Social Identity Mapping as a Tool for Data Collection in Adolescent Recovery: A Mixed Methods Adaptation Study

Abstract

Background

Social group memberships and group behaviors are pivotal in supporting or inhibiting adolescents’ addiction recovery. Yet most research on social processes has relied on traditional surveys that cannot capture the complex nature of evolving social networks. Social Identity Mapping in Addiction Recovery (SIM-AR) asks respondents to create a visual representation, or map, of their social network, social groups, and social influences. The SIM-AR thus provides a novel, interactive, and detailed means to collect social network information, and does so in a way that is accessible and meaningful to participants.

Objectives

This study aimed to pilot and adapt the SIM-AR with adolescents and young adults in addiction treatment and recovery programs so it might be used as a standardized tool for data collection.

Design

This was a multi-method study that piloted the SIM-AR approach in both paper-pen and online (oSIM-AR) formats, and used this feedback to update the approach.

Methods

Participants (N = 35; Ages 14-19; 71% male) were recruited from addiction treatment and recovery programs to complete a facilitated SIM-AR/oSIM-AR, brief survey, and interview (November 2020-January 2023). We engaged participants in an iterative process so the approach could be modified: 28 completed the paper-pen version and seven completed the online version. Descriptive qualitative analysis was used to examine participant interviews.

Results

Adolescents found the SIM-AR process acceptable, “fun,” and “helpful.” The paper-pen version took ∼30 minutes (M=32.52, SD=11.07) and the online version took <15 minutes (M=13.75, SD=4.31). The paper-pen version generated longer reflection interviews than did the online format (M=25.75, SD=7.46 vs. M=9.46, SD=1.07 min.), suggesting the approaches prompted different degrees of reflection.

Conclusion

Support was found for the SIM-AR as an interactive, easy-to-use, and time-efficient measurement tool with adolescents. We plan to implement the SIM-AR in recovery programs to examine whether these programs generate network change.

Registration

https://osf.io/8vdcp/?view_only=02544c530b0746dd812a78b9d8c008b3; https://osf.io/4mrp6/?view_only=dca62d77b1dd4e829ea15f3ba1a90611.

Introduction

In the US, over 2 million adolescents experience the onset of a substance use disorder (SUD) annually, leading to a variety of short- and long-term negative consequences for their development, well-being, and future prospects. ¹ Adolescence is a pivotal developmental period and a vulnerable time for the onset of SUD. Alcohol and other drug (AOD) exposure affects the developing brain, slowing the maturity of important brain regions,^2-4 which can compromise the achievement of developmental milestones. Early and/or heavy substance use can lead to SUD and dependence. Despite evidence-based intervention and SUD treatments, adolescents with SUD often suffer rapid relapse to use.^5,6 One key challenge undermining stable recovery is that the surrounding social environment can pose many risks for adolescents attempting recovery. Alternatively, when the appropriate supports are in place, the social environment can support and serve as a protective factor for recovery.

Social influence is a major driver of health risk behaviors including alcohol and other drug use.^7-11 A person’s social identity is defined by how they view themselves in their social worlds, i.e., the groups to which they feel they belong, and it is developed through self-categorization processes. ¹² That is, through observing group norms, values, and behaviors, an individual perceives behaviors that will be acceptable and perceived as indicating belonging to that group. ¹³ Additionally, some groups may have more influential or central leaders, i.e., individuals who set the tone for the group norms and behaviors. ¹⁴ Thus, a person’s social identity can change depending on the social setting in which they find themselves. A sense of social identity, or belongingness, is especially important in adolescence, as the person looks to their peer groups to define themselves. ¹⁵

Depending on how much time is spent with each of these groups and the group composition,^16,17 different patterns of high- or low-risk behavior may emerge. The risks for engaging in substance use are greater for adolescents within homes where family members use substances or do not monitor their child’s behaviors, and risks are also greater when they have peers and friends who use substances.^18-21 Adolescents with family members who use or provide alcohol, and adolescents who spend more time with close friends or peers who use substances, are at substantially greater risk of substance use, developing a SUD, and having increased difficulty with recovery. ¹⁶

The social identity model of recovery (SIMOR) and social identity model of transition (SIMOT) propose that through engaging with treatment or other recovery supports, social networks should begin to shift to more abstinent or recovery-supportive groups.^22,23 These shifts to more recovery-oriented networks can aid recovery by enhancing one’s ability to cope, boosting self-efficacy and motivation for sobriety, and by establishing behavioral norms and models for social functioning that do not rely on substance use, ultimately leading to shifts in one’s social identity.^22-26

Group membership can be complex during adolescence, representing a liminal stage of identity development, during which time adolescents are essentially “trying on” different social groups and identities, and learning how to define themselves in the context of these groups.^12,27,28 Friends and peers are often highly influential during this period, as they look to these external social groups to begin to differentiate themselves from their family identity. Indeed, this defining process is often undertaken as a way to distinguish new and alternative social pathways, whether relative to family, schooling, or other establishment figures. This can make social networks and norms more difficult to track among adolescents because it can involve identity shifts as adolescents are asked by parents and providers to change relationships viewed as problematic to their recovery. Group memberships may also be disrupted with frequent entry and discharge from treatment programs.^29-31

Existing adolescent social network measures, including structured interview and survey approaches (e.g., Important People Inventory ³² ), provide valuable information regarding peer substance use exposure and relationship risk. However, these approaches are generally less equipped to capture the dynamic and overlapping group identities that characterize adolescent social worlds, including how adolescents situate themselves across multiple peer contexts, how groups interact with one another, and how social identities shift during recovery.

