Powell (1997: 105) stresses the importance of pre-testing. A pre-test gi ves the researcher an opportunity to identify questionnaire items that tend to be misunderstood by the participants. The pre-test offers certain advantages beyond helping refine the data collection instrument. It can permit a preliminary testing of the hypothesis, point out a variety of problems, not anticipated, relating to design and methodology, facilitate a practice run of the statistical procedure to be used and perhaps even indicate that the final study may not produce any meaningful results and therefore should be rethought or
abandoned (Powell 1 997: 105). Babbie ( 1 995: 1 53) stresses that, no matter how careful one is in the design of a data collection instrument such as a questionnaire, there is always the possibility, indeed the certainty, of error. Pre-testing should be used on people who are at least relevant. Dane ( 1 990: 127) describes the purpose of pre-testing as to
"fine tune the instrument in much the same way that a bench check allows a technician to evaluate a part before installing it".
An important purpose of pre-testing was to devise a set of codes or response categories for each question which will cover, as comprehensively as possible, the full range of responses which may be given in reply to the questions in the main investigation. In addition, the time factor was taken into consideration when pre-testing. For this to work effectively, the pre-testing was done on individuals who represented a variety of people, which the main study was intended to cover.
Pre-testing was done at the University of Natal, where the researcher is a student. Ten people who are involved in records management were involved in the process. The pre
testing took three days. The researcher went through all the responses to the pre-test and found that there were corrections that needed to be made to the instrument. Those who were pre-tested raised the issue of some of the questions being too technical. The researcher realised it was not a major problem, since she was going to be there to conduct face to face interviews. If there was a need to explain those technical words, she would be able to explain them. When it came to the time factor, it was found that an interview took about twenty minutes. The corrections were made and the instrument was successfully administered.
3.9 Interviewing Process
After the corrections from the pre-testing were checked and corrected, the interview schedule was compiled (see Appendix IT). Though the researcher introduced herself to the respondents, as it was a face-to-face interview, an introductory letter explaining the purpose of the research was attached to the schedule. A letter from the Information Studies Programme (see Appendix Ill) introduced the researcher and confirmed that she
was currently carrying out a study on e-records management practices in the public sector in Lesotho. This was done so that the researcher would be given any necessary assistance to facilitate the conduct of the study. The researcher made some preliminary visits to some of the nineteen government ministries and was assured of their co-operation. She observed that all the ministries were accessible, since they were all located in the capital city, Maseru.
Before leaving SO,llth Africa, the researcher made appointments with some of the personnel in the ministries, through e-mail and the telephone. This made it easier to carry out the interviews. Before an interview could take place, the person in charge, that is the registry supervisor, introduced the researcher to all the records management staff and the interviewer was asked to introduce herself in front of the whole staff. The interviews were first held on 1 8 August 2003 and ended on 22 August 2003. On the first day the researcher interviewed four respondents. The following day, 19 August, she interviewed six individuals and, on the 20 August, three. On the 2 1 August the researcher interviewed four respondents and three people were interviewed on 22 August. Each interview took almost twenty minutes, which the respondents found very fair, since people do not want to be kept for longer than half an hour. Questions for the archivist dealt mainly with management and legal issues, in terms of policy implementation and guidelines governing e-records in the public sector in Lesotho.
3.10 Data Analysis
Data comes in many shapes and forms. The role of analysis is to bring data together in a meaningful way and enable researchers to interpret or make sense of it (Saunders 2000:
77). According to Kothari ( 1 990: 1 5 1 ), "there is a distinction between data processing and analysis". He explained that, technically speaking, processing implies editing, coding, classification and tabulation of collected data, so that they are amenable to analysis. The term analysis refers to the computation of certain measures, along with searching for patterns of relationships that exist among data-groups (Kothari 1990: 15 1).
Powell ( 1 997: 67) stated that data is analysed to describe the characteristics of the units
of analysis, to make predictions about specific relationships, and to test associated relationships.
