Due to lack of research in this area this is an exploratory investigation. The qualitative method is likely to be the most appropriate to explore the aims and objectives of this research as it is essential in areas where there is little knowledge and opens up new areas of research (May and Pope, 1995). Experimental and quantitative methods are less well suited to answer these research questions as they do not examine the experiences and actions of people’s life experiences in much depth (Gribich, 1999). Qualitative research is often prerequisite to quantitative research, particularly in an area that has received little or no previous investigation as it can generate new concepts and theories of practice (Carpenter, 1997).
Qualitative research methodology focuses on individuals’ lived experiences as they are presented in thoughts, ideas, feelings, attitudes and perceptions. In addition, the research approach emphasizes human behaviour and social interaction. It explores the quality of a phenomenon, not the quantity. The aim of qualitative methodology is to develop new knowledge based on participants’ own beliefs and experiences, not on pre-de¬?ned, testable hypotheses. It is inductive rather than deductive, and it is interpretative rather than predictive.
The design is ¬‚exible, iterative and emergent and therefore requires of the researcher an ability to change and adapt the research process in accordance with emerging results. Qualitative research is thus different from quantitative research as it allows for ¬‚exibility throughout the research process. Several data collection methods can be used, such as individual interviews, focus group discussions, or participant observations, in order to gain a deeper understanding of health, illness and rehabilitation. It can be used in combination with quantitative studies, but also as a research method of its own.
In health research, the qualitative methodology has gained increasing credibility during the last decade. However, it is not yet frequently used in rehabilitation research. As rehabilitation, outcomes are dependent on people’s attitudes, thoughts and motivation regarding the rehabilitation process, and as the rehabilitation process in itself builds on social interaction, studies with a qualitative design could become useful tools in the development and improvement of rehabilitation.
Creswell (15) de¬?nes qualitative research in the following manner:
Qualitative research is an inquiry process of understanding
based on distinct methodological traditions of inquiry that
explore social human problems. The researcher builds a
complex, holistic picture, analyses words, reports detaileviews
of informants and conducts the study in a natural setting.
(15: p. 15).
Morse (16) describes the research methodology as a tradition that is interpretative and which deals with the social world and how this world is interpreted, understood, experienced and produced by human beings. She also emphasizes that qualitative
methodology uses research designs that are ¬‚exible and sensitive to the social context in which the study is performed.
In accordance with Creswell, Morse states that qualitative methodology involves methods of analyses and explanations that are complex, detailed and contextual. The World Health Organisation (4) argues that it is important to describe culture and behaviour of people and groups of people and that the analysis should focus on the viewpoints addressed by those being studied.
One basic assumption in qualitative methodology is that realities are multiple and socially constructed (17). This means that they will vary between different groups of people and in different social settings. They are time and context bound. This is a social constructionist perspective (6). Realities are experienced differently depending on who is experiencing and judging them. Therefore, it is the researcher’s obligation to ¬?nd these differences, not to ¬?nd a single truth. The view of reality as multiple and changing is an ontological assumption.
Another essential assumption adhered to the qualitative tradition is the view that the researcher and the informants (i.e. people subjected to investigation) interact with each other (17). The research process goes on between the two and they will in¬‚uence each other. This is different from, and contradictory to, the notion of a neutral and distanced observer claimed in the positivistic tradition.
A third assumption is that qualitative research is inductive, time and context bound and requires an emergent study design. The emergent design means that the researcher should be ¬‚exible and sensitive to developing ideas, themes, questions and theories throughout the whole research process (17). This is a methodological assumption. Embracing this assumption makes it dif¬?cult for the researcher to compare and estimate associations between different social contexts.The researcher looks rather for the uniqueness of a social process or phenomenon.
STRENGTHS OF QUALITATIVE RESEARCH
Qualitative methods are useful when the issues of interests do not seem amenable to quantification (Skyes et al. 1992). Kitzinger (1994) argued that qualitative research highlights the respondents’ attitudes and priorities and can encourage open conversation and facilitate the expression of ideas and experience that might be left undeveloped in an interview. Skyes et al. (1992) argued that qualitative research methods will provide an indication of the range of health need in the area, the relative importance individuals attached to them and ideas about how they can be met. They further state that qualitative research can provide an opportunity for people to express views about services currently available and contribute ideas on the kind of services they would like. As qualitative research is responsive to individual situations, qualitative research can help local people to feel actively involved in the purchasing process rather then the passive providers of the information (Skyes et al., 1992). They further stated that the health need and priorities of less accessible groups in an area or of minority groups may also be most effectively tackled through qualitative methods.
