|
Volume 9
|
Qualitative Methods |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
In this Issue Qualitative Methods and Philosophies Hermeneutics and Nursing Research: History, Processes, and Exemplar Qualitative Methods and Applications
Qualitative Methods in Research
Other Original Research
|
Using Data from a Qualitative Study of Adolescent Fatherhood to Create Population-Specific Operational Definitions for the Kinscripts Conceptual Framework Constance M. Dallas, PhD, RN The authors gratefully acknowledge the National Institutes of Health, National Institute for Nursing Research Grant # RO1NR007767-01A1 Contact information: Constance Dallas, PhD, RN, Associate Professor, Department of Health Systems Science, College of Nursing, University of Illinois at Chicago, Room 952, Dallascm@uic.edu, 312-413-7859 Using Data from a Qualitative Study of Adolescent Fatherhood to Create Population-Specific Operational Definitions for the Kinscripts Conceptual Framework1 To be useful in research or in a clinical setting, a theory or conceptual framework must be derived from study data or must exhibit fit with and faithfulness to the data. Creating population-specific operational definitions is an early step in adapting an a priori theory to a target population and ensuring “fit”.2 Population-specific operational definitions based on data are particularly important with African American fathers and their families because most family research is atheoretical and primarily descriptive. The purpose of this article is to describe how we used qualitative data from a study of adolescent fatherhood to create operational definitions for the three domains of the kinscripts conceptual framework:1 kin-work, kin-scription, and kin-time. The operational definitions developed from the data for each of these conceptual domains are specific to unmarried, low-income African American adolescent fathers. Constructing data-based theory is particularly important with African American fathers and their families because most family research is atheoretical and primarily descriptive, such as the work of Edin and Fefalas.3 According to Sandelowski2 any theory or conceptual framework must ultimately be derived from study data or must exhibit fit with and faithfulness to the data before it can be used in research or applied in clinical situations. An early step in adapting an a priori theory to a target population and ensuring “fit” is to create population-specific operational definitions that can later be used to develop concepts specific to the target population, thereby increasing the fit of the theory for that population.2 One concern in using theory to guide data collection is that the oretical definitions of concepts are abstract and, therefore, completely independent of a specific time or place.4,5 Operational definitions derived from theoretical definitions provide empirical markers to guide data collection. In comparison to theoretical definitions, operational definitions explicate the characteristics of theoretical concepts and describe what operations, or in this case behaviors, allow one to determine if a specific theoretical concept exists in a particular situation4,5 this point is particularly important when theories are used to study families of color because most theories were developed for Caucasian populations and lack cultural specificity. The purpose of this article is to describe how we used qualitative data from a study of adolescent fatherhood to create operational definitions specific to the population of unmarried, low-income African American adolescent fathers. We present a description of the kinscripts conceptual framework, an overview of the adolescent fatherhood study and a description of how we used study data to develop population-specific operational definitions. Background Conceptual Framework We utilized the kinscripts conceptual framework, developed by Burton and Stack,1 to guide data collection and analysis of the adolescent fatherhood study. The kinscripts conceptual framework was developed from over 20 years of ethnographic data with low-income African American families, the target population of the adolescent fatherhood study. We chose the kinscripts conceptual framework, which permits researchers to, “organize and interpret qualitative observations of the temporal and interdependent dimensions of family life course transitions, the creation and intergenerational transmissions of family norms, and the dynamics of negotiation, exchange, and conflict within families as they construct their life course”.1 Burton and Stack based this framework on ecological theory and life course theory.6-8 Both the ecological and life course theories stress the importance of having a flexible definition of family to accommodate the variety of family structures (Table 1) common among African American families,9 including our target population. Additionally, we focused on the challenges in the environmental context in which these families function, on parenting over time in the context of adolescent developmental stage, and on examining fathers within their environmental context from multiple perspectives. The kinscripts conceptual framework1 consists of three conceptual domains defined by Burton and Stack: kin-work (work to sustain the family over time), kin-time (timing and sequence of life course transitions) and kin-scription (recruitment for kin-work). We developed general definitions of each of the conceptual domains at the beginning of data collection: kin-work (work