The statistical approach was adopted for data analysis. Powell (1985: 145) points out that the basic purpose of statistical analysis is to summarize observations or data in such a manner that they provide answers to the hypothesis or research question. Statistical methods are applied for descriptive purposes and for statistical inference. The researcher used descriptive statistics, which mainly dealt with the tabulation of data, graphs and the calculation of descriptive measures.
Before data was analyzed, it was evaluated. This process is sometimes called data editing (Cohen, Manion and Morrison 2000: 265). The purpose is to check for ambiguity, completeness, comprehensibility, internal consistency, relevance and reliability (Powell 1997: 63). The researcher checked the data in line with errors in recording, handwritten notes and to check whether or not there was an answer to every question. The following section describes how data was coded.
3.1 1 Data Coding
According to Powell (1985: 160), SPSS is the most widely available statistical package.
The SPSS system is a comprehensive, relatively easy-to-use computer program for statistical analysis, report writing, tabulation and general-purpose data management.
SPSS provides simple tables to multivariate analysis. The researcher resorted to using this package because of its features, which include a colour graphic package and a conversational software system that allows interaction between the data and the user (Pow ell 1985 : 160). A coding key was drawn, in which numerical values were assigned to all closed-ended questions such as 'yes' and 'no' responses. Data was entered on a data matrix design, using the SPSS for Windows Version.
A combination of content analysis and qualitative coding was used to interpret the open
ended type of responses. The method consisted of two steps: the first step involved identifying different concepts as they appear in the responses to questions in the set of
interview schedules and, secondly, steps involved in sorting the concepts into categories (Fielding 1993: 227).
3.12 Evaluation of the Research Methodology
As mentioned in previous chapters, the purpose of the research was to investigate the management of e-records in the public sector in Lesotho. The investigation was conducted in order to provide information for planners and policy-makers. The whole idea was to make meaningful recommendations for managing e-records and archives for access and preservation.
The study utilized the survey research design. As descriptive research, the research was designed to answer questions that would describe and compare variables to a particular standard, or summarise the relationships between a particular standard, or summarise the relationships between two or more variables. The survey design was utilized because of its strengths insofar as it is economical and there is a rapid turn-around time in data collection (Creswell 2003: 154).
The major limitation is that survey research cannot provide information on cause-effect relations such as experimental studies. In addition, all survey methods are handicapped by non-response (Goyder 1987: 63). Non-response is not only affected by the percentage of the units of analysis, but by the way respondents respond to questions, the uneven impact of interview schedule structure and question wording. Pre-testing the instrument before the actual interviewing was one of the ways to minimize the ambiguity of the data collection instruments.
There is no single best way of collecting data. The method chosen depends on the nature of the research questions posed and the specific questions that one wants to ask respondents. The aim of the method was to obtain valid and reliable data and to answer the research questions, which can be used as the basis for credible conclusions.
Though self-administered questionnaires are considered to be a highly structured data collection method, the investigator has no control over the conditions in which the data are elicited (Wilson 1996: 102). It may not be the designated respondent who actually completes the questionnaire, it may be a group effort, or the designer may complete the questionnaire in any order that the respondent likes, despite the careful ordering of the questions. Interview schedules in this study were selected to avoid such problems linked to the questionnaires. Where clarification was needed, the researcher was able to provide it and the respondent was motivated to continue to answer further questions.
Researchers should understand strategies, approaches and techniques available, to justify the choices they make. Since there was no sampling the researcher wanted to get as much information as possible. The data collection method applied was appropriate in this case.
3.13 Summary
The research methodology was discussed. The population of the study was described.
Data collection methods were explained and the data collection instrument was discussed.
The advantages and disadvantages of interviews as a data collection technique were given and the design and structure of the schedule used were discussed. A description of the interview process was made. Other issues examined in Chapter Three included the validity and reliability of data collection methods and the analysis of data. The chapter concluded by evaluating the research methodology.
CHAPTER FOUR: PRESENTATION OF FINDINGS AND INTERPRETATIVE DISCUSSION