WEAKNESSES OF QUALITATIVE METHODS
One disadvantage is, of course, that it is a much younger research tradition than the quantitative one, at least within the ¬?eld of health and medicine. Therefore it is not as tested as quantitative methods. It is sometimes regarded as very time-consuming, which is another disadvantage.
In qualitative research methods the interviewer may ignore important nonverbal communication or may accept comments at their face value, presenting narrative rather then interpretative analytical reporting (Dodds et al. 1996). Britten et al. (1995) argued that qualitative methods cannot be use for statistical relationships between variables. Dodds et al. (1996) argued that translation, transcription and process of sifting and digesting the data can be time consuming.
Semi structured in depth interviews were chosen as the most appropriate method of data collection for this study. Smith (2001) describes the ‘natural fit’ that exists between qualitative research and semi structured interviews, since this method allows much more flexibility then the more conventional structured interview, questionnaire or survey.
This method was chosen in place of other qualitative techniques such as focus groups because female Muslim participants may be reluctant to share their personal experiences of accessing to physiotherapy services. In a focus group situation they may be uncomfortable with a mixed group, where their anonymity and confidentiality could not be guaranteed. It can lead to increased insight into people’s thoughts, feelings, and behaviour on important issues (Bowling, 2002). The open ended nature of the question defines the topic under investigation and provides opportunities for both interviewer and interviewee to discuss some topics in detail. In a semi structured interview, the interviewer also has the freedom to probe the interviewee to elaborate on an original response or to follow a line of inquiry introduced by the interviewee.
There are disadvantages of using interviews as data collection methods. Interviewing is very time consuming and it may be difficult to control the length of the interview (Denscombe, 1998). In order to maintain the structure of the interview, the researcher took time to develop interviewing skills, such as keeping the questions clear and focused before the interview took place. This is because an interview is an activity requiring careful preparation, much patience, and considerable practice if the eventual rewards are to be worthwhile (Cohen and Manion, 2000).
Bowling (2002) highlights the potential threat of interview bias and the importance of reducing this through recognition of the potential researcher effect. Interview bias involves the use of leading, dual or biased questions. To avoid this neutral probes and prompts will be utilized.
The sample selection used in a study affects the reliability and external validity of the research therefore, it is important that the most appropriate sample method is chosen out of the wide variety available. Robson (2000) states,” Sampling is closely linked to the external validity or generaliseability of the findings in an enquiry”.
Sampling in qualitative methodology differs from sampling in quantitative research (4). Whereas probabilistic, random sampling is a gold standard in quantitative studies, non- probabilistic samples are preferred in qualitative research.
This means that the choice of study subjects is purposive and strategic. As the design is emerging along with data collection and analysis, the purposive sample allows for ¬‚exibility and changing sampling strategies throughout the research process. The sample should theoretically be representative of the study population, but it is not representative from a statistical point of view (17, 22, 23).
The World Health Organisation (4) has suggested several sampling techniques for qualitative methodology, of which I will mention 4. Maximum variation means that the chosen informants are different from each other in as many aspects as possible. It is important to decide the inclusion criteria so that the variation is captured.
Another technique is called snowball or chain sampling. This can be used when we do not know how to reach the people that we want to include in a study. We can then start by interviewing 1 person and by the end of the interview, ask the informant to point out another person who is similar or different from him and whom he believes will provide further information. A homogeneous sample is used when we search for people who are similar in certain aspects.
Sample size in qualitative research is necessarily small due, in part, to the complexity of the data. Bowling (2000) supports this and highlights that sufficient size is reached when, upon the judgement of the researcher, concurrent themes and issues are emerging from the participant. Size is therefore flexible, though guided by the time and resources available (Silverman, 2001).
Participants were selected using a non-probability purposive sampling strategy (Hudson, 2003). This was seen as most appropriate for this study as it involves selecting a sample from a population which meets the inclusion criteria of the research study. This sample will be a non-probability sample because there will be no intention to make a statistical generalisation on any population beyond the sample interviewed. Purposive sampling involves the researcher using their own judgement to achieve a particular purpose, to satisfy the needs of the study and make the sample theoretically representative (Robson, 2000). Recruitment of participants will be via posters in four different mosques in ladies and gents places of prayers.
The inclusion and exclusion criteria for the sample are in table below
Pakistani origin migrated to the UK post world war II
Born in the UK
Live in north-east of England
Live other then north-east of England
Age between 60-80 years
Age below 60 years
Can speak urdu and English
Can not speak urdu
Have knee joint pain for 3 or more years not due to injury, road traffic accident or any trauma
Knee joint pain due to injury road traffic accident or truma
Experience physiotherapy treatment in Pakistan and the UK
Never have physiotherapy in Pakistan
In response to the poster in the mosques, seven men and four women contacted researcher by telephone. Out of eleven people, two women and two men meet the inclusion criteria and agreed to take part in study.
