Clinical Data Mining With the Listening Guide: An Approach to Narrative Big Qual

Abstract

We developed a novel approach to narrative Big Qual research that combines Carol Gilligan and Lyn Mikel Brown’s Listening Guide with Irwin Epstein’s clinical data mining. We adapted the voice-based research methodology of the Listening Guide for use with a corpus of clinical case notes drawn from an integrated data system (IDS) of a social service intervention serving families in an immigrant enclave. This methodological innovation was inspired by the insight that the Listening Guide can be used to trace and name the layering of meaning within any narrative, whether that narrative reflects the experience of an individual person or, as in this case, the community and everyday life of a social service intervention. Critically, this approach pivots on theorizing the subject as the collective of the intervention itself, as narrated by case managers, who can be understood as narrating subjects within the cultural, figured world of the intervention. In the context of a larger process and outcome evaluation, marrying these two approaches provided context, texture, and depth to supplement existing data sources like self-report survey data and participant observation, and offered a glimpse inside the “black box” of the intervention. We adapted the Guide through three readings of the clinical case notes: once for stanza structure, once inspired by the I-Poem technique but modified for these third-person narratives, and once with an eye to the contrapuntal voices of the inner and outer worlds of the intervention. As a methodological innovation this approach represents an advance in Big Qual and a promising approach to conducting narrative research on large qualitative data sets within mixed methods studies.

Introduction

This paper discusses an emergent qualitative method developed by our research team in the domain of big qualitative data or Big Qual (Brower et al., 2019), which combines Lyn Mikel Brown and Carol Gilligan’s (1991, 1992) Listening Guide with Irwin Epstein’s (2009) clinical data mining (CDM) methodology. Emergent qualitative methods can help researchers answer novel questions, bridge disciplinary boundaries, and leverage new types of data for knowledge production (Hesse-Biber & Leavy, 2013). They have the potential to open up new ways of knowing, helping us see familiar material with fresh eyes. Big Qual refers to the use of large qualitative data sets within a mixed methods study to add breadth, depth, and context to an analysis. The Listening Guide is a voice-based research methodology in which narrative data, typically interview transcripts, are read from multiple angles and ultimately organized into poetic form as a mode of analysis and meaning making. Clinical data mining involves leveraging existing agency data for research purposes, allowing practitioners to refine service delivery and researchers to conduct studies without additional data collection expenses.

Here, we offer a proof of concept for the adaptation of the Listening Guide to clinical case notes from a common administrative database shared by stakeholders of an economic mobility social service intervention. We draw on a data set of 191 narrative text-based case notes exported from an integrated data system (IDS hereafter), a computer-based platform for sharing administrative data across stakeholders, increasing the capacity for information sharing between workers, management, and funders and creating new opportunities for macro research across service sectors that were previously siloed (Culhane et al., 2018).

The aim of the paper is to describe the process of adapting the Listening Guide for use with this very different type of qualitative data, including the challenges, possibilities and limitations perceived and encountered by the researchers. We describe this process through three readings of the data. The first reading, for structure, highlighted the tensions of implementation as the demands of the economy and the preset terms of the intervention’s logic model collide with the lived reality of families. The second reading, focused on women’s domestic difficulties, helped us see how case managers work closely with female clients to navigate domestic situations and cultivate more equal partnerships around financial decision-making. In the third reading, we adapted the I-Poem technique to construct subject-verb poems that attuned us in a less mediated, more emotional way to the relational work unfolding in the clinical encounter.

The paper is structured as follows. First, we review the literature on the Listening Guide, both its psychoanalytic roots and its subsequent uptake among critical sociologists, and we situate Listening Guide research within narrative inquiry more broadly. Next we describe the context of our research, data sources, and analytic processes. Our data analysis section contains three readings of the data and discussions of these readings. Finally, we reflect on the implications of this work for other potential practitioners of Big Qual and CDM, including the challenges and limitations of the method.

Literature Review

The Listening Guide that informs our method was originally developed by feminist psychologists Lyn Mikel Brown and Carol Gilligan (1991, 1992) and additional detailed descriptions of the steps of the method have been published since (Doucet, 2018; Doucet & Mauthner, 2008; Gilligan et al., 2003; Sorsoli & Tolman, 2008; Woodcock, 2016). The method involves a systematic way of working with transcripts or other narrative material, listening three or four times with different types of attention.

As originally formulated, the first listening is for plot. The second listening is for the self—the “I” speaking in the story. The goal of the second listening is to open oneself up to be in dialogic relationship with the speaker, on the premise that all knowledge is situated, filtered through individual subjectivity, and that a listener’s establishment of a psychic, emotional relationship with the speaker is an avenue toward knowing. Brown and Gilligan characterize these two ways of listening as responsive. The final two listenings, they explain, are to be resistant, that is, listenings done “against the conventions of the dominant culture” (1991, p. 47). In the original context in which the method was developed, Brown and Gilligan, interested in narratives of relationship and relational conflict, listened on the one hand for voices of love and care, and on the other, for voices of justice, equality, reciprocity, and fairness. They did this to highlight the built-ness of gender relations, and the ways that conventionally gendered and aged modes of interaction divorce people from aspects of themselves to the detriment of their well-being and wholeness. They did this also to illustrate how patriarchy, defined as a system of domination that orders individuals in hierarchical relation to one another on the basis of gender and age (men over women over children) might be visible in the ways relational conflicts are narrativized. Resistant listening, then, is listening done to tease out particular tensions. It is targeted listening that operationalizes the conceptual framework.