needed to ensure the survival of the baby), kin-time (shared understandings about the appropriate sequence of role transitions into parenthood and grandparenthood) and kin-scription (the process of socializing individuals to new roles, specifically what and who facilitates and hinders the kin-work of these adolescent fathers). Description of the adolescent father study. We briefly describe study methods although a more detailed description of the methods is found elsewhere.10-13 This study was approved by the University of Illinois at Chicago Institutional Review Board. We utilized qualitative, longitudinal case study design to examine the evolution of the paternal role for unmarried, low-income, African American adolescent fathers. Twenty-six kinship systems (adolescent father, his parents, the adolescent mother and her parents) were recruited using a variety of strategies, including radio ads on predominantly African American radio stations, recruitment flyers and face-to-face recruitment at community agencies. Shortly after enrollment one kinship system was lost to follow up and a second was withdrawn when the adolescent father was killed. Therefore, we interviewed a total of 24 kinships systems when each adolescent mother was in her third trimester of pregnancy and again when the adolescent parents’ baby was 1, 6, 12, 18 and 24 months of age for a total of six data points.10,11 Adolescent parents were between 14 and 19 years of age; 75% of the adolescent fathers were 17 and 19 years of age.14 We provide more detailed information about the study sample elsewhere.15 Our inclusion criteria specified that both adolescent parents report wanting the father to remain involved with his child.11 This was the first birth for both adolescent parents, all of whom were unmarried at enrollment and who remained unmarried for the duration of the study. At least one parent of the adolescent parent (their baby’s grandparent) had to participate in the study since the kinship system, and not individual kinship system members, was the “case” or unit of analysis for this study. We did not differentiate biological from surrogate parents during data collection or data analysis. Prior to the adolescent father’s death there were a total of 111 enrolled participants: 25 adolescent fathers, 25 adolescent mothers, 50 paternal and maternal grandmothers and 11 paternal and maternal grandfathers. Variations in household structure included an 18 year-old adolescent father who lived alone and an adolescent couple living with the paternal family in a five-generation household.14 Over 95% of the interviews were conducted in the subject’s home. Multiple research team members were sent to the home so that each adolescent parent and grandparent could be individually interviewed simultaneously. Each person in the kinship system was interviewed separately at each data point using a semi-structured interview. Our interview guide included open-ended questions based on preliminary studies, including a pilot study.11 The questions in each interview guide for each person in the kinship system were similar but the wording was adapted for each role. Interview questions were designed to elicit the perspectives of each kinship system member regarding the paternal behaviors of the adolescent father in each kinship system, specifically: the timing of role transitions to parenthood and grandparenthood for kinship system members (kin-time), the kin-work of the adolescent father and the kin-scription strategies that facilitated or inhibited the kin-work. Each interview was audio-recorded and lasted one-two hours each. Interviews were professionally transcribed. Methods What follows is a description of the iterative process we used to create population-specific operational definitions from the study data. The actual population-specific operational definitions are included in Tables 4-6 and described later. For this process, we used approximately 65 (slightly more than 10%) kinship system member interviews of the 580 interviews conducted.14 Developing data management templates. We began this process by developing data management templates in order to organize the large qualitative database. These templates included the descriptive codes, the individual case summaries (ICS), and kinship case summaries (KSS). The descriptive codes were derived from the broad conceptual domains of kin-work, kin-time, and kin-scription, the interview guide, and the data. Our provisional code book was based on codes from preliminary research and the interview guide. We revised the codes based on incoming data. The final codebook of descriptive codes was version eight. These codes were one of three types: (a) substantive areas, such as parenting behaviors; (b) the three conceptual domains and (c) kinship member whose behavior was being reported. The template for the major topical areas of the individual case summary (ICS; Table 2) were developed from the interview guide and from patterns in the data for each conceptual domain noted to emerge during data collection. The ICS was used to summarize relevant data from each individual transcribed interview. Comparatively, the major topical areas of the kinship case summary (KSS; Table 3) template was developed to summarize information at the kinship system level and focused on similarities and differences among kinship members. We also examined differences between the adolescent father’s family (paternal family) and the adolescent mother’s family (maternal family). Since the kinship