DATA COLLECTION TOOL
Aide memoir re. Interviews can be found in appendix 3. The proposed method for data collection is a series of in-depth semi-structured interviews as this method is useful and relevant tool in qualitative research (Bailey, 1997). Research interviews generate rich data about perceptions, feelings, experiences, motives, attitudes and knowledge among individuals. The qualitative research interview is usually performed on a conversational basis, meaning that the researcher uses rather loose, broad and open- ended questions or interview themes in a thematic interview guide. The aim is to encourage the informant to talk. Kvale (23) states that a good qualitative research interview is one where the informant talks and the researcher is silently listening most of the time and only probes on things she/he hears during the interview. Sometimes the qualitative researcher may use a semi-structured interview form with more or less closed questions. It is however, not common to use only structured and closed format of questions similar to questions in a questionnaire.
The respondents were given a choice in the setting of the interview in order to make them feel as comfortable as possible, two participants chose the community centre and two chose to be interviewed in their own home for reasons of privacy. A series of open questions will be used to ensure the discussions stay within the overall aims of the study. These will be especially useful, as a novice researcher without the expertise to guide free discussion will carry out the interviews (Harvey-Jordan and Long 2001).
The decision to tape record the interviews was taken in order that the interviewer could concentrate on listening and responding to the interviewees rather than be distracted by trying to write down what had been said. Recording the interview will help the discussion flow because the interviewer does not have to write down the response to one question before moving on to the next. In note-taking there is an increased risk of interviewer bias because the interviewer is likely to makes notes of the comments which make immediate sense or are perceived as being particularly relevant or interesting. Tape recording ensured that the whole interview was captured and provides complete data for analysis so cues that were missed the first time can be recognised when listening to the recording.
Reliability and validity are the notions of quantitative studies rather then qualitative research. According to Silverman (2000) qualitative research does not aim to produce reliable data but aims to be consistent and dependable and to be representative of the data. Credibility, transferability, dependability and confirmability are the criteria commonly used to evaluate the trustworthiness of qualitative research (Lincoln and Guba, 1985).
Credibility in qualitative research relates to the term internal validity. This is more often associated with quantitative research where causal relationships between variables are sought, and it refers the truth of the study (Krefting, 1991). In qualitative research, where the beliefs that multiple realities exist, the truth of a study relates to how accurately the phenomena being studied is represented. This is reflected in the researcher’s willingness to include contradictory statements rather then to present a one sided argument from the data (Silverman, 2000). Credibility is the term used in qualitative methodology to answer questions about the truth value. As one of the basic assumptions in qualitative methodology is that realities are multiple, credibility refers to the researcher’s ability to capture these realities. Has he or she really understood and described the informants well enough? Would it be possible for other people to recognize themselves, or the context that we describe? Several techniques have been developed in order to increase credibility in a qualitative study, of which the most frequently used are prolonged engagement, triangulation, peer debrie¬?ng and member checking. Prolonged engagement refers to our efforts to really understand and become acquainted with the social context under study and the people in this context. When writing the report, researchers can describe how they made these efforts.
The job of the researcher is to accurately represent the multiple realities revealed by the participants (Koch, 1994). Krefting (1991) suggests that more sensitive information may be offered as a rapport is developed over time. The participants should be more honest in their answers with the assurance of anonymity from the researchers (Rubin and Rubin, 1995).
Researcher bias can be reduced by peer examination, that is the research findings and process can be discussed with impartial colleagues that may also check the categories developed from the data to enhance the truth of the research (Krefting, 1997).
A further strategy employed to ensure the credibility was triangulation (May and Pope, 1995) by the author’s field and the researcher’s interviewer notes and from a comprehensive literature review (Carpenter, 1997). Triangulation means that we try to view the research problem from different angles, for example by engaging several data collection methods, a team of researchers with different professional background or the use of different theories to “mirror” the developing results. Throughout the study, the researcher also engaged in continual self-reflection to maintain an awareness of how existing assumptions regarding the phenomenon might conceal new experiences of the phenomenon or prejudice my findings.
Peer debrie¬?ng refers to a technique where the emerging concepts, themes or model are presented to colleagues outside the project, for instance in a seminar discussion. Do they ¬?nd the results relevant, reasonable and logical? We can also check the emerging results by asking our informants, i.e. member checking. Do they recognize themselves in the descriptions? Sometimes we write a summary of the interviews and send it to the informants or we go back to them and present parts of the interpretations of the material (17).
