View in EDS HTML Full Text PDF Full Text

Method to the Madness: Tracking and Interviewing Respondents in a Longitudinal Study of Prisoner Reentry

Saved in:
Bibliographic Details
Title: Method to the Madness: Tracking and Interviewing Respondents in a Longitudinal Study of Prisoner Reentry
Language: English
Authors: Fahmy, Chantal (ORCID 0000-0002-6010-8457), Clark, Kendra J., Mitchell, Meghan M., Decker, Scott H., Pyrooz, David C.
Source: Sociological Methods & Research. Aug 2022 51(3):1282-1324.
Availability: SAGE Publications. 2455 Teller Road, Thousand Oaks, CA 91320. Tel: 800-818-7243; Tel: 805-499-9774; Fax: 800-583-2665; e-mail: journals@sagepub.com; Web site: http://sagepub.com
Peer Reviewed: Y
Page Count: 43
Publication Date: 2022
Sponsoring Agency: US Department of Justice, Office of Justice Programs
Contract Number: 2014MUCX0111
Document Type: Journal Articles
Reports - Research
Descriptors: Longitudinal Studies, Correctional Institutions, Institutionalized Persons, Males, Law Enforcement, Program Effectiveness, Surveys, Interviews, Juvenile Gangs, Group Membership, Attrition (Research Studies), Crime, Telecommunications, Meetings, Social Media
Geographic Terms: Texas
DOI: 10.1177/0049124119875962
ISSN: 0049-1241
Abstract: After nearly four decades of growth, the number of people held in U.S. prisons has begun to decline. In an era of decarceration, social scientists need to understand prisoner reentry experiences. Longitudinal studies are one strategy to accomplish this goal. Yet, the retention of a formerly incarcerated population across waves of interviews is challenging due to their transient lifestyles and limited support systems, which may be further complicated by gang involvement. This article details the longitudinal follow-up procedures used in the LoneStar Project--a multiwave study of 802 males first interviewed in prison and reinterviewed twice in the year postrelease--to build rapport, complete interviews, and minimize attrition. We then evaluate the effectiveness of our procedures on important outcomes including interview yields, appointments, and incoming calls. Results indicate that any outgoing contact with respondents via appointment reminders and other reciprocal modes of contact lead to greater project engagement and a greater likelihood of interview completion. We conclude with relevant takeaways for researchers seeking to maximize survey participation with hard-to-reach populations.
Abstractor: As Provided
Entry Date: 2022
Accession Number: EJ1343366
Database: ERIC
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
    Url: https://content.ebscohost.com/cds/retrieve?content=AQICAHj0k_4E0hTGH8RJwT4gCJyBsGNe_WN95AvKlDbXJGqwxwFh0dLZX0Ar9MukYn7WqCmsAAAA4zCB4AYJKoZIhvcNAQcGoIHSMIHPAgEAMIHJBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDKVWGL4svsc7AXSyAAIBEICBm65ogmYPiEX1YME7vWH3Un2hHEnczw3Rx1hMkjVghHPSpoOgC0Nn-Zh-BYPXFQNPTq1eS0KxLCScP8h1-ehsxLaCNrmMD5XRGsra5i9LjYe9prLrtqdDCBJsMxVJcd6gBDAT4oyLJxpbwVf60erIIMo91BgyllQeyFs_t7INDmHlBCJIyge2MHcGxWX_wiRTGVldDJRHL-2YeaiN
Text:
  Availability: 1
  Value: <anid>AN0157637593;som01aug.22;2022Jun28.05:23;v2.2.500</anid> <title id="AN0157637593-1">Method to the Madness: Tracking and Interviewing Respondents in a Longitudinal Study of Prisoner Reentry </title> <p>After nearly four decades of growth, the number of people held in U.S. prisons has begun to decline. In an era of decarceration, social scientists need to understand prisoner reentry experiences. Longitudinal studies are one strategy to accomplish this goal. Yet, the retention of a formerly incarcerated population across waves of interviews is challenging due to their transient lifestyles and limited support systems, which may be further complicated by gang involvement. This article details the longitudinal follow-up procedures used in the LoneStar Project—a multiwave study of 802 males first interviewed in prison and reinterviewed twice in the year postrelease—to build rapport, complete interviews, and minimize attrition. We then evaluate the effectiveness of our procedures on important outcomes including interview yields, appointments, and incoming calls. Results indicate that any outgoing contact with respondents via appointment reminders and other reciprocal modes of contact lead to greater project engagement and a greater likelihood of interview completion. We conclude with relevant takeaways for researchers seeking to maximize survey participation with hard-to-reach populations.</p> <p>Keywords: longitudinal; survey research; prisoner reentry; retention; interviewing; gang membership; tracking</p> <p>Prisoner reentry is among the most pressing issues facing criminal justice reform. Indeed, around 95 percent of people admitted to prison are eventually released and nearly 80 percent of those released are rearrested within six years, commanding the attention of policy makers and researchers alike ([<reflink idref="bib1" id="ref1">1</reflink>]; [<reflink idref="bib36" id="ref2">36</reflink>]). For smart decarceration initiatives to be effective and sustainable, research in this area must be practical and empirically driven ([<reflink idref="bib23" id="ref3">23</reflink>]). Major public and private investments have been made to study the reintegration processes and outcomes of former prisoners, such as the Serious and Violent Offender Reentry Initiative (SVORI) and the Safe Streets and Second Chances longitudinal studies.[<reflink idref="bib5" id="ref4">5</reflink>] The research produced from these studies is incredibly valuable for two reasons. First, their methodology involves gathering survey rather than official data, the latter of which is typically limited to crime-specific outcomes. Successful reentry is more than desistance from crime. Second, they focus specifically on prisoner and reentry issues, particularly employment and housing. Representative samples and tailored questions allow researchers to paint a richer picture of the prisoner reentry process than is possible from large-scale, general longitudinal studies such as the National Longitudinal Survey of Youth (NLSY97).</p> <p>Sustaining study participation among a prisoner reentry population across multiple waves of interviews presents serious challenges. Transient lifestyles, frequent contact with the criminal justice system, and limited postrelease support systems make it difficult to secure high rates of retention ([<reflink idref="bib19" id="ref5">19</reflink>]; [<reflink idref="bib38" id="ref6">38</reflink>]; [<reflink idref="bib64" id="ref7">64</reflink>]; [<reflink idref="bib69" id="ref8">69</reflink>]). The lack of information on how to track and interview this difficult-to-study population creates a gap in knowledge about transitions from prison to the streets, especially in terms of strategies used to gather good data and minimize attrition in a longitudinal study. In addition, the study of this hard-to-reach population may offer insights for other groups who are difficult to find and interview, such as high-risk families, substance users, homeless persons, and adolescents ([<reflink idref="bib7" id="ref9">7</reflink>]; [<reflink idref="bib28" id="ref10">28</reflink>]; [<reflink idref="bib30" id="ref11">30</reflink>]; [<reflink idref="bib31" id="ref12">31</reflink>]; [<reflink idref="bib33" id="ref13">33</reflink>]; [<reflink idref="bib44" id="ref14">44</reflink>]). Given the continued interests in understanding prisoner reentry, particularly in an era of decarceration, it is necessary to examine the methods underlying studies of prisoner reentry.</p> <p>This article examines the longitudinal methods employed in the tracking and interviewing of respondents in the LoneStar Project, or the Texas Study of Trajectories, Associations, and Reentry. An added goal of this article is to determine whether these methods were more or less successful with gang members, a particularly challenging group to study. In outlining their procedures in a two-city network study of gang members, [<reflink idref="bib22" id="ref15">22</reflink>]:2) observed that "there remains little published guidance on field methods to aid researchers in longitudinal studies of gang processes." The LoneStar Project data are ideal to offer guidance on effective techniques to retain study respondents, particularly gang members, owing to a large sample of inmates (<emph>N</emph> = 802, 45 percent classified as gang members) and detailed data gathering on tracking and interviewing procedures. We divide the article into four parts. First, we outline the value of longitudinal research to the sociological study of crime and prisoner reentry. Second, we describe the procedures used in the LoneStar Project. Third, we explore the efficacy of various procedures on our ability to track and retain our respondents. In doing so, we evaluate the correlates of three outcomes: interview yields, appointments, and incoming calls. Finally, we provide guidance for researchers interested in the tracking and retention of gang and nongang members from prison to the "free world," as our respondents termed it.</p> <hd id="AN0157637593-2">Longitudinal Research in Criminology</hd> <p>The value of longitudinal research in criminology has been debated extensively (see, e.g., [<reflink idref="bib6" id="ref16">6</reflink>]; [<reflink idref="bib29" id="ref17">29</reflink>]; [<reflink idref="bib49" id="ref18">49</reflink>]). It is now indisputable, however, that longitudinal studies produce crucial information about continuity and change in problem behaviors over the life course, increase confidence in causal inferences with observational data, and open up new lines of inquiry ([<reflink idref="bib54" id="ref19">54</reflink>]). It is precisely this ability to document life-course transitions that is the virtue of longitudinal research. It brings validity to conclusions about sociogenic effects of new environments (e.g., arrests, joining a gang, prisoner reentry) that are not present in static designs ([<reflink idref="bib56" id="ref20">56</reflink>]; [<reflink idref="bib59" id="ref21">59</reflink>]).</p> <p>The value added by a longitudinal research design is also apparent in the study of prisoner reentry. Several studies have featured an in-prison interview followed by the tracking of respondents as they exit prison and return to their communities (e.g., SVORI, Boston Reentry Study). Longitudinal methods are important to understand prisoner reentry because of the inconsistent and erratic lifestyle of many ex-prisoners ([<reflink idref="bib68" id="ref22">68</reflink>]). They constitute a special population by which researchers can "test" the rigor of their longitudinal methodological approach, comparable to other challenging populations of interest to social scientists.</p> <p>The key to longitudinal studies of problem behaviors is the ability to track and reinterview people over scheduled time points. The effort invested in longitudinal studies in criminology is rarely made public. Instead, such knowledge is stored as artifacts in the boxes and hard drives of investigators, and even then, such information is only unearthed to revisit or extend substantive findings ([<reflink idref="bib47" id="ref23">47</reflink>]; [<reflink idref="bib59" id="ref24">59</reflink>]). Even when a blueprint of study methods is provided (see, e.g., [<reflink idref="bib3" id="ref25">3</reflink>]; [<reflink idref="bib22" id="ref26">22</reflink>]), it is unclear which procedures or protocols are statistically more or less effective than other procedures. Moreover, much of what is known about how to track particular groups is known from the public health literature following substance abusers ([<reflink idref="bib37" id="ref27">37</reflink>]; [<reflink idref="bib67" id="ref28">67</reflink>]), high-risk youth ([<reflink idref="bib41" id="ref29">41</reflink>]; [<reflink idref="bib46" id="ref30">46</reflink>]), and female sexual assault survivors ([<reflink idref="bib65" id="ref31">65</reflink>]). Ex-prisoners are undoubtedly a different population, facing a unique risk set in the reintegration process. Therefore, understanding how high-quality data captured at multiple waves with high retention rates are gathered will help create a superior body of evidence for theory and practice.</p> <p>In particular, low retention rates present challenges to both external and internal validity in longitudinal research. The group of respondents retained may not be an accurate reflection of an original representative sample, which threatens the external validity and generalizability of the findings. Moreover, findings from a study with low retention rates might dilute the accuracy of the study population's status and needs (internal validity), decreasing the value of such research. Therefore, any large-scale longitudinal undertaking needs to take into consideration methods for tracking and retaining respondents as well as the consequences of nonrandom sample attrition.</p> <p>The management and tracking of hard-to-reach populations presents many challenges, including residential mobility and instability, fear of providing personal information (especially those who engage in illegal behavior), and general disinterest in the study at hand ([<reflink idref="bib35" id="ref32">35</reflink>]; [<reflink idref="bib53" id="ref33">53</reflink>]).[<reflink idref="bib6" id="ref34">6</reflink>] Maintaining contact over time becomes more difficult when the respondent is actively trying to avoid the criminal justice system ([<reflink idref="bib52" id="ref35">52</reflink>]). For example, they deliberately evade calls from unknown phone numbers and are hesitant to communicate ([<reflink idref="bib4" id="ref36">4</reflink>]). Criminally involved persons are important to track and interview since they tend to be most "at risk" of nonparticipation by virtue of their race/ethnicity, gender, and socioeconomic status. Gang membership may further exacerbate these issues. They may be more likely to perceive researchers (and outsiders in general) as untrustworthy and conceivably involved with law enforcement ([<reflink idref="bib4" id="ref37">4</reflink>]; [<reflink idref="bib19" id="ref38">19</reflink>]; [<reflink idref="bib22" id="ref39">22</reflink>]). Despite these concerns, [<reflink idref="bib45" id="ref40">45</reflink>] found that gang members were just as likely as nongang members to participate in prison research and provide valid responses, standing in contrast to others who find that gang members are more secretive and harder to access by researchers (e.g., [<reflink idref="bib25" id="ref41">25</reflink>]; [<reflink idref="bib55" id="ref42">55</reflink>]).</p> <p>The major challenge in retaining individuals in panel studies of prisoner reentry exists upon release from prison, especially if the initial in-prison interview was difficult to secure ([<reflink idref="bib26" id="ref43">26</reflink>]; [<reflink idref="bib40" id="ref44">40</reflink>]). From a prisoner's perspective, participating in prison interviews may be a break from the daily monotony behind prison walls and can provide psychological satisfaction ([<reflink idref="bib15" id="ref45">15</reflink>]; [<reflink idref="bib21" id="ref46">21</reflink>]). However, once released, their level of commitment and interest in the study may dwindle considerably as they have alternative demands and preferences to fill their time during reintegration ([<reflink idref="bib20" id="ref47">20</reflink>]; [<reflink idref="bib35" id="ref48">35</reflink>]; [<reflink idref="bib53" id="ref49">53</reflink>]).</p> <p>[<reflink idref="bib22" id="ref50">22</reflink>]:3-4) note they could not find "any published article or report that specifically focused on methods for identifying and maintaining contact with gang members over time." The lack of research on imprisoned gang members and documentation of such research when it does occur is problematic because it could be a guide for future research on prisoner reentry. This takes on added significance in light of the fact that gang members constitute approximately 15 percent of the 1.5 million state and federal prisoners, the vast majority of whom will eventually reenter communities across the country ([<reflink idref="bib34" id="ref51">34</reflink>]; [<reflink idref="bib57" id="ref52">57</reflink>]). With limited research to serve as a guide, LoneStar Project staff relied heavily on protocols used by other comparable studies who tracked respondents after they exited prison. Table 1 contains a list of these strategies used to minimize attrition in previous research.</p> <p>Graph</p> <p>Table 1. Strategies to Increase Respondent Retention.