and so being able to directly recruit young people does not eliminate the risks around sampling bias; careful attention to respectful relationship management in the initial approach is still required.

In the current study, all potential participants were identified first by an organization. In those cases where programs were delivered in a group context (such as alternative education classes or group-based therapy programs), researchers presented the research to the young people in their group settings and invited then to complete an “indication of interest” form with contact details that researchers would then follow up individually. This allowed young people to either decline outright or to be open to an initial approach by researchers and decide later whether or not they would participate. Where programs were delivered on an individual basis (such as individual therapy), program staff made the initial approach providing an information sheet and indication of interest form to complete that was then passed onto researchers for follow up. Researchers were selected for their enthusiastic commitment to the research and this passion was critical in these initial meetings.

Because we had a full list of names of all eligible young people in each locality, we could follow up with staff when it appeared that a young person may not have been approached. By matching names across agency lists, we could identify young people who were the clients of multiple services, and thus, if a particular staff member or organization was not comfortable approaching a specific young person, another agency could be asked to make the initial approach on our behalf. This method maximized the chances that every eligible young person would at least be asked if they would meet a researcher to talk about participating. Through this method, we reduced organizational gatekeeper bias and were able to achieve a 98.5% recruitment rate.

Retaining and Maintaining Active Engagement in the Study

The issues researchers confront in trying to retain participants in longitudinal research have been well canvassed. Researchers have reported that tracking challenges are accentuated when research focuses on young people in general and with marginalized youth in particular (Aaltonen, 2016; Diviak, Wahl, O’Keefe, Mermelstein, & Flay, 2006; Hanna et al., 2014; Hardgrove et al., 2015; Vangeepuram et al., 2016; Ward & Henderson, 2003). Widely diverging retention rates have been reported with rates dropping to less than 30% when the focus is on marginalized youth (Bonevski et al., 2014; Diviak et al., 2006; Hardgrove et al., 2015; Hooven et al., 2011; Vangeepuram et al., 2016, Wright et al., 2015). Clearly, such low rates have serious implications for data quality in both quantitative and qualitative research (Abrams, 2010; Bonevski et al., 2014; Hanna et al., 2014; Vangeepuram et al., 2016).

While it may be challenging to retain the active engagement of marginalized young people in research, there are examples of successful projects which contain valuable lessons for researchers. Successful projects appear to use multiple retention and engagement strategies rather than relying on a single approach (Wright et al., 2015). For instance, such projects typically feature high levels of researcher agility such as quick adaptations when it is clear that a particular strategy is not successful, they also require researchers to be aware of the developmental needs of the young person and the way that these may influence their responsiveness to research (Taylor, 2009; Ward & Henderson, 2003). Early in their development, new technologies were embraced by researchers who saw in them the potential to increase retention rates. However, they have produced mixed results (Bull et al., 2013; Gwin Mitchell et al., 2015). It is clear from more recent reports that young people appropriate these technologies for their own purposes, and in so doing, they may lose their value as tracking tools for researchers. For instance, initially cell phones were found to be useful tracking aids; however, as a primary technique, they are now reported to be less effective because of the high turnover of these devices and the fact that young people share their phones with each other, leaving researchers unclear who they may be communicating with (Ward & Henderson, 2003). Furthermore, to be useful as a tracking device, researchers need access to dedicated research cell phones (Bull et al., 2013; Gwin Mitchell et al., 2015), so that they can respond promptly to young people and so that they can actively manage confidentiality and privacy issues. This, of course, has financial implications for research, which may not be able to resource the allocation of cell phones to all field staff.

New technologies also change over time and so even if researchers specify at the outset how they will use a particular technology, as a longitudinal project develops, the ways they will be used to support ongoing active involvement may well change. For instance, researchers have used SNS and related applications to locate young people (Bull et al., 2013; Gwin Mitchell et al., 2015). However, over time, the particular applications, as well as the ways in which they could be used, given privacy and ethical requirements placed on research, changed (Bull et al., 2013; Gwin Mitchell et al., 2015). While new technologies hold potential for assisting in finding young people, there have also been reports of precipitous drops in attrition rates when used as the primary tracking technique (Bull et al., 2013).