Transferability means the ability to generalise from the findings of the study. Lack of generalisability is one of the criticisms of qualitative enquiry as, due to the uniqueness of each qualitative research situation, it is impossible to replicate results (Holloway and Wheeler, 1996). Shepard et. al. (1993) suggests that within phenomenological research, generalising is not the job of the researcher; instead the reader is required to judge how well the study applies to other situations familiar to them. The job of the researcher is to attempt to describe the particular phenomenon in such a depth as to enable the reader to generalise.
As qualitative samples are small, non- probabilistic and as the research deals with detailed, in-depth analyses rather than large-scale population-based studies, it is not possible to generalize the ¬?ndings using traditional statistical inference.
There are mainly two different stances here within the qualitative paradigm. The ¬?rst says that the qualitative researcher aims at obtaining analytical generalization. Qualitative methodology strives to capture human life, thought, inter- action and social contexts. Therefore, the knowledge obtained from these studies should be transferable to other similar social contexts (17). The developed theory should ¬?t and be applicable beyond the study population from a theoretical point of view. Theory competence is therefore a prerequisite in qualitative methodology (23). The second stance claims that qualitative researchers never deal with the question of generalizability. They claim that the detailed, in-depth and small-scale descriptions are good enough and can alone present a view of the world. The readers of qualitative publications are the ones judging the value and the applicability of the ¬?ndings (17).
Dependability refers to the epistemological notion that researcher and study subjects are interrelated and interacting with each other, thus also in¬‚uencing each other. And as perceived realities are constantly changing, questions of replicability are not in focus. Dependability instead relates to the ability of the researcher to be ¬‚exible and change perspective in accordance with the emerging process. Dependability asks whether the same findings would be achievable if the study were repeated, mirroring the term reliability used in quantitative research. However, it has already been discussed, that inherent in qualitative research is the assumption that multiple realities exist in the complex, ever changing context of an individual’s life. If this is the case it is not relevant for the qualitative research to produce the same results time after time. Lincoln and Gaba (1985) recognise that instead, the qualitative research process should be logical, traceable and clearly documented. The technique proposed by Lincoln & Guba (17) is called audit trail and refers to a strategy whereby the research process is documented and described in detail and preserved for eventual audits. It should thus be possible for outsiders to follow all steps and decisions in this process.
The Con¬?rmability refers to the researcher’s ability to be neutral to data. Con¬?rmability is also checked by an audit trail, this time meaning that the auditor should be able to ¬?nd the derived qualitative results well grounded in data (17). Conformability requires the findings and interpretations of the study to be explicit within the data (Carpenter, 1987). Finlay (1999) states that fundamental to the process of analysis of a phenomenological study is to stay true to the data, and for themes to arise from the data rather than be imposed on it.
Informed, written consent to initial participation in the study is necessary and will be gained before commencing the project and prior to each interview (Kvale, 1996). To ensure participants are fully informed prior to consent being given, any questions raised need to be answered by the researcher. The interviewees were assured of anonymity and confidentiality. They did not have to answer a question that made them feel uncomfortable.
The participants consented to the interviews being recorded, translated and transcribed, before being analysed. Both the recording and transcription will be kept locked away during the project and destroyed afterwards to maintain anonymity and confidentiality. Prior to analysis the participants will be allowed to read the transcript and may clarify meanings of statements, edit the transcripts or completely remove statements so that the transcription is true and accurate and the participant is not misquoted (Rubin and Rubin, 1995).
In the interviews participants spoke in Urdu as well as in English. The tapes would be given to an independent translator for the transcription and translation of the responses which were in Urdu. Transcripts would then be given to participants to read. They could change or delete anything they do not want to include, before researcher read them. Analysis will involve the construction of themes exploring the values, experiences and meanings within the data (Sarantakos, 1988). The analysis process is dynamic with themes and key ideas arising from the data, rather then being imposed on it (Carpenter, 1997).
The initial step in qualitative data analysis is for the researcher to familiarise themselves with the data. The transcript will be read repeatedly and initial ideas around the general themes and concepts will be made (Finlay, 1999). Relevant and interesting quotes from the texts will be written onto index cards, marked on the back of the index cards there will be marked a clear system to trace quotes back to the original transcript. The aim of this is to sort the data from a complex whole into manageable units (Lincoln & Guba, 1985). This is a time consuming process but Bailey (1997) highlights that in this process lays the foundation for the rest of the analysis and so careful and accurate indexing will improve the overall quality of the study.