</p> <p> <ephtml> <table><thead><tr><th>Strategy</th><th>References</th></tr></thead><tbody><tr><td>Financial incentives</td><td><xref ref-type="bibr" rid="bibr7">Bolanos et al. (2012)</xref>, <xref ref-type="bibr" rid="bibr8">Bootsmiller et al. (1998)</xref>, <xref ref-type="bibr" rid="bibr9">Boys et al. (2003)</xref>, Cohen et al. (2013), <xref ref-type="bibr" rid="bibr14">Collins et al. (2000)</xref>, <xref ref-type="bibr" rid="bibr20">Dillman (2000)</xref>, <xref ref-type="bibr" rid="bibr31">Graziotti et al. (2012)</xref>, <xref ref-type="bibr" rid="bibr42">McLaughlin et al. (2014)</xref>, <xref ref-type="bibr" rid="bibr44">Meyers et al. (2003)</xref>, and <xref ref-type="bibr" rid="bibr50">Mutti et al. (2014)</xref></td></tr><tr><td>Flexible scheduling of interviews</td><td><xref ref-type="bibr" rid="bibr30">Grape et al. (2018)</xref> and <xref ref-type="bibr" rid="bibr31">Graziotti et al. (2012)</xref></td></tr><tr><td>Personalization of all contacts</td><td><xref ref-type="bibr" rid="bibr2">Anseel et al. (2010)</xref>, <xref ref-type="bibr" rid="bibr20">Dillman (2000)</xref>, and <xref ref-type="bibr" rid="bibr71">Yammarino et al. (1991)</xref></td></tr><tr><td>Advance notice for interviews</td><td><xref ref-type="bibr" rid="bibr2">Anseel et al. (2010)</xref> and <xref ref-type="bibr" rid="bibr20">Dillman (2000)</xref></td></tr><tr><td>Respondent interest in study topic</td><td><xref ref-type="bibr" rid="bibr20">Dillman (2000)</xref>, <xref ref-type="bibr" rid="bibr35">Hunter, Puente, and Salo (2003)</xref>, and <xref ref-type="bibr" rid="bibr53">Pickett et al. (2014)</xref></td></tr><tr><td>Sponsorship/funding of research</td><td><xref ref-type="bibr" rid="bibr2">Anseel et al. (2010)</xref> and <xref ref-type="bibr" rid="bibr20">Dillman (2000)</xref></td></tr><tr><td>Persistence in communication with respondents by research team</td><td><xref ref-type="bibr" rid="bibr17">Cotter et al. (2005)</xref>, <xref ref-type="bibr" rid="bibr30">Grape et al. (2018)</xref>, <xref ref-type="bibr" rid="bibr31">Graziotti et al. (2012)</xref>, and <xref ref-type="bibr" rid="bibr44">Meyers et al. (2003)</xref></td></tr><tr><td>Visiting respondent homes</td><td><xref ref-type="bibr" rid="bibr7">Bolanos et al. (2012)</xref>, <xref ref-type="bibr" rid="bibr33">Hobden et al. (2011)</xref>, and <xref ref-type="bibr" rid="bibr43">Menendez, White, and Tulsky (2001)</xref></td></tr><tr><td>Reminder postcards</td><td><xref ref-type="bibr" rid="bibr7">Bolanos et al. (2012)</xref>, <xref ref-type="bibr" rid="bibr8">Bootsmiller et al. (1998)</xref>, <xref ref-type="bibr" rid="bibr9">Boys et al. 2003</xref>), <xref ref-type="bibr" rid="bibr33">Hobden et al. (2011)</xref>, and <xref ref-type="bibr" rid="bibr58">Rossman et al. (2011)</xref></td></tr><tr><td>Social media to find respondents</td><td><xref ref-type="bibr" rid="bibr7">Bolanos et al. (2012)</xref> and <xref ref-type="bibr" rid="bibr51">Mychasiuk and Benzies (2011)</xref></td></tr><tr><td>Public records to find respondents</td><td><xref ref-type="bibr" rid="bibr7">Bolanos et al. (2012)</xref>, <xref ref-type="bibr" rid="bibr13">Coen, Patrick, and Shern (1996)</xref>, <xref ref-type="bibr" rid="bibr16">Cotter et al. (2002)</xref>, Farrington (2003), <xref ref-type="bibr" rid="bibr33">Hobden et al. (2011)</xref>, and <xref ref-type="bibr" rid="bibr58">Rossman et al. (2011)</xref></td></tr><tr><td>Rapport with family/friends</td><td><xref ref-type="bibr" rid="bibr16">Cotter et al. (2002)</xref>, <xref ref-type="bibr" rid="bibr28">Gilmore and Kuperminc (2014)</xref>, <xref ref-type="bibr" rid="bibr30">Grape et al. (2018)</xref>, <xref ref-type="bibr" rid="bibr31">Graziotti et al. (2012)</xref>, <xref ref-type="bibr" rid="bibr33">Hobden et al. (2011)</xref>, <xref ref-type="bibr" rid="bibr43">Menendez et al. (2001)</xref>, and <xref ref-type="bibr" rid="bibr70">Wright et al. (1995)</xref></td></tr><tr><td>Contact card</td><td><xref ref-type="bibr" rid="bibr7">Bolanos et al. (2012)</xref>, <xref ref-type="bibr" rid="bibr8">Bootsmiller et al. (1998)</xref>, <xref ref-type="bibr" rid="bibr9">Boys et al. (2003)</xref>, <xref ref-type="bibr" rid="bibr13">Coen et al. (1996)</xref>, <xref ref-type="bibr" rid="bibr30">Grape et al. (2018)</xref>, and <xref ref-type="bibr" rid="bibr44">Meyers et al. (2003)</xref></td></tr></tbody></table> </ephtml> </p> <hd id="AN0157637593-3">Current Focus</hd> <p>The LoneStar Project consisted of survey-based interviews with 802 prisoners within days prior to their release from Texas prisons (Wave 1) and follow-up interviews at two waves postrelease. Although the survey instrument differed slightly, there were around 30 sections, 70 domains, and over 50 constructs at each wave. These items ranged from commonly surveyed demographics and theoretical constructs to sensitive information about in-prison experiences (e.g., victimization and misconduct) along with postrelease criminality and gang involvement. The Wave 1 sample consisted of a stratified random sample of all inmates released between April and December 2016 from the Huntsville unit, the busiest and largest of the five Texas Department of Criminal Justice's (TDCJ) release centers. Owing to the original purpose of the study—to evaluate gang members' experiences while incarcerated and upon reentry—TDCJ-identified gang members were oversampled by a factor of 5 and comprised 46 percent of the sample.[<reflink idref="bib7" id="ref53">7</reflink>][<reflink idref="bib45" id="ref54">45</reflink>] document the Wave 1 data collection processes in detail, with an emphasis on gang and nongang differences in study methodology. This article builds upon their work by shifting the focus from Wave 1 to prisoner reentry data collection procedures that occurred in Waves 2 and 3, where there were sharp divergences in method (i.e., phone-based rather than face-to-face) and context (i.e., community-based rather than prison-based).</p> <p>We extend the literature on tracking high-risk populations in multiple ways. In the first section—<emph>Follow-up Interview Procedures</emph>—we describe the methods and procedures used in our longitudinal data collection which successfully retained 66 percent of respondents at Wave 2 and 64 percent at Wave 3. In the second section—<emph>Evaluating the Empirical Efficacy of Procedures</emph>—we statistically examine the effectiveness of the procedures used as they relate to interview yields, appointments, and incoming calls (each of which are investigated by classification as gang and nongang status). In the third section—<emph>Looking Back to Look Forward</emph>—we conclude with lessons learned from our longitudinal research with a high-risk and hard-to-reach population. Previously, researchers have provided general suggestions about longitudinal procedures, such as using a database to manage contact information, but these general suggestions are rarely paired with the day-to-day specifics necessary for implementation and are never fully disclosed. Consequently, this study provides a much-needed "blueprint" for researchers by disclosing and testing the effectiveness of the procedures and methods used to track and interview high-risk populations.</p> <hd id="AN0157637593-4">Method</hd> <p>Although the criminological literature relying on longitudinal research designs typically only reports the results generated from surveys of study participants, our focus here is on the operational data gathered as part of the LoneStar Project. This is essentially the metadata we acquired through our efforts to prepare for data collection, track respondents, and complete interviews. In our first section—<emph>Follow-up Interview Procedures</emph>—we heavily rely on descriptive statistics when sharing our longitudinal research methods, procedures, and resources to provide the reader with a comprehensive understanding of the scope and magnitude of our processes.</p> <p>In the second section—<emph>Evaluating the Empirical Efficacy of Procedures</emph>—we use multivariate statistics to explore the effectiveness of our procedures across the 91 weeks of longitudinal data collection, which functions as our unit of analysis.[<reflink idref="bib8" id="ref55">8</reflink>] We aim to explain weekly variation in three dependent variables: <emph>interview yields</emph> or the number of interviews completed in a week (<emph>M</emph> = 11.16, <emph>SD</emph> = 6.18, range = 26), <emph>appointments</emph> or the number of interview appointments made with study participants in a week (<emph>M</emph> = 25.07, <emph>SD</emph> = 22.77, range = 106), and <emph>incoming calls</emph> or the number of calls received from study participants (<emph>M</emph> = 66.26, <emph>SD</emph> = 54.79, range = 376).</p> <p>Figure 1 provides a visual representation that summarizes our research efforts. It captures how the project unfolded including wave start and end dates, the frequency of calls, completed interviews, retention rates, and historical and holiday events.</p> <p>Graph: Figure 1. Weekly retention, interview yields, outgoing calls, and events throughout data collection (N = 91). Outgoing calls include both appointment and cold calls. This estimate does not include reminder or returned calls. Weeks 64–66 correspond with Hurricane Harvey.</p> <p>We account for a variety of week-level independent variables. Many measures represented project staff efforts to secure interviews. Those included what we termed <emph>cold calls</emph> (<emph>M</emph> = 110.99, <emph>SD</emph> = 75.71, range = 327)—or calls to respondents who were due for an interview. They also included the number of <emph>appointments</emph> which were scheduled with respondents. <emph>Reminders</emph> (<emph>M</emph> = 8.68, <emph>SD</emph> = 12.40, range = 49) were sent to everyone who had an upcoming interview and those contacts were reported. Similarly, any <emph>incoming calls</emph> (<emph>M</emph> = 66.26, <emph>SD</emph> = 54.79, range = 376) made by respondents and <emph>outgoing calls</emph> (<emph>M</emph> = 189.92, <emph>SD</emph> = 131.53, range = 662) made by project staff were documented. We also included a measure of <emph>overdue interviews</emph> (<emph>M</emph> = 305.88, <emph>SD</emph> = 111.03, range = 494) to determine the proportion of our respondents who were overdue for their scheduled interviews. Within these models, we also incorporated a measure of criminal activity via <emph>arrests</emph> (<emph>M</emph> = 2.60, <emph>SD</emph> = 2.08, range = 9), based on official records from the Texas Department of Public Safety. This helped to determine the proportion of respondents who had been involved with the criminal justice system after reentry. Interviewer efforts and measures of arrests were combined with dichotomous measures for <emph>holidays</emph>[<reflink idref="bib9" id="ref56">9</reflink>]—including Easter Sunday, Independence Day, Thanksgiving, Christmas, and New Year<bold>'</bold>s Day, and <emph>Hurricane Harvey</emph>. Hurricane Harvey was dummy coded (Harvey = 1) during the three weeks (weeks 64, 65, and 66) of the hurricane affecting many respondents in southeast Texas. We introduce lagged effects of prior week counts for the following variables: <emph>prior arrests</emph>, <emph>prior cold calls</emph>, <emph>prior appointments</emph>, <emph>prior reminders</emph>, <emph>prior incoming calls</emph>, and <emph>prior outgoing calls.</emph> These analyses allow for the scientific exploration of the effectiveness of research procedures for increasing interview yields among a high-risk population.</p> <hd id="AN0157637593-5">Follow-up Interview Procedures</hd> <p>The sections below describe the technology used, staffing, and procedures relied upon during the follow-up data collection for the LoneStar Project. The sample statistics below shed light on the magnitude and scope of the procedures used throughout data collection and result in 12 important takeaways garnered from this project.</p> <hd id="AN0157637593-6">Technological Infrastructure</hd> <p>Longitudinal research greatly benefits from a strong technological foundation, including proper management, accuracy, and security of data ([<reflink idref="bib2" id="ref57">2</reflink>]). We relied on multiple forms of technology. First, phone-based interviews were used to facilitate follow-up surveys. The sample size, project funding, scheduling challenges, and geographic distribution of the individuals after release in a state as large as Texas precluded face-to-face follow-up interviews. Consequently, 91 percent of follow-up interviews were conducted over the phone and 9 percent of interviews occurred face-to-face (all of which were reincarcerated respondents). We relied on <emph>Google Voice</emph> for all phone, text, and e-mail contacts. <emph>Google</emph> is one of the largest, most reliable companies in the technology industry with a highly stable online data infrastructure. Google Voice is free and a single online account allowed all project staff and interviewers to place calls using the same phone number.[<reflink idref="bib10" id="ref58">10</reflink>] This was advantageous because our interviewers were spread across multiple physical locations and could work from anywhere with Internet access. The online account also tracked all incoming and outgoing calls, text messages, and voice mails, which allowed for the creation of a real-time database. However, Google Voice lacks customer support (e.g., a helpline) and we found their online support forums to be inadequate for several of the issues we encountered.</p> <p>Second, our research design required flexible and secure approaches to data collection. Because we had a team of 28 interviewers dispersed across three states gathering highly sensitive data from a marginalized population, it was necessary to build an encrypted, secure server to accommodate these demands. This server allowed for data to be entered and accessible to project staff across multiple universities simultaneously. To prevent data loss, the server was backed up nightly and two technical support specialists were available to remedy any issues.</p> <p>Third, prior to administering the survey, interviewers needed to know who to call, when to call, and what numbers to dial. Although multiple databases were used to generate a call list and maintain up-to-date information for respondents and their collaterals, this represented a serious challenge in our longitudinal data collection procedures.[<reflink idref="bib11" id="ref59">11</reflink>] For Wave 2 interviews, we built a database in <emph>Microsoft Access</emph>. Respondent data was added to our "call list" five days prior to a respondent's scheduled follow-up interview. The call list, appointments, and call outcomes were maintained through <emph>Blaise</emph> (Statistics Netherlands), an add-on for computer-assisted telephone interviewing (CATI). It became evident that Blaise and Access were ineffective because project staff were updating contact information in multiple places, concurrent users could not access or edit the data, and it was challenging to export contact information between software. Not long into the Wave 2 interviews we abandoned the Access system and created a new database in Blaise. Housing the contact information in Blaise made it easy to transfer data into and out of the CATI system, but many of the issues remained (e.g., data were hard to access, data changes were hard to track, and system functionality—such as graphs and reporting—was limited). Additionally, issues arose with the CATI program (e.g., appointments were deleted, contact information was lost, collateral numbers were not called, refused respondents reappeared on the call list, and tracking call outcomes was tedious). Based on these concerns, we explored alternative contact management and CATI options to implement for Wave 3.</p> <p>For Wave 3, all contact information, calls, and appointments were managed through <emph>ACT!</emph>, a client relationship management (CRM) database. With ACT! it was easy to import and export data, allow multiple users to make simultaneous updates, restrict data access by user (e.g., project staff vs. interviewers), and create queries. In contrast to the many CRM systems available that require cloud-based servers, ACT! allowed for the local installation and storage of data, which was needed, owing to confidentiality issues surrounding the sensitive nature of the data. In ACT!, a cold call query was created to include all respondents who were due for an interview, had not refused, were not incarcerated, and did not already have an upcoming appointment. Upcoming appointments were placed on a task list and could be assigned to a specific interviewer. All call outcomes were tracked through a streamlined process and any errors in data entry could be edited. Interviewers had more freedom and control over the ACT! system, allowing them to look at the respondent's history file and see when he was called, what number was called, and the outcome and notes surrounding the calls. Added transparency ensured that each respondent number was called at least five times before moving on to the next number.