Social Identity Mapping (SIM^33-36) is an alternative to the traditional survey approach to address these gaps by using a visual and interactive mapping process designed to capture group membership, identification, inter-group relationships, perceived recovery support, and substance use norms simultaneously within a broader social identity framework. To date, the tool has been used in a range of different contexts, from workplaces to universities, and its methodological delivery allows for the capture of measures of importance for each context (including the integration of questions tailored for different kinds of participants). For the purpose of this study, participant questions were adapted to include measures quantifying the status of group members’ substance use behaviors (i.e., to examine group norms related to substance use). They were also adapted to ensure the capture of other related characteristics important to understanding relationships in recovery, such as how much inter-relational conflict is present and how much different group identities may have in common.^37-39

In contrast to a typical survey approach for acquiring social network information, the SIM process asks respondents to create a visual representation of their social networks, including individuals or groups that present support for, or barriers to, recovery. In light of addiction research and theory indicating the importance of situating the self amidst a one’s social group connections,^22,34,40 for this study we integrated a visual point in the maps to indicate the participant’s own self (a “Me” circle). This encourages respondents to view their social networks, and their influences, specifically in relation to themselves. Respondents are then asked to portray their various groups (e.g., home, school, friends, sports teams, etc.) graphically around the “Me” node at the center of the map. They are then asked to indicate the importance of each group, how much they feel they identify with each group, the perceived substance use behaviors of groups, and to characterize the degree of conflict between groups and the perceived levels of commonality between groups (i.e., how much groups might have in common with each other). These variables were selected to operationalize key social identity elements during the recovery processes.^22,23 For example, perceived group substance use reflects the behavioral expectations associated with different group identities, while ratings of group importance and identification indicates the relative salience of different social identities during recovery. Similarly, indicators of conflict and commonality between groups captures tensions between recovery-supportive and substance-using social identities, and the degree to which adolescents perceive compatibility across different social groups or sub-networks.

Although the SIM has been used internationally to collect social network data, including with samples in recovery,^41-43 the method was not designed specifically for use with adolescents, who have different experiences and understandings of their social world compared to adults. There is also currently no standardized protocol for the administration of the SIM-AR with adolescent research participants.^33,34,41 Consistent with lived experience approaches to instrument development, our work to adapt the SIM-AR for adolescents sought to ensure that the measure reflected youths’ understandings of their social worlds and their recovery experiences. Engaging adolescents in the refinement process may improve the developmental relevance, acceptability, and usability of social network assessment tools in this population. This is particularly important for a visualization tool that can be used to engage young people in an understanding of the role of social networks in SUD and recovery processes.

Aims

This report presents a description of the process of developing the SIM-AR method for use with youth with SUD. The aims of the present pilot study were to pre-test and refine SIM-AR administration protocols and delivery with adolescents. Data collection was conducted using the SIM-AR, and qualitative interviews were used to examine respondents’ experience and reflections on the process. Quantitative usability indicators (e.g., time to facilitate) and qualitative interview data were examined complementarily to evaluate different aspects of feasibility and acceptability. The objective was to produce a standardized administration of the SIM-AR that could be easily implemented in future research with adolescents. As a secondary goal, the study aimed to explore whether there were aspects of the SIM-AR process that could be leveraged in a novel intervention focused on examining adolescent recovery social supports and initiating motivation for increasing them. Finally, although we did not set out to compare the paper-pen and online versions, this study also provides a natural context for exploring and comparing both modalities.

Methods

The internal review boards for the University of Connecticut (UConn) and for Massachusetts General Hospital (MGH) approved all human subject protocols for the study (UConn: 7/1/2020 and MGH: 02/19/21). The study protocol was registered twice on the Open Science Framework (OSF): (1) the overall study, which focused on the paper-pen version of the SIM-AR (https://osf.io/8vdcp/?view_only=02544c530b0746dd812a78b9d8c008b3, 10/4/2021), and (2) the portion of the study that used the electronic version, online social identity mapping (oSIM; https://osf.io/4mrp6/?view_only=dca62d77b1dd4e829ea15f3ba1a90611, 9/27/2022).

Inclusion Criteria and Recruitment

From November 2020 to January 2023, we purposively sampled youth ages 12-19 who were currently attending treatment or engaged in a recovery support service (e.g., Recovery High School [RHS]^44,45) for their AOD use, or had recently seen a provider for their AOD use. Although we planned to limit the sample to only adolescents receiving outpatient treatment, this criterion was expanded when recruitment became too difficult with the COVID-19 restrictions. The Principal Investigator (PI) directly contacted local treatment centers and RHSs for recruitment and posted online advertisements to the broader community (e.g., through the MGH Rally online recruitment page). A total of 43 youth were interested and ultimately, 35 youth agreed and consented to participate in this study (81%). Participants under 18 years of age were consented with a parent, and provided assent. All participants received a $35 gift card for their participation. All participants completed a brief demographic survey, a facilitated SIM-AR, and an interview. See Table 1 for participant characteristics and Figure 1 for an overview of the study timeline.

Table 1.

Participant Characteristics

​	M (SD) or %
Age	16.97 (1.54)
Race/Ethnicity
African American	11.43%
American Indian	2.86%
Hispanic	22.86%
White	85.71%
Other	8.57%
Not reported	2.86%
Gender
Female	22.86%
Male	71.42%
Non-binary	5.71%
Other	2.86%
Not reported	0%

Figure 1.