Like all data analysis methods, the Listening Guide is not a neutral technique, but by necessity carries the epistemological, ontological, and theoretical assumptions of the researchers who developed it, as well as the sometimes different assumptions of the researchers who take it up (Doucet & Mauthner, 2008; Mauthner & Doucet, 1998, 2003). Together and separately, Andrea Doucet and Natasha Mauthner advocate for those who use it to be explicit about the epistemological and ontological concepts of subjects that are informing their research practices and accounts. They build on gaps in the original method by drawing on reflexivity scholarship (Doucet & Mauthner, 2008). Brown and Gilligan’s (1991) guide was designed in the transitional years between second and third wave feminism as part of a broad identity politics project to identify, excavate, and raise up voices, particularly those of girls, women and members of other marginalized groups, and it reflects the scholarship of the era. Namely, it promotes relational understandings of the world, lending itself to analyses that highlight interdependence over individualism and autonomy, and asks for a limited kind of reflexivity from researchers, focused on accounting for one’s social location and emotional responses to transcripts, but not about one’s theoretical presuppositions.

In its original iterations, the Guide also does not theorize the relationship of voices to subjects, but is predicated on a politics of “giving voice” as if voices have a substance, can be given and taken, and are not constituted in and through the research encounter but rather exist independently of the researcher. The original Guide imagines the researcher as an instrument who uncovers voices within a narrative, voices that are conceived of as pre-existing but rendered invisible by positivist frameworks, and a narrative that is seen as unproblematically representing the experience of the speaker, critiques that Brown (1994) has thoughtfully engaged.

The original Guide is also silent on an important debate within feminist theory about the nature of subjectivity and the extent to which subjects are located in, versus constituted by, social, cultural, and discursive contexts. In their own (2008) adaptation of the Listening Guide, Doucet and Mauthner extend the original method by proposing the concept of the narrated subject as a way of holding together the critical and culturally constructed subjects, arguing that the multiple readings of the Guide offer a way to operationalize this concept. Narrated subjects construct their identities through the stories they tell, they are intelligible in relation to other subjects, and their narratives reflect the constraints and affordances of broader social forces and dominant discourses.

Doucet and Mauthner (Doucet, 2018; Mauthner, 2017) also urge researchers taking up the Listening Guide to be explicit and self-critical about the assumptions built into the method, and for that matter, any other method used to build knowledge. Without exercising a deeper form of reflexivity, taking methods themselves as objects of investigation, and considering the ways that particular research practices produce particular kinds of knowledge, researchers risk unintentionally re-inscribing representationalist biases in their work.

As techniques of narrative analysis, the tools in the toolkit of the Listening Guide are ultimately ways of analyzing the stories people tell. The Guide has been widely adapted by scholars across disciplines, and has been used to investigate such topics as: women’s experiences of depression as gendered and raced (Beauboeuf-Lafontant, 2007), older couples’ meaning-making around cancer diagnoses and end of life (Gardner, 2008), the interplay between romance and notions of sexual equality in how young women narrativize sexual experiences (Milnes, 2004), fathering narratives (Doucet, 2018), and women’s experiences of corporate workplace transition (Balan, 2005). Taking up Doucet and Mauthner’s call for reflexivity, Johnston (2015), studying transgender students’ experiences of gender discrimination, violence, and stigma on campus, explores how cisgenderism shapes transgender research and narrative analysis. This call is also taken up by Petrovic and colleagues, who apply the method within the scholarship of teaching and learning to analyze personal critical reflection papers written by dietetics students (Petrovic et al., 2015).

All of these studies rely on stories, and from a certain perspective, stories are all we have. Narrative research proposes that we can systemically analyze the stories people tell as a way of knowing about the world. Narrative inquiry encompasses a broad range of techniques and traditions across disciplines, united by five common analytical stances (Chase, 2010). First, narrative researchers treat narrative as a distinct form of discourse, that is, as retrospective meaning making, wherein attention is paid to sequences and consequence and “events are selected, organized, connected, and evaluated as meaningful for a particular audience” (Riessman, 2008, p. 1). Second, narratives are viewed as verbal action; narratives do work, they accomplish goals. They are stories with a purpose. Riessman continues, “Narratives do not mirror, they refract the past. Imagination and strategic interests influence how storytellers choose to connect events and make them meaningful for others” (p. 6). Third, narrative is regarded as enabled and constrained by a range of social resources and circumstances, as the stories available to be told vary across time, place, and context. Fourth, narratives are viewed as socially situated interactive performances, that is, as “joint productions of narrative and listener,” (Chase, 2010, p. 215), whether that listener is present in person for naturally occurring face-to-face talk, or simply imagined by the narrator. Finally, many narrative researchers regard themselves as narrators, whether in co-producing narratives with research participants, or in developing interpretations of others’ narratives. Acknowledgment of this is critical as it can help protect against a researcher selectively retelling the story of the data based on his or her own imagination and strategic interests. While narrative can offer a useful window into both individual sense-making and social dynamics, with each of these five analytical stances come limitations as well as affordances for which practitioners of narrative inquiry must account.