system was the unit of analysis, researchers compiled data from the ICS for each kinship system member for each data point to develop the KSS. In particular, they noted patterns of similarities and differences within the kinship system at each data point with emphasis on changes over time. Both types of summaries (ICS and KSS) were guided by a description of case summaries.16 Data were entered at the descriptive rather than interpretive level for both types of case summaries.17 The processes of developing the final version of the descriptive codes, the individual and kinship case summary templates, and subsequent data management and analysis activities, were conducted over a two year period during monthly or biweekly meetings by the research team of five investigators. The research team was multidisciplinary and members contributed academic, research and clinical expertise in adolescent development, nursing, psychology and medical sociology. The team included both Caucasian and African American members with extensive experience with underserved and vulnerable populations. The clinical, academic and racial/ethnic diversity of the research team enhanced the group’s ability to fully examine the complex phenomenon by providing multiple perspectives. It is also essential to note that the investigators actually performed most of the interviews during the first year of data collection and were immersed in the data prior to actually beginning data analysis. Input from our consultants, Dr. Linda Burton, one of the originators of the kinscripts framework, and Dr. Kathleen Knafl, our qualitative methods expert consultant, provided ongoing support in assisting the team to stay true to the kinscripts framework and in maintaining systematic rigor during analysis. Components of the conceptual domains. After developing data management templates to organize the large qualitative database, we constructed components (a grouping of similar data under one term) for each of the kinscripts framework1 conceptual domains. This step began by linking descriptive codes with the appropriate kinscripts conceptual domain. The step was followed by the systematic selection of interview transcripts for the first three data points (prenatal, one, and six month interviews) for five kinship systems (for approximately 65 individual interviews). The kinship systems were selected based on the completeness of the data sets and variation in the amount and level of contact between the adolescent father and his baby as recorded in the ICS and KSS. For each conceptual domain, we reviewed all previously coded data for all members of the five kinship systems linked with that conceptual domain. Two research team members divided the data from the kinship systems and worked independently to review the coded data segments to verify their conceptual relationship to the domain. This process was repeated for each of the kinscripts conceptual domains. The beginning work to define the components came from two different sources: preliminary themes of kin-work from other work12 and patterns of data noted in the KSS. An additional strategy proved very useful in identifying the components for kin-scription and kin-time. We created a matrix as a working grid for outlining all of the components for each of these domains. The matrix contained the kinship members (in the columns) and a preliminary listing of components derived from the patterns of data delineated in the KSS template (in the rows). Investigators used similar techniques while working independently. These techniques were used to review coded data for a selected group of transcripts and to complete and expand the preliminary components listed in the matrix. Findings were then presented to the entire research team to further develop the components. During meetings the team referred back to the theoretical foundations of the kinscripts conceptual framework to guide the completion of the process. Findings were presented to the entire research team to develop the operational definitions for each of the components that were specific to our study of unmarried, low-income African American adolescent fatherhood. Results Population-specific operational definitions. Operational definitions of Kin-work. Burton and Stack1 define kin-work as “the work that families need to accomplish to survive over time” (p. 175). Kin-work is a broad construct that incorporates survival tasks, such as providing food and shelter, caring for dependents, procreating, and maintaining connections among family members. Based on the data kin-work was categorized into six components: being physically and emotionally present; providing direct care; teaching and guiding; providing material support; providing indirect support, such as buying food for the family in the home where the baby resides and general kin-work. (See Table 4 for the components for kin-work and their operational definitions.) Much of the work on fatherhood was performed with adult, married, Caucasian fathers and has focused on interactions between fathers and their children, using the classic definition of paternal involvement developed by Lamb, Pleck, Charnov, and Levine.18 Lamb et al. identify three components of paternal involvement, interaction, availability and responsibility. We did identify overlap for some of our components, interaction (providing direct care, teaching and guiding), availability (being physically and emotionally present) and responsibility (providing material support). We identified two additional