[<reflink idref="bib12" id="ref60">12</reflink>]</p> <p>Computer-assisted personal interviewing (CAPI) software from <emph>Blaise</emph> housed the survey instrument and was used to collect data for all follow-up interviews—phone-based for nonreincarcerated interviews and in-person for reincarcerated interviews. This software allowed interviewers to effectively and quickly record respondents' answers. Complex skip patterns were created based on each respondent's answers—primarily generated from the Wave 1 interview—which facilitated survey customization to ensure respondents were given questions tailored to their specific reentry experiences.[<reflink idref="bib13" id="ref61">13</reflink>]</p> <p>Key Takeaway #1</p> <p> <emph>Many technological systems (e.g., a secure server, Blaise, Google Voice, Dropbox, and ACT!) will be used, adapted, or discarded throughout the project, which highlights the need for continued evaluation of the functionality of systems used for data collection and project management.</emph> </p> <hd id="AN0157637593-7">Staff Management</hd> <p>Wave 1 data collection in the prisons ended on December 10, 2016; Wave 2 data collection occurred between May 30, 2016 and May 9, 2017; and Wave 3 data collection occurred between February 7, 2017 and February 26, 2018. This meant that for some months, project staff resources were split between two distinct data collection efforts (see Figure 1). The follow-up data collection involved a devoted team, which consisted of two PIs, one site coordinator, six doctoral students, and 28 interviewers. Wave 2 and Wave 3 interviewers were trained (during Wave 1 data collection) on interviewing techniques and the interviewing software through three in-person large-scale trainings and numerous one-on-one sessions. Training manuals[<reflink idref="bib14" id="ref62">14</reflink>] and the presentations were shared through Dropbox. After face-to-face instruction, interviewers were given two weeks to practice on a test server. Following the trial period, interviewer evaluations were completed and scored by project staff using a standardized protocol measuring data entry, survey administration, and proper interviewing procedures. Interviewers were required to pass their evaluations prior to interviewing. Follow-up interviewers also received training on accessing the server, using Google Voice, building rapport over the phone, and calling procedures. Staffed shifts were determined a week in advance using <emph>SignUp Genius</emph>, a free application which allows users to "sign up" for shifts on a weekly basis. Additionally, the use of a free group messaging service, <emph>Group Me</emph>, was used to maintain contact with interviewers should questions arise.</p> <p>Hiring and training of interviewers was ongoing throughout Wave 2 and Wave 3 data collections. Because many of the interviewers were students—some graduated, moved, or had schedule changes—project staff needed to adapt and hire additional personnel. During Wave 3, the number of reincarcerated interviews substantially increased, consistent with the findings in research on recidivism. These interviews were challenging to staff because they often required full-day or multiday trips across Texas, as well as the complexity of scheduling interviews with incarcerated respondents. Wave 1 interviewers were instrumental in completing reincarcerated interviews because of their residence in Texas. In some instances, though, taking a flight or driving from Arizona or Colorado was more feasible for Wave 2 and Wave 3 interviews, given the size and geographic distribution of over 100 prisons across the state of Texas.</p> <p>Key Takeaway #2</p> <p> <emph>Successful data collection is highly contingent on devoted and capable project staff and interviewers. Invest in training, resources, and communication to ensure quality data are collected and that project tasks run smoothly.</emph> </p> <hd id="AN0157637593-8">Preparation for Data Collection</hd> <p>Several procedures from in-prison data collection were designed to facilitate successful follow-up interviews. First, respondents were compensated for completing either follow-up interview after their release, as financial incentives have been proven to increase retention rates ([<reflink idref="bib9" id="ref63">9</reflink>]; [<reflink idref="bib50" id="ref64">50</reflink>]). Second, effort was made to build rapport with the respondents at the Wave 1 (baseline) interview ([<reflink idref="bib8" id="ref65">8</reflink>]; [<reflink idref="bib13" id="ref66">13</reflink>]). Interviewers were trained to greet the respondent with a handshake, a welcoming smile, and to listen intently to their stories.[<reflink idref="bib15" id="ref67">15</reflink>] Third, the survey instrument was strategically designed for topics that may be challenging to talk about—specifically criminality and gang involvement ([<reflink idref="bib72" id="ref68">72</reflink>]). Prior to these sensitive sections, the interviewer would remind respondents about confidentiality and ensure that their information would not be shared with criminal justice officials. Additionally, these sections appeared later in the survey to ensure that interviewers had gained rapport with the respondent prior to discussing challenging topics.</p> <p>Fourth, a "contact card" was completed at the end of each survey, consistent with established practices in longitudinal data collection ([<reflink idref="bib16" id="ref69">16</reflink>]; [<reflink idref="bib41" id="ref70">41</reflink>]). Information was collected regarding respondents' prior addresses, release address, phone numbers, e-mail addresses, social media handles, and up to five collateral contacts—including the name, relationship to the respondent, phone numbers, and home addresses ([<reflink idref="bib7" id="ref71">7</reflink>]; [<reflink idref="bib51" id="ref72">51</reflink>]). The reliability of this information varied greatly. Some respondents maintain no collateral contact information (5 percent) or did not know the city or location where they would be paroled. Others left their contact materials in their "house" (i.e., prison cell) at the time of the interview. Despite these challenges, 15 percent of respondents provided one contact, 26 percent provided two contacts, and 48 percent of respondents provided three or four long-standing addresses or phone numbers from memory. Regardless of the information provided, respondents were encouraged to contact the project (via Google Voice) upon reentry to update any contact information. In fact, 392 respondents contacted the project 574 times to update their contact information—about 59 percent of these updates came from nongang members. Respondents received a study envelope in their release packet in hopes of creating sustained contact between project staff and the respondent. The release packet contained a brochure about the project (see Online Appendix A), a LoneStar Project business card, and a reloadable gift card for incentives.</p> <p>Key Takeaway #3</p> <p> <emph>Follow-up interview success may be contingent on prior waves of data collection; therefore, researchers should focus on building rapport, strategically placing sensitive research questions, ensuring confidentiality, garnering interest in the project, and collecting valid contact information.</emph> </p> <hd id="AN0157637593-9">Reminders and Additional Contacts</hd> <p>To encourage participation, maintain contact with respondents, and aid in setting up interview appointments, staff sent reminders, holiday wishes, and birthday cards to respondents ([<reflink idref="bib9" id="ref73">9</reflink>]; [<reflink idref="bib30" id="ref74">30</reflink>]; [<reflink idref="bib33" id="ref75">33</reflink>]). Based on the respondent's preferred contact method (i.e., phone, text, e-mail, or postcard), interview reminders were sent out one to two weeks prior to their scheduled interview date. During Wave 2, 520 reminder contacts were initiated and 286 reminder contacts were initiated for Wave 3 (see Table 2).[<reflink idref="bib16" id="ref76">16</reflink>] Similar processes existed for birthday cards. In total, 563 birthday wishes—in the form of postcards, phone calls, texts, and e-mails—were sent to respondents who had completed a follow-up interview. Respondents frequently noted that they enjoyed receiving "happy birthday" wishes and one respondent mentioned that the project was the only "happy birthday" wish he received that year. Finally, 245 holiday cards were sent out for the New Year (2018), thanking respondents for their participation, wishing them the best in 2018, and reminding them of their last chance to interview (see Figure 1 for trends).</p> <p>Graph</p> <p>Table 2. Incoming and Outgoing Contacts for Wave 2 and Wave 3 Interviews.</p> <p> <ephtml> <table><thead><tr><th rowspan="2">Contacts</th><th colspan="2">Wave 2 (<italic>n</italic> = 532)</th><th colspan="2">Wave 3 (<italic>n</italic> = 514)</th></tr><tr><th>Incoming</th><th>Outgoing</th><th>Incoming</th><th>Outgoing</th></tr></thead><tbody><tr><td>Total calls</td><td>1,701a</td><td>7,496</td><td>1,192a</td><td>8,138</td></tr><tr><td> Cold calls</td><td /><td>3,013</td><td /><td>7,094</td></tr><tr><td> Appointments</td><td /><td>1,557</td><td /><td>744</td></tr><tr><td>Total voice mails</td><td>791a</td><td>2,693</td><td>559a</td><td>3,334</td></tr><tr><td colspan="5">Cold call outcomes</td></tr><tr><td> Completed interview</td><td /><td>34</td><td /><td>152</td></tr><tr><td> No answer/busy</td><td /><td>429</td><td /><td>1,621</td></tr><tr><td> Left message with person</td><td /><td>—</td><td /><td>758</td></tr><tr><td> Appointment made</td><td /><td>195</td><td /><td>408</td></tr><tr><td> Refused</td><td /><td>22</td><td /><td>24</td></tr><tr><td> Voice mail</td><td /><td>2,139</td><td /><td>3,200</td></tr><tr><td> Disconnected</td><td /><td>180</td><td /><td>879</td></tr><tr><td> New number</td><td /><td>—</td><td /><td>25</td></tr><tr><td> Partial</td><td /><td>4</td><td /><td>27</td></tr><tr><td> Other</td><td /><td>14</td><td /><td>—</td></tr><tr><td colspan="5">Appointment outcomes</td></tr><tr><td> Completed interview</td><td /><td>463</td><td /><td>362</td></tr><tr><td> No answer/busy</td><td /><td>134</td><td /><td>89</td></tr><tr><td> Left message with person</td><td /><td>—</td><td /><td>19</td></tr><tr><td> Appointment made</td><td /><td>294</td><td /><td>86</td></tr><tr><td> Refused</td><td /><td>10</td><td /><td>24</td></tr><tr><td> Voice mail</td><td /><td>554</td><td /><td>134</td></tr><tr><td> Disconnected</td><td /><td>87</td><td /><td>6</td></tr><tr><td> Partial</td><td /><td>54</td><td /><td>33</td></tr><tr><td> Other</td><td /><td>15</td><td /><td>—</td></tr><tr><td>Total text conversations</td><td>455a</td><td>—</td><td>827a</td><td>—</td></tr><tr><td>Reminders</td><td /><td>332</td><td /><td>291</td></tr></tbody></table> </ephtml> </p> <p>1 <emph>Note</emph>: — = field was either not applicable or not obtainable.</p> <p>2 <sups>a</sups>An additional 496 incoming calls, 2,260 outgoing calls, 288 voice mails, and 276 incoming text message conversations occurred between February 7, 2017 and May 9, 2017. This was during the Wave 2 and Wave 3 simultaneous data collection, which made it impossible to distinguish which calls and texts were associated with each wave. Therefore, they have been left out of the reported estimates in this table and in the text.</p> <hd id="AN0157637593-10">Interview Contact Procedures</hd> <p>Other than reminders and postcards, contacts designed to schedule or complete an interview occurred through two means: cold calls and appointment calls.[<reflink idref="bib17" id="ref77">17</reflink>] In total, there were 7,496 outgoing calls for Wave 2 and 8,138 for Wave 3 data collection (see Figure 1 and Table 2).</p> <hd id="AN0157637593-11">Cold calls</hd> <p>Five days prior to each respondent's interview due date, respondents with <emph>any</emph> phone numbers were added to a "cold call" list which interviewers called from daily.[<reflink idref="bib18" id="ref78">18</reflink>] If no respondent information existed, collaterals (when available) were called with the expectation to provide updated contact information for the respondent. Call procedures involved interviewers: (<reflink idref="bib1" id="ref79">1</reflink>) reviewing the respondent's profile to determine which number to call, (<reflink idref="bib2" id="ref80">2</reflink>) reviewing any prior call notes, (<reflink idref="bib3" id="ref81">3</reflink>) calling the number, (<reflink idref="bib4" id="ref82">4</reflink>) tracking the call outcome (i.e., a "completed interview," "no answer," "disconnected," "appointment made," "left message with person," "refusal," "a partially completed interview," "voice mail," or "busy"), and (<reflink idref="bib5" id="ref83">5</reflink>) increasing the manual call counter.[<reflink idref="bib19" id="ref84">19</reflink>] If a phone number had five "unsuccessful" call attempts (i.e., "no answer", "voice mail") or if it was marked as "disconnected," the interviewer would indicate that number was no longer working and it would be skipped from subsequent calls.[<reflink idref="bib20" id="ref85">20</reflink>] If a respondent had no other numbers on file, they were marked as "exhausted" and moved to the retention list. Interviewers made 3,013 total cold calls during Wave 2 and 7,094 cold calls during Wave 3 (see Table 2).</p> <hd id="AN0157637593-12">Appointments</hd> <p>As interviewers cold called respondents and scheduled appointments, they recorded appointments with the ACT! task list and on an "interviewer schedule"[<reflink idref="bib21" id="ref86">21</reflink>] in <emph>Microsoft Excel</emph>. Based on respondent requests, a reminder text was sent the night before the interview and many respondents (<emph>n</emph> = 332 in Wave 2; <emph>n</emph> = 291 in Wave 3) requested this service. Also, rescheduled appointments were relatively high and represent the iterative process involved with completing an interview. Out of the 1,557 initial appointments made in Wave 2, 1,094 did not result in an interview at the scheduled time[<reflink idref="bib22" id="ref87">22</reflink>] and Wave 3 had even fewer appointments (<reflink idref="bib744" id="ref88">744</reflink>) which can be seen in Table 2.</p> <hd id="AN0157637593-13">Reciprocal contacts</hd> <p>Throughout all stages of data collection, effort was made to build rapport with respondents through interviewing techniques, being available to answer any questions, and occasionally just being there to listen. All respondents' calls, texts, and e-mails to the Google Voice account were returned by a project staff member (see Table 2). There were 1,701 incoming calls and 455 incoming text conversations for Wave 2 and 1,192 incoming calls and 827 incoming text conversations for Wave 3.[<reflink idref="bib23" id="ref89">23</reflink>] The number of incoming text conversations increased greatly between Wave 2 and Wave 3, likely because respondents had increased access to phones and technology as they spent more time out of prison.</p> <hd id="AN0157637593-14">Adding to the Madness</hd> <p>The most difficult part of this project was initiating and maintaining contact with respondents postrelease. This is largely a consequence of the nature of this high-risk, hard-to-reach population. To combat contact challenges, we relied on numerous platforms.</p> <hd id="AN0157637593-15">Social media</hd> <p>Scholars have maintained successful contact with respondents through the use of social media such as <emph>Facebook</emph> ([<reflink idref="bib7" id="ref90">7</reflink>]; [<reflink idref="bib51" id="ref91">51</reflink>]). Consequently, a personal Facebook page with an alias was created. Due to issues related to privacy settings and ambiguity in searches of popular names, social media was not useful for contacting respondents. Of our nine attempts, only one interview was completed, two people never responded but read the message, and six people never read the message.</p> <p>Key Takeaway #4</p> <p> <emph>The use of social media searches may minimally increase interview yields which does not equate to the time invested for searching and locating respondents. However, social media may be beneficial to other cohorts in panel studies such as youth or other groups who actively engage with social media platforms.</emph> </p> <hd id="AN0157637593-16">Phone access</hd> <p>Cell phones can be helpful in easing the reentry process and also in facilitating continued study participation ([<reflink idref="bib24" id="ref92">24</reflink>]; [<reflink idref="bib66" id="ref93">66</reflink>]).