Methodology

Context

Data is drawn from an intervention based in a collaborative of social service agencies serving a financially strained immigrant enclave in a large East coast city. The lead organization responsible for case management is a nonsectarian Christian youth development organization that offers afterschool and summer camp programs in addition to family-facing services. Collaborating organizations include a community college, an entrepreneurship development program, and a work-skills nonprofit. The goal of the intervention is to improve families’ financial capability, with lasting, ideally intergenerational effects. ¹ The intervention serves a community of 155 primarily Arabic-speaking young families, mostly Christian but some Muslim, from a single country in the Middle East/North Africa region. ² Although some of the partnering agencies are faith-based and many of the participants describe themselves as religious, religious identity or expression is not a requirement for participation and is not an explicit component of programming.

As a partner in the Ascend network at the Aspen Institute, the intervention is based on a two-generation model and aims to develop family wellbeing by addressing children’s needs and parents’ needs in concert. To qualify as potential participants in the intervention, families must present as two involved adults along with their minor children. ³ Offerings are driven by families’ self-identified goals, in line with the social work code of ethics. Central to the program is adult English language instruction, and at least one parent from each family must enroll in ESL classes, held at the community college for parents of school-aged children, or at the lead agency headquarters if childcare is needed. School children can enroll in the Center’s after school program. Interested adults may enroll in the office skills and business development courses offered by partnering organizations once their English language skills permit. At the center of the intervention is a robust case management program. All families receive case management from an Arabic-speaking case manager, meeting at least monthly, or more often as needed.

The city in which the intervention is based has historically been a key point of entry for immigrants, and has a diverse population with respect to race, ethnicity, and country of origin, with 41.7% of the population identifying as foreign born on the most recent census. The majority of families reside in a tightly bounded geographic area within the city, which in the pre-COVID-19 era was experiencing rapid gentrification and income volatility. Across the two zip codes where most families live, the large majority of housing units (76.2%) are renter occupied apartments in multi-unit buildings. Among those residents for whom poverty status is known, a majority (58.5%) are at twice the poverty level.

A diverse mixture of recent immigrants reside in these neighborhoods, drawn by the area’s reputation for low-cost housing. However, as real estate pressures in the larger metro area have intensified, these pressures have spilled over into the city in which the intervention is located. In 2016, median gross rent as a percentage of household income was 32.2% and 28.9% respectively in the relevant zip codes, meaning that the average renter in these areas is rent-burdened, defined as spending 30% or more of household income on rent.

Case Notes as Data

The data set for the paper consists of a three-month window of case notes (January 2019 through March 2019) exported by the researchers from the Zoho IDS. ⁴ The IDS functions as a common platform for the collaborating agencies, linking individual level data to allow the agencies and funder to monitor and track how services are being used and to avoid duplicating one another’s efforts. Proponents of IDSs argue that such systems promote efficiency in service delivery and facilitate the generation of data to inform evidence-based decision making, while others caution that these systems can introduce complex and thorny legal issues around privacy protections (Culhane et al., 2010).

There were 191 text-based entries across 91 households, 144 of which were written by two Arabic speaking part-time workers who are themselves members of the community. As the case notes are the property of the agency and clients consent to case notes and evaluation when they enroll in the agency programming, we did not seek the verbal or written informed consent of the individuals represented in the case notes. However, we obtained consent from the agency leadership and staff as part of the Quality Assurance/Quality Improvement evaluation and chose to use a method that maintains the same degree of confidentiality of clients afforded by the case managers themselves. We cleaned the data in a text editor, removing client IDs, metadata including date, time spent, and type of visit, as well as personally identifying information such as family names. Grammar, punctuation, and spelling irregularities and errors were common in the data set, and we retained these where it did not interfere with the reader’s understanding. First names were also retained in the cleaned data set, and pseudonymized later. All names in this manuscript have been changed for privacy. ⁵ All materials were stored on secure servers.

We first conducted a thematic analysis. ⁶ During this process, we began to experience the notes as narratives and became interested in working with them in a way that preserved their internal integrity. Familiar with the techniques of the Listening Guide from previous research (Fontaine, 2015), we decided to experiment by applying them in this new context. It was not a straightforward process. These are not Brown and Gilligan’s narratives of self and relationship divulged in a clinical therapeutic context, but documentation of monthly case management meetings, crafted by front line workers with varying levels of formal education, professional preparation, and English language proficiency, recorded to document case managers’ interactions with families as part of their paid employment for an audience of funders, supervisors, colleagues, and research staff.

Changing the Subject

As we began to contemplate and then do this work, we wrestled with tensions inherent in the adaptation of a method designed for a certain type of data and analysis for our rather different purposes. While there is great variety in the topics that researchers have investigated with the Listening Guide, the data sources themselves are traditionally narrative interview transcripts and focus group transcripts created for the express purpose of research. The unit of analysis is typically the individual, whose long-form personal narratives are analyzed as a window into the ways in which meaning-making processes both reflect and resist social structures and discourses. There is good reason for this. The method is rooted in psychodynamic theory and aims to surface the multiple perspectives and accounts that are often simultaneously present in the stories people tell about themselves.

We build on this tradition with our insight that the protocol can also be used to name and understand the tensions and competing accounts within any narrative.