components, providing indirect support and general kin-work, not previously described but are consistent with the kinscripts conceptual definition of kin-work.1The remaining behavioral codes identified by our research team. Operational definitions of kin-scription. Burton and Stack1 define kin-scription as “the process of assigning kin-work to family members” (p. 175). They note that “rather than accept the attempts of individuals to set their own personal agendas, families are continually rounding up, summoning, or recruiting individuals for kin work.” Kin-scription, as conceptualized by Burton and Stack, is an ongoing process. The strategies may be verbal and/or non-verbal, direct or indirect. Because of the potential conflict between personal and family agendas, power is brought into play in the kin-scription process, leading to a degree of tension in the kin-scription process. Burton and Stack propose that some kin are more easily recruited than others, and that women are socialized to value the maintaining of kin ties, which may make them more susceptible to being kin-scripted for kin-work. They also noted the presence of negative kin-scription, defined as excluding a particular kin member from kin-work, either in a specific area or a complete rejection of that person as a performer of kin-work. We identified multiple instances of negative kin-scription in the data which had been described, but not explicitly defined, in the kinscripts framework. Although instances of negative kin-scription were common, they were less numerous than those of positive kin-scription. We developed three categories of components for kin-scription: kin-scription within the kinship system (positive and negative), kin-scription outside of the kinship system (positive and negative) and kin-scription evaluation. Kin-scription within the kinship system (positive and negative) involved kin-scription strategies performed by members of the kinship system. We identified nine components for this conceptual domain (Table 5). These components were: role assignment, strategies for teaching kin-work, tangible support, relationship with father/father surrogate, prior experiences with childcare as a way to learn the role, emotional support to adolescent father, emotional connectedness of the adolescent father to the baby, adolescent father’s self-initiated learning strategies and negative kin-scription. We identified the components of self-kin-scription, which are kin-scription behaviors performed by the adolescent father himself, and were not previously addressed in the kinscripts framework. Burton and Stack1 defined kin-scription as the efforts of other members of the kinship system directed toward a particular individual, however, in our data kinship system members described instances of adolescent fathers performing self-kin-scription for kin-work by deliberately seeking out learning experiences related to his performance of kin-work. These instances included reading about pregnancy or babies, watching videos about pregnancy or babies, attending prenatal visits, and even volunteering to care for a friend’s or relative’s baby. We felt that this was a conceptually important element of the overall kin-scription process. Self-kin-scription appeared to have the same intention as kin-scription by other family members but was initiated by the adolescent father. The presence of both positive and negative kin-scription highlighted the need to develop the category of component to evaluate the effects of kin-scription strategies described below. Kin-scription outside of the kinship system. Kin-scription strategies were also performed by persons outside of the kinship system and included both positive kin-scription by non-kin and negative kin-scription. Instances of kin-scription performed by non-kin included descriptions of health professionals, teachers and organizations, such as schools or neighborhood groups and, in particular, the adolescent father’s peer group. Kin-scription efforts by persons outside of the kinship systems were reported to be performed to either reinforce or to counter the kin-scription efforts of kinship system members. Kin-scription evaluation , The third category of components for kin-scription, was defined as the perceived receptivity of the target to kin-scription strategies and reflected the degree to which the kin-scription strategies were successful in recruiting a kinship system member to actually perform the desired kin-work. Not all interviews included these types of statements. The target’s receptivity could be described as being either positive or negative. Examples of adolescent father’s receptivity to kin-scription strategies included statements, such as: “he’s learning real well” or “he just doesn’t listen”. Evaluative statements were usually made by grandparents. A few adolescent fathers did describe the impact of kin-scription on their definition or acceptance of their kin-work; e.g., socialization to their role. We determined that these additional elements should be included after consulting with Dr. Burton. A list of the components for kin-scription and their operational definitions are summarized in Table 5. Operational definitions of kin-time. Burton and Stack1 define kin-time as the temporal guides or scripts that exemplify the family members’ shared understanding about when and in what sequence life course transitions and kin-work are to take place. For this study, it is the family’s norms (spoken or unspoken rules) about the adolescent’s life course