Study Timeline of Procedures and Adaptations. Note. Rally = Online community recruitment board. RHS = Recovery High School. TX = Treatment

Sample Size and Information Power

When designing the study and planning for the appropriate sample size, we considered the concept of information power, or the idea that the more information the sample holds relevant to the study aims and questions (e.g., by virtue of their experience, characteristics, etc.), the lower number of participants is needed.^46,47 As a result, this study’s purposive sampling strategy included youth attending a variety of treatment and addiction recovery supports to ensure we gathered diverse adolescent perspectives. This suggests there may be adequate power for the study focus; however, given that we were recruiting from programs focused on supporting recovery, the results may not generalize to non-treatment seeking youth, or to youth who have not engaged with substance use programming.

Data Collection and Measures

We started with the 2018 SIM-AR tool published by Beckwith and colleagues as a foundation for the study, ³⁴ but this publication did not include a facilitator guide, and queries with the authors indicated that no specific guide for conducting SIM-AR had been produced. We therefore developed a guide for use with the first SIM-ARs, based on recommendations from the SIM-AR developers and our own experience interviewing youth. The guide included a brief introduction to the SIM-AR task and then guided participants through the creation of the SIM-AR, step-by-step. The guide was used for all study sessions. Following the SIM-AR exercise, participants were given a short break and then administered a semi-structured interview.

Early study visits were conducted by the study’s PhD-level female PI with each participant individually over Zoom. Later study visits also comprised several small groups conducted over Zoom (when participants were students at the same RHS). As COVID-19 restrictions lifted, the PI and study team conducted study visits onsite at two RHSs in both individual and group formats. School staff supported the study visits by sharing the study recruitment materials with students and letting us know if they wanted to be approached for a study visit and by finding a private room for the interview/focus group to take place. All interviews were recorded and transcribed. Study visits conducted over Zoom were recorded and transcribed.

Survey Measures

The brief demographic survey consisted of modified questions from the NIH All of Us demographic battery ⁴⁸ asking for participant age (years), ethnicity (check all that apply: African American, American Indian, Asian, Hispanic, White, Other), and gender (male, female, non-binary, other), as well as the start date of school (if currently attending) and start date of treatment (if currently in treatment).

SIM-AR and oSIM-AR

Participants were told that we were interested in understanding their social networks, and that they should depict those social networks or groups using a map. Participants were first asked to draw a circle with “ME” in the center of their map (the online version started with this already in place). Both versions of the mapping process then captured the following variables about an individual’s social network, albeit in slightly different ways, as this report will detail: social groups (number and name/type); social group size, importance, level of identification with the group, alcohol use, drug use; time spent with group; conflict between groups and between the self and groups; and commonality between groups. One of the map versions we piloted also included a measure of centrality of each group member within the group, which was used to assess the behavior of group leaders or influential group members (labeled “Centrality”; described further below in SIM, Version 3.0 (Circle) ). See Table 2 for a listing of each variable collected on the SIM-AR and oSIM-AR and how it was collected.

Table 2.

SIM-AR and oSIM-AR Versions: Types of Data Collected and Location on SIM-AR/oSIM-AR

​	V 1.0 and 2.0 (SIM-AR square)	V 3.0 (SIM-AR circle)	V 4.0 (oSIM-AR)
Social Group (descriptive group label, e.g., family, school friends, work)	Written on post-it	Written on post-it	Typed on circle
Number of group members a	Written on post-it	Written on post-it	Answer to question
Alcohol use of each group member	Sticker dot, Color coded	Sticker dot, Color coded	​
Options: Heavy, casual, no use, in recovery
Alcohol use across the group (perceived group average) a	​	​	Answer to question
Options: Heavy, casual, no use, in recovery
Substance use of each group member	Sticker dot, Color coded	Sticker dot, Color coded	​
Options: Heavy, casual, no use, in recovery
Substance use across the group (perceived group average)	​	​	Group dot, Color coded
Options: Heavy, casual, no use, in recovery
Centrality of each group member	​	Sticker dot placed on one of two concentric circles within each group’s post-it	​
Options: More central (inner circle), less central (outer circle)
Days spent with group in past month	Written on post-it	Written on post-it	Answer to question
Range: 0-30
Level of importance of group a	Select size of post-it	Select size of post-it	Entire group dot changes size
Level of identification with group	Written on post-it	Written on post-it	Answer to question
Range: 1-7
Conflict between self and each group	Color coded line between groups	Color coded line between groups	Color coded line between groups
Conflict between each group
Options: None, Low, High b
Commonality between self and each group	Color coded line between groups	Color coded line between groups	​
Options: None, Low, High
Commonality between each group
Options: None, Low, High

Note. Empty boxes indicate that element is not collected on that map type. When “answer to question” is indicated, it means the participant provides the answer and then the question disappears for the that group so it is not visually represented on the map, unless the participant opens the question again.

^aIn the oSIM-AR, there are 5 options for group importance, while in the SIM-AR there are only three.

^bIn the oSIM-AR, there is also an option for Medium Conflict.

The paper-pen version of the SIM-AR process used the following materials: example map and key, notebook paper for brainstorming groups, large white paper for creating the SIM-AR, pen for naming groups and adding information about each group, sticky notes for adding each group to the SIM-AR (three sizes), colored pencils for drawing conflict and commonality lines between groups, and circle stickers for indicating the alcohol and drug use of group members (five colors, two sizes).