In this instance, we sought to highlight the interplay between the inner and outer worlds of the intervention, between the emergent difficulties facing families and the preset terms of the intervention’s logic model. At the time we were conducting this analysis we were not familiar with Doucet and Mauthner’s (2008) concept of the narrated subject, so we took inspiration in conceptualizing our subject from other sources. We drew on ecology and systems theory, which sees urban environments like that of this community social service intervention as both an amalgam of self-contained and autonomous entities and well as an entity in its own right (Evans, 2011; Zaghloul, 2018). We connected this to a social work ecosystems approach that “views individuals and environments as constantly interacting with and adapting to one another” (Ahmed et al., 2017, p. 49). We also noted consistencies with the concept of figured worlds (Holland et al., 1998) within anthropological theory, understood as spheres of human engagement that “take shape within and grant shape to the coproduction of activities, discourses, performances and artifacts” (p. 51). Reflecting on these various theoretical constructs with the benefit of hindsight, we suggest that the case notes are best understood as the narratives of narrating subjects (Doucet & Mauthner, 2008) within the figured world (Holland et al., 1998) of the intervention.

In this instance the milieu of the intervention contains the entire social environment or ecosystem of the community, including clients, front line social workers, administrators, and community members working and living in this urban space. We saw the subject as the immigrant enclave around the intervention that is conscious of itself as an entity, and the case managers as the self-conscious representatives of that collective subject. We also understood that as self-conscious representatives their narratives—like all narratives—would be partial and selective, emphasizing certain points rather than others, perhaps protective of the privacy of its members. We did not see the case managers as simply representing themselves or the families in their essential nature, but rather as telling a certain kind of story about the families and about their work.

Given the differences between the type of data the Listening Guide was designed to work with, and the type of data we were analyzing, we were creative in its implementation, understanding it as a resource kit for drawing out layers of meaning. While our earlier thematic analysis was conducted with the Dedoose software package, here we worked more simply, with a word processor. We worked with the texts in three ways. In one reading we broke blocks of text into stanzas and lines in order to grasp their structure and plot. In another reading we adapted the technique of the I-Poem. As case notes were written in the third person, we read for named subjects and further revised our stanza and line breaks, deciding to begin each line with a statement or action by one of the parties, either client or case manager (CM) or in some cases another family member. This process helped us to see the case notes as a conversation between clients and case manager and to tune in to the emotional level of the conversation.

In yet a third reading we focused our attention on what struck us as the germ, or core, of the note. Here we listened both for what was said, and for what went unsaid, in the knowledge that case notes contain only the most basic outline of the clinical encounter. Perhaps because of the context of the encounter and the life stage most families were embroiled in, these stories tended to center on the domestic space: troubles with children and husbands, financial affairs, sicknesses and injuries, and job and housing matters. These stories of stresses and shocks, conflicts and dramas in the lives of clients, are the inner world of the intervention. We read these in juxtaposition with references to the outer world of the intervention, those moments when the case managers seemed to be speaking directly to the funder and supervisors in terms sanctioned by the intervention.

This three-step process resulted in a collection of three- and four-stanza poems. We share selected examples here, although we present our examples in a somewhat different sequence, with the I-Poems last, in the interest of clarity.

Analysis: The Poetry of Domestic Life

Reading 1: Stanzas, or Reading for Structure

In our first reading, we reformatted the notes from their original narrative text blocks to a poetic form with stanzas. In doing this we noticed that they shared a roughly common structure. The following example is typical:

Mariel

4  pm, Mariel came today to case management session,

Mariel return back reading log for her daughter Lucy.

Mariel said she filled job application to work in Papayas and

She waiting for them to call her.

CM told Mariel that good step to take right now special Mariel daughters old enough to go home by themselves from the school

CM wish Mariel a good luck and pray with her that God guide her and the best for her benefits

Mariel said that Manal will go to the Dickerson high school and Mina and Lucy doing good at school

CM encourage Mariel to use the English Language more to prepare herself for that job

Mariel said her husband Ghobrial doing good at work

CM create incentive reading plan with Mariel for her kids.

Mariel next app will be in 4/30 @5 pm

It begins with a time stamp, noting the hour of the appointment and a statement of who is in attendance. Next is a report of the clients’ progress toward their self-identified goals. This could involve an update on attendance and performance in the ESL, office skills, or business development classes and updates on job searches. This section also often includes a reference to children’s reading logs, which the lead agency requires all families with school age children to complete each month. These are all touchpoints in the figured world of the intervention. Following these updates there is often discussion of any other challenges or issues present in the clients’ family, which we discuss below as the “germ” of the note. The case notes generally conclude in a manner that mirrors their opening, with the outlining of concrete steps to be taken and a note of when the next appointment is scheduled. Next steps are frequently expressed in terms of the funders’ preferred metrics, behaviorally-inflected artifacts like reading logs and reading incentive plans.

Reading 2: The “Germ,” or Reading Domestic Encounters

Our second reading focused on the “germ” of the session. While these sections are often relatively short, they reflect much longer conversations, as case notes compress an hour-long meeting into a handful of sentences. Perhaps because of the context of these sessions, held during the day while children are at school and husbands are at work, the topics are typical of women’s talk in the community and cultural group; matters of childbearing and childrearing, husbands, money, health, housing, and schooling. These discussions are inflected with the priorities of the intervention, including women’s education and paid participation in the labor market as pathway to improved family economic security. The pair of case notes below express this focus on the beneficial impact of egalitarian marital relationships on family upward mobility:

Nailah

3 pm, Nailah came today for case management,

She is enjoying her class and

She said she learned good amount through the Management classes,

Nailah is going to start a cooking class by June 26 until Sep 26, and after that

She will continue ESL class and will try to find a part time job.

