In terms of the study sample, in contrast to quantitative research, qualitative sampling aims to gain depth and understanding of specific individuals in a specific setting. (Mays & Pope 2000) Thus the sample is deliberately biased, (Greenhalgh 1997) to achieve this aim purposive sampling is often used. (Johnson and Waterfield 2004) This involves selecting participants that the researcher thinks will deliver a depth and richness of information directly relevant to the research question. (Johnson and Waterfield 2004) Davey et al (2004) selected fourteen participants eight of which were recruited from two local Parkinson support groups in Hampshire and six from participants of a previous study. All caregivers were marital partners of the Parkinson’s patient and ranged from 44 to 79 years of age, with am average age of 69.9 years. In terms of the number of participants this would seem to be adequate to produce validity as saturation (a point where extra participants do not reveal extra data themes) occurred after the first five interviews. (Davey et al 2004) Qualitative samples should never be selected purely on the basis of convenience (Greenhalgh 1997) and should reflect a full range of diversity of participants within a specific setting. (Popay et al 1998) In addition the sample should, if appropriate, attempt to recruit participants that may contradict or modify the analysis. (Mays & Pope 2000) Therefore with regard to the Davey et al (2004) sample it would seem to be slightly limited in terms of the diversity of participants, with regard to the marital status to the patient of all caregivers (all married to the patient and may be more emotionally connected) and in the older age of participants (younger caregivers may be more able to cope). We may not go into more detail at this point as the diversity of caregiver characteristics, individual ages and abilities were not outlined or discussed specifically in the paper. However, it must be noted that although flaws within the sample, this does not detract from the validity of the papers conclusions within this setting only the transferability of the concepts/ findings to differing settings. (Mays & Pope 2000) Therefore we could conclude that in terms of the sample the conclusions are valid, and help to directly answer the research question.
With regards to methodology, in depth semi-structured interviews were employed to explore caregiver’s views and experiences. This means that a loose structure was employed during the interview, which consisted of open ended questions in which emerging concepts may be explored in more detail and for further clarification. (Mays & Pope 2000) This approach has obvious advantages when trying to explore and probe for themes and experiences that a structured interview does not possess and leads us to conclude that this approach was indeed appropriate to answer the research question lending credence to the papers validity. However, credibility and dependability and thus validity could have been further enhanced by a process known as triangulation. (Greenhalgh 1997, Johnson & waterfield 2004, Mays & Pope 2000) According to Mays and Pope (2000) triangulation is a process whereby data is accrued from differing data collection methods, adding depth and thoroughness too the raw data. This greatly improves validity (Greenhalgh 1997) For example, with the Parkinson’s paper, collecting data from a focus group, in addition to interviews may well have enhanced internal validity. (credibility and closeness to truth) This would have allowed the caregivers the chance to interact with each other, helping to probe and harness information. However, it must be noted that in terms of validity, triangulation techniques add to credibility and in terms of the research question in depth semi-structured interviews were appropriate and valid in answering the research question.
Furthermore, to conduct a valid qualitative interview and undertake efficient data analysis takes considerable skill, (Mays & Pope 2000) particularly with regard to the self-awareness of the interviewer (Koch 1993) for example the researcher should be aware of monitoring their interview technique, not asking leading questions and giving respondents enough time to elucidate. (Mays & Pope 2000) This may be aided by keeping a field journal where interactions and reactions are noted and reflected on. (Koch 1993). Which leads us to a concept that many strategists call reflexivity. This may be considered in terms of the background, experiences, and views that the researcher naturally brings with them when conducting research. This subjectivity if made explicit and transparent at the outset may enhance credibility, giving the researcher an enhanced instinctual awareness of an environment or situation and consequently the experienced and self-aware researcher becomes an incisive research tool. (Mays & Pope 2000) These subjective preconceptions are not the same as bias unless the researcher does not identify them in the paper. (Malterud 2001) Greenhalgh (1997) clarifies further to suggest that a researchers’ expertise in the field to be examined enhances credibility but only if their backgrounds are described in detail so that the results may be interpreted accordingly by the reader. This leads us to the most significant flaw in the paper we are reviewing. Unfortunately we are not made aware of the researcher’s perspectives, or even which of the researchers conducted the interviews. The backgrounds, areas of experience or expertise of the researchers are not stated or even discussed, leaving us without transparency in terms of researcher bias. Thus, many authors would argue that this aspect of non disclosure does indeed detract from the validity of the papers’ conclusions. (Malterud 2001, Greenhalgh 1997) however, as we shall discuss in conclusion to this essay, this does not necessarily invalidate the papers’ conclusions, as the reader may make validity inferences and meaning from the transcripts in the results section. Furthermore, Greenhalgh (1997) suggests that it is the data analysis section where sense can be made of the raw data.