For virtual study visits, schools hosting the study visit and individuals participating at home were mailed study packets in which all the materials needed for the study visit were included. The oSIM-AR is an online program, so it only requires that participants can access a device with internet access. Once in the portal, participants point and click to add groups to their social network and are prompted to add details about each group they list on their map.

Interview

The interview guide was developed through consultation with the research team. Participants were told that our study team had a tool developed for adults and wanted to get their feedback so we could develop the social identity mapping as a tool for data collection to better understand their social networks during a recovery journey. The interview involved asking questions about participants’ reflections on their completed SIM-AR/oSIM-AR, followed by questions about the process of completing it. Although this was not a formal cognitive interviewing study, some interview prompts were informed by cognitive interviewing approaches ⁴⁹ and explored how adolescents interpreted key SIM-AR components and understood the mapping prompts. Table 3 includes the semi-structured interview guide and provides additional information on how the questions were used to collect data to inform future use of the method.

Table 3.

Interview Questions and Their Purpose to Examine Potential Adaptations

Interview topics and questions	Aim of question
Process-Oriented Questions
1. How did you find the process of completing the map?	• Feedback on instructions for structured facilitation guide • Feedback on materials needed to create SIM-AR • Participant acceptability of the approach
2. Were there things you found difficult about completing the SIM?	• Feedback on instructions for structured facilitation guide • Feedback on materials needed to create SIM-AR
3. When I asked you to draw compatibility lines, what sorts of things did you consider?	• Check on internal validity of SIM-AR components • Assess whether core concepts are easy to understand – or if instructions should be modified
4. When I asked you to draw conflict lines, what sorts of things did you consider?
*Similar style of questions asked for other SIM components.
5. Are there important things to know about your relationships and social networks that the SIM did not capture?	Assess whether: • There are additional key concepts to include on the map • There other key concepts that cannot be included on the map but would be good to ask about in a questionnaire
6. Are there important things to know about your relationships and recovery supports that the SIM did not capture?
7. Is there anything else you’d like to tell me about your experience in completing the SIM today that I have not asked about?	Capture any other feedback
Reflection-oriented questions
1. When you look at your finished map, what do you see, and what does it make you consider about yourself?	• Explore whether the SIM-AR process leads to reflection and the potential for intervention development
2. Did you learn anything new about your relationships and social networks through creating your map?
3. What does your map tell you about the recovery journey you are on?
4. Did anything in your map surprise you?
5. If you think ahead to 6 months from now, what would you like your map to look like? a. And what would you not want it to look like 6 months from now?
6. Do any of the groups have a positive influence on your life? (If yes: which ones and how?)	• Assess whether there is additional information about group influence that needs to be captured outside of the substance use characteristics of group members captured
7. Do any of the groups have a negative influence on your life? (If yes: which ones and how?)
8. Now, consider for each group, how easy or difficult it would be for you to leave that group. By leave a group, that could be that you stop returning text messages or phone calls or stop spending time with members of that group. Let’s start with the group you think would be easiest to leave.	• Collect more nuanced data on how the importance and identity of groups captured on the maps might play out in real life
9. If I had asked you to complete a social identity map 6 months ago, do you think your map would look different? a. If YES, how so?	• As the pilot is cross-sectional, assess if they can self-identify changes that have occurred • Consider how to examine network changes due to the COVID pandemic restrictions

Data Management

Ultimately, we had several versions of the paper-pen SIM-AR (described in the results as “SIM, Version 1.0 – 2.0 (Square)” and “SIM, Version 3.0 (Circle)”) and one version of the oSIM-AR.

Paper-Pen SIM-AR Data

We initially asked participants to mail the completed SIM-AR in a postage-paid envelope, but after one went missing and others had areas for clarification, we asked participants to take a picture of their map and email the image to us. This procedure allowed us to view the SIM-AR during the break between the SIM-AR exercise and the interview and immediately follow up with clarifying questions. It also reduced participant burden and meant that they could keep their map if they wished. The data provided by the completed SIM-ARs (full list in Table 2) were entered into a REDCap database instrument built for this project by one research coordinator and checked by another. (The REDCap XML database structure is freely available for download and use for data entry on our OSF project page.) Questions or discrepancies in data entry were discussed by the research team and resolved via consensus.

oSIM-AR Data

Once submitted by the participant, data from the oSIM-AR were automatically transformed into variables, including basic summary variables (e.g., total number of groups in a participant’s network) and could be exported from the online portal as CSV and Excel files. ⁵⁰

Interview Data

The interview transcriptions were entered into an Nvivo project for data management. ⁵¹ The analysis of the interviews used a Descriptive Qualitative Approach, which is used when seeking “to discover and understand… a process, or the perspectives” of individuals involved.^52,53 This approach is appropriate where the researcher is looking to understand discrete categories or concepts.

Adaptation Process

Administration of the SIM-AR and oSIM-AR was facilitated by the guide developed for this study. This guide was iteratively modified as the facilitators developed experience with the procedures and received participant feedback, with most modifications happening in the initial piloting of the study. Figure 2 provides the different iterations of the SIM-AR within both the paper-pen and the online versions. Table 4 provides a detailed summary of the adaptation process.

Figure 2.

SIM and oSIM Adaptations. (A) Indicates the original SIM with legend ( ³⁴ ); (B) indicates the final adapted version of the SIM using square post-its (SIM 2.0, Square); (C) indicates the final version of the SIM using circles to capture group member centrality (SIM 3.0, Circle); (D) indicates the final adapted oSIM

Table 4.