Nailah has a good husband

who encouraged her all the time to learn and communicate and

she has the passion to do that,

Her children are doing good at school,

CM asked her to send him a copy of their report cards,

CM told Nailah that the registration for the new class will be on August,

Nailah expressed her appreciation for the program and the services.

Nailah next app will be 2/18@4 pm.

Reva

4 pm, CM met with Reva today for case management,

Reva is enjoying the classes and the program

Reva doing very well in her class and the business class too,

Reva and her husband are a good example for families that looking to improve their life,

They attend the business class together,

Reva passed the driver license road test and now

They looking to buy a car,

CM show them how to search online for a car,

Reva is very happy with this step and her children as well

Reva brought the reading logs and

CM gave her the new ones for Kata and Bishoi,

Kata and Bishoi are doing very well at school,

Reva will continue in the program for the fall class,

CM told her that the registration will be on August,

Reva expressed her appreciation for all services

Reva next app will be 3/27@5 pm.

The two case notes follow a common structure. Both can be broken into three stanzas. The first stanza is comprised of updates on the family’s use of services provided by the intervention (ESL, money management, and entrepreneurship classes) in addition to the outside cooking class that Nailah is pursuing. In the second stanza, the germ of the case note, the case manager moves on to summarize the emotional center of the conversation. In these two notes, this takes the shape of normative judgements about the clients’ marriages. The case manager declares that “Nailah has a good husband” and “Reva and her husband are a good example of families that looking to improve their life.” The remainder of the second stanzas contain evidence to support these claims. Nailah’s husband supports her goal of improving her ability to speak English so that she can communicate outside of the immediate enclave, and Reva and her husband share a goal of her learning to drive and of purchasing a family vehicle, which will enable her to move with ease and independence outside of the immediate geographical area. Again, the third and final stanzas revisit the terms of the logic model, reporting on each family’s progress toward achievement of their stated goals and the ways in which the children’s achievement mirrors that of their parents. Finally, both case notes conclude with case managers reporting that clients expressed gratitude for the services and noting the day and time of the next appointment.

In contrast with the “good” husbands and families of these two notes, other case notes present as cautionary tales.

Hana

5 pm, Hana came today to case management session.

Hana said her kids don’t like to read.

Hana said she doing okay in class.

Hana has Problem with her husband and

She was really sad when she was taking about it because

He is gambling and that make her life horrible in home even

He promise her to stop but he didn’t that’s make her really sad even

Her kids at home notice that.

Hana cry a lot because he lost a lot of money.

CM told her to talk with him again and told him how bad this habit and he has to quit and how it destroy his money and his family too.

CM advised Hana to speak with uncle and let him speak with husband maybe it will work and he will see how this habit effect his family.

CM pray with Hana and her Husband.

CM create incentive reading plan with Hana with her kids.

Hana next appt will be in 3/28@4 pm.

Nasrin

Nasrin came today to case management session.

Nasrin said she will start the Arabic class next week 2/27@4:30 pm.

Nasrin complain about her husband Girgis

He didn’t give her enough money to spend and also when she needs any money He questioned her about how she will spend that.

He hardly give her money enough and

She said she work a lot and

She deserve to be happy and to get what she needs.

CM advice Nasrin to take one day off from work during the weekdays to be able to go to get her driver license.

CM encourage Nasrin to open Bank account to put her money then she will able to use her money when she needed.

CM told Nasrin if she needed any help with that to call her.

CM create incentive reading plan with Nasrin for her son Youseff.

Nasrin next appt will be in 3/12@5 pm

Broken into lines and stanzas, the parallel structure of these two case notes is remarkable. Both were written by the same case worker, a part time employee whose family is also enrolled in the intervention. In accordance with the logic model, they begin with statements of who is in attendance, followed by updates on children’s academic performance and parents’ participation in the program’s language learning offerings. This first stanza is notably sparse and formulaic in comparison to the rest of the note. It is interesting, however, that the case worker reports on Hana’s children’s lack of interest in the reading incentive program, one of the program’s central behavioral interventions. Hana’s matter-of-fact dismissal of this priority (“Hana said her kids don’t like to read”) can be understood as an act of quiet resistance against the intervention’s unswerving focus on this particular metric. The case manager’s reporting of her defiance may be a way of tacitly acknowledging a mismatch between the logic model of the intervention and the actual needs of its clients.