Davey et al (2004) utilised an iterative or sequential analysis, grounded theory approach to the analysis of the transcribed interviews. This means that the initial interviews were analysed for emergent themes or for refining avenues for further investigation for subsequent interviews. (Mays & pope 2000) This again is perfectly congruous and appropriate with regard to the aims and objectives of the study. Grounded theory as a form of content analysis involves coding incidents within the data to identifying emerging categories or themes which can then be confirmed and qualified with further evidence or matched themes. (Mays &Pope 2000) So, once again the method of analysis is entirely appropriate and congruous with the research question, to extrapolate caregiver’s views and experiences. However, this process again needs skilful application in the recognition of phrases, incidents or types of behaviour specific to the environment under investigation. (Mays & Pope 2000) which brings us back to the concept of reflexivity. As we are not aware of the background or expertise in the field of researchers we are not able to relate this information to their interpretations or prejudices. In addition many authors state that the method analysis should be systematic and include some form of quality control. (Johnson & Waterfield 2004, Mays & Pope 2000) This means that the data should be coded by more than one researcher to ensure similar interpretations of the data. (Greenhalgh 1997) The first five interviews in the Parkinson’s paper were coded independently by different members of the research team who then met and formed a consensus on viewpoints and emerging themes. Davey et al (2004) Thus, lending extra credibility to an already appropriate method analysis. (Greenhalg 1997, Johnson & Waterfield 2004) This could have been further enhanced by respondent validation or member checking, where participants review the collected data or the emergent themes/ viewpoints with a view to being able to recognise themselves (Johnson & waterfield 2004) Although participants were, at the end of the interview, allowed to add any other concerns they were not shown nor were they involved the in refining the completed analysis, which would have added further validity (Johnson & Waterfield 2004, Mays & Pope 2000) This does not detract from the appropriateness or the credibility of the data analysis and would only have conferred further validity.
A further drawback found in the study, that dovetails with reflexivity is the concept of an audit trail. Even if the reader did not agree with the research findings, they should be able to follow the readers’ deductions and decision pathways in a systematic and analytical way. (Koch 1993) Common faults found in qualitative research are inadequate descriptions of the data analysis methods or decision trails made by the researchers. (Mays & Pope 2000) Mays and Pope (2004) elucidate further by suggesting that: “by the end of the study, it should be possible to provide a clear account of how early, simpler systems of classification evolved into more sophisticated coding structures and thence into clearly defined concepts and explanations for the data collected. “ (Mays & Pope 2000:95) It is possible, to a certain extent, to infer a tenuous decision trail from the transcripts presented in the paper, this helps the reader to infer validity by looking at the raw data and at times many points are made in one paragraph which doesn’t seem to be supported by the data, however, there is data on show to support the main themes, but not in all cases, thus a detailed audit trail may have helped in clarification and interpretation.
We may conclude that the primary research question was clear and succinct. In terms of answering the research question, we have shown that the research papers’ sample was adequate and appropriate to the task, and the data collection and analysis techniques were also appropriate and congruent to the primary research question. The flaw in the representation of the sample only limits transferability and does not impinge upon the validity of the paper. With regard to the methods used, highlighted flaws would help to increase validity and again do not detract from the papers’ credibility. In terms of the lack of background disclosure and the unclear audit trail, we have shown that some authorities suggest these flaws could indeed detract from the validity of the papers’ conclusions and at least demand further investigation. However as Davey et al (2004) illustrate in their discussions their findings and conclusions concur with similar studies within the realm of Parkinson’s disease and indeed with similar studies relating to the impact of falls on caregivers with other conditions, which Greenhalgh (1997) purports helps with conclusion validity. Greenhalgh (1997) also suggests that when assessing validity, the data interpretation should concur with common sense and the themes uncovered in the Davey et al (2004) paper, causes and consequences of falls, managing falls, information about falls and the impact on caregivers would certainly slot into that pattern. With these conclusions in mind we may suggest that the conclusions in the paper are valid and do indeed answer the primary research question.
References
Davey C, Wiles R, Ashburn A, Murphy C 2004 Falling in Parkinson’s disease: the impact on informal caregivers. Disability and Rehabilitation 26 (23): 1360-1366
Greenhalgh T 1997 How to read a paper. BMJ publishing, London
Johnson R, Waterfield J 2004 Making words count: the value of qualitative research. Physiotherapy Research International 9 (3): 121-131
Koch T 1994 Establishing rigour in qualitative research: the decision trail. Journal of Advanced Nursing 19: 976-986
Malterud K 2001 Qualitative research: standards, challenges and guidelines. Lancet 358: 483-488
Mays N, Pope C 2000 Qualitative research in healthcare 2nd edition. BMJ books, London
Popay J, Rogers A, Williams G 1998 Rationale and standards for the systematic review of qualitative literature in health services research. Qualitative Health Research 8: 341-351