SIM-AR and oSIM-AR Adaptation Process

Version	Change	Rationale	Source of feedback
Original → SIM-AR 1.0 (prior to piloting)	1. Added days spent on average in a month	Ensure we were capturing appropriate data points for research	Review of literature
	2. Added level of identification with group rating
SIM-AR 1.0 → SIM-AR 2.0 (square)	1. Placement of alcohol and substance use dots for each member (small dot placed on large dot)	Improve understanding of social group member use of substance	Facilitator review of data after mapping with 5 participants
	2. Change color of conflict lines – make them all jagged instead of using shape of line (squiggly v. jagged) to differentiate conflict levels	Difficult to understand different conflict levels	Facilitator observations and discussion with participant - participant suggestion to use colors instead to reduce confusion when viewing the map.
	3. Change color of commonality lines – instead of using single and double lines to represent degree of commonality between groups	Be consistent with approach for conflict; Improve visualization of map	Facilitator observation, participant feedback that the single and double lines (along with other map elements) made the map look messy.
SIM-AR 3.0, Circle	1. Use circle-style post-its instead of squares	Capture influential group members, specifically their AOD use behaviors	Early participant feedback about the importance of a specific person within a group; review of literature
	2. Add concentric circles within the post-it to represent centrality - group members are placed on these lines
oSIM-AR	Added the following rating questions for the components not captured on the oSIM that were captured in the SIM-AR:	Ensure consistency between the paper-pen and online version	Experience with the paper-pen version; collaboration with University of Queensland
	1. Group size: added a rating question for each group asking how many members were in that group.
	2. Capture alcohol use separately from other substances: used the rating option to add a question about the group’s level of alcohol use and kept the circle color option to change the group color based on the substance use level.
	3. Add questions for each group on (1) level of identification and (2) time spent.
	4. Used the comment box to prompt participants to share how much each group supported their recovery.

SIM, Version 1.0 – 2.0 (Square)

Before we piloted the SIM-AR using the original published version by Beckwith and colleagues, ³⁴ we added new prompts for participants to indicate during the mapping how many days they spent with each group on average and how much they felt that they identified with the group (scale 1-7 with 1 being the least and 7 the greatest). These variables were important to gather from a research perspective, where the amount of time spent has been an important predictor of outcomes ¹⁶ and self-identification with the group is also important to capture when considering social identity change. ⁴³

After data collection with five participants using SIM Version 1.0, three areas were improved, resulting in SIM Version 2.0 (see Figure 2B): (1) drug and alcohol of group members, (2) conflict lines, and (3) commonality lines. In SIM Version 1.0, alcohol and substance use for each group member were depicted separately. ³⁴ This meant it was unclear if the same person with, for example, high alcohol use, had a similar pattern of drug use. In Version 2.0, we asked participants to first place the alcohol use dots (larger) for each group member on the SIM-AR and then place the smaller substance use dots (smaller) for each group member on top of the larger dot. Thus, the perceived alcohol and drug use behaviors for the same group member were indicated on the same placement. Lines indicating low or high conflict between groups or group members were drawn in the same color for Version 1.0. The instructions were for them to be drawn squiggly with low conflict and jagged with high conflict, a difference that became difficult to distinguish depending on each participant’s artistic style. One participant suggested color-coding them, and we subsequently updated the approach for Version 2.0 so participants would use different colors for conflict levels (orange = low, red = high) and draw them as jagged. For Version 2.0 we also adapted the commonality lines to be color-coded for the degree of commonality (green = low, blue = high), to simplify instructions and be consistent with our approach to capturing conflict.

SIM, Version 3.0 (Circle)

During the interviews using earlier versions, several participants commented on the influence of a specific person on their recovery experience. Although the SIM-AR collects information about group members’ level of substance use, it does not capture the extent to which group AOD use norms were driven by a group leader or influential group members (i.e., group member centrality). We revised the SIM-AR to capture this dynamic and piloted this version. Circle-shaped Post-its in three sizes (matching the three sizes used to indicate group importance in the square SIM-AR) were printed with two sets of concentric circles on them. The inner concentric circle represented group leaders or those considered central to the group. The outer concentric circle represented members of the group who were considered less central to the group. Participants were asked to place sticker dots for each member on the circle based on how central to the group they considered that member.

Although this solution captured group member centrality, it was difficult to facilitate. Another challenge with the use of the circle-shaped Post-it notes is that although the different sizes were clear when placing them on the SIM-AR when we took a photo of the completed SIM-AR for study documentation, the sizing differences (representing group importance) disappeared visually. The person entering data thus used a ruler to ensure they were coding the correct circle size. We eventually determined that this was too much to ask of participants at once, i.e., to consider each group member’s substance use and select the appropriate color, then when placing the sticker dot for that member on the group circle, think about how central they are to the group and place them on the appropriate circle line. Because we were using stickers, they could not easily move a group member to a different central point after reflection. One participant suggested bingo chips as an alternative approach. Although that is an option that could work in an in-person setting, it would mean that taking the map with them would prove difficult unless they took a picture.

oSIM-AR

The oSIM tool was developed to easily capture social network data, but the version available for licensing did not capture some of the information we had on the SIM-AR. We collaborated with the team at the University of Queensland to adapt the oSIM to our specific study questions. Many of the SIM-AR aspects were easily incorporated in the oSIM: (1) groups are added with the click of a button allowing the participant to type in the group name; (2) participants rated the importance of the group and the circle size changed accordingly; (3) participants rated the substance use of the entire group and the color changed (we used the same color options to indicate risk level of use in the paper-pen version); (4) the conflict lines were drawn by making connections between circles, asking the participant to rate the level of conflict, and the lines changed color (we used the same color options to indicate conflict level in the paper-pen version); (5) each group has both a rating question section and a comment section such that researchers can add questions about groups on a sliding scale or add an overall question for the participant to type in a short answer. There remained some differences between the paper-pen and online approaches, some of which we used to revise the program through adding new rating questions (see Table 4).