[<reflink idref="bib24" id="ref94">24</reflink>] Some respondents applied to receive a Federal Communications Commission funded cell phone shortly following their release if they made less than US$17,665 in the previous year. Staff from the Restorative Justice Ministries Network of Texas provided personnel at the bus stop a few blocks from the Huntsville Unit who would provide cell phone instructional sheets and information to those taking a <emph>Greyhound</emph> bus to their residence ([<reflink idref="bib10" id="ref95">10</reflink>]).[<reflink idref="bib25" id="ref96">25</reflink>] Following approval, a phone would be mailed to the applicant's location and phones could be activated on one of the eight wireless carriers. For those discharged from prison or paroled to a personal residence, these phones were an invaluable resource because many respondents used them to complete their follow-up interviews. Often, carriers allowed individuals to have cell phone access free of charge for up to one year following their release, but these cell phones—particularly phones with camera/video capabilities—were not always permitted in halfway houses or transitional centers. Other respondents were able to secure a personal cell phone through family members, their employer, or from their previous phone plan; however, these often use prepaid minutes. With the average interview length just under 90 minutes, many respondents were concerned with not having enough minutes to complete their interview. Consequently, we would complete a portion of the interview and wait until minutes were replenished to finish the interview. It was common for a respondent's phone to be out of minutes (i.e., not able to receive calls) for one call attempt and then accessible during the next call attempt. To account for this, it was important to continue calling a number until we were confident that it could not be reached.[<reflink idref="bib26" id="ref97">26</reflink>] Additionally, this concern could have been remedied by providing phone minutes in lieu of or along with financial incentives for respondents with access issues, a topic that others have explored ([<reflink idref="bib62" id="ref98">62</reflink>]).</p> <p>Key Takeaway #5</p> <p> <emph>Cell phone access and available minutes are vital resources for continued study involvement with over-the-phone interviews.</emph> </p> <hd id="AN0157637593-17">Halfway houses and transitional centers</hd> <p>Nearly one-quarter of respondents were paroled to transitional centers or halfway houses. This group faced several challenges when attempting to complete their interviews due to privacy policies, cell phone restrictions, scheduling conflicts, and crowded and noisy facilities. Often, the halfway house was unable to confirm or deny a respondent's residence or allow the project to contact the respondent due to confidentiality. Therefore, we relied on the respondent to contact the project when it was time for their interview. In a handful of instances, a project staff member was able to contact a respondent's case manager and set up a time for the respondent to use an office phone to complete their interview. In other instances, case managers indicated that they could only let the respondent use their office phone if they first verified it with the respondent's parole officer. Because it was important that the project remain separate from criminal justice entities in the eyes of respondents, project staff avoided all interactions with parole officers. However, having an established agreement with the halfway house or transitional center prior to data collection would have eased access concerns.</p> <p>Key Takeaway #6</p> <p> <emph>Halfway and transitional housing may present issues to respondent participation due to confidentiality restrictions and limited access to telephones.</emph> </p> <hd id="AN0157637593-18">Institutional review board (IRB)</hd> <p>Through 91 consecutive weeks of follow-up data collection, there were fewer than 10 instances when respondents contacted the IRB or a notice of potential harm was documented. It was important to keep the IRB updated on any potentially adverse events—whether they affected the interviewer, respondent, or both—to ensure ethical standards were consistently being met. We made it a point to check in periodically with the IRB to ensure an open line of communication. However, minor issues that were not considered to be a question of ethics or risk were left at the discretion of project staff and principal investigators. The open communication helped project staff easily resolve issues once they arose. One instance was a case of impersonation where a respondent's ex-girlfriend used a male friend to call the project, complete our respondent's interview, and receive payment.[<reflink idref="bib27" id="ref99">27</reflink>] After having a conversation with our respondent, the IRB was notified of a concern about a confidentiality breach. After investigation, the IRB case was closed, the project implemented a new protocol which required respondents to verify their TDCJ identification number prior to starting an interview, and the actual respondent was interviewed and compensated.</p> <p>A second instance occurred during a reincarcerated interview when one respondent exposed himself to an interviewer toward the conclusion of the interview. Although project interviewers were trained specifically on how to handle this type of situation, the interviewer used her discretion, ignored his behavior, and finished the interview. Upon completion, the interviewer immediately called project management, project staff then contacted the IRB, and took steps to ensure the interviewer was supported and allotted the time necessary to cope with the experience, with the option to terminate her interviewer role. In contrast, the IRB was not concerned with this incident because their interests lie with the safety and confidentiality of study respondents.</p> <p>Lastly, there were six times respondents mistakenly called the IRB directly with mundane requests such as setting up an appointment or asking about reimbursement. When this occurred, project staff worked quickly to return these calls and resolve any issues.</p> <p>Key Takeaway #7</p> <p> <emph>Open communication with the IRB is vital to study success and handling study-related concerns, especially unexpected issues. Keeping the IRB abreast of any potentially difficult situations—even if they were later deemed mundane—will facilitate excellent correspondence and trust by the IRB.</emph> </p> <hd id="AN0157637593-19">Additional call concerns</hd> <p>Other concerns that arose involved building rapport with collaterals, difficult respondents, and anonymous incoming contacts.[<reflink idref="bib28" id="ref100">28</reflink>] Project staff and interviewers often spoke with collaterals when attempting to contact respondents. Establishing credibility with these collaterals while maintaining respondents' confidentiality was challenging. For example, one respondent provided his future employer's phone number at the Wave 1 interview as the best way to contact him. When it was time to contact this respondent for Wave 2, it was important that we demonstrate enough credibility and professionalism with the employer to get him to pass along our message without divulging information that could jeopardize the respondent's job. Specifically, we ensured that training materials were clear enough to comprehend that under no circumstances were we to reveal to any collateral contacts—be it employers, church leaders, or family members—that a study participant had been incarcerated, as accidentally disclosing information could possibly jeopardize his economic or social standing.</p> <p>There were many instances when interviewers and project staff had to deal with problematic respondents who were under the influence of illegal substances or alcohol, falling asleep, acting or speaking in a sexually explicit manner, and/or being rude or aggressive. In one instance, a respondent masturbated while on the phone with a staff member. The staff member terminated the call and notified the appropriate personnel of the incident. From that point on, this respondent was only communicated with via text message by project staff.[<reflink idref="bib29" id="ref101">29</reflink>]</p> <p>Lastly, there were 259 instances where the project received incoming contacts from numbers that we were not able to associate with a respondent. This typically occurred when the respondent had not provided any contact information at the Wave 1 interview and then called from a new number without leaving a voice mail. Other similar instances occurred when a respondent used another person's phone to contact the project and did not leave a voice mail, got a new number and did not leave a voice mail, or left a voice mail that was inaudible. In these instances, a note was made of the contact and return contacts were attempted.</p> <p>Key Takeaway #8</p> <p> <emph>Challenges will appear throughout any data collection enterprise with a reentry population, but these challenges exist with any hard-to-track population. Preparing for difficult circumstances underscores the importance of interviewer training and close communication with project staff and oversight agencies to expediently remedy such concerns.</emph> </p> <hd id="AN0157637593-20">Completed Interviews</hd> <p>A substantial amount of effort went into all contacts with respondents prior to completing an interview. On average, 4.5 calls were placed (range from 1 to 49) before respondents completed their interview for Wave 2 and 7.8 calls were placed (range from 1 to 59) prior to respondents completing Wave 3 (see Table 2). Based on these data, we identified a diminishing return of calls on interview yields—after 1 call at the individual level and 243 calls at the weekly level. Of course, these diminishing returns do not translate neatly to practice because every addition to our retention numerator was critical. Approximately 50 percent of our sample completed interviews within four calls; whereas, the other 50 percent completed interviews after four calls with one person requiring 59 phone calls for completion. It is clear that persistence pays off and that diminishing returns will vary based on project goals.</p> <p>Weekly estimates of interview yields are presented in Figure 1. On average, we completed 11 interviews a week for Wave 2 and 9 interviews a week for Wave 3, but weekly yields varied from 0 to 27. The figure also depicts how calls and interview yields were relatively unaffected by holidays and historical events (e.g., Hurricane Harvey). In addition to completed interviews, there were 71 instances in Wave 2 and 58 instances in Wave 3 when the interview could not be completed during a single setting (i.e., partial interviews; see Table 3). Consequently, interviewers and project staff worked to schedule follow-up appointments, sometimes requiring up to eight sessions to complete an interview. All but 17 of these respondents completed their interviews by the end of Wave 2 and there remained 19 partial interviews at the end of Wave 3.</p> <p>Graph</p> <p>Table 3. Descriptions of Survey Outcomes for Wave 2 and Wave 3 Interviews.</p> <p> <ephtml> <table><thead><tr><th>Survey Information and Outcomes</th><th>Wave 2 (<italic>n</italic> = 532)</th><th>Wave 3 (<italic>n</italic> = 514)</th></tr></thead><tbody><tr><td>Partial interviewsa</td><td>58</td><td>60</td></tr><tr><td>Refusalsb</td><td>28</td><td>35</td></tr><tr><td> Does not know/no reason</td><td>13</td><td>16</td></tr><tr><td> Does not have time</td><td>6</td><td>2</td></tr><tr><td> Does not need the money</td><td>2</td><td>1</td></tr><tr><td> Not happy with payment</td><td>5</td><td>2</td></tr><tr><td> Other/not interestedc</td><td>2</td><td>13</td></tr><tr><td> Gang coded</td><td>—</td><td>1</td></tr><tr><td colspan="3">Call attempts</td></tr><tr><td> Mean</td><td>4.47</td><td>7.82</td></tr><tr><td> Median</td><td>3</td><td>4</td></tr><tr><td> Mode</td><td>1</td><td>2</td></tr><tr><td> Standard deviation</td><td>5.56</td><td>8.90</td></tr><tr><td> Minimum</td><td>1</td><td>1</td></tr><tr><td> Maximum</td><td>49</td><td>59</td></tr><tr><td colspan="3">Survey length (in minutes)</td></tr><tr><td> Mean</td><td>1:25:37</td><td>—</td></tr><tr><td> Standard deviation</td><td>24:24</td><td>—</td></tr><tr><td> Minimum</td><td>24:02</td><td>—</td></tr><tr><td> Maximum</td><td>3:02:38</td><td>—</td></tr></tbody></table> </ephtml> </p> <ulist> <item>3 <emph>Note</emph>: Survey length data were not available for Wave 3 because Wave 3 procedures did not allow for this calculation. — = field was either not applicable or not obtainable.</item> <item>4 <sups>a</sups>For Wave 2 and Wave 3, respectively, 10.9 percent and 11.7 percent of respondents had partial interviews at some point; however, by the end of data collection only 3.2 percent and 3.7 percent of interviews remained partials. <sups>b</sups>For Wave 2 and Wave 3, respectively, 5.3 percent and 6.8 percent of respondents refused to participate in the project. <sups>c</sups>"Other" for Wave 2, "not interested" for Wave 3. <sups>d</sups>One respondent simply said "gang code" when asked for his reason for refusal.</item> </ulist> <p>Key Takeaway #9</p> <p> <emph>Although on average four (Wave 2) to eight calls (Wave 3) were necessary for interview completion, we advise against placing upper bounds on the number of calls made to respondents, as each addition to the numerator is important.</emph> </p> <hd id="AN0157637593-21">Refusals and Attrition</hd> <p>At the end of Wave 2 data collection, 28 respondents had refused to participate in data collection (see Table 3), which were explained as (<reflink idref="bib1" id="ref102">1</reflink>) not having time (<emph>n</emph> = 6), (<reflink idref="bib2" id="ref103">2</reflink>) not needing the money (<emph>n</emph> = 2), (<reflink idref="bib3" id="ref104">3</reflink>) not being happy with payments (<emph>n</emph> = 5),[<reflink idref="bib30" id="ref105">30</reflink>] (<reflink idref="bib4" id="ref106">4</reflink>) being uncomfortable or annoyed with the questions (<emph>n</emph> = 2),[<reflink idref="bib31" id="ref107">31</reflink>] and (<reflink idref="bib5" id="ref108">5</reflink>) no specified reason (<emph>n</emph> = 13). After a respondent refused to continue participation, he was removed from the call list and from our retention efforts.</p> <p>For Wave 3 data collection, a more systematic and comprehensive method for tracking refusals was implemented. This was useful to ensure interviewers did not accidentally attempt to contact a person after they had refused to participate; however, it was more difficult to track the reason for refusal. During Wave 3, 35 respondents refused to participate, of which 24 of these refusals came from reincarcerated respondents. Although most respondents did not give a reason, the majority who did stated they were no longer interested in the project (see Table 3).[<reflink idref="bib32" id="ref109">32</reflink>] After data collection, the combined (Wave 2 and Wave 3) retention rate was 78%.</p> <p>Key Takeaway #10</p> <p> <emph>It is possible to maintain relatively high retention rates and low refusal rates with a largely gang-involved and/or high-risk sample of returning prisoners.</emph> </p> <hd id="AN0157637593-22">Payment</hd> <p>When an interview was completed, each respondent was asked to verify their <emph>Walmart</emph>[<reflink idref="bib33" id="ref110">33</reflink>] card number. Verified card numbers were electronically loaded (US$35 for Wave 2, US$70 for Wave 3) within 36 hours using a Walmart corporate gift card account.[<reflink idref="bib34" id="ref111">34</reflink>] Respondents who lost their card were mailed a replacement; however, these respondents only received US$30 (Wave 2) or US$65 (Wave 3) for their interview to compensate for the US$5 it cost the project for the lost card.[<reflink idref="bib35" id="ref112">35</reflink>] A total of 247 respondents lost at least one card throughout Wave 2 and Wave 3 data collection—38 of those respondents needed two replacement cards, and 3 respondents were given 3 cards each. Replacement cards were initially mailed to respondents with the money loaded onto the card. However, there were many instances where respondents moved residences, or the card got lost in the mail, or was intercepted by another person, and the money was lost. Given these issues, a new procedure was implemented for Wave 3 where replacement cards were sent with only US$5 on them and respondents had to text or call the project to receive the remainder of the payment.[<reflink idref="bib36" id="ref113">36</reflink>]</p> <p>Key Takeaway #11</p> <p> <emph>The ease of electronic fund transferring is beneficial because respondents can be paid quickly following the interview and any payment disputes could be tracked through an online system.