After these perfunctory updates, the core of the two case notes are the female clients’ marital and financial struggles. In both cases the women identify their husbands’ fiscal attitudes as the source of their difficulties; Hana’s husband gambles, wasting their family’s resources, and Nasrin’s husband unreasonably limits (in her view) her access to money. In both instances, the case manager gives voice, at length, to the women’s complaints. While written from a third-person perspective, limiting the reader’s feeling of intimacy with Hana and Nasrin, the use of demonstrative language (“that make her life horrible,” “he questioned her”) suggests a degree of bonding and identification between the case manager and client. The case manager uses repetition for emphasis and cites Hana’s emotional display (“she was really sad,” “that’s make her really sad,” “Hana cry a lot”) as evidence of the severity of her situation. Similarly, in the narrative of Nasrin’s situation, the case manager uses repetition (“He didn’t give her enough money to spend,” “he hardly give her money enough”), perhaps indicating the amount of time in session that was spent discussing the issue or perhaps as a persuasive device. The case manager translates Nasrin’s complaint into a discourse of deservingness: “and she said she work a lot / and she deserve to be happy / and to get what she needs.” This is an interesting construction of the problem in that it positions the wife as entitled to access to resources which will make her happy.

The third stanza of the case notes again mirror one another, and consist of the case manager’s strategic recommendations to the women to help solve their dilemmas: to appeal to their husbands, to try to find common ground, to appeal to other male family members whose counsel might be more persuasive, to pray together, to take a series of concrete steps to come into one’s own self-sufficiency and thus shift the power dynamic in the relationship.

And in the fourth stanza, the case manager returns, obligatorily, to the requirement of the logic model, dutifully noting the creation of a new reading incentive plan and scheduling a follow up appointment in one month’s time. The juxtaposition in tone and detail between the creative, problem solving, emotionally-invested, rapport-rich core of the second and third stanzas, and the flatness of the first and last stanzas suggest that much of the real work of the sessions is done without regard for the requirements of the logic model.

In each of these notes the man is the primary economic actor and the wife is responsible for running the household but dependent on him for money. All three families are facing common structural difficulties. They are pressed by the high cost and low quality of housing, transformations in the labor marketplace that make the single-earner family model less and less viable, and a cultural climate of inhospitability toward even lawfully-residing immigrants. Men in this community are usually the primary wage earners, often in cash economies, while women care for children. In each of these notes the woman is expressing frustration and dissatisfaction with this traditional model of marriage in which the male is a primary authority and the female a dependent subsidiary to him. In each the female client and the case manager together seem to be collaborating, trying on different ways of narrativizing the conflict, and different justifications for a more equal partnership: for Hana, it pivots on emotional strain, and for Nasrin on deservingness.

These are women growing in their awareness that there may be another way, but who lack experience in renegotiating the marital contract. The experience of migration, of moving from one part of the world to another, changes the context within which a married woman with children experiences herself as a wife and mother and opens up new ways to be a woman in the world. While over half of migrants are women (UN DESA, 2013), women’s experiences are underexamined within migration studies (Pedraza, 1991). Women who migrate face the challenge of navigating the space between their “home culture” (i.e., the prevalent culture in their country of birth) and the “host culture” (i.e., the prevalent culture in their country of reception). Upon arriving in the host culture, women become aware of new ways of doing womanhood through observing new gender role patterns as well as taking advantage of opportunities for education and paid employment outside the home (Espín, 1999). Perhaps marital dynamics that were customary in their home countries now chafe, fit less comfortably when pressed up against the variety and diversity of family structures encountered as recent immigrants in a large, diverse, East coast city. Perhaps the structures now seemed negotiable, whether through divorce as some women contemplated, or through discussion and diplomacy, as case managers tended to advise. Women and case managers worked together through these challenges, brainstorming ways to address them through planning, negotiation, and problem solving.

Reading 3: I-Poems, or Tuning into Feeling

Presenting these narratives as I-Poems helps distill even further their structure and reveals their shifting emotional valences. Constructing I-Poems typically involves tuning in to the first-person voice in a narrative and crafting poems with stanzas and verses out of a longer narrative by initiating line breaks with each I-statement. That is, the researcher creates a version of the narrative that highlights its internal logic and the tension between the concurrently held conflictive positions by “separating each I phrase (subject and verb) from the narrative and listing it in the order of appearance…with each ‘I’ starting a new line of the poem and stanza breaks indicated where the I shifts direction or where a singer might pause for breath” (Gilligan & Eddy, 2017, p. 78).

Since these case notes are composed in a distant third person, we decided to adapt this process by creating line breaks with references to different subjects. In the examples that follow, I-Poems of the last two notes discussed, we have three main subjects: the presenting (female) client, the husband, and the case manager (CM):

Nasrin

Nasrin came today

Nasrin said

Nasrin complain

He didn’t give her

He questioned her

He hardly give her

She said

She work

She deserve

She needs.

CM advice

CM encourage

CM told

CM create

Hana

Hana came today

Hana said

Hana said

Hana has Problem

She was really sad

He is gambling

He promise her

He didn’t

Her kids at home notice

Hana cry

CM told her

CM advised

CM pray

CM create

Both narratives are structured by the case manager in a way that grants voice and visibility to tensions in these women’s marriages and families. They are, like the rest of the notes in the sample, narratives told from a particular point of view, from one woman to another, and as such they reflect women’s concerns: matters of kids doing well in school or not, kids’ medical issues, husbands’ employment, and the domestic negotiations and power struggles of marriage.

In both, the client comes to the appointment, offers a general update about the members of her family, and then begins confiding in the case manager about marital difficulties. The women vent, complain, worry, and cry, and the case manager listens, advises, encourages, and prays. There is an intimacy to these poems, as if, as Edwards and Weller (2012) point out, the researcher is now “standing alongside” subjects rather than “gazing at” them, as thematic analysis encourages. The case managers offer practical suggestions aimed at helping women solve these problems in their marriages.