Qualitative Analysis and Positionality

One graduate student researcher who had not engaged in the study visit reviewed the transcripts and drafted the initial codebook. The codebook was then discussed by three members of the study team that included the study PI, BA-level intern, and graduate student and refined prior to moving on to coding transcripts in an iterative process with regular study team meetings. The codebook had both substantive codes related to adolescent recovery, which are reported elsewhere,^31,54 and the following codes on the method and adaptation process, of focus in this report: (1) feelings about the SIM-AR process or its results; (2) SIM-AR process including feedback and recommendations; (3) SIM-AR materials.

Results

Methodological Differences in Administration Between SIM-AR and oSIM-AR

Facilitation

Our study protocol required the use of the facilitator guide for the mapping process. From our experience administering the mapping to participants, the paper-pen version was more likely to require a facilitator present. (We had one enthusiastic participant complete a map prior to the study visit using the materials and the template with little error. Since we mailed them only one set of materials, during the study visit, we had them update the map with the facilitator prior to the reflection interview). Participants completing the oSIM-AR seemed able to do it quickly and easily, even trying to jump ahead of the facilitator. With either type of administration (SIM-AR or oSIM-AR) participants could keep the completed maps (i.e., complete the map and then take a picture to send to the research team). Those administered the oSIM-AR could save a PDF copy of the completed map before submitting it electronically to the research team.

Time Required

For participants doing the paper-pen version, the mapping portion lasted 19-63 minutes (M = 32.52, SD = 11.07), and the interview lasted 10-45 minutes (M = 25.75, SD = 7.46). For participants doing the online version, the mapping portion lasted 9-20 minutes (M = 13.75, SD = 4.31), and the interview lasted 8-10.54 minutes (M = 9.46, SD = 1.07). Thus, the oSIM-AR took about half the amount of time as the SIM-AR, and its reflection interview also took about half the time, despite using the same structured interview guide.

Qualitative Perspectives on Key SIM-AR/oSIM-AR Elements

Below, we present qualitative findings from the SIM-AR/oSIM-AR process and post-exercise interviews related to several key components of the mapping exercise. We focus on these elements because they help demonstrate whether the map was capturing the social and recovery-related constructs it was theoretically designed to assess, while also informing the feasibility of the adapted approaches. Exemplar quotations illustrating these findings are provided in Table 1 in the Appendix.

Importance and Identification

There were variations in how participants discussed importance of versus identification with groups, indicating that these are distinct elements to capture. For a lot of participants, the importance of groups was determined by how they would feel in their absence, how much they engaged with them, the kind of impact the groups had on them, the kind of respect they got from them, and their feelings towards them. In contrast, for many participants, how much they identified with a group was determined by how much time they spent together, how much they engaged and talked to one another, and how accepting they felt a person or group was of them.

Conflict

Adding conflict lines between the groups depended on the type of relationships (e.g., family versus friends) and the level and nature of the interaction with the group members, including how well they got along or if the relationship was “rocky.” When asked about how they determined between high versus moderate conflict, participants indicated that the red line (high conflict) would be for more severe conflicts – “fighting” or “big arguments”– whereas the orange line (moderate conflict) would be for “just arguing” or bickering.

Commonality

For most of the participants, commonality lines were drawn considering the characteristics, values, personality types, comfort level, proximity, interests, and relatability of group members. They also considered who got along with whom and the extent to which they did (or whether they would get along if they did not currently interact at all or regularly). For many participants, sober friends and family were compatible, whereas friends who used drugs were less compatible with family.

Group Member Centrality

Centrality of group members (very central versus less central) was only captured through the “ SIM, Version 3.0 (Circle)” for 14 participants. Although we would not recommend using this particular version due to logistical issues, participants still had valuable reflections about the centrality aspect of social networks. To determine a group member’s centrality, participants considered their level of activity in the group (e.g., organizing), whether they mattered to others, who is most popular, and the amount of time spent with group members and knowledge of their behaviors. However, for six participants, their responses suggested that instead of indicating each member’s centrality to the group (as instructed), they focused on how central each member was to their own lives.

Group Placement Around “Me.”

The facilitator’s instructions were intentionally flexible, allowing participants to move their groups where they wanted them on the maps in relation to the “ME” written in the center (for both the SIM-AR and oSIM-AR approach). When we asked participants about how they elected to place their groups, some participants indicated that the placement of circles was dependent on the level of closeness with those groups, while others stated it was random. The unsystematic nature of group placement militates against using distance from the center as a research variable. It may be possible, however, to add directions to participants to place groups in such a way that closeness to center is a meaningful variable, e.g., the closer to the center, the more important the group.

Qualitative Perspectives Related to the SIM-AR/oSIM-AR Experience

In addition to responding about the specific elements of the mapping, participants had several reflections on the mapping process and suggestions for potential areas for improvement. A table of exemplar quotes is included in the appendix.