</emph> </p> <hd id="AN0157637593-23">Concerted Retention Efforts</hd> <p>Not every call reached a respondent, nor was every phone number reliable. Inevitably, numbers were disconnected or inaccurate which resulted in many respondents being inaccessible. Nonworking numbers were removed from a respondent's profile and saved as an "old number" on their ACT! profile. Individuals were then removed from the cold call list if they had no working phone numbers or if all of their numbers had been exhausted (i.e., each number called more than five times). Concerted retention efforts ([<reflink idref="bib7" id="ref114">7</reflink>]; [<reflink idref="bib16" id="ref115">16</reflink>]; [<reflink idref="bib33" id="ref116">33</reflink>]; [<reflink idref="bib58" id="ref117">58</reflink>]) then began on this subsample, assuming they had not reached the Wave 2 attrition cutoff (≥150 days post–Wave 1 interview).[<reflink idref="bib37" id="ref118">37</reflink>]</p> <p>For Wave 2, eight staff members were assigned a retention workload, using an Excel spreadsheet to track these more intensive efforts.[<reflink idref="bib38" id="ref119">38</reflink>] The retention list included 261 respondents whose profiles were reviewed by project staff. These efforts were successful in securing an additional 42 interviews, but involved 50 e-mails, 179 postcards, 61 social media or <emph>Intelius</emph> (a public records repository) searches, and numerous phone calls to each respondent.</p> <p>Slightly different retention procedures were used for Wave 3 because ACT! had the functionality to track all retention outcomes directly, but the requirements necessary to be placed on the retention list did not change (i.e., no working phone numbers or all numbers had been exhausted). A list of 291 retention respondents were divvied up between eight project staff. For Wave 3, project staff simultaneously called and texted all available phone numbers, sent postcards to all addresses, e-mailed, and conducted social media searches. After concerted retention procedures had concluded, 304 postcards were sent to <emph>all</emph> respondent addresses in a final attempt to secure interviews. Over 2,470 calls were made, 176 postcards, and 60 e-mails were sent, resulting in an additional 76 completed interviews.</p> <p>For 12 respondents who were deemed the most challenging to track down, we hired a private investigative (PI) firm at US$175 per locate.[<reflink idref="bib39" id="ref120">39</reflink>] The information generated from the firm was extensive. Project staff then made phone calls, sent text messages, and mailed out postcards and certified letters to the updated contact information. After 77 calls and multiple mailings, we secured two completed interviews, one refused to participate,[<reflink idref="bib40" id="ref121">40</reflink>] and one person was deceased.[<reflink idref="bib41" id="ref122">41</reflink>] For the remaining eight respondents, the only contact made was with their collaterals.</p> <hd id="AN0157637593-24">Reincarcerated Interviews</hd> <p>At times, collaterals would share that the respondent was not released or had returned to jail or prison. In Wave 2, we relied on collateral reports of reincarceration. According to collaterals, 32 respondents were incarcerated, of whom 18 were validated with official data from TDCJ, whereas the whereabouts for the remaining 14 were identified via collaterals. Given that approximately 21 percent of all TDCJ released prisoners return to prison within one year ([<reflink idref="bib39" id="ref123">39</reflink>]), reincarcerated follow-up interviews were expected and prior agreements had been made with TDCJ and the IRB to conduct them. In total, only two respondents were interviewed in prison for their Wave 2 interview.</p> <p>Since information provided by collaterals was not always accurate and contact information from the prison system was often delayed, we needed an alternative source of reincarcerated data. Project staff initiated searches of publicly available data from TDCJ's website and vinelink.com (a national database designed for victim inquiries into perpetrator whereabouts) to identify respondents who were reincarcerated in jails and prisons throughout Texas. From May 2017 through March 2018, 32 percent of our sample was reported to be in jails and 22 percent of our sample had been in TDCJ custody (i.e., prison). If a respondent was found in the reincarcerated data, his profile was flagged as "temporarily suspended" and no contacts would be attempted until he was released. Instead, project staff would attempt to interview him at the TDCJ or jail facility. Coordinating these interviews was a time intensive process: gaining access to the facility, finding an available interviewer, and coordinating travel, including multi-interview trips, although the interview yield was worth the effort. Overall, we completed 23 interviews in county jails and 71 interviews at TDCJ facilities. These efforts were critical in securing longitudinal data for our respondents, because some respondents had not completed a Wave 2 interview at the time of their Wave 3 reincarcerated interview.</p> <p>Key Takeaway #12</p> <p> <emph>Although focused retention efforts and reincarcerated data searches are time-consuming and expensive, they can substantially increase retention rates with criminally involved populations, rendering them a success.</emph> </p> <p>The proceeding sections addressed the first focus of our article, which was to share the technologies, tools, and procedures relied upon throughout follow-up data collection. These are useful for researchers who are planning to implement longitudinal data collection efforts. The procedures alone do not equate to effectiveness, which is why the next section is centered on statistical efficacy.</p> <hd id="AN0157637593-25">Evaluating the Empirical Efficacy of Procedures</hd> <p>The goal of this section is two-fold. First, we explored which strategies are most likely to yield successful outcomes as measured by interview yields, appointments, and incoming calls. Based on prior research, and what we have described above, we would hypothesize that retention procedures (i.e., comprehensive contact card, total calls and reciprocal contacts, establishing rapport early on, and advance notice of interviews) would contribute to interview yields, more scheduled appointments, and more incoming calls to the project (see Table 1). To analyze these techniques, we estimated a series of models regressing weekly incoming calls, appointments, and interview yields on contemporaneous and lagged measures. Second, we determined whether some measures are better correlates of interview yields for gang members than nongang members in the study. If we find that some measures are better predictive of the outcomes than others, it lends credence to the assertion that longitudinal researchers should invest in specific strategies of data collection for high-risk subsamples.</p> <hd id="AN0157637593-26">Interview Yields</hd> <p>For the purpose of our study, interview yields were the most important outcome because without an interview, we lacked any and all data on the respondent during that particular wave. During Wave 2, we completed 532 interviews with a retention rate of 66 percent, and during Wave 3, we completed 514 interviews, which equated to a retention rate of 64 percent. However, we sought to examine if any of the follow-up procedures significantly influenced our weekly interview yields. We found that appointments, number of rearrests, cold calls, and incoming calls were all associated with increased weekly interview yields (see Table 4). The number of appointments set with respondents for the current week was strongly predictive of interview yields. This is perhaps unsurprising since those who take the time to schedule an appointment (or who are incarcerated and more willing to be interviewed out of boredom) are much more likely to follow through and complete the interview. Number of rearrests and incoming calls the current week and number of rearrests and cold calls the prior week were also positively associated with weekly interview yields. The positive association with rearrest may seem counterintuitive; however, once respondents were reincarcerated it was easier to secure an interview because they were confined to one physical space and thus locatable. Furthermore, the positive association between interview yields and reciprocal calls demonstrates the importance of ongoing communication between project staff and respondents which is required to secure interviews.</p> <p>Graph</p> <p>Table 4. Multivariate Ordinary Least Squares Regression Models Predicting Weekly Interview Yields, Weekly Appointments, and Weekly Incoming Calls.</p> <p> <ephtml> <table><thead><tr><th rowspan="2">Variables</th><th colspan="2">Weekly Interview Yields</th><th colspan="2">Weekly Appointments</th><th colspan="2">Weekly Incoming Calls</th></tr><tr><th><italic>b</italic></th><th>(<italic>SE</italic>)</th><th><italic>b</italic></th><th>(<italic>SE</italic>)</th><th><italic>b</italic></th><th>(<italic>SE</italic>)</th></tr></thead><tbody><tr><td colspan="7">Current week</td></tr><tr><td> Number of rearrests</td><td>0.575**</td><td>(0.170)</td><td>0.170</td><td>(0.648)</td><td>3.524</td><td>(1.882)</td></tr><tr><td> Cold calls (per 100) a</td><td>0.323</td><td>(0.523)</td><td>2.279</td><td>(1.979)</td><td>29.360***</td><td>(2.875)</td></tr><tr><td> Appointment calls (per 10)</td><td>1.899***</td><td>(0.297)</td><td>—</td><td>—</td><td>—</td><td>—</td></tr><tr><td> Reminders</td><td>0.015</td><td>(0.031)</td><td>0.134</td><td>(0.118)</td><td>0.920**</td><td>(0.309)</td></tr><tr><td> Overdue interviews</td><td>−0.002</td><td>(0.004)</td><td>−0.029*</td><td>(0.015)</td><td>−0.020</td><td>(0.037)</td></tr><tr><td> Incoming calls (per 10)</td><td>0.215**</td><td>(0.069)</td><td>0.359</td><td>(0.260)</td><td>—</td><td>—</td></tr><tr><td colspan="7">Prior week</td></tr><tr><td> Number of rearrests</td><td>0.378*</td><td>(0.166)</td><td>0.408</td><td>(0.630)</td><td>0.489</td><td>(1.850)</td></tr><tr><td> Cold calls (per 100)</td><td>1.054*</td><td>(0.518)</td><td>−2.113</td><td>(1.959)</td><td>−7.793</td><td>(4.208)</td></tr><tr><td> Appointment calls (per 10)</td><td>−0.126</td><td>(0.309)</td><td>7.874***b</td><td>(0.770)</td><td>—</td><td>—</td></tr><tr><td> Reminders</td><td>0.018</td><td>(0.031)</td><td>0.022</td><td>(0.117)</td><td>0.437</td><td>(0.325)</td></tr><tr><td> Incoming calls (per 10)</td><td>0.084</td><td>(0.073)</td><td>−0.386</td><td>(0.276)</td><td>0.722</td><td>(1.117)</td></tr><tr><td colspan="7">Other controls</td></tr><tr><td> Hurricane Harvey</td><td>−0.256</td><td>(1.743)</td><td>3.108</td><td>(6.636)</td><td>−4.105</td><td>(19.410)</td></tr><tr><td> Holidays</td><td>−0.790</td><td>(1.063)</td><td>0.150</td><td>(4.055)</td><td>−14.460</td><td>(11.790)</td></tr><tr><td>Constant</td><td align="center" colspan="2">1.088</td><td align="center" colspan="2">11.280</td><td align="center" colspan="2">5.931</td></tr><tr><td><italic>R</italic>2</td><td align="center" colspan="2">0.817</td><td align="center" colspan="2">0.802</td><td align="center" colspan="2">0.701</td></tr></tbody></table> </ephtml> </p> <ulist> <item>5 <emph>Note: N</emph> = 91. Standard errors in parentheses. Bivariate statistics are available upon request. — = field was either not applicable or not obtainable.</item> <item>6 <sups>a</sups>Weekly incoming calls uses outgoing calls rather than cold calls and appointment calls separately to avoid issues of multicollinearity while maximizing model fit. <sups>b</sups>Coefficient differed statistically between gang and nongang members. While appointment calls the prior week significantly increased weekly appointments for both gang and nongang members, the coefficient was about 1.75 times higher for nongang members (<emph>b</emph> = 7.74) than gang members (<emph>b</emph> = 4.44).</item> <item>7 *<emph>p</emph> ≤.05. **<emph>p</emph> ≤.01. ***<emph>p</emph> ≤.001.</item> </ulist> <hd id="AN0157637593-27">Appointments</hd> <p>Our next outcome of interest was appointments. From the previous section, we know that appointments were the strongest predictor of interview yields, so it was important to determine which items were associated with an increased likelihood of appointments. There were 2,301 appointments calls made throughout the duration of the project (see Table 2). The results indicated that only overdue interviews during the current week and appointment calls the week prior were significantly associated with appointments (see Table 4). The lagged association with appointment calls may be due to rescheduled appointments. Many respondents had appointment calls (with the intention of getting an interview completed that day) but may have forgotten, leading to rescheduled appointments made for the following week. Regarding overdue interviews, the results are in the opposite direction of what one would expect. This may be due to the fact that once overdue, respondents forget about or are less concerned with making up that interview in the future.</p> <hd id="AN0157637593-28">Incoming Calls</hd> <p>The final outcome of interest was incoming calls, which demonstrated a respondent's willingness to engage with project staff (see Figure 1 for trends). There were 15,634 outgoing calls made throughout the duration of the project—some in forms of reminders, cold calls, and staff callbacks (see Table 2). The high volume of outgoing calls resulted in many returned calls from respondents and collaterals with a total of 2,893 incoming calls—all of which were returned by project staff. Outgoing calls and reminders for the current week were both positively related to weekly incoming calls. This suggests that multiple forms of targeted outgoing contacts influence reciprocal incoming contacts. This is especially important considering how weekly appointments and interview yields were predicated on various forms of reciprocal contacts between the project and respondents (see Table 4).</p> <p>Because every effort was made to build rapport throughout data collection, project staff diligently returned every phone call, even investing time in having casual conversations with respondents. We prioritized returned calls because they helped foster rapport, allowed for project staff to check in with respondents, and allowed us to update contact information. Incoming calls suggest that our respondents were willing to reach out and engage in the research process.</p> <p>Overall, it is evident that reaching out to respondents in some way prior to their interviews—via reminder texts and outgoing calls—is a fruitful endeavor for minimizing attrition. Inevitably, these outgoing forms of contact lead to reciprocal communication which produces appointments and subsequently completed interviews. Although primarily through indirect effects, the message is still clear: Incorporating as many modes of contact with respondents (and their collaterals) as possible over the course of the study will likely yield a higher retention rate.</p> <hd id="AN0157637593-29">Gang Differences Across Interview Yields, Appointments, and Incoming Calls</hd> <p>The supplemental purpose of this section was to explore if the efficacy of procedures varied by gang status. We find very few differences within our sample (see Table 4). In fact, the only statistically significant difference between gang and nongang members across the three dependent variables—interview yields, appointments, and incoming calls—is that appointment calls in the prior week significantly increased weekly appointments for all respondents, but the coefficient was nearly two times greater for gang members. The lack of statistical significance in our overall findings between gang and nongang members points to the fact that inmate status appears to trump gang status. In other words, reintegrating prisoners—gang members or not—constitute a hard-to-reach sample of which individual differences may not matter as much to determine the best methodologically sound procedures. Thus, our verdict is that all high-risk samples should be evaluated based on the project's particular needs and how researchers approach their project, rather than the specifics of <emph>who</emph> they are trying to retain.</p> <hd id="AN0157637593-30">Discussion</hd> <p>A study of this magnitude, especially one that deals with a hard-to-reach population, requires considerable flexibility, patience, persistence, preparation, and collaboration. For the LoneStar Project, the collaboration necessary to continue data collection required cooperation among three universities (Arizona State University, Sam Houston State University, and University of Colorado Boulder) with different IRBs, budgeting processes, personnel rules, disciplines, and cultures. Project staff members were in regular communication (multiple times a day) to handle concerns and leverage resources and expertise. Although we have outlined 12 core takeaways from this research, below we arrive at three key conclusions in the longitudinal study of prisoner reentry.