These dynamics are made more visible by stripping the notes down to their subjects and verbs. By doing so we can perceive a shift in intensity between the portions of the notes that contain updates on the official business of the intervention, marked by verbs like “came here” and “said,” and the emotional center of the narratives, marked by more descriptive verbs like “question,” “promise,” “cry,” and “deserve.” The more intense and evocative language deployed by case managers in these sections suggests these are emotionally demanding sessions and that the empathy offered by these female case managers to their female clients addresses major needs.

Discussion

We have reported in this paper on the adaptation of the Listening Guide protocol for analyzing clinical case notes found in an online database that is shared across organizations collaborating to deliver services to working poor immigrant families. This methodological innovation represents a new technique for surfacing marginalia, an advance in Big Qual, and promising approach to conducting narrative research on large qualitative data sets within mixed methods studies. We undertook this line of inquiry in parallel to a process and outcome evaluation including a quasi-experimental survey-based study that uses established standardized scales to measure behavior change across the domains of adult wellbeing, child wellbeing, healthy relationships, and financial stability. We adopted this method because we wanted to glimpse inside the “black box” of the intervention and see how more assimilated community members, the case managers, worked with new community members, the clients, to guide and help them adapt to their new environment.

As critical qualitative researchers working as part of an interdisciplinary mixed methods team, we were eager to develop inventive ways to supplement the data our team was already gathering from periodic self-report surveys and ethnographic fieldwork. We did not aim for triangulation between our quantitative and qualitative approaches, as each are informed by distinct paradigmatic assumptions, but rather sought to “allow expression of different facets of knowledge or experience (Bazeley, 2004, p. 4). The existence of a language barrier between most members of the research team and the Arabic-speaking community meant that we could not easily participate in the social life of the intervention without disrupting it. We were also inspired by the work of other critical mixed methods researchers who advocate tuning our attention as researchers to the ways that participants may speak back to the researcher and his/her tools. For instance, Sarah McClelland (2016) advocates attending to “notes or verbal comments spontaneously offered by participants” and others (Castro Baker et al., 2018; Stoudt, 2016) suggest that engaging with comments made on the margins of written surveys can help us better understand survey respondents’ meanings and intentions. We propose further that the case notes in the IDS can be understood as marginalia, contextualizing, pushing back on, and fleshing out additional data sources, like that of our survey. While our preliminary survey results paint a promising portrait of community resilience (Castro Baker, Grinnell-Davis, et al., 2020; Castro Baker et al., 2019), the case notes provide additional texture, highlighting the variety and volatility that lurks just below the surface. We were inspired by the transformative-emancipatory paradigm within mixed methods research (Mertens, 2003; Teddlie & Tashakkori, 2015), which places central importance on understanding the lives and experiences of marginalized groups and understands these lives and experiences as structured by asymmetric power relations.

The case notes are also an instance of Big Qual data. Big Qual refers to the use of large data sets involving either primary or secondary data, typically analyzed by teams of researchers as part of funded mixed methods research projects (Brower et al., 2019). Data for Big Qual projects may be newly collected from the field as in the study described by Brower and colleagues (2019), downloaded from social media sources like Facebook and Twitter, culled from qualitative responses to open-ended survey questions, or repurposed from digital archives or various kinds of administrative databases for reuse and secondary analysis (Davidson et al., 2019). The term likely derives from Big Data, which danah boyd and Kate Crawford (2012) define as a socio-technical phenomenon in which massive quantities of information produced by and about people, things, and their interactions are recruited for analysis, often by very powerful computers, to generate insights and build knowledge. While a recent paper proposed bounding the term Big Qual to refer to either primary or secondary data with at least 100 participants (Brower et al., 2019) with the recognition that the criteria is necessarily provisional, others have noted that that “bigness” of a qualitative data set may also be expressed in terms of its richness, density, the sophistication of sampling methods, and the kinds of questions that are asked of the data (Bisel et al., 2014).

Our analytical move of leveraging a pre-existing body of data gathered for non-research purposes, known as clinical data mining (Epstein, 2009), has benefits and drawbacks. In clinical data mining, researchers “systematically retrieve, codify, analyze, and interpret available qualitative and/or quantitative information concerning clients’ characteristics and needs, services, and interventions received, and outcomes achieved derived from available clinical records for the purpose of reflecting back upon on the practice and policy implications of their findings” (Epstein, 2015, p. 501). It is critiqued by some for producing “less than perfect ‘gold-standard’ research studies” with randomized, controlled experiments regarded as the gold standard (Epstein, 2009, p. 9). CDM suffers reputationally from its association with practice, which is ordered in negative hierarchical relation to “pure” research. Most CDM studies to date have been quantitative, as researchers have found it challenging to identify qualitative clinical data sets rich and textured enough to yield robust findings. ⁷ Indeed, the rise of market-driven philosophy and practice in social welfare policy and practice has meant that nonprofit organizations are increasingly pressed into focusing on quantifiable performance outcomes (Abramovitz & Zelnick, 2015, 2018), and as a consequence, line workers face pressures to prioritize efficiency over depth of service and are not incentivized to keep the kind of extensive records that would seem to lend themselves to qualitative CDM.

