CHAPTER 3 RESEARCH METHODOLOGY
The previous chapter discussed literature relating to the study. It covered related studies in records management which relate to the purpose of records management, electronic records management, electronic document management, legislative framework governing records management in South Africa, electronic records disaster management planning, records management education, introduction of IT in records management, training and different processes of records management in the public sector and the role of records in improving service delivery. This chapter takes the reader through the methodology applied in conducting this study. The scientific study research methodology brings about the paradigms used. It entails the paradigms and design of the research, population sampled and studied, sampling methods used as well as the data collection methods and instruments. I concur with Ngulube (2005a:128) that the research procedure is all about the population, sampling method, instrumentations, data processing and treatment of statistics where necessary because without some of these there would be no research.
There are several research methods that are applied in conducting the scientific research. The research methods can either be identified or distinguished as quantitative method or qualitative method. In this study, the researcher used the mixed research method. That is, the researcher used both quantitative and qualitative methods (Myers 1997; Creswell 1994; Johnson and Christensen 2004; Fidel 2008). Certain problems and challenges for certain research topics need a combination of both methods by their nature to ensure validity of the results or findings (Bryman 1988:173; Cohen, Manion and Morrison 2000:112). The mixed method improved the quality of the research by minimising biases, limitations and weaknesses. This is because the disadvantages of one method was closed by the advantages of the other and vice versa (Johnson and Christensen 2004; Matveev 2002; Creswell 2003).The rationale for using the qualitative research methodology is that it explores information in the form of quality, such as explanations, descriptions and narratives (Fidel 2008:265; Gorman and Clayton 1997:28). The qualitative research method gave participants an opportunity to give their thoughts, interpretations and understanding by describing and explaining the situation in their environment. It concentrated much on the context of what is studied to provide an understanding of the political, social, psychological, economical and cultural condition of the environment under the study (Anderson and Arsenault 1998:119-134). A qualitative research study does not lead the researcher to statistical results or procedure but rather the findings relate to lives, experience, behaviour, emotions and feelings, organisational functioning and social movement, and still some of the information would also used for statistical purposes based on census and background (Strauss and Corbin 1998: 11). The qualitative method was used to view the experience of the participants about the condition or problem being studied and explored the reasons for their kind of response to the situation (Creswell 1994:2).
The quantitative research method explored and measured the situation basing on statistical information such as how many people supported or did not support certain issues or statements and interpret the results (Fidel 2008; Creswell 1994; Powell 1997; Matveev 2002). According to Terre Blanche, Durrheim and Painter (2006:138) “data are basic material with research work. It comes from observation and can take a form of number (numeric or quantitative data) or language (qualitative data)”.
A research design is an action plan for the research conducted. It covers the population or sample studied, design type – whether exploratory or correlational or experimental or descriptive, data collection duration and reliability and validity of threats (Hernon and Schwartz 2009:1). This study was conducted as an exploratory survey. A research design is used to give a guideline about how cheap, simply and economically the data would be collected and analysed in relation to the purpose of the research. In other words, it offers the researcher the simplest and most affordable way of conducting a research beforehand (Terre Blanche, Durrheim and Painter 2006: 34). The design of the research depends on its questions and the data, to give a way for questions and data connections and show the tools and procedures relevant for use in answering the research questions. This primary procedural plan includes five major characteristics, which are the strategy used, conceptual framework, study population and subject, as well as the tools and procedure adopted for data collection and analysis (Punch 2006:48).
Therefore, the research design answers to the questions regarding the followed research strategy, research framework, population provided data, data collection procedure, and data analysis methods. Bless, Smith and Kagee (2006:71) define a research design as “the specification of the most adequate operations” conducted with the aim to test a particular hypothesis under the study or to answer the research questions.
“One of the major steps in survey designs is to define the population before collecting the sample” (Ngulube 2005a:46). The population is a group of elements sharing the same sentiment. It is a large pool from which our sampling elements are drawn and to which the researcher generalised the findings. The study might focus on an organisation or institution but what the researcher is interested in is the people belonging to that organisation (Black 1999:111; Babbie, Haley and Zaino 2003:112; Ngulube 2005a:129; Welman and Kruger 2001:46). “Theoretically speaking, population encompasses all the elements that make up our unit of analysis” (Terre Blanche, Durrheim and Painter 2006: 133). Welman and Kruger (2001:46) and Ngulube (2005a:129) stated that population entails object, group, organisation, human products exposed to a certain condition, and that was the condition under this research. In most cases the data is collected from the victims of the problem by the researcher. Lewis and Ritchie (2003:49) argue that in selecting a research setting and population, the researcher depends on the research question and existing literature to identify a population which is in a position to provide relevant, rich and comprehensive information.