</p> <p>First, use data to inform the preparation and implementation of the project. By gathering and storing the metadata that was used by the research staff every day (e.g., interview schedule, call outcomes, incoming calls, interview yields, reincarcerations), we were better able to track our progress, successes, and missteps to better guide continued access to respondents. Constantly evaluating our processes allowed us to document which procedures were beneficial and which could be discarded for future use. This provided a roadmap of our own project as well data to test the effectiveness of our procedures, enabling us to identify best practices on how others can improve upon our actions. For instance, our findings regarding the utility (or lack thereof) of social media, and Facebook specifically, to track respondents is at odds with previous research ([<reflink idref="bib7" id="ref124">7</reflink>]; [<reflink idref="bib51" id="ref125">51</reflink>]) since social media was not an effective tool to reduce attrition in our sample.</p> <p>Second, perseverance pays off. The amount of time, work, and energy in the form of phone calls, text messages, e-mails, postcards, meetings, and brainstorming sessions is to be expected for a project of this nature. The LoneStar Project staff's mindset was that many hours invested to secure even <emph>one</emph> interview was invaluable. We nonetheless attempted to report our interview yields based on each separate process so researchers can determine whether these efforts are worth their time and resources. From our perspective, keeping this end goal in mind—that every interview counts—also had the added benefit of actively engaging staff with the project. Consistent with previous research that we chose to model, persistent and constant communication with respondents was certainly found to enhance retention in the LoneStar Project ([<reflink idref="bib17" id="ref126">17</reflink>]; [<reflink idref="bib30" id="ref127">30</reflink>]; [<reflink idref="bib31" id="ref128">31</reflink>]). In fact, if we could only give one piece of advice for retention, it would be to have persistent communication with respondents.</p> <p>Finally, flexibility is critical. Leaving prison is a monumental and stressful life-course transition in respondents' daily lives, routines, and roles ([<reflink idref="bib68" id="ref129">68</reflink>]). Because of their inconsistent and largely unpredictable schedules, researchers must be patient with their time and ready to interview at odd hours of the day and/or in atypical settings. Due to financial, safety, and geographic restraints, we were unable to visit respondent homes to complete interviews as previous research has purported to be extremely beneficial to reduce attrition ([<reflink idref="bib33" id="ref130">33</reflink>]; [<reflink idref="bib43" id="ref131">43</reflink>]). This flexibility on the part of the research team likely would have boosted our retention rate. For some ex-prisoners, however, participating in an interview at all was not a priority; they are in the midst of enjoying their newfound freedom and attempting to put the pieces of their lives back together. Understanding the worldview of reintegrating former prisoners, and the challenges they face, helped LoneStar Project staff establish rapport and trust, while still gathering the necessary data.</p> <p>In conclusion, the current study sheds light on and determines the effectiveness of the tracking and retention procedures used by staff for the federally funded LoneStar Project in order to help inform future research on prisoner reentry in an era of decarceration. Our article aimed to detail <emph>all</emph> strategies used with a criminally involved sample in a digestible and easily emulated format. The era of mass decarceration cannot be fulfilled simply by using agency-level data or other administrative records to add to our evidence base in prisoner reentry research. Hearing ex-prisoners'—gang and nongang members alike—perspectives on the challenges that lie before them upon community reentry propels examination to continue in novel ways in this area of research. With a mere handful of studies employing data collection strategies similar to the LoneStar Project, it is imperative the strategies used to increase retention be discussed openly and for future research endeavors. We hope scholars with similar data collection strategies use our study as a guide and continue to add to the limited knowledge on tracking high-risk populations in the free world.</p> <hd id="AN0157637593-31">Supplemental Material</hd> <p>Appendix_A_release_brochure.pdf</p> <p>Supplemental Material, Appendix_A_release_brochure - Method to the Madness: Tracking and Interviewing Respondents in a Longitudinal Study of Prisoner Reentry</p> <p></p> <p>Supplemental Material, Appendix_A_release_brochure for Method to the Madness: Tracking and Interviewing Respondents in a Longitudinal Study of Prisoner Reentry by Chantal Fahmy, Kendra J. Clark, Meghan M. Mitchell, Scott H. Decker and David C. Pyrooz in Sociological Methods & Research</p> <p></p> <hd id="AN0157637593-32">Notes</hd> <ref id="AN0157637593-33"> <title> Notes </title> <blist> <bibl id="bib1" idref="ref1" type="bt">1</bibl> <bibtext> The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.</bibtext> </blist> <blist> <bibl id="bib2" idref="ref57" type="bt">2</bibl> <bibtext> The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was supported by Grant No. 2014-MU-CX-0111 awarded by the National Institute of Justice, Office of Justice Programs, U.S. Department of Justice. The research contained in this document was coordinated in part by the Texas Department of Criminal Justice (723-AR15). The contents of this article, including its opinions, findings, and conclusions, are those of the authors and do not necessarily reflect the views or policies of the U.S. Department of Justice or the Texas Department of Criminal Justice.</bibtext> </blist> <blist> <bibl id="bib3" idref="ref25" type="bt">3</bibl> <bibtext> Chantal Fahmy https://orcid.org/0000-0002-6010-8457</bibtext> </blist> <blist> <bibl id="bib4" idref="ref36" type="bt">4</bibl> <bibtext> Supplemental material for this article is available online.</bibtext> </blist> <blist> <bibl id="bib5" idref="ref4" type="bt">5</bibl> <bibtext> For the Serious and Violent Offender Reentry Initiative, 69 sites received anywhere from US$500,000 to US$2 million each for a total of US$100 million; funded by the U.S. Departments of Justice, Labor, Education, Housing and Urban Development, and Health and Human Services. Safe Streets and Second Chances was funded US$4 million for its pilot study by the Charles Koch Foundation.</bibtext> </blist> <blist> <bibl id="bib6" idref="ref16" type="bt">6</bibl> <bibtext> Other scholars cite racial minority status ([11]; [27]; [61]), low-income ([60]; [61]), males ([12]; [18]), and low education ([40]) as those who are "at risk" and less likely to be retained.</bibtext> </blist> <blist> <bibl id="bib7" idref="ref9" type="bt">7</bibl> <bibtext> Texas Department of Criminal Justice's (TDCJ) determines affiliation with gangs using classification identifiers such as self-admission, tattoos, informants, mail, property, and gang hand signs. Individuals who met three or more of these identifiers are considered confirmed gang members whereas those who had at least one but not three identifiers are considered suspected gang members ([32]; [63]). For the purpose of our study, TDCJ-identified gang members constituted anyone with nonzero levels of affiliation (confirmed, suspected, and former gang members).</bibtext> </blist> <blist> <bibl id="bib8" idref="ref55" type="bt">8</bibl> <bibtext> The entire data collection period was 97 weeks, but we only model 91weeks of follow-up interviews as the first 6 weeks were solely dedicated to Wave 1 (in-prison) interviews.</bibtext> </blist> <blist> <bibl id="bib9" idref="ref56" type="bt">9</bibl> <bibtext> Denoted holidays are those that had the greatest bearing on interviewers' and respondents' availability to interview.</bibtext> </blist> <blist> <bibtext> In the middle of Wave 2 data collection, over 70 calls were being placed daily. Google Voice has an unstated limit on the number of calls allowed per day. Based on our patterns of usage, we exceeded that capacity, necessitating the creation of a second account to meet interviewing demands.</bibtext> </blist> <blist> <bibtext> There are many client relationship management (CRM) systems available, such as <emph>SalesForce</emph>, but almost all of these CRMs require storing data on cloud-based servers. Given the sensitivity surrounding our population, and the lack of transparency on the security of cloud-based storage, we relied upon standalone software that could be installed on our server.</bibtext> </blist> <blist> <bibtext> Research on longitudinal follow-ups reports that diminishing returns on investment occur after five calls for each number ([17]; [37]; [60]). Based on this research, we chose to contact each number five times before moving onto the next respondent number. See discussion below regarding diminishing returns in the "Completed Interviews" subsection.</bibtext> </blist> <blist> <bibtext> For example, if a parolee missed his Wave 2 interview, at Wave 3 we would ask him, "What are your parole conditions?" However, if the respondent answered the question regarding parole conditions during the Wave 2 interview, at Wave 3 we would ask, "Have your parole conditions changed from ([import] parole conditions from Wave 2)?" If yes, then the interviewer would indicate which conditions had change and how they changed.</bibtext> </blist> <blist> <bibtext> Training manuals can be viewed on https://<ulink href="http://www.researchgate.net/project/The-LoneStar-Project">www.researchgate.net/project/The-LoneStar-Project</ulink>.</bibtext> </blist> <blist> <bibtext> At the conclusion of Wave 1 interviews, many respondents had built rapport with the interviewers and were excited about the opportunity for payment and the possibility to talk with the interviewer again. Scheduling challenges prohibited respondents from <emph>always</emph> completing follow-up interviews with the same interviewer, although project staff tried to accommodate requests as they arose.</bibtext> </blist> <blist> <bibtext> From April through September (2016), there was an issue with the database incorrectly exporting reminder information so not every respondent received a reminder notification prior to October.</bibtext> </blist> <blist> <bibtext> Although we relied on Access-, Blaise-, and Blaise-computer-assisted telephone interviewing to organize our contact information and call list for Wave 2 interviews, the sections below outline our procedures for ACT!—which was adapted after numerous concerns occurred with the prior systems (see subsection "Technological Infrastructure" for a review).</bibtext> </blist> <blist> <bibtext> Interviewer shifts were four hours to allot time to complete two interviews. However, if no interviews were scheduled, the interviewer was instructed to cold call 20 people.</bibtext> </blist> <blist> <bibtext> The manual call counter indicated how many times a number had been called. Interviewers manually increased this number following each contact.</bibtext> </blist> <blist> <bibtext> Past research emphasizes the importance of attempting multiple contacts with respondents before considering them "exhausted" (i.e., unreachable; [17]; [37]; [60]). The current study utilized this advice by having interviewers attempt the first five contacts with respondents then having project staff complete more intensive retention efforts that are discussed later in this article.</bibtext> </blist> <blist> <bibtext> Using SignUp Genius, most interviewers staffed the same set of shifts each week based on their class schedules and other obligations. The interview schedule—detailing interviewer shifts and respondent appointments—was housed in a <emph>Dropbox</emph> folder that could be accessed by all interviewers and project staff. Interviewers would check the interview schedule prior to each shift to determine whether they would be making cold calls, completing an appointment, or doing both.</bibtext> </blist> <blist> <bibtext> If the respondent did not answer the first appointment call attempt, the interviewer would leave a voice mail, wait for the respondent to call back, and if no call was returned within 5–10 minutes, the interview would try one (or two) more time(s). Then, the interviewer documented the final call outcome in Blaise or ACT! (see Table 2). The interviewer would also document whether the appointment was completed, failed, or rescheduled for another date.</bibtext> </blist> <blist> <bibtext> <emph>Google Voice</emph> does not display the total number of incoming texts. Rather, this number represents the total number of text conversations initiated by respondents. Therefore, one text conversation could contain many incoming and outgoing text exchanges.</bibtext> </blist> <blist> <bibtext> Initially, we wanted to provide cellular phones to respondents upon release, but cost prohibited this resource.</bibtext> </blist> <blist> <bibtext> The Huntsville bus stop has been the subject of numerous documentaries, given its long-standing place in Texas reentry experiences. To learn more about the history of the bus stop, see [10].</bibtext> </blist> <blist> <bibtext> Throughout Wave 2 data collection, many respondents were housed in one specific transitional center where one respondent had a phone which they allowed many other respondents to use to contact the project. While this was useful because it gave those without a phone means to contact the project, it also presented some issues when Wave 3 data collection began and multiple respondents had indicated this number was the best way to contact them despite the fact they had typically moved on from the transitional center.</bibtext> </blist> <blist> <bibtext> The impersonator provided our respondent's birthdate, which interviewers used to validate his identity.</bibtext> </blist> <blist> <bibtext> Multiple times, ex-prisoners contacted the project asking to join the study after hearing about the project from friends. Although flattered, we explained that respondents could not be added to later waves and wished them well.</bibtext> </blist> <blist> <bibtext> In the few instances where serious issues were reported, the principal investigators (PIs) forwarded the reports to the Institutional Review Boards and provided adequate support to the staff member who experienced the interaction.</bibtext> </blist> <blist> <bibtext> Some respondents misunderstood the payment schedule and felt they were supposed to be paid for their Wave 1 interview in prison and never were.</bibtext> </blist> <blist> <bibtext> Only two respondents refused because they were uncomfortable or annoyed by the questions. This refusal rate is very low since the survey included multiple sensitive topics. Interviewers were trained to reiterate the confidential nature of our project to respondents while also reminding them that these calls were not recorded. We suspect that these reassurances helped to alleviate concerns respondents may have had about disclosing sensitive information.</bibtext> </blist> <blist> <bibtext> Over the phone, refusals are challenging to track because many respondents would say, "I don't want to participate" and immediately hang up. Interviewers were thus in a challenging position because they wanted to get a reason for refusal but did not want to contact the respondent against their wishes.</bibtext> </blist> <blist> <bibtext> Walmart reloadable gift cards were used to disseminate incentives. We chose Walmart because they are a large retailer in many cities across Texas and reloadable gift cards allowed respondents to keep one card throughout the duration of the project, which was electronically loaded with money.</bibtext> </blist> <blist> <bibtext> This compensation was similar, but slightly less, compared to other criminological studies offering incentives such as payment on a graduated scale ([48]) or US$50 for each interview with a US$250 maximum ([69]).</bibtext> </blist> <blist> <bibtext> <ulink href="http://Walmart.com">http://Walmart.com</ulink> requires a minimum of US$5 to be loaded on every reloadable gift card purchased by a corporate account. Therefore, every gift card that was given to a respondent had US$5 loaded on the card. When they lost their card, they also lost this US$5 which was deducted from the payment on their replacement card.</bibtext> </blist> <blist> <bibtext> Some respondents never contacted the project for full payment. Project staff continued to call these respondents multiple times to initiate their payment, yet data collection ended with 109 respondents who never returned our calls.