Regardless of the value of new technologies, successful retention ultimately depends heavily upon the quality of the relationship researchers establish with participants. Successful projects all report that being able to recruit enthusiastic and energetic researchers and then resource and support them out in the field for extended periods of time are the most important determinants of high retention rates. Thomson and Holland (2003) described a “devolved” approach where researchers each had direct responsibility for finding and retaining a specific cohort of young people. Ward and Henderson (2003) reported success in using an approach characterized by strong relationships between researchers and young people. Other researchers have also reported that a focus on personalized, trust-based relationships and allocating responsibility for tracking specific participants to specific researchers produced good results (Shahabi et al., 2011).

A critical part of this relationship management is the exercise of care in making contact because of the high likelihood that significant life events will occur between contacts. For instance, Ward and Henderson (2003) found that when making first contact, they needed to avoid pleasantries such as inquiring about a participant’s young child because in the interim, this child may, in fact, have been removed by child welfare services. They recommended brief initial phone calls focused on the research and further suggested careful pacing of interviews, so that researchers could judge at what point they asked particular questions in order to avoid provoking emotional or traumatic memories. Moving from a first contact to the completion of the interview may take a considerable length of time, as researchers adapt and respond to the fluctuating circumstances that characterize the lives of marginalized young people (Author, 2014b). Reports of successful projects also emphasize the critical value of record keeping, including the regular updating of contact details, having multiple points of contact, and multiple modes of finding youth (physical addresses, cell phone and landline numbers, e-mail addresses, and other Internet-based applications; Bonevski et al., 2014; Gwin Mitchell et al., 2015; Taylor, 2009; Thomson & Holland, 2003; Wright et al., 2015).

While all of the above factors are common to any longitudinal study, research with marginalized young people may also be particularly vulnerable to “researcher factors”; the emotional impact that doing the research has on researchers (Gustavson and Cytrynbaum, 2003). If not addressed, researcher factors have a significant potential to undermine projects because they can immobilize researchers leaving them unconfident about continuing to track youth or fearful for their personal safety (Bonevski et al., 2014). It is important, therefore, to train researchers in specific retention-related skills such as how to persist, to creatively problem solve, to use networks effectively to locate youth, and to manage themselves effectively when faced with confronting situations, so that young people will feel comfortable and positive when approached (Hooven et al., 2011).

Research with vulnerable populations will inevitably have an emotional impact on researchers (Sanders, Munford, Liebenberg & Henaghan, 2014). Part of managing the relational space of such research thus involves attending to these researcher factors. Some authors encourage intense reflexivity (see, e.g., Taylor, 2009). Others argue for a more pragmatic and task-focused approach that nonetheless provides some spaces for researchers to reflect upon the impact the research has upon them, while retaining a primary focus on locating and interviewing young people (McLeod and Yates, 1997). Taylor (2009) highlights the need for researcher reflexivity around “use of self.” She suggests that while these issues are more often taken up in qualitative accounts, they are nonetheless of importance in tracking youth for quantitative research because the ways in which researchers use themselves in the tracking process will influence whether or not young people choose to stay involved and allow themselves to be found. On the other hand, Mcleod and Yates (1997) talk of a “distant-formal research relationship” that avoids intensely introspective reflection on the self in favor of managing researcher safety in pragmatic ways. Support is given and opportunities to reflect are offered, but nonetheless the focus remains on the young person. Longitudinal research introduces an additional layer into research relationships because they endure over time. McLeod and Yates (1997) suggest this calls for some degree of reflexive scrutiny to ensure that the relationships stay research focused and do not stray into young people’s own lives. Thomson and Holland (2003) found that providing a space for researchers to reflect, write, and talk about the research was valuable both as a way of supporting researchers and for deepening data analysis.