The index cards will then be categorised based on their contents. One by one the cards will be allocated themes and cards with similar themes will be grouped together. This process will continue for all the index cards with the creation of initial and vague categories for all of the units of data. Once all the data has been assigned, the categories will be given more rigid titles or rules for inclusion. Each of the units of data will be then be reviewed in these groups and any overlaps in categories will be eliminated (Lincoln and Guba, 1985).
The assigned categories will then be used to inform later data collection to establish more data within each of the themes. Links between the categories will then be made. The data collection process ends when all the data have been exhausted and the categories have been saturated. The categories and concepts arising from data can then be presented (Pope and Mays, 2000).
Critical Appraisal of Methods
QUALITATIVE VS. QUANTITATIVE
Due to the descriptive nature of qualitative research, it has often being criticised for being subjective and therefore biased (Denzin and Lincoln, 1998). Quantitative research explores casual relationships asking how many and how much. This scientific evidence has long been the focus of medical research and Carpenter (1997) highlights that this medical model of research has formed the theory base for physiotherapy. Yerxa (1991) argues that although qualitative research is less precise, it is more real. Bailey (1997) supports this view stating that therapists can further understand their clients in relation to their disability by using qualitative research.
Unstructured, or open interviews are based on every day conversation, they follow a line of enquiry where the research participants leads the discussion within the boundaries of the research question (Blaxter et al., 2001). The open interview can be used to explore the subjective experiences of the participants. Some disadvantages to open interviews are that the researcher will not know how long will be needed to cover all the information (Burns, 2000).
Kvale (1996) describes the semi-structured interview as one that has a sequence of themes to be covered alongside a question that can be used. The researcher can clarify any misinterpretations and inconsistencies in responses, and questions can be repeated and worded differently to elicit a truer response (Bowling, 1997). Rubin and Rubin (1995) state that most types of interviews follow the combination of unstructured and semi-structured phases where at different points the topic of discussion is led either by the researcher or the participants. The themes of the interviews and questions scheduled in this study have been derived from the literature review. The interview schedule is contained in the appendix 4.
Due to the nature of the project there are many ethical considerations to be made. Participation in a research project should be voluntary and the decision to participate needs to be based on an understanding of the purpose and scope of the study and the main features of the research design (Kvale, 1996). Burns (2000) highlights that if any determent is threatened or reward offered for non-participation or participation, then the consent given is not voluntary. Informed consent should include the participant’s right to withdraw from the study any time (Burns, 2000). In this study these factors are made explicit within the consent forms contained in appendix.
Denscombe (1998) highlights that the data collected is and should be treated as a genuine reflection of the participant’s thoughts and feelings. The opportunity for participants to withdraw any comments made in the interviews will be given as is stated clearly in the consent forms. The participants will be given a copy of the transcripts to proof read and from this, may remove or clarify any statements they made. Rubin & Rubin (1995) advocate this process arguing that, although it may reduce the accuracy or impact of the final project, the protection of the participants should be the primary focus of the researcher
Kvale, (1996) highlights that close interpersonal interactions in interviews allow a deep exploration of the subject. However, such a close relationship between researcher and participants could result in a biased approach to the presentation of the results from the researcher. Kvale, (1996) stated that unacknowledged bias in research can invalidate the results of the interview. For this reason the potential bias in this study has been highlighted and steps that will be taken to reduce the bias include keeping a reflective diary regarding the influences of the value and beliefs system of the researcher (Koch, 1994) and a colleague checking of the results (Krefting, 1991).
Content analysis uses the transcripts of the interviews to develop the themes and concepts so that the analysis stays true to the data collected, reducing the subjective influences of the researcher on the analysis process. Direct quotes can be used and as a result the data is kept in its original form, providing a rich source of information. Continual analysis of the data throughout the research process allows themes to be modified and created throughout the study (Carpenter, 1997). However, although the themes are founded in the data they are also to an extent created by their researcher who may exercise their own personal bias when developing these (Koch, 1994). To reduce this bias, an independent person with an understanding of the analysis process will be asked to review the index cards and derive themes from these, and any discrepancies will be addressed in discussion with the researcher (Carpenter, 1997).
All interviews will be audio-taped. This will enhance the quality of the data collected in two respects. Clearly an audiotape recording will provide a more accurate record of what has been then a mere recollection or hand written notes of the author (Sacks, 1992). Secondly and of equal importance, is that the tapes can be replayed and the transcripts improved over time, allowing new themes to emerge from the data that may not have been apparent from the initial reading. One criticism levelled at the use of audiotapes is that they do not accommodate facial expressions (Silverman, 2001). For the purpose of this study, it is not anticipated that this will compromise the credibility of the data.
Tapes will be given to an independent translator and the participants asked to read transcriptions of the tapes in order to increase the validity of the study. Transcription of the data and analysis process is also time con