</bibtext> </blist> <blist> <bibtext> An attrition cutoff point of 150 days was established for Wave 2 to allow for a methodological distinction between waves. At 150 days, respondents were closer to the Wave 3 follow-up than they were to the Wave 2. Although 16 Wave-2 interviews were completed past 150 days (ranging from 151 to 222) the overwhelming majority (97 percent) were within the 150-day cutoff point.</bibtext> </blist> <blist> <bibtext> Efforts involved are (1) calling and texting all available phone numbers (at least three calls per number), (2) sending postcards to known addresses, (3) e-mailing (at least two e-mails sent), (4) conducting social media searches, (5) conducting online searches using Intelius and <emph>Lexis Nexis</emph> to find new numbers, (6) calling any numbers that a respondent had called from, and (7) requesting parole contact information (i.e., addresses and phone numbers).</bibtext> </blist> <blist> <bibtext> The rate of US$175 per locate was discounted from typical costs for this service. Selection criteria for the PI searches were intentional. Respondents were targeted if they (1) were living in the Houston area (because the PI firm was familiar with that location), (2) were current gang members (because they were underrepresented in Wave 3 completed interviews), and (3) identified as black or Hispanic (because they were underrepresented in Wave 3).</bibtext> </blist> <blist> <bibtext> This respondent was "on the run," had a dozen aliases on file in state records, and was alarmed that we found him.</bibtext> </blist> <blist> <bibtext> There were six instances where we were notified and able to confirm that a respondent had died. Although it is likely that more than six respondents passed away during our data collection period, we have not yet identified those individuals ([5]). Collaterals, online obituaries, and news reports were used to confirm the deaths: Two died in fatal automobile accidents, one died from an accidental overdose, two were killed in drive-by shootings, and one was shot and killed while attempting to rob a residential dwelling.</bibtext> </blist> </ref> <ref id="AN0157637593-34"> <title> References </title> <blist> <bibtext> Alper Mariel, Durose Matthew R., Markman Joshua. 2018. 2018 Update on Prisoner Recidivism: A 9-year Follow-up Period (2005-2014). Washington, DC: U.S. Department of Justice, Bureau of Justice Statistics.</bibtext> </blist> <blist> <bibtext> Anseel Frederik, Lievens Filip, Schollaert Eveline, Choragwicka Beata. 2010. " Response Rates in Organizational Science, 1995–2008: A Meta-analytic Review and Guidelines for Survey Researchers." Journal of Business and Psychology25(3):335–49.</bibtext> </blist> <blist> <bibtext> Arnold Tom. 2014. "Longitudinal Studies of Criminal Behavior: What We Have Learned about Collecting Data on the Life Course." The DLC Criminologist2(2):9–23.</bibtext> </blist> <blist> <bibtext> Becker Kirsten, Berry Sandra H., Orr Nate, Perlman Judy. 2014. "Finding the Hard to Reach and Keeping Them Engaged in Research." Pp. 619–41 in Hard-to-survey Populations, edited by Tourangeau R., Edwards B., Johnson T. P., Wolter K. M., Bates N.. New York, NY: Cambridge University Press.</bibtext> </blist> <blist> <bibtext> Binswanger Ingrid A., Stern Marc F., Deyo Richard A., Heagerty Patrick J., Cheadle Allen, Elmore Joann G., Koepsell Thomas D.. 2007. " Release from Prison—A High Risk of Death for Former Inmates." The New England Journal of Medicine356(2):157–65.</bibtext> </blist> <blist> <bibtext> Blumstein Alfred, Cohen Jacqueline, Farrington David P.. 1988. " Criminal Career Research: Its Value for Criminology." Criminology26(1):1–35.</bibtext> </blist> <blist> <bibtext> Bolanos Franklin, Herbeck Diane, Christou Dayna, Lovinger Katherine, Pham Aurora, Raihan Adnan, Rodriguez Luz, Sheaff Patricia, Brecht Mary-Lynn. 2012. " Using Facebook to Maximize Follow-up Response Rates in a Longitudinal Study of Adults Who Use Methamphetamine." Substance Abuse: Research and Treatment6:1–11.</bibtext> </blist> <blist> <bibtext> Bootsmiller Bonnie J., Ribisl Kurt M., Mowbray Carol T., Davidson William S., Walton Maureen A., Herman Sandra E.. 1998. " Methods of Ensuring High Follow-up Rates: Lessons from a Longitudinal Study of Dual Diagnosed Participants." Substance Use & Misuse33(13):2665–85.</bibtext> </blist> <blist> <bibtext> Boys Annabel, Marsden John, Stillwell Garry, Hatchings Kevin, Griffiths Paul, Farrell Michael. 2003. " Minimizing Respondent Attrition in Longitudinal Research: Practical Implications from a Cohort Study of Adolescent Drinking." Journal of Adolescence26(3):363–73.</bibtext> </blist> <blist> <bibtext> Burnett John. 2015. " Look at This: The Bus Station." National Public Radio, April.</bibtext> </blist> <blist> <bibtext> Butler James, Quinn Sandra C., Fryer Craig S., Garza Mary A., Kim Kevin H., Thomas Stephen B.. 2013. " Characterizing Researchers by Strategies Used for Retaining Minority Participants: Results of a National Survey." Contemporary Clinical Trials36(1):61–67.</bibtext> </blist> <blist> <bibtext> Claus Ronald E., Kindleberger Lisa R., Dugan Mary C.. 2002. " Predictors of Attrition in a Longitudinal Study of Substance Abusers." Journal of Psychoactive Drugs34(1):69–74.</bibtext> </blist> <blist> <bibtext> Coen Anita Saranga, Patrick Diane C., Shern David L.. 1996. " Minimizing Attrition in Longitudinal Studies of Special Populations: An Integrated Management Approach." Evaluation and Program Planning19(4):309–19.</bibtext> </blist> <blist> <bibtext> Collins Rebecca L., Ellickson Phyllis L., Hays Ron D., McCaffrey Daniel F.. 2000. " Effects of Incentive Size and Timing on Response Rates to a Follow-up Wave of a Longitudinal Mailed Survey." Evaluation Review24(4):347–63.</bibtext> </blist> <blist> <bibtext> Copes Heith, Hochstetler Andy, Brown Anastasia. 2013. "Inmates' Perceptions of the Benefits and Harm of Prison Interviews." Field Methods25(2):182–96.</bibtext> </blist> <blist> <bibtext> Cotter Robert B., Burke Jeffrey D., Loeber Rolf, Navratil Judith L.. 2002. " Innovative Retention Methods in Longitudinal Research: A Case Study of the Developmental Trends Study." Journal of Child and Family Studies11(4):485–98.</bibtext> </blist> <blist> <bibtext> Cotter Robert B., Burke Jeffrey D., Stouthamer-Loeber Magda, Loeber Rolf. 2005. " Contacting Participants for Follow-up: How Much Effort Is Required to Retain Participants in Longitudinal Studies? " Evaluation and Program Planning28(1):15–21.</bibtext> </blist> <blist> <bibtext> Crisanti Annette S., Case Brian F., Isakson Brian L., Steadman Henry J.. 2014. " Understanding Study Attrition in the Evaluation of Jail Diversion Programs for Persons with Serious Mental Illness or Co-occurring Substance Use Disorders." Criminal Justice and Behavior41(6):772–90.</bibtext> </blist> <blist> <bibtext> Curtis Ric. 2010. "Getting Good Data from People That Do Bad Things: Effective Methods and Techniques for Conduting Research with Hard-to-reach and Hidden Populations." Pp. 141–58 in Offenders on Offending: Learning about Crime from Criminals, edited by Bernasco W.. London, England: Willan.</bibtext> </blist> <blist> <bibtext> Dillman Don A.2000. Mail and Internet Survey: The Tailored Design Method. 2nd ed. New York, NY: John Wiley & Sons.</bibtext> </blist> <blist> <bibtext> Edens John F., Epstein Monica, Stiles Paul G., Poythress Norman G.. 2011. " Voluntary Consent in Correctional Settings: Do Offenders Feel Coerced to Participate in Research? " Behavioral Sciences & the Law29(6):771–95.</bibtext> </blist> <blist> <bibtext> Eidson Jillian L., Roman Caterina G., Cahill Meagan. 2016. " Successes and Challenges in Recruiting and Retaining Gang Members in Longitudinal Research: Lessons Learned from a Multisite Social Network Study." Youth Violence and Juvenile Justice15(4):396–418.</bibtext> </blist> <blist> <bibtext> Epperson Matthew W., Pettus-Davis Carrie, Grier Annie, Sawh Leon. 2018. "Promote Smart Decarceration." Pp. 181–203 in Grand Challenges for Social Work and Society, edited by Fong R., Lubben J., Barth R. P.. New York, NY: Oxford University Press.</bibtext> </blist> <blist> <bibtext> Fatherhood Research & Practice Network. 2016. Using Cellphone Technologies in Fatherhood Programs and Research. Philadelphia, PA: Center for Policy Research.</bibtext> </blist> <blist> <bibtext> Fong Robert S., Buentello Salvador. 1991. " The Detection of Prison Gang Development: An Empirical Assessment." Federal Probation55(1):66–69.</bibtext> </blist> <blist> <bibtext> Fox Kathleen A., Lane Jodi, Turner Susan F.. 2018. Encountering Correctional Populations: A Practical Guide for Researchers. Oakland: University of California Press.</bibtext> </blist> <blist> <bibtext> George Sheba, Duran Nelida, Norris Keith. 2014. " A Systematic Review of Barriers and Facilitators to Minority Research Participation among African Americans, Latinos, Asian Americans, and Pacific Islanders." American Journal of Public Health104(2):16–31.</bibtext> </blist> <blist> <bibtext> Gilmore Devin, Kuperminc Gabriel P.. 2014. " Testing a Model of Participant Retention in Longitudinal Substance Abuse Research." American Journal of Evaluation35(4):467–84.</bibtext> </blist> <blist> <bibtext> Gottfredson Michael R., Hirschi Travis. 1987. " The Methodological Adequacy of Longitudinal Research on Crime." Criminology25(3):581–614.</bibtext> </blist> <blist> <bibtext> Grape Annette, Rhee Hyekyun, Wicks Mona, Tumiel-Berhalter Laurene, Sloand Elizabeth. 2018. " Recruitment and Retention Strategies for an Urban Adolescent Study: Lessons Learned from a Multi-center Study of Community-based Asthma Self-management Intervention for Adolescents." Journal of Adolescence65:123–32.</bibtext> </blist> <blist> <bibtext> Graziotti Ann L., Hammond Jane, Messinger Daniel S., Bann Carla M., Miller-Loncar Cynthia, Twomey Jean E., Bursi Charlotte, Woldt Eunice, Nelson Jay Ann, Fleischmann Debra, Alexander Barbara. 2012. " Maintaining Participation and Momentum in Longitudinal Research Involving High-risk Families." Journal of Nursing Scholarship44(2):120–26.</bibtext> </blist> <blist> <bibtext> Hill Cece. 2009. " Gangs/Security Threat Groups." Corrections Compendium34(1):23–37.</bibtext> </blist> <blist> <bibtext> Hobden Karen, Forney Jason Curtis, Durham Kathleen Wyszacki, Toro Paul. 2011. " Limiting Attrition in Longitudinal Research on Homeless Adolescents: What Works Best? " Journal of Community Psychology39(4):443–51.</bibtext> </blist> <blist> <bibtext> Hughes Timothy, Wilson Doris James. 2004. Reentry Trends in the United States. Washington, DC: Bureau of Justice Statistics.</bibtext> </blist> <blist> <bibtext> Hunter Jennifer, Puente Manuel De La, Salo Matt. 2003. Comparative Ethnographic Research on Mobile Populations. Washington, DC: U.S. Census Bureau, Statistical Research Division.</bibtext> </blist> <blist> <bibtext> James Nathan. 2015. Offender Reentry: Correctional Statistics, Reintegration into the Community, and Recidivism. Washington, DC: Congressional Research Service.</bibtext> </blist> <blist> <bibtext> Kleschinsky John H., Bosworth Leslie B., Nelson Sarah E., Walsh Erinn K., Shaffer Howard J.. 2009. " Persistence Pays off: Follow-up Methods for Difficult-to-track Longitudinal Samples." Journal of Studies on Alcohol and Drugs70(5):751–61.</bibtext> </blist> <blist> <bibtext> Lattimore Pamela K., Steffey Danielle M.. 2009. The Multi-site Evaluation of SVORI: Methodology and Analytic Approach. Washington, DC: National Institute of Justice, Office of Justice Programs.</bibtext> </blist> <blist> <bibtext> Legislative Budget Board Staff. 2017. " Overview of Criminal Justice Correctional Population Projections, Recidivism Rates, and Costs per Day." Report provided by the Legislative Budget Board, Texas Department of Criminal Justice, Texas Department of Public Safety, February 23, 2017.</bibtext> </blist> <blist> <bibtext> Martinez-Ebers Valerie. 1997. " Using Monetary Incentives with Hard-to-reach Populations in Panel Surveys." International Journal of Public Opinion Research9(1):77–86.</bibtext> </blist> <blist> <bibtext> McCuller William J., Sussman Steve, Holiday Karen, Craig Sande, Dent Clyde W.. 2002. " Tracking Procedures for Locating High-risk Youth." Evaluation & the Health Professions25(3):345–62.</bibtext> </blist> <blist> <bibtext> McLaughlin Patricia, White Nick, King Kate, Hann Robert G., Williams Robert J., Schopflocher Donald, West Beverly, Flexhaug Trevor. 2014. QLS Front-Line Retention Manual: Methods for Achieving a 94% Cohort Retention Rate in Longitudinal Research. Report prepared for the Ontario Problem Gambling Research Centre.</bibtext> </blist> <blist> <bibtext> Menendez Enrique, White Mary Castle, Tulsky Jacqueline P.. 2001. " Locating Study Subjects: Predictors and Successful Search Strategies with Inmates Released from a U.S. County Jail." Controlled Clinical Trials22(3):238–47.</bibtext> </blist> <blist> <bibtext> Meyers Kathleen, Webb Alicia, Frantz Jeanne, Randall Mary. 2003. " What Does It Take to Retain Substance-abusing Adolescents in Research Protocols? Delineation of Effort Required, Strategies Undertaken, Costs Incurred, and 6-month Post-treatment Differences by Retention Difficulty." Drug and Alcohol Dependence69(1):73–85.</bibtext> </blist> <blist> <bibtext> Mitchell Meghan M., Spooner Kallee, Jun Wu, Pyrooz David C., Decker Scott H.. 2018. " Survey Research with Gang and Non-gang Members in Prison: Operational Lessons from the LoneStar Project." Trends in Organized Crime1–29. doi:10.1007/s12117-018-9331-1.</bibtext> </blist> <blist> <bibtext> Morrison Theodore C., Wahlgren Dennis R., Hovell Melbourne F., Zakarian Joy, Burkham-Kreitner Susan, Richard Hofstetter C., Slymen Donald J., Keating Kristen, Russos Stergios, Jones Jennifer A.. 1997. " Tracking and Follow-up of 16,915 Adolescents: Minimizing Attrition Bias." Controlled Clinical Trials18(5):383–96.</bibtext> </blist> <blist> <bibtext> Moule Richard K., Decker Scott H.. 2013. " 'Hidden in Plain Sight': Locating the Men and Women of the 1954 Boston Special Youth Program." Journal of Qualitative Criminal Justice and Criminology1(2):267–91.</bibtext> </blist> <blist> <bibtext> Mulvey Edward P., Schubert Carol A., Piquero Alex R.. 2014. Pathways to Desistance—Final Technical Report. Washington, DC: National Institute of Justice, Office of Justice Programs.</bibtext> </blist> <blist> <bibtext> Murray Glenn F., Erickson Patricia G.. 1987. " Cross-sectional versus Longitudinal Research: An Empirical Comparison of Projected and Subsequent Criminality." Social Science Research16(2):107–18.</bibtext> </blist> <blist> <bibtext> Mutti Seema, Kennedy Ryan David, Thompson Mary E., Fong Geoffrey T.. 2014. " Prepaid Monetary Incentives-predictors of Taking the Money and Completing the Survey: Results from the International Tobacco Control (ITC) Four-country Survey." Sociological Methods & Research43(2):338–55.</bibtext> </blist> <blist> <bibtext> Mychasiuk R., Benzies K.. 2011. " Facebook: An Effective Tool for Participant Retention in Longitudinal Research." Child: Care, Health & Development38(5):753–56.</bibtext> </blist> <blist> <bibtext> Passetti Lora L., Godley Susan H., Scott Christy K., Siekmann Melissa. 2000. " A Low-cost Follow-up Resource: Using the World Wide Web to Maximize Client Location Efforts." American Journal of Evaluation21(2):195–203.</bibtext> </blist> <blist> <bibtext> Pickett Justin T., Metcalfe Christi Falco, Baker Thomas, Gertz Marc, Bedard Laura. 2014. " Superficial Survey Choice: An Experimental Test of a Potential Method for Increasing Response Rates and Response Quality in Correctional Surveys." Journal of Quantitative Criminology30(2):265–84.</bibtext> </blist> <blist> <bibtext> Posick Chad, Rocque Michael. 2018. Great Debates in Criminology. New York, NY: Routledge.</bibtext> </blist> <blist> <bibtext> Pyrooz David C., Decker Scott H.. 2019. Competing for Control: Gangs and the Social Order of Prisons. Cambridge, England: Cambridge University Press.</bibtext> </blist> <blist> <bibtext> Pyrooz David C., McGloin Jean Marie, Decker Scott H.. 2017. " Parenthood as a Turning Point in the Life Course for Male and Female Gang Members: A Study of within-individual Changes in Gang Membership and Criminal Behavior." Criminology55(4):869–99.</bibtext> </blist> <blist> <bibtext> Pyrooz David C., Mitchell Meghan M.. 2019. " The Use of Restrictive Housing on Gang and Non-gang Affiliated Inmates in U.S. Prisons: Findings from a National Survey of Correctional Agencies." Justice Quarterly1–27. doi:10.1080/07418825.2019.1574019.</bibtext> </blist> <blist> <bibtext> Rossman Sheli B., Roman John K., Zweig Christin H., Lindquist Christine H., Rempel Michael, Willison Janeen Buck, Mitchell Downey P., Fahrney Kristine. 