Feelings About the Process

The mapping seemed to help many participants understand their feelings about their network, groups, their recovery process, or how they need to focus on themselves first to become better before they help others. Eleven participants specifically described the mapping process as helpful, in that it made them more aware of how people are placed in their life, the role these people play in recovery, and how these people support or hinder them. Indeed, participants described how the mapping gave them insights into patterns that they never thought about in their relationships and gave them new insights, including the following examples: that they should focus on their goals more and who supports those goals, who they should be around for better recovery, enjoying reflecting on closeness with others, the realization that they had very distinct groups, and happy to see they had made positive social changes. Two participants commented that they wanted to keep the map to share with others (e.g., therapist).

Feedback on the Process

The process-related feedback from participants who completed the oSIM was primarily about liking it and finding it easy. Five participants mentioned that the level of support being provided by group members should be part of the SIM-AR, along with how group members’ substance use or history of use may impact their own level of use. We subsequently added an open-ended question for them to answer about each group concerning how the group supports their recovery (oSIM; added to reflection interview guide for paper-pen version). Only a single participant indicated that it was boring, but during the interview, that person also acknowledged that the process helped them gain a new perspective. Several participants completing earlier versions of the map also discussed parts of the process they did not like, in that some of the components (e.g., conflict lines) made it messy, which led to adaptations (described earlier).

Materials

This code related primarily to the paper-pen version of the SIM-AR. Most participants liked using the sticker dots because it made it easier to group people, although it was time-consuming for some participants with large networks. For example, because of the high casual use in one participant’s network, they ran out of the yellow sticker dots they were mailed, an issue that is easily resolved when doing this in person as the facilitator can bring extra. Four participants reported that they found the included map example, legend, and example group names helpful in starting the mapping and following along during the session.

Discussion

In this paper we describe the process we used to test and adapt an interactive tool for social network data collection with adolescents experiencing SUD recovery. In total we recruited 35 participants and then made adaptations both to the paper-pen version (SIM-AR) and a subsequent online version (oSIM-AR). Our experience was that participants were enthusiastic about completing the SIM, while at the same time we were able to capture detailed social network data from adolescents with minimal confusion, and a clear sense of ownership and engagement on the part of the participant. Data derived from the SIM process included: number of groups and group members, substance use behavior of groups, time spent with groups, level of importance of and identification with groups, perceived between-group commonality and conflict. The follow-up interview included more detailed questions about how groups support or undermine recovery, thus providing a further lens for understanding social risks and opportunities. The centrality of group members was also assessed, but this version of the map was difficult to implement and may only offer minimal value. As a secondary by-product of our approach, we also discovered that we could easily share the findings with participants by allowing them to keep the completed maps, which was valued by several participants. The resulting map was easy for them to digest and reflect on, demonstrating the way this unique, and visually appealing, presentation of data could also improve the accessibility of research findings to non-scientists, and open it up as a means of developing coproduced data collection. Considering the time our adolescent participants contributed to this study, and to similar studies, having the ability to own your own creation and data (e.g., by allowing participants to keep the map) is an important ethical value to consider accounting for whenever possible.

Considerations for Choosing SIM-AR or oSIM-AR

Despite the overlap in the kinds of data that were collected in both the paper-pen and online approaches, there are some subtle differences worth discussing. First, in the paper-pen version, each individual member’s level of AOD use can easily be captured using the colored sticker dots. This allows for the generation of precise substance use ratios across individual network members. In the online version, participants can only indicate AOD use at the group level. This results in substance use ratios at the group level across the network. Although more information can be tied to individual members in the paper-pen version, literature suggests that it is perceived group behavior that is most important in understanding how social identity influences one’s own behavior.^40,55,56 Thus, it is not clear that member-level information adds substantial quantitative predictive value to recovery processes unfolding over time. Yet, several of our participants suggested that we did not collect enough information about the perceived support of specific, strongly-influential group members. As this study did not examine what elements need to be captured at the individual member level versus the group level, future studies should compare which level of fine-grained detail might have more explanatory power.

The paper-pen version requires more materials, takes longer, and requires more post-processing (data checking and entry) than the electronic version. The paper-pen version also seems more likely to need a facilitator present. These are all important elements to consider when planning a research study and accounting for time and resources. These may also be important resource considerations for busy clinical environments or for community programs considering implementing this approach. Yet, the paper-pen version also seemed to generate more critical insight and reflection during the post-mapping interviews and could also potentially be used by a skilled practitioner to help intervene in addressing social network influences during recovery, or to support the development of a network-focused intervention.^57,58 Although this study was not designed to determine the reasons for this difference, it is possible that the more tactile and visually interactive nature of the paper-pen process encouraged greater reflection. Indeed, participatory qualitative research has suggested that interactive and visual engagement processes may facilitate reflection, meaning-making, and discussion of complex experiences. ⁵⁹ However, as assessment reactivity is a potential issue to contend with in psychosocial research,^60-62 the study aims should be considered when selecting one of these data collection tools.