The population of this study was drawn from two hospital units within the Limpopo Province of South Africa. The hospital units targeted were the Records Management Unit and the Information Management Unit. These units were relevant because they use patients’ files on a daily basis to discharge their duties. The population was drawn from the hospitals detailed in Appendix 1.
The sampling frame provides information required about the selection of the samples. It also provides a detailed foundation where the research sample can be drawn, and for a population that is enough for a high quality selection of the participants (Lewis and Ritchie 2003:88; Johnson and Christensen 2004:197). In other words:
Sampling frame is a complete list on which each unit of analysis is mentioned only once. Unless such a sampling frame is borne in mind, it is impossible to judge the representativeness of the obtained sample properly (Welman and Kruger 2001:47-48).
Welman and Kruger (2001:47-48) underscore that the sampled participants served as representatives of the whole population, which is the sampling frame because they are considered to be the entire population. The population representatives also differ in terms of the knowledge or capacity in the provision of information about the problem being investigated and the way it affects them. Individual representatives within the same population understood the same issues differently because they provided different opinions and displayed different knowledge about the same issues or problems. As noted by Fowler (2002:14) and Ngulube (2005a:134), the sample frame is capable of limiting the possibility of generalising the population. Johnson and Christensen (2004:193) argue that the sample frame assists in ensuring the validity of the research results.
The researcher arranged a list of all categories of the identified population of the study to facilitate a random selection of individual participants. This is known as sample framing (Leedy and Ormond 2005:183). The population of this study was located in hospitals within the five districts of the Limpopo Province of South Africa. Each hospital uses MS Office Excel spreadsheets (called staff establishment). The staff establishments list all employees of the institution, according to their units and position/level or designation. The staff establishment spreadsheets were also used to stratify and randomly select employees from different levels in the records management unit and information management unit to participate in this study (Powell and Connaway 2004: 100). This study is framed according to Appendix 1.
The sample size entails the number of participants chosen from the whole population. The researcher focused on the population of the study when drawing a sample. The selected members or part of the entire population is called the sample (Bless and Higson-Smith 2000:84; Rowley 2002:19; Ngulube 2005a:130; Nachmias and Nachmias 1996:201). The sample drawn made a sample size. “There are no rules for sampling size” (Ngulube 2005a:130). The results of the study may be affected by the sample size and plan in terms of the type, level and generalisation of the study conclusion. A small sample reduces readers’ interest to use the results and a large number of the sample gives confidence to the results, although it can be very expensive to handle. The sampling method determines the validity and reliability of the research conclusion (Ngulube 2005a:132-134; Marshal and Rossman 1995). Leedy (1997:211) underscores that sampling is not necessary for a population that is not more than 100. Terre Blanche, Durrheim and Painter (2006:49) underline that in a research where the number of sampled participants is too little; it is not advisable to use random sampling because that sample may not be representative enough for the entire population of the study. This is also the case with a large number of sampled participants in non-random sampling.
The sampling method, discussed under sampling procedure, was used because it helped to save time, money and energy. Thus the researcher avoided moving around from office to office collecting data from all the people throughout the whole country (Nachmias and Nachmias 1996:201; Miles and Huberman 1994:27 in Punch 2006:50). Seaberg (1988); Neuman (2000:217); Grinnell and Williams (1990) in Ngulube (2005a:134) underscore that a minimum of 10% of the sample, especially for a large population, is good enough to draw a valid and reliable data. In this study, out of 100% (324) of the total population of the study, the researcher sampled a total sample of 65% (210). From the 100% (210) of the total sample, the records management sample was 74% (155) and the information management sample was 26% (55). The number of the sample was guided by the representation of each category out of a total of 324 members of staff employed in hospitals in the Limpopo Province. This difference in sample per category resulted from the fact that the records management category had a greater total population of 258 (79%) and the information management category had lesser total population of 66 (21%). Sixty-five per cent (65%) of the population was accounted for by 210 participants. The sample was drawn, as illustrated in Table 3.1, using the sampling procedure described in the next section. The researcher used the Raosoft sample size calculator available on (http://www.raosoft.com/samplesize.html) to calculate confidential level and margin of error, as in Figure 3.1. Using the Raosoft sample size calculator, out of the total population of 324, the total sample of 210 made a confidential level of more than 95% and less than 4,10% margin of error.
Welman and Kruger (2001:64); Onwuegbuzie, Jiao and Bostick (2004: 107) argue that in sampling “the smaller the total population, the relatively larger the sample should be to ensure satisfactory results”, but the lower the total population of the unit the lesser the percentage out of the total sample. This helps to minimise the standard error. Appendix 1 illustrates sampling in both numbers and percentages to clarify the fact above. The sample was drawn from the way the records management units administer the records and the information management units use records for statistics every day. Table 3.1 gives details of the sample sizes per district. The data was compiled from Appendix 1.