2011. The Multi-site Adult Drug Court Evaluation: Study Overview and Design, Volume 1. Washington, DC: Urban Institute, Justice Policy Center.</bibtext> </blist> <blist> <bibtext> Sampson Robert J., Laub John H.. 1997. "A Life-course Theory of Cumulative Disadvantage and the Stability of Delinquency." Pp. 146–70 in Developmental Theories of Crime and Delinquency, edited by Thornberry T. P.. New York, NY: Routledge.</bibtext> </blist> <blist> <bibtext> Sandler Jenna, Feinberg Emily, Silva Christine, Chen Ning, Silverstein Michael. 2013, November 5. "Predictors of Phone Call Attempts to Complete Follow-up Assessments in Longitudinal Mental Health Research." Presented at the Annual Meeting of American Public Health Association.</bibtext> </blist> <blist> <bibtext> Senturia Yvonne D., Mortimer Kathleen McNiff, Baker Dean, Gergen Peter, Mitchell Herman, Joseph Christine, James Wedner H.. 1998. " Successful Techniques for Retention of Study Participants in an Inner-city Population." Controlled Clinical Trials19(6):544–54.</bibtext> </blist> <blist> <bibtext> Sugie Naomi F.2018. " Utilizing Smartphones to Study Disadvantaged and Hard-to-reach Groups." Sociological Methods & Research47(3):458–91.</bibtext> </blist> <blist> <bibtext> Texas Department of Criminal Justice. 2007. Security Threat Groups: On the Inside. Report provided by the Texas Department of Criminal Justice. Retrieved from https://<ulink href="http://www.tdcj.texas.gov/documents/Security%5fThreat%5fGroups%5fGRAD.pdf">www.tdcj.texas.gov/documents/Security%5fThreat%5fGroups%5fGRAD.pdf</ulink>.</bibtext> </blist> <blist> <bibtext> Tourangeau Roger. 2014. "Defining Hard-to-survey Populations." Pp. 3–20 in Hard-to-survey Populations, edited by Tourangeau R., Edwards B., Johnson T. P., Wolter K. M., Bates N.. Cambridge, England: Cambridge University Press.</bibtext> </blist> <blist> <bibtext> Ullman Sarah E.2011. " Longitudinal Tracking Methods in a Study of Adult Women Sexual Assault Survivors." Violence Against Women17(2):189–200.</bibtext> </blist> <blist> <bibtext> Vincent Kathryn B., Kasperski Sarah J., Caldeira Kimberly M., Garnier-Dykstra Laura M., Pinchevsky Gillian M., O'Grady Kevin E., Arria Amelia M.. 2012. " Maintaining Superior Follow-up Rates in a Longitudinal Study: Experiences from the College Life Study." International Journal of Multiple Research Approaches6(1):56–72.</bibtext> </blist> <blist> <bibtext> Walton Maureen A., Ramanathan Chathapuram S., Reischl Thomas M.. 1998. " Tracking Substance Abusers in Longitudinal Research: Understanding Follow-up Contact Difficulty." American Journal of Community Psychology26(2):233–53.</bibtext> </blist> <blist> <bibtext> Western Bruce. 2018. Homeward: Life in the Year after Prison. New York, NY: Russell Sage Foundation.</bibtext> </blist> <blist> <bibtext> Western Bruce, Braga Anthony A., Hureau David M., Sirois Catherine. 2015. " Study Retention as Bias Reduction in a Hard-to-reach Population." Proceedings of the National Academy of Sciences113(20):5477–85.</bibtext> </blist> <blist> <bibtext> Wright James D., Allen Tupper Lampton, Devine Joel A.. 1995. " Tracking Non-Traditional Populations in Longitudinal Studies." Evaluation and Program Planning18(3):267–77.</bibtext> </blist> <blist> <bibtext> Yammarino Francis J., Skinner Steven J., Childers Terry L.. 1991. " Understanding Mail Survey Response Behavior: A Meta-Analysis." Public Opinion Quarterly55(4):613–39.</bibtext> </blist> <blist> <bibtext> Yan Ting, Cantor David. 2019. " Asking Survey Questions about Criminal Justice Involvement." Public Health Reports134(Supplement 1):46–56.</bibtext> </blist> </ref> <aug> <p>By Chantal Fahmy; Kendra J. Clark; Meghan M. Mitchell; Scott H. Decker and David C. Pyrooz</p> <p>Reported by Author; Author; Author; Author; Author</p> <p></p> <p>Chantal Fahmy is an assistant professor in the Department of Criminology and Criminal Justice at the University of Texas at San Antonio and received her PhD in Criminology and Criminal Justice from Arizona State University. Her research interests focus on prisoner reentry and reintegration, social support and health, and institutional corrections. Her recent work has been published in Criminal Justice and Behavior, Environment and Behavior, and Journal of Adolescent Health.</p> <p>Kendra J. Clark is a graduate student in the department of Sociology at the University of Colorado Boulder. She received her BS in Psychology and Sociology from Baker University and her MA in Sociology from the University of Colorado Boulder. Her current work examines areas of incarceration, reentry, recidivism, and gangs. Her research is supported by a NSF Graduate Research Fellowship and she worked as a project staff member for the LoneStar Project—an NIJ funded study examining gangs and reentry experiences.</p> <p>Meghan M. Mitchell is an assistant professor in the Department of Criminal Justice at the University of Central Florida. She received her BS and MS from North Dakota State University and PhD in Criminal Justice from Sam Houston State University. Her current work examines the areas of corrections, subcultures, reentry, and inequalities. She has published in Justice Quarterly, Journal of Criminal Justice, Deviant Behavior, Criminal Justice Policy Review, Trends in Organized Crime, and Journal of Ethnicity in Criminal Justice.</p> <p>Scott H. Decker is foundation professor emeritus in the School of Criminology and Criminal Justice at Arizona State University. His main research interests are in gangs, violence, and active offenders. He is a Fellow in both the American Society of Criminology and the Academy of Criminal Justice Sciences. He is the author of Life in the Gang: Family, Friends and Violence (Cambridge, 1996), Policing Immigrants: Local Law Enforcement on the Front Lines (University of Chicago, 2016) and Competing for Control: Gangs and the Social Order of Prisons (Cambridge, 2019).</p> <p>David C. Pyrooz is associate professor of Sociology and Faculty Associate of Problem Behavior and Positive Youth Development at the Institute of Behavioral Science at the University of Colorado Boulder. His research interests are in the areas of gangs and criminal networks, incarceration and prisoner reentry, and developmental and life-course criminology. His recent books include Confronting Gangs: Crime and Community (Oxford, 2014), The Handbook of Gangs (Wiley-Blackwell, 2015), and Competing for Control: Gangs and the Social Order of Prisons (Cambridge, 2019).</p> </aug> <nolink nlid="nl1" bibid="bib36" firstref="ref2"></nolink> <nolink nlid="nl2" bibid="bib23" firstref="ref3"></nolink> <nolink nlid="nl3" bibid="bib19" firstref="ref5"></nolink> <nolink nlid="nl4" bibid="bib38" firstref="ref6"></nolink> <nolink nlid="nl5" bibid="bib64" firstref="ref7"></nolink> <nolink nlid="nl6" bibid="bib69" firstref="ref8"></nolink> <nolink nlid="nl7" bibid="bib28" firstref="ref10"></nolink> <nolink nlid="nl8" bibid="bib30" firstref="ref11"></nolink> <nolink nlid="nl9" bibid="bib31" firstref="ref12"></nolink> <nolink nlid="nl10" bibid="bib33" firstref="ref13"></nolink> <nolink nlid="nl11" bibid="bib44" firstref="ref14"></nolink> <nolink nlid="nl12" bibid="bib22" firstref="ref15"></nolink> <nolink nlid="nl13" bibid="bib29" firstref="ref17"></nolink> <nolink nlid="nl14" bibid="bib49" firstref="ref18"></nolink> <nolink nlid="nl15" bibid="bib54" firstref="ref19"></nolink> <nolink nlid="nl16" bibid="bib56" firstref="ref20"></nolink> <nolink nlid="nl17" bibid="bib59" firstref="ref21"></nolink> <nolink nlid="nl18" bibid="bib68" firstref="ref22"></nolink> <nolink nlid="nl19" bibid="bib47" firstref="ref23"></nolink> <nolink nlid="nl20" bibid="bib37" firstref="ref27"></nolink> <nolink nlid="nl21" bibid="bib67" firstref="ref28"></nolink> <nolink nlid="nl22" bibid="bib41" firstref="ref29"></nolink> <nolink nlid="nl23" bibid="bib46" firstref="ref30"></nolink> <nolink nlid="nl24" bibid="bib65" firstref="ref31"></nolink> <nolink nlid="nl25" bibid="bib35" firstref="ref32"></nolink> <nolink nlid="nl26" bibid="bib53" firstref="ref33"></nolink> <nolink nlid="nl27" bibid="bib52" firstref="ref35"></nolink> <nolink nlid="nl28" bibid="bib45" firstref="ref40"></nolink> <nolink nlid="nl29" bibid="bib25" firstref="ref41"></nolink> <nolink nlid="nl30" bibid="bib55" firstref="ref42"></nolink> <nolink nlid="nl31" bibid="bib26" firstref="ref43"></nolink> <nolink nlid="nl32" bibid="bib40" firstref="ref44"></nolink> <nolink nlid="nl33" bibid="bib15" firstref="ref45"></nolink> <nolink nlid="nl34" bibid="bib21" firstref="ref46"></nolink> <nolink nlid="nl35" bibid="bib20" firstref="ref47"></nolink> <nolink nlid="nl36" bibid="bib34" firstref="ref51"></nolink> <nolink nlid="nl37" bibid="bib57" firstref="ref52"></nolink> <nolink nlid="nl38" bibid="bib10" firstref="ref58"></nolink> <nolink nlid="nl39" bibid="bib11" firstref="ref59"></nolink> <nolink nlid="nl40" bibid="bib12" firstref="ref60"></nolink> <nolink nlid="nl41" bibid="bib13" firstref="ref61"></nolink> <nolink nlid="nl42" bibid="bib14" firstref="ref62"></nolink> <nolink nlid="nl43" bibid="bib50" firstref="ref64"></nolink> <nolink nlid="nl44" bibid="bib72" firstref="ref68"></nolink> <nolink nlid="nl45" bibid="bib16" firstref="ref69"></nolink> <nolink nlid="nl46" bibid="bib51" firstref="ref72"></nolink> <nolink nlid="nl47" bibid="bib17" firstref="ref77"></nolink> <nolink nlid="nl48" bibid="bib18" firstref="ref78"></nolink> <nolink nlid="nl49" bibid="bib744" firstref="ref88"></nolink> <nolink nlid="nl50" bibid="bib24" firstref="ref92"></nolink> <nolink nlid="nl51" bibid="bib66" firstref="ref93"></nolink> <nolink nlid="nl52" bibid="bib62" firstref="ref98"></nolink> <nolink nlid="nl53" bibid="bib27" firstref="ref99"></nolink> <nolink nlid="nl54" bibid="bib32" firstref="ref109"></nolink> <nolink nlid="nl55" bibid="bib58" firstref="ref117"></nolink> <nolink nlid="nl56" bibid="bib39" firstref="ref120"></nolink> <nolink nlid="nl57" bibid="bib43" firstref="ref131"></nolink>
Header DbId: eric
DbLabel: ERIC
An: EJ1343366
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Method to the Madness: Tracking and Interviewing Respondents in a Longitudinal Study of Prisoner Reentry
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Fahmy%2C+Chantal%22">Fahmy, Chantal</searchLink> (ORCID <externalLink term="https://orcid.org/0000-0002-6010-8457">0000-0002-6010-8457</externalLink>)<br /><searchLink fieldCode="AR" term="%22Clark%2C+Kendra+J%2E%22">Clark, Kendra J.</searchLink><br /><searchLink fieldCode="AR" term="%22Mitchell%2C+Meghan+M%2E%22">Mitchell, Meghan M.</searchLink><br /><searchLink fieldCode="AR" term="%22Decker%2C+Scott+H%2E%22">Decker, Scott H.</searchLink><br /><searchLink fieldCode="AR" term="%22Pyrooz%2C+David+C%2E%22">Pyrooz, David C.</searchLink>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="SO" term="%22Sociological+Methods+%26+Research%22"><i>Sociological Methods & Research</i></searchLink>. Aug 2022 51(3):1282-1324.
– Name: Avail
  Label: Availability
  Group: Avail
  Data: SAGE Publications. 2455 Teller Road, Thousand Oaks, CA 91320. Tel: 800-818-7243; Tel: 805-499-9774; Fax: 800-583-2665; e-mail: journals@sagepub.com; Web site: http://sagepub.com
– Name: PeerReviewed
  Label: Peer Reviewed
  Group: SrcInfo
  Data: Y
– Name: Pages
  Label: Page Count
  Group: Src
  Data: 43
– Name: DatePubCY
  Label: Publication Date
  Group: Date
  Data: 2022
– Name: SourceSuprt
  Label: Sponsoring Agency
  Group: SrcSuprt
  Data: US Department of Justice, Office of Justice Programs
– Name: NumberContract
  Label: Contract Number
  Group: NumCntrct
  Data: 2014MUCX0111
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: Journal Articles<br />Reports - Research
– Name: Subject
  Label: Descriptors
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Longitudinal+Studies%22">Longitudinal Studies</searchLink><br /><searchLink fieldCode="DE" term="%22Correctional+Institutions%22">Correctional Institutions</searchLink><br /><searchLink fieldCode="DE" term="%22Institutionalized+Persons%22">Institutionalized Persons</searchLink><br /><searchLink fieldCode="DE" term="%22Males%22">Males</searchLink><br /><searchLink fieldCode="DE" term="%22Law+Enforcement%22">Law Enforcement</searchLink><br /><searchLink fieldCode="DE" term="%22Program+Effectiveness%22">Program Effectiveness</searchLink><br /><searchLink fieldCode="DE" term="%22Surveys%22">Surveys</searchLink><br /><searchLink fieldCode="DE" term="%22Interviews%22">Interviews</searchLink><br /><searchLink fieldCode="DE" term="%22Juvenile+Gangs%22">Juvenile Gangs</searchLink><br /><searchLink fieldCode="DE" term="%22Group+Membership%22">Group Membership</searchLink><br /><searchLink fieldCode="DE" term="%22Attrition+%28Research+Studies%29%22">Attrition (Research Studies)</searchLink><br /><searchLink fieldCode="DE" term="%22Crime%22">Crime</searchLink><br /><searchLink fieldCode="DE" term="%22Telecommunications%22">Telecommunications</searchLink><br /><searchLink fieldCode="DE" term="%22Meetings%22">Meetings</searchLink><br /><searchLink fieldCode="DE" term="%22Social+Media%22">Social Media</searchLink>
– Name: Subject
  Label: Geographic Terms
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Texas%22">Texas</searchLink>
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1177/0049124119875962
– Name: ISSN
  Label: ISSN
  Group: ISSN
  Data: 0049-1241
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: After nearly four decades of growth, the number of people held in U.S. prisons has begun to decline. In an era of decarceration, social scientists need to understand prisoner reentry experiences. Longitudinal studies are one strategy to accomplish this goal. Yet, the retention of a formerly incarcerated population across waves of interviews is challenging due to their transient lifestyles and limited support systems, which may be further complicated by gang involvement. This article details the longitudinal follow-up procedures used in the LoneStar Project--a multiwave study of 802 males first interviewed in prison and reinterviewed twice in the year postrelease--to build rapport, complete interviews, and minimize attrition. We then evaluate the effectiveness of our procedures on important outcomes including interview yields, appointments, and incoming calls. Results indicate that any outgoing contact with respondents via appointment reminders and other reciprocal modes of contact lead to greater project engagement and a greater likelihood of interview completion. We conclude with relevant takeaways for researchers seeking to maximize survey participation with hard-to-reach populations.
– Name: AbstractInfo
  Label: Abstractor
  Group: Ab
  Data: As Provided
– Name: DateEntry
  Label: Entry Date
  Group: Date
  Data: 2022
– Name: AN
  Label: Accession Number
  Group: ID
  Data: EJ1343366
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=eric&AN=EJ1343366
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1177/0049124119875962
    Languages:
      – Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 43
        StartPage: 1282
    Subjects:
      – SubjectFull: Longitudinal Studies
        Type: general
      – SubjectFull: Correctional Institutions
        Type: general
      – SubjectFull: Institutionalized Persons
        Type: general
      – SubjectFull: Males
        Type: general
      – SubjectFull: Law Enforcement
        Type: general
      – SubjectFull: Program Effectiveness
        Type: general
      – SubjectFull: Surveys
        Type: general
      – SubjectFull: Interviews
        Type: general
      – SubjectFull: Juvenile Gangs
        Type: general
      – SubjectFull: Group Membership
        Type: general
      – SubjectFull: Attrition (Research Studies)
        Type: general
      – SubjectFull: Crime
        Type: general
      – SubjectFull: Telecommunications
        Type: general
      – SubjectFull: Meetings
        Type: general
      – SubjectFull: Social Media
        Type: general
      – SubjectFull: Texas
        Type: general
    Titles:
      – TitleFull: Method to the Madness: Tracking and Interviewing Respondents in a Longitudinal Study of Prisoner Reentry
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Fahmy, Chantal
      – PersonEntity:
          Name:
            NameFull: Clark, Kendra J.
      – PersonEntity:
          Name:
            NameFull: Mitchell, Meghan M.
      – PersonEntity:
          Name:
            NameFull: Decker, Scott H.
      – PersonEntity:
          Name:
            NameFull: Pyrooz, David C.
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 08
              Type: published
              Y: 2022
          Identifiers:
            – Type: issn-print
              Value: 0049-1241
          Numbering:
            – Type: volume
              Value: 51
            – Type: issue
              Value: 3
          Titles:
            – TitleFull: Sociological Methods & Research
              Type: main
ResultId 1