An Ethnographic Case Study Approach to Studying the Process of Child Abuse Investigation in the United Kingdom
Jon Prosser
ABSTRACT: This article is concerned with methodological issues related to studies into investigative practices following accusations of child abuse. Part One describes an ethnographic case study approach adopted by a study conducted in the UK and explores how research design was in part molded by the circumstances and context in which the study took place. Part Two takes a broader perspective and identifies some of the problems and benefits which may ensue as a result of adopting a qualitative approach. It is apparent that adopting a qualitative strategy, particularly an ethnographic one, is problematic. However, whilst ethnographic case study work brings with it significant methodological problems it offers valuable insights into the process of child abuse investigation carried out by protection agencies.
Part One
Prosser and Lewis (1992) conducted a study between 1989 and 1991. They considered 30 cases of families who claimed to have been wrongly accused of child abuse. The central aim of the study was not to establish the guilt or otherwise in each case but to identify families' perceptions (children, mothers, fathers, and grandparents) of the process of child abuse investigation. What is presented here is not a prescriptive or "ideal" model of how such a study should be conducted, but an attempt to record one approach. This approach reflects some of the difficulties, uncertainties and mistakes made by a research team who used a sociological framework in studying how accusations of child abuse were investigated.
CLICK ON COMMENTS TO READ MORE...
Subscribe to:
Post Comments (Atom)
About Me
- Colombia
- Pamplona, Nsder, Colombia
- Master in Teaching English as a Foreign Language.
Background
ReplyDeleteThe beginnings of the study had a profound effect on the way in which it was conducted. The work came about not out of free-standing academic interest, but because a client organization provided the financial resources and focus. The client organization "PAIN" (Parents Against Injustice) is a British organization which provides support for parents and children in cases where parents are accused of child abuse which they refute. PAIN is recognized as a group which supports and fights for the rights of such families. They commissioned our research group (within an independent academic institution) to carry out a study into the process of child abuse investigation in the UK. Whilst the client organization provided access to a large sample of cases which later formed a data base for the study, it was agreed that the project would be conducted without interference by the client and that the final report was to be based on findings grounded in the data.
Research Design
Given that the focus of the study was to be the process of child abuse investigation which reflected families' perceptions, a qualitative approach appeared most befitting. A case study strategy, with its emphasis on the examination of one setting, was considered appropriate since it placed action and events in context. In order to take account of the issue of uniqueness, to deepen understanding of the phenomena under study, and to make findings more predisposed to theory generation and generalization (Miles & Huberman, 1986), a multi-site case study approach was adopted. This enabled issues emerging from one case to be compared and contrasted with issues from other cases. It was hoped that this would enhance not only the validity of findings and the scope of constructs developed, but also contribute to the robustness of the understandings established.
Consideration was initially given to a research design which would explore the multiple perspectives of all participants and compare families' views with those of investigators, for example social workers, police, and doctors. The notion of adopting a participant-observer role and collecting first-hand data from "new" cases where the family claimed innocence was appealing.
Early attempts to establish agreement on a research design with bodies charged with investigating cases of possible child abuse in order to gain access to the investigators' world proved futile. Instead, we decided to take a sample of PAIN's cases which had reached a conclusion and to "backward map," drawing on the memories of both families and investigators. Again this proved fruitless. The research team had felt ethically bound to mention to the investigating bodies that the sponsoring organization was PAIN and this effectively closed many doors. Barred access is not a new phenomena to those in sociological research who wish to study powerful organizations as Taylor (1989) points out:
ReplyDeleteYou will find many ethnographic studies of relatively powerless groups, such as school-children, the sick and handicapped, gays or dope smokers, but very few, if any, of powerful groups such as leading politicians, senior civil servants or military chiefs. This is not because sociologists are not interested in power and how it is used. They are. It is simply because sociologists (and other potential observers) are not normally granted access to centers of power. The only knowledge we have of what goes on in some powerful groups comes when one of the participants makes disclosures and even then, in the UK at least, there are problems in getting such revelations published. The problem of access illustrates that what can be achieved through participant observation is strictly limited (pp. 66-67).
An "ideal" ethnographic approach therefore was not a plausible option. Although there was consolation in the possibility that a participant-observation role could have eroded the "natural" behavior of families and investigators, a third, less satisfactory design, was chosen out of necessity rather than preference. It was decided to consider cases from PAIN files and to accept the limitations that would occur as a result of focusing down on families' perceptions without recourse to comparison of alternative, possibly conflicting interpretation of events provided by investigators.
Because of limited time and resources, and knowing that investigation into cases may take years to complete, only cases that had reached a conclusion were included in the sample. This meant that much of the data were based on an historical perspective, i.e. families looking back at their cases.