The study was initially targeted to cover six units of the hospitals, namely
Chief executive officers
A validity and reliability study for the questionnaire was conducted. The result of the study was that only information and records management units are relevant to participate in the study, especially for the completion of the questionnaire, because the content is only relevant for information and records management participants. The study then changed the whole focus to information and records management units in the hospitals of Limpopo in five districts, and the questionnaire was also amended to be in line with the change.
Sampling is “the selection of research participants from an entire population, and involves decision about which people, setting, events, behaviour, and/or social process to observe” (Terre Blanche, Durrheim and Painter 2006:49). Selected participants participated as representatives of the whole population. Eventually the researcher generalised the findings of the whole research as if it was drawn from the whole population (Slater, 1990: 40; Terre Blanche, Durrheim and Painter 2006:49; Johnson and Christensen 2004:197; Mouton 2002:110; Hernon and Schwartz 2009:1). In conducting sampling, the population was clearly defined and the sample was systematically drawn (Mouton 2002:110; Hernon and Schwartz 2009:1) to avoid studying every person who is doing anything within the whole geographical area (Nachmias and Nachmias 1996:201; Miles and Huberman 1994:27 in Punch 2006:50).
The sampling procedure was the foundation of the study and the population characteristic was “fully discussed” (Ngulube 2005b:46). The researcher used the probability sampling method known as the stratified random sampling for quantitative data collection. It was used to select people based on random procedure and this sampling was used by grouping or separating participants into non-overlapping groups according to their districts and fields of work (see table 3.1 and Appendix 1). The researcher then applied simple random within the grouped population in each institution (Zou 2006:1; Fuller 1993:1; Johnson and Christensen 2004:207; Burton, Croce, Masri, Bartholomew and Yefremian 2005:104). It is with stratified sampling that standard errors were reduced as this technique controls variance proportions (Sapsford 1999:70). This is because all the units or categories of the population were covered in the sample.
TABLE OF CONTENTS
Table of contents
List of tables
List of figures
List of appendices
List of abbreviations and acronyms
CHAPTER 1: BACKGROUND OF THE STUDY
1.2. Background to the study
1.3. Research problem
1.4. Justification for the study
1.5. Significance of the study
1.6. Scope and limitations of the study
1.7. Ethical considerations
1.8. Definition of key concepts
1.9. Research methodology
1.10. Organisation of the thesis
CHAPTER 2: CONTEXTUAL OVERVIEW OF RECORDS MANAGEMENT AND ELECTRONIC RECORDS
2.2. The benefit of proper records management in the public sector
2.3. The purpose of records management in the public sector
2.4. Records management models
2.5. Introduction of information technology (IT) in records management
2.6. Managing electronic records
2.7. The nature/characteristics of a dependable records
2.8. Electronic records management
2.9. Records management education and training
2.10. Legislative framework governing medical records management in public health sector of South Africa
2.11. The role of records management in promoting Batho Pele principles and improving service delivery
2.12. Records disaster management planning and electronic records
2.13. Appraisal, retention and disposal of records
2.14. Public health services and records management practice
CHAPTER 3: RESEARCH METHODOLOGY
3.2. Research methodology
3.3. Research design
3.4. Data collection methods (including instruments)
3.5. Data analysis and presentation
3.6. Evaluation of the research methodology
CHAPTER 4: PRESENTATION OF THE FINDINGS OF THE STUDY
4.2. Questionnaire findings
4.3. Observation and interview findings
CHAPTER 5: INTERPRETATION OF THE FINDINGS OF THE STUDY
5.2. Electronic records management in the public health sector
5.2.1. Knowledge and understanding
5.3. The state of records management in the public health sector of South Africa in Limpopo Province
5.4. The major causes of missing files in the public health sector of South Africa in Limpopo Province
5.5. E-records and e-health awareness
5.6. Availability of the necessary infrastructure
5.7. Availability of capacity for e-health and e-records in terms of human resources
5.8. Identification of existing guideline documents like policies and legislative framework governing e-records
5.9. Improvement of records management and e-health readiness
CHAPTER 6: SUMMARY, CONCLUSIONS AND RECOMMENDATIONS
6.5. Proposal for further research
6.6. Last summary
List of references
GET THE COMPLETE PROJECT
RECORDS MANAGEMENT IN SUPPORT OF SERVICE DELIVERY IN THE PUBLIC HEALTH SECTOR OF THE LIMPOPO PROVINCE IN SOUTH AFRICA