Through this exercise we have endeavored to show that even given the relative linguistic sparsity of these narrative accounts when compared to the lengthy and robust first-person accounts typically used in Listening Guide research, narrative case notes and other types of qualitative administrative data can still be productively mined by creative-minded and committed researchers.

Researchers working with clinical data sources, secondary data, and Big Qual data must remain ever aware of the conditions that structure the production of the raw material they are analyzing. In our case this meant attending to how the IDS structured social relations and the content of the case notes. We realized that given the vulnerability of the population served and the sensitivity of the matters discussed, case managers were strategic in their decisions of what to include and exclude from case notes, cognizant of their power to shape how clients are seen in the eyes of funders, program administrators, research staff, and other social workers. Their case notes, like all narratives, are constructed in a particular context, for a particular audience, and for a particular purpose, including as a condition of their continued paid employment. The common structure of the notes may reflect, in part, the nature of much social work practice in the current moment, in which digital tools and technologies standardize work practices and products (Phillips, 2019). But case managers also bend the notes to their own purposes, approaching this job requirement as an opportunity to describe the value they bring as workers, whether by connecting families to resources, commiserating with them, offering guidance, or counseling and praying with them. They also contain case managers’ normative judgements about families, as in the example of Reva and her husband who are together deemed “a good example for families looking to improve their life.” Finally, these are stories of struggle and collective resilience, stories that make us see the gravity and the persistence of families’ difficulties as well as their humanity.

These narratives are necessarily partial. As researchers engaged in an exploratory methodological project we allowed ourselves to be guided by interest and instinct. Looking back retrospectively on the experience, we note that we may have focused more on narratives of domestic strife and negotiation than on the other major thematic concern of the notes, children’s health. Both were common in the data set, but the former overlapped more with our own preoccupations, and (perhaps because of this) we perceived an emotional resonance in them.

In addition, in part because the case notes function as records of conversations between young mothers enrolled in the intervention and Arabic-speaking female translators, they are more useful for learning about the experiences of women than those of men. Although men periodically attend case management sessions with their wives, the notes tend to emphasize the women’s perspectives, perhaps because of the rapport that has developed over time between case managers and clients. Men’s struggles, to the extent they appear in the notes, are reported in relation to how they affect female clients. Furthermore, men’s shifting ways of taking up their traditional roles, including care of children, is not captured in the case notes. The notes thus present a partial picture of how gender roles and relations are refigured in the space of the social service intervention, specifically reflecting women’s interests and meaning making.

They also, paradoxically, reveal the successes of the model as well as its limitations. As families see improvements in their economic status, new problems can emerge within households. Consistent with the literature on migration and women’s identity, we see that from their newfound proximity to a very different host culture, women are becoming aware of alternative ways of inhabiting womanhood. We heard commonalities in the experiences of these women, feelings of frustration and dissatisfaction with the pressures they face and the labor of the relational identity work they are engaged in. While it is not possible for us to grasp the full complexity of their experiences through the window of the case notes, as these are necessarily a selective reporting of the conversations between case managers and clients, we are able to glimpse at hints of their domestic dissatisfaction and the perhaps unintended consequences of migration on their marital relationships, relationships with children, and senses of self.

There are several practical considerations that may be useful for other research teams to consider. First, qualitative CDM leverages existing resources for research purposes, permitting mixed methods research teams to add additional sources of data without incurring expenses associated with new data collection, or needing to coordinate with agency partners. This is a distinct advantage when working with social service agency partners who are already facing pressures to maximize service delivery on tight budgets, and who may not welcome the additional responsibility of facilitating inquiries by researchers whose priorities and agendas may differ from their own. Second, as agencies face increasing pressure to comply with pre-set logic models driven by austerity and cost-saving measures (Abramovitz & Zelnick, 2015, 2018), this method provides a way to understand how staff and clients experience the eco-system of an intervention and creatively adapt or resist it (Castro Baker et al., 2020). Finally, for our geographically-distributed research team, working with case notes in the IDS permitted the involvement of members of the research team whose participation otherwise would have been cost prohibitive, simply because of transportation costs and budgetary constraints.

Compared to more traditional approaches using shared codebooks and computer-assisted qualitative data analysis software, the multiple listening structure of the Guide permits researchers to work more nimbly with data and adapt it for their needs. Critically, researchers interested in doing work in this vein should familiarize themselves with the rich body of methodological writing about the Listening Guide and its limitations. This means grappling with the imbrication of the method with the project of second wave feminism, and more specifically, with its implicit conflation of voice and subject. Researchers must then articulate their own operative theory of the subject and its relation to the narrative content. In addition to these forms of formal reflexivity, researchers should also acknowledge the ways in which they have acted as narrators.

As a matter of process, we approached the process intuitively, developing and revising our approach as we went. As a research method there is a radical element to it. It is a slow method for the digital age. Reading the case notes one is left with a sense of them as variations on a form, iterations of a theme, kaleidoscopic snapshots of one woman confiding in another. This is data mining but in a very different sense. We took the case notes as artifacts, produced in a particular context and for a particular purpose, and worked to breathe new life into them so that we could examine them from novel angles to hear a different story, one that pushes back against taken-for-granted truths about the doing of direct service work, one that goes against the grain of the dominant culture of efficiency, productivity, and linear progress.