5.6 VALIDITY AND RELIABILITY
5.6.2 Pilot Test (Pre-test)
According to De Vos, (2002:368), pre-testing of the questionnaire involves trying it out on a small number of people who have characteristics similar to those of the target population.
109 A pilot study was conducted to pre-test the questionnaire. Pre-testing refers to testing the questionnaire on a small number of the sample of respondents, to identify and eliminate potential harmful questions. The purpose is to ensure that the questionnaire meets the researcher‟s expectations in terms of information that it obtains.
Pre-testing the research instrument during the survey development stage was done through a pilot study covering 10 respondents. The results of the pre-test led to the test-retest reliability. This reliability measure used the same measurement scale a second time under nearly the same conditions. The correlation between the answers to the first and second tests were then examined and found acceptable. The results of the pre-test led to some corrections to the questionnaire.
5.8 ETHICAL CONSIDERATIONS
Ethics deal with the development of moral standards that can be applied to situations in which there can be actual or potential harm to any individual or a group. They are of particular concern to the researcher because their success is based on public cooperation (Roberts-Lombard, 2002:19).
Researchers have some general obligations to people who provide data in research studies which include the obligation not to harm, force or deceive participants (Chodokufa, 2009:14). Participants should be willing and informed and the data or information they provide must be held in utmost confidence (Tustin et al., (2005).
Ethics were crucial for the successful accomplishment of this research work. This also helped to reduce research errors that could have arose because other people who were supposed to be part of the research have been excluded or refused to participate.
5.9 DATA PREPARATION
This is the process of checking the quality of the data gathered and converting it into an electronic format that can be read and manipulated by computer software (Cant et al., 2005:149; Roberts-Lombard, 2002:149). Data preparation seeks to
110 ensure that high quality data is available for statistical analysis. According to Roberts-Lombard (2002:149) the quality of statistical results is in most cases the product of careful exercise in the data preparation phase. The steps which were followed in the data preparation process for this study include validation, editing, coding, data entry and data cleaning (Cant et al., 2005:150; Roberts-Lombard, 2002:149). The steps that are followed in data preparation process are discussed separately below.
5.9.1 Data validation
Data validation is the very first stage of data preparation which involves the examination of raw data for the purpose of insuring that information collected is accurate. Validation can be defined as the process of determining, to the extent possible, whether a surveys‟ interviews or observations were conducted correctly and free of fraud or bias (Cant et al, 2005:187). The purpose of validation is to determine whether any shortcuts were taken during the fieldwork. The goal of validation is mainly to detect interviewer fraud or the failure of the interviewer to follow important fieldwork instructions.
5.9.2 Data editing
Editing involves a critical examination of the completed questionnaire in terms of compliance with criteria for collecting meaningful data, and to deal with questionnaires that are not completed (Cooper and Schindler, 2003:236). The inspection of questionnaires to make modifications or corrections is termed editing (Roberts-Lombard, 2002:150; Cant et al., 2005:151). Editing can also be described as the review of the questionnaires with the objective of increasing precision and accuracy. As a process, editing consists of screening the questionnaires to identify illegible, incomplete, inconsistent or ambiguous questions (Roberts-Lombard, 2002:150; Cant et al, 2005:151). In this study, questionnaires were checked thoroughly for ambiguities, omissions, inconsistencies and other errors.
111 According to Terre Blanche and Durrheim (2002:98), coding involves applying a set of rules to the data to transform information from one form to another. Zindiye (2008:236) further explain coding as the assigning of a number or symbol to the answers so that responses can be grouped into limited categories. Coding was also described by Chodokufa (2009:89) as converting the questionnaire into numeric form in order to allow for quantitative data analysis.
Coding involves assigning numbers or other symbols to answers so that responses can be grouped into a limited number of classes or categories (Roberts-Lombard, 2002:151). The classifying of data into limited categories sacrifices some data but is necessary for efficient analysis. Instead of requesting the word male or female in response to a question that asks for the identification of one‟s gender, the codes “M” or “F” can be used (Cooper & Schindler, 2006; Roberts-Lombard, 2002:151). The purpose of coding is to transform respondents‟ answers to survey questions into codes or symbols that can easily be entered into and read by a statistical analysis software package.
In this study, two approaches to coding were done. The first was pre-coding which refers to assigning codes to response options before field work began and hence printing the relevant codes on the questionnaire. Pre-coding was done to dichotomous questions by assigning numbers to possible answers, for example, 1 for Yes and 2 for No. Final coding was done during data preparation to establish a codebook describing each variable in the dataset.
5.9.4 Data entry
Data entry can also be referred to as data capturing which was described by Roberts–Lombard (2002:151) as the transfer of data from any acceptable data collection instrument (questionnaire in this case) into the computer. Cant et al.,
(2005:161) describes data entry as a process includes the tasks that are involved with the direct input of the coded data into some specific software package that will then be used to manipulate and transform the raw data into useful information. In this study data was entered into Microsoft excel before it was analysed.
112 The process of checking coded and entered data for errors before starting data analysis is called data cleaning. Data cleaning is very important especially in cases where data coding and entry has been done manually, as in this case. Cant et al., (2005:163) argues that where data coding and entry has been done manually, there is no doubt that there may be errors. A thorough data cleaning process was done before analysis.
5.10 DATA ANALYSIS
Data analysis is the application of reasoning to understand data that have been gathered (Zikmund & Babin, 2010:66). It is a process of gathering, modelling and transforming data with the goal of highlighting useful information, suggesting conclusions, and supporting decision making. This section deals with data analysis procedures. Data analysis usually involves reducing accumulated data to a manageable size, developing summaries, looking for patterns and applying statistical techniques. It also includes the interpretation of research findings in the light of the research questions, and determines if the results are consistent with the research hypotheses and theories. The choice of the methods of statistical analysis depends on the type of question to be answered, the number of variables, and, the scale of measurement. The type of question the researcher is attempting to answer is a consideration in the choice of the statistical technique. Based on this factor, the researcher may be concerned about the central tendency of a variable or distribution of that variable. Data analysis for this study included descriptive statistics and Pearson Chi-square test and Independent samples T- test.
Since the question and scale of measurement require that the significance of entrepreneurship education on performance of SMMEs be established, the appropriate method to use to analyse the data was the Chi-square test and the T- test as well as descriptive statistics.