trajectory, including childbearing and parenting. It comprises the shared understanding of kinship system members about the age and circumstances in which the adolescent is expected to become a parent and the age that the adolescent’s parents expect to become grandparents. These family’s norms, guides, or scripts related to the timing of childbearing and parenting are unique to each adolescent parent’s family of origin so there is potential for disagreement within the kinship system. As such, these norms, guides, or scripts may not reflect the norms, guides, or scripts of the dominant culture or the norms, guides, or scripts of other families who reside in the same neighborhoods or communities. When a family member’s behavior is consistent with the family’s shared temporal guides or scripts, or kin-time, that behavior is synchronous with the family’s kin-time and may be considered on-time behavior. However, when a family member’s behavior violates the family’s shared temporal guides or scripts, the family member’s behavior is asynchronous with the family’s kin-time and is considered to be off-time behavior. Thus adolescent pregnancy can be either synchronous or asynchronous with the family’s kin-time. Additionally, kinship system members may have different temporal guides or scripts for adolescent childbearing, parenthood, and grandparenthood. Our further examination of kin-time revealed four components: emotional responses to the adolescent pregnancy and parenting, readiness for the adolescent pregnancy and parenting, and life course transitions into the parenthood and grandparenthood role. A list of the components for kin-scription and their operational definitions are summarized in Table 6. We provided more depth to the kin-time domain and have described it as a process of life course transitions. We relied upon empirical data to show the impact of parenthood during adolescence on the life course trajectory of kinship system members and each individual’s readiness for transition to new roles. Discussion We contribute to the literature a detailed description of a method of using qualitative data to create population-specific operational definitions as an early step in the process of adapting an a priori theory. The knowledge of this method is important for all researchers who want to apply a theory developed from data on middle-class Caucasians to a target population of color for whom the theory was not specifically developed. Presently fatherhood literature focuses almost exclusively on the father’s interaction with his children.19 Utilizing the kinscripts conceptual framework promoted a broader examination of fatherhood that included the contextual environment, timing of development (kin-time) and socialization strategies (kin-scription)to get kin-work done. We hope this information can be used to reconceptualize unmarried, low-income, African American adolescent fathers, who are generally viewed negatively. We were also able to reflect changes over time in our operational definitions, such as impact of pregnancy and parenting on life trajectory, changes in expected and actual kin-work and the adolescent father’s receptivity to kin-scription. Limitations Several limitations may affect our findings. We chose to perform individual, rather than family-level interviews, to ensure confidentiality and promote disclosure. We recognize that data from family-level interviews would provide different types of data and permit us to interpret the data with the family as the unit of analysis. Nevertheless, examining the individual-level interviews of the kinship system has provided valuable information for future research to explore dynamics with the kinship system as the unit of analysis. Since we used approximately 10 percent of our interviews, it is possible that with additional analysis, modifications to our operational definitions may occur. We doubt, however, that any modifications would be extreme for two reasons: first, our operational definitions were consistent across the interviews used and second, we purposively selected the kinship systems based on differences in the adolescent father’s level of involvement. Finally, we used self-reported data from interviews only and did not incorporate observation of any behaviors. Future research will include interviews as well as observations of fathers’ interaction with their children. Conclusions Although rigorous and time-consuming, qualitative data can be utilized to develop operational definitions in the early steps of revising and constructing theory for specific population. The iterative and rigorous process we have described to develop population-specific operational definitions may prove helpful for researchers and clinicians who work primarily with culturally diverse populations. Operational definitions specific to these populations may permit a more comprehensive understanding of the population which would most likely lead to better health care and outcomes. Table 1: African American Family Structures as Described by Billingsley (1992)
Table 2. Major Topical Areas in the Individual Kinship System Member Case Summary (ICS) for Each Data Point (prenatal, 1, 6, 12, 18 and 24 months)
Table 3. Major Topical Areas in the Kinship System Case Summary (KSS) for Each Data Point (prenatal, 1, 6, 12, 18 and 24 months)
Table 4. Operational definitions of Kin-work
Table 5. Operational Definitions of Kin-scription
Table 6. Operational Definitions of Kin-time:
References
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||