PARTH: An Approach to Securing the Active Engagement of Marginalized Young People in Longitudinal Research

In the study, the PARTH principles, (outlined in Figure 1) played a significant role in keeping researchers focused on retention efforts and on ensuring the active participation of the young people across the six cycles of data collection. A commitment to research as a conduit for giving marginalized youth a voice in policy and service delivery had been a prerequisite for researchers to work in the project. We sought researchers who were passionate about involving marginalized young people in research and who expressed a desire to act as enthusiastic “ambassadors” for the project. However, as researchers came to terms with the intense commitment and effort that was required to relocate youth, it became clear that there was a disconnect for some between the principles they had espoused when coming into the project and the deeper perspectives they held concerning marginalized young people and the role of the researcher. It became clear that researchers were not necessarily immune to the insidious effects of prevailing pejorative public discourses regarding this group of young people that characterized them as difficult, irresponsible, lacking a social conscience, and unlikely to want to stay involved in the research (Hardgrove et al., 2015). Some researchers, for instance, attributed the difficulties they faced in finding young people as evidence that young people did not want to participate in the research. However, as the project developed, it became clear that this was not the case and that if researchers were able to acknowledge the complex lives of these young people and to adapt search strategies and persevere that young people were, in fact, keen to participate. Consider, for example, the following discussion between researcher and participant, following an 8-month search:

So here we are at the interview, and it is great to see you!

Yeah, I know, I gave you a hard time to find me didn’t I?

It’s surely taken a while to get to today, but you have had a lot going on, I could see from Facebook some of the things that happened for you and I knew you needed to deal with those things, they are more important than the research, we can wait for it to be the right time for you aye?

Oh, well, I shoulda turned up to one of those interviews I agree to, I know I shoulda. And you sat outside my house all day waiting for me, and I knew you were there. You are so patient, you always find me, you always wait for me. Why do you do that?

Well you are really important to me, you are one important piece in my big puzzle, and I need you to help me complete the puzzle. But I also know that sometimes your life is really complicated and you need to take care of yourself and your bubba first. You always tell me that you want to stay in the study and so I figure I just need to be patient and wait for it to work for you. That’s not too much for me to do. I figure it is my job to find you not your job to be found!

I’m pleased that you found me, I really wanna to tell you what’s been happening since our last interview. It’s sweet, I get a chance to tell my story, it might help someone else not fall into so many traps as I did aye?

Yup, it is very cool that you share your story with us, it is important, and you have a lot to contribute, a lot of wisdom, a lot of knowledge.

Such exchanges were common occurrences in interviews that had taken time to secure. Despite this, some interviewers suggested that after three attempts to locate youth or after three “no shows” for an interview, a young person should be removed from the study. The challenges researchers faced in locating young people and in dealing with the confronting living circumstances of many young people in the study made some researchers reluctant to make contact with young people or keen to prematurely give up looking for them. These types of perspectives interfered with researcher capacity to persevere and required careful management by the research leaders who themselves managed a cohort of young people. The research leaders’ active role in the research (tracking young people, conducting interviews, and managing data input and analysis) meant that they were able to understand the issues that interviewers faced in the field and the feelings they needed to manage in researching the lives of marginalized young people (Ferguson & Walker, 2014).

Repeated failures to locate youth, multiple broken appointments, and being repeatedly exposed to the confronting living circumstances of marginalized young people can have a powerful cumulative emotional impact on researchers. It is, however, important to separate out these types of researcher factors from genuine challenges in locating youth and securing interviews, because they call for different responses. Researcher factors require focused attention upon the perspectives researchers take to their work. For example, researchers were supported to manage their emotional reactions to tracking challenges and to being confronted with young people’s circumstances by using field diaries, regular debriefing, and close supervision. Like McLeod and Yates (1997), this support was focused tightly on the key research tasks which were relocating and interviewing young people. The study operated from positive youth development and strengths-based perspectives, and support for researchers was grounded in these perspectives. These required that researchers maintained a hopeful orientation to finding young people and supporting their participation in the research. These perspectives also meant that researchers needed to be mindful that when young people gave permission to be relocated, this implied a promise that the researcher would find them.

The strong team focus provided a forum for addressing the genuine challenges in locating young people. Here, researchers could share information on successful tracking techniques and this pragmatic support helped them adapt and respond quickly to the unpredictable and fluctuating circumstances of the young people. Action-oriented strategies such as “drive-bys,” that is, days spent by researchers, often in pairs, driving around the areas young people were known to frequent looking for them and leaving contact details with people nominated by youth as contact points, were used throughout the study. More passive, office-based strategies such as SNS and cell phones were also used, within privacy limits, as tracking methods. Unlike research reported elsewhere (Bull et al., 2013; Gwin Mitchell et al., 2015), cell phones proved to be extremely valuable. This was particularly so when number portability became available and young people could migrate their phone numbers between devices. Of course, because there was no way of knowing that the young person we were seeking was actually responding to an SNS or text message, the first contact was relatively oblique, for example, “[agreed code name] is looking for you again.”