ReplyDeleteThere were at least two basic problems to consider. Firstly, because families had sought PAIN's support, it was probable that they had received a strongly negative experience of the investigative process (accepting there are few positive experiences). This meant that, not only was the sample to a degree self-selecting, but also that it was biased in favor of those families with strongly negative views. Secondly, there was a possibility that families' perceptions would be a reflection of memories, distorted as a result of their negative experiences. These factors, as far as possible, were taken into account and perceptions were compared and contrasted by method, researcher, and source triangulation (Denzin, 1970). This approach was taken to establish a degree of internal validity within cases and external validity across cases.
Figure 1 represents a model of the research design illustrating the various key stages. The data were collected by three researchers — one female (R1) and two male (R2 and R3) — from a sample of 30 cases. A modified analytic induction strategy was used (McCall & Simmons, 1969) to collect and analyze data, develop an understanding of the process of child abuse investigation, and test that understanding rigorously.
In order to achieve this there was a need, particularly during the initial "open" phase, to standardize the team's recording of data. The team recorded day-to-day activities of participants (and researchers) in a field diary. The data were organized under 4 subheadings: (1) Contextual information, which provided information about participants, sites, and settings; (2) Analytic Memos, in which a record of systematic thinking about the data were made. Such memos contained (a) new concepts which emerged, (b) emergent hypotheses which required testing, and (c) further data collection required in the future in order to "ground" the emergent concepts more fully; (3) A Record of Data, where information from participant observation of situations, events, interactions, activities, including descriptions and quotations, were recorded; (4) Methodological notes, which identified the team's concerns with regards processes and procedures associated with the collection of data in the field.
ReplyDeleteThis approach to the recording of data not only helped individual field workers to identify how working hypotheses emerged from the data and to check their work for validity and reliability, but enabled co-workers to compare, contrast, and cross-check their work and the work of others in the team. It was not intended that a common method of organizing data be used by the research team (for example using predetermined categories). Although such an approach may have ensured a degree of reliability across the team, it may have been at the cost of premature coding and inflexibility.
It can be seen in Figure 1 that in stage one, the "open" phase, each researcher collected data and identified significant issues and explanations for the cases they had studied. During the second stage, emergent findings from the research team were compared and contrasted, in order to identify common, interesting, or unusual features. In stage three, the team re-entered the field and, by focusing on key issues established earlier in the study and reflecting on data from further cases, interpretations were refined or reformulated.
ReplyDeleteAlthough the above process is common in qualitative research, one aspect of the process was more unusual and possibly controversial. It is normal, prior to conducting any research, to review the literature. In this study this was not done. The research team had not been involved in child abuse research prior to this project (child welfare and education were their forte). Rather than consider this a disadvantage the team decided it was advantageous. By not reviewing the literature and thus retaining the "virgin" state of their knowledge, the team hoped that assumptions, presumptions, and prejudices which may accompany field workers into the field would be limited, if not avoided. This did mean that some findings which surprised us were commonplace to experienced researchers in this area.
It was only during the second stage, when emergent substantive issues were established, that the search for the insights from other studies began. On reflection, whilst there were some advantages of this approach, this may have been a mistake. The study came up with new concepts such as "system abuse," but other features might have been identified earlier and explored in greater depth, if sensitizing issues had been established by reviewing the literature prior to entering the field.
The Sample
ReplyDeleteSampling, as with the research design for this study, was problematic. In the UK a difficulty for researchers studying child abuse is identifying a population from which a sample may be drawn. This task is particularly difficult for those working in a statistical paradigm or focusing on families who have been wrongly accused, since records have not been kept by an impartial or a government agency. No attempt is made to differentiate between those who are guilty and those who are innocent. There is no equivalent of the USA's National Center on Child Abuse and Neglect and hence no reliable figures that are comparable to those given by Besharov (1986, 1988).
Access
ReplyDeleteAccess to families and individuals within families was less of a problem than anticipated. Initially, contact was made by letter explaining that the study was supported by PAIN and outlining its aims. Family members were reticent in the beginning. On a number of occasions the scampering of feet could be heard when researchers first knocked on the door. We were told the children habitually ran and hid when strange or "official" looking people arrived. Access to children was often only gained once parents, the gatekeepers to much of the data, were confident that we could be trusted. Time was taken to socialize with families to gain acceptance and an understanding of contextual features. The ability to develop relationships quickly is an important skill for the case study worker as Nisbet and Watt (1982) point out:
In case study you as a researcher have the luxury and the dilemma of being your own "chief instrument." Ultimately the success or failure of your efforts will depend on your ability to develop good personal relationships
FOR A COMPLETE TEXT, GO TO THIS LINK
ReplyDeletehttp://www.ipt-forensics.com/journal/volume7/j7_3_1.htm