Limitations

This is an exploratory pilot study that did not gather longitudinal data to examine the predictive utility of social networks and recovery outcomes. There may also have been some facilitator influence on participant responses, including social desirability bias or a desire to appear engaged in the process, which may have particularly influenced positive feedback regarding the mapping experience. We attempted to minimize this risk by using a standardized facilitator guide and explicitly soliciting both positive and negative feedback. Additionally, the sample was predominantly white and male. Future studies should focus on recruiting a diverse cohort with greater gender balance. Although the sample size was modest, our purposive sampling approach to recruit adolescents from treatment and recovery services in various stages of recovery contributed to high information power.^46,47 Yet, given that this was primarily a help-seeking sample, the results may not generalize to youth outside of these types of settings. The study visits required access to a computer or tablet for data entry (for the oSIM). This may present barriers to participants without this access if attempting to conduct remote study visits, a barrier that may be more pronounced for youth if they live in homes without devices or online access. However, this approach removed some participation barriers, in that adolescents could participate without securing transportation and study visits could be flexibly scheduled. Similarly, partnering with RHSs facilitated recruitment and allowed participation in settings that were convenient and familiar to students.

Conclusions

In sum, in this study, we successfully adapted the social identity mapping tool for adolescents to include a standardized facilitator guide and to have both paper-pen and online options. We found that the visual and hands-on aspects of this approach were well-received, in part because the approach has been tailored to the adolescent’s maturational level, and offers an alternative and personalized approach to engaging unique learning and interaction styles. Our future steps are to implement the use of SIM-AR in settings seeking to promote recovery-supportive networks to examine how successfully these programs generate change in networks over time to support adolescent recovery. For example, programs could implement the SIM-AR early in a young person’s engagement and assess again within a few months to what degree social network composition has changed. Based on this pilot work, this approach could be feasible both individually with youth or in group settings (e.g., as part of a group workshop). Additional future steps also involve working with clinicians and youth to further develop the SIM approach beyond a data collection tool into a potential intervention module.

Appendix.

Appendix.

Table 1. Exemplar Quotes from SIM-AR/oSIM-AR Adaptation Study

Feedback About Specific Map Components
Importance	204 and 205 (oSIM): How they would feel if that person wasn’t there
	207 (oSIM): Yeah, pretty much that, and then the respect that they give me, how much they want to actually see me do good in life.
Identification	204 (oSIM): How much I’m with them, who I’m always talking to, the ones I’m talking about; just what I’m doing that I can see in the group, basically.
	205 (oSIM): Probably how accepting they are of me.
Conflict	141: Thought about how many unresolved conflicts there were or how many, considered like who had wronged who and if someone in the group had wronged me or a friend, or did something to someone, then I would put a red/orange line.
	204 (oSIM): What makes me stressed and makes other people in the group stressed.
Commonality	103: I had my sober friends compatible with my family. So maybe like, if I have a few of them come talk to my mom and, like, talk to my family about things, they’ll understand before I bring my drug addict friends over and have a sit-down talk with them.
	124: That was probably a little bit the hardest part, because it’s hard to think of commonality, what people have in common. I was more thinking about how much people would like each other’s company, like how much-- you know what I mean? But I feel like, yeah, that’s-- so my main thing is like how much would each group enjoy each other, how much would I enjoy this group, or something like that.
	142: I guess my school friends, I have a lot in common with them. We grew up in the same place, kind of have the same sort of thing going on. I guess the people who I feel the most comfortable around are probably the people you have most in common with.
Group Member Centrality	P138: I kind of did just who’s the most popular in the group.
	P139: I did it by who I talk to the most, really, or who I would hang out with the most.
	140: For example, this one is my brother… So I kind of put him non central because I don’t be with him all of the time… Same thing for my friends right here. If they’re in the circle, I see them pretty often.
	P144: So I thought about how much I interact with each person on a day-to-day basis. I thought about how important each one was to me, or how much they’ve done for me, or how much time we’ve spent together. I also kind of thought about who I liked more versus other people. Who I kind of preferred to hang out with as opposed to maybe other people who I’m almost always in the mood to hang out with versus other times. Because I’m an extrovert, but there are definitely times where I have to be in a certain mood or certain energy level to hang out with certain people, even if they’re my friends. And so that was another thing to consider. Regardless of my state of mind or something like that, who I can still hang out with if I’m maybe not feeling the best or something like that.
	P145: Who I’m closer with, who I feel closer with, and sometimes who I spend more time with because I feel super close. There is one person in my close friends’ group who I don’t spend a whole lot of time with, but I feel fairly close to, kind of the same as other people. But I see her like once a month, and I see all of my other friends twice a week. So I was like, “I feel pretty close to her,” just because we still stay interacting. But in general, I’m closer with my other friends just because I see them more. So it’s a mix of how I feel and then time spent.
Group placement around “Me”	204: I just kind of did it randomly.
	205: I was really considering the [inaudible]. But I would say my dad’s-- I’m definitely closer to my dad’s side. INTERVIEWER: Okay. And so is that maybe closer to you? P205: I might have done that subconsciously, honestly, yeah.
	207: I actually did it so whoever’s really close to me really wants to see me do good, better than them. That’s who I put closer to me, and then people that are more conflicts, less trying to see me do well, or at least that’s how it seems, I pushed them away from me.
	208: I just put it randomly.

ORCID iD

Emily A. Hennessy https://orcid.org/0000-0002-5146-5823

Ethical Considerations

The Uconn IRB and the Massachusetts General Hospital IRB approved all human subject protocols for the study (Protocol #: 2021P000078).

Consent to Participate

Written consent was received by participants prior to initiating any study activities. Minors were consented with an adult parent.

Data Availability Statement

Given the small sample size and the unique characteristics of the participants, data beyond what is reported in this manuscript will not be made publicly available in order to minimize the risk of potential breaches of confidentiality.*