The study adopted a youth-focused relational approach to tracking and retention. By the end of the second year, the research field team had reduced from 30 to a core group of six researchers who were either employed full time on the project or released on a part-time basis from NGOs for the research. This team maintained a strong commitment to completing the research, to achieving the very best possible retention rate, and to ensuring that young people remained actively engaged in the study. Each interviewer had responsibility for tracking and interviewing a subgroup of young people across the 6 years and, as others have noted, this relational approach was an important facet of successful retention (Thomson & Holland, 2003; Ward & Henderson, 2003). It meant that not only were researchers able to build strong relationships with the young people on their list, but also that they became familiar with the people in each young person’s network and they could draw on this familiarity when finding young people. It also meant that over the course of the study, the researchers were able to develop important skills in supporting young people to actively participate in research, and these skills were able to be shared with other researchers and also with youth workers and others who appreciated learning new strategies for supporting marginalized young people.

Across the study, approximately 60% of the young people were relatively easy to find, the remaining 40% absorbed most of the tracking efforts. As the study progressed, it became apparent that this 40% was comprised of two different groups. One group lived very precarious lives. They had very tenuous or fractured connections to their families and they were often evading the justice system. Enduring relationships were unusual among this group and a network of contacts thus needed to be recreated at each interview. These young people spent their days engaged in very basic survival activities, such as stealing food and finding safe places to sleep. The young people did not generally engage with new technologies; they had neither the resources to purchase devices nor the networks that would support them in accumulating the technical skills to make their use meaningful. Thus, cell phones and SNS were not effective tracking tools. We found these young people in one of two ways. Drive-bys around their last known neighborhoods and searching for leads as to their current location enabled us to begin to build a picture of their current geographical zone which we would then search. The other strategy involved searching in the prison system. Protocols were established with prison authorities that allowed us to request a name search if we could provide proof that the young person had given us permission to ask for a search and if we could demonstrate reasonable grounds to suspect they were in prison. Reasonable grounds included their names appearing in the court reports or elsewhere in local media, or if a contact had told us they were in prison. While family members might not know the current location of a young person, they often knew when they were in prison. Searching through the prison system required meticulous record keeping both relating to permission forms signed and dated by the young person and in terms of keeping full names, aliases, and an accurate date of birth, as these comprised the search criteria.

The other group of young people who were hard to find were more tech-savvy and relatively better resourced. They had more extensive and enduring social networks often, but not always, involving family. These networks would protect them but would also pass our messages along. While it could take time to find these young people, unlike the other group who would disappear completely for long periods of time, it was as if these youth were hidden in plain view. Getting to an interview came to resemble a game of hide and seek. We would turn up to houses to be informed they had just left, receive a text message from them shortly after suggesting an alternative meeting place, where they would often fail to materialize. We could often see their activities on SNS and would send them a private message and receive a reply that again led to a dead end. We would wait outside nominated houses for them to turn up, suspecting all the while that they might be inside. This time-consuming process was challenging to sustain because it was resource intensive and dispiriting. However, persevering with this searching and positively engaging with anyone who would talk to us and pass messages along always ultimately resulted in an interview. We came to understand these delicate engagement processes as manifestations of young people’s agency where they determined the length of the search and the manner in which it would be resolved. Our role was to engage positively in the process while still maintaining careful ethical oversight on the process to ensure that young people were not put under pressure to agree to participate when they did not wish to do so. Interviewers were trained to check carefully for any signs of reluctance and to take the initiative in such situations by suggesting that the interview not take place. This was complemented by researcher supervision meetings where each case was discussed and decisions were made about whether or not to continue to look for particular youth. As Aaltonen (2016) notes, there is a clear tension here that needs careful ongoing management to ensure that researchers do not stop looking for a young person who wants to be found but equally that they do not put young people who do not wish to participate under pressure to agree. It is in the day-to-day management of researchers out in the field that these issues are resolved, what Aaltonen (2016) refers to as ethics in practice. Maintaining a positive youth-centered, strengths perspective was critical here. Rather than interpreting young people’s actions as irresponsible, time wasting or manipulative, we chose to define these engagements as positive evidence of young people taking control over their circumstances; the research provided them with opportunities to exercise their agency. In engaging in the process, we demonstrated our commitment to the young person, communicating to them that they were important and worth the wait. Humility was important in this process; we needed to remind ourselves that were we in the same circumstances as these young people, we might well respond in a similar way.

As noted above, careful record keeping was a critical part of successful retention. This was the case not only for engagement with the prison system, it applied across the project. Being able to track back and identify people who had been helpful in previous searches was important. Accordingly, all contact details from every round, along with detailed notes on how each young person was located at each wave, were kept in a secure central database that was updated regularly. This reservoir of information was invaluable in relocating young people; sometimes, for instance, a cell phone that had been inactive for a year suddenly became active again. Young people were asked to provide at least three separate ways of locating them including, for instance, family members, foster parents, friends, service providers, and neighbors. Youth also signed consent forms indicating permission for us to search through organizations, and these forms were stored for the duration of the study. We asked young people to rank the order in which they wanted us to make contact with nominated people and organizations, so that they knew the order in which we would be working through their contact lists. At each phase of the study, phone numbers, addresses, e-mails, cell phones, and SNS addresses of as many people as the young people were comfortable giving us were captured and we regularly used all of these contact points in the tracking process. At different points in the process, different modes were effective. For instance, Facebook and Beebo were useful during Years 2–4. However, toward the end of the study, they were less helpful as some young people tired of them and others realized the risks that having their information in the public arena posed for them.

Important throughout, all of these processes were not only the relational practices that researchers established with young people but also with the organizations that supported the research throughout all of the phases of the study. This was important not only for the data collection phases but also for the data analysis phase and the translation of research findings into research products that could be used by organizations to advance their practice. All of the organizations were excited to be involved in research that could contribute to evidence-informed practice with marginalized young people.

Conclusion

The PARTH principles explored in this article were critical to successful sampling, recruitment, retention, and meaningful participation of young people in the study. These principles provided a comprehensive reference point for managing and supporting researchers to stay focused on the young people and to adapt their strategies in finding young people to the daily realities of these young people’s lives. Most importantly, the PARTH principles facilitated a change in the researcher’s orientation to the research and to young people who lived complex and challenging lives. Rather than referring to the young people as “hard to find” and difficult to engage in research, successful recruitment and retention occurred when researchers took ownership of the process and its attendant challenges. The PARTH principles provided a ready-reference point for this orientation to the research and to the participants. This shift in focus and responsibility for finding young people, from the young person to the researcher, meant that attention could focus exclusively upon the strategies that might best facilitate locating the young person. We observed that when the focus shifted onto the young person, researchers were more likely to talk in defeatist terms such as “this young person does not want to be found.” PARTH principles, on the other hand, did not produce this pessimistic outlook and instead attention focused on identifying what strategies might work including asking the researcher to think creatively about the strategies that could be used to find young people. The conversations then became “what might it take to find this young person?” and “what needs to be put in place to facilitate that?” This shift in focus also meant that when the young person was eventually found, the researcher was in a positive rather than frustrated frame of mind and this resulted in better interviews. What this project demonstrated was that even when undertaking research with marginalized youth, that researcher factors were stronger determinants of successful retention than were participant factors and that by controlling researcher factors and providing effective support to researchers, youth issues became more manageable.

The PARTH principles created a research process that enabled the participation of young people in the study on their own terms. This was important to the research both as a principle and in terms of data quality. It was critical that young people actively engaged in the interviews and trusted the interviewer, so that they would be comfortable with honestly sharing their experiences. Thus, PARTH principles were not only helpful in sustaining researchers through the challenges of tracking and finding young people, their application also provided valuable analytical material for the research. The processes of negotiating interviews and navigating around tracking obstacles provided valuable insights into how the young people in the study lived their lives, the ways in which they created meaning, the developmental issues they confronted alongside the risks they needed to manage, the types of adult engagement that were and were not effective in engaging them, and the ways in which they needed to be sometimes accessible and at other times to be hidden in a plain view.