CHAPTER 3: RESEARCH METHODOLOGY
This chapter presents the methodology that was used in the study. Methodology is the science of finding out and it is a subfield of epistemology, which is the science of knowledge (Baxter and Babbie, 2004:4). Indeed, utilising an appropriate methodology is not only the right path to the truth, but also the duty of every researcher (Damian and David, 2007:54). Moreover, the ability to blend and use methods, which are appropriate, is an important issue for researchers to realise and incorporate into their research (Knox, 2004:124). The methodology covers issues that relate to the type of data collected, the way it was collected, and analysed. The chapter is organised as follows, first it discusses the research paradigms relevant to the study, followed by the design of the study. It outlines the design of the instruments and variables used. The analytical methods employed are also presented and discussed. Measures that were taken to ensure validity, reliability and ethical standards in the study are also highlighted.
To recap, the research objectives of the study are stated again. The first objective was to determine the level of access to and usage of ICTs by communal farmers in Seke and Murewa districts of Zimbabwe. The second objective was to evaluate how access to ICTs influence climate change awareness amongst the communal farmers. The third objective was to investigate the contribution of ICTs in climate change adaptation amongst these communal farmers. Based on these objectives the next section presents the research paradigms.
THE RESEARCH PARADIGM
du Plooy (2009:19) defined a research paradigm as “a set of shared basic beliefs about how researchers view that which they study”. Morgan (2007:50-54) identified four different versions of the paradigm concept, which are paradigms as worldviews, as epistemological stances, as shared beliefs among members of a specialty area and as model examples of research. All versions treat paradigms as shared belief systems that influence the kind of knowledge researchers seek and how they interpret it, however, what distinguishes them is the level of generality of that belief system (ibid). This study adopts and discusses three relevant paradigms namely, the post normal science (PNS), pragmatism, and mixed methods.
Each of these is discussed in the next sections.
Post Normal Science (PNS)
Post-Normal Science (PNS) is a new conception of the management of complex science-related issues, which focuses on aspects of problem solving in which there are high stakes and uncertainty, decisions are urgent, and there are plural and conflicting value systems (Funtowicz and Ravetz, 2003:1). This research paradigm grew out of the realisation that the traditional scientific paradigm had inadequacies. For example, Westra (1997:238) noted that complexity and unpredictability are here to stay, and PNS makes the previous paradigms obsolete. Funtowicz and Ravetz (2003:2) further argue that the expectations of regularity, simplicity and certainty in the traditional ‘normal’ scientific mind-set, could inhibit our understanding of new problems and appropriate methods for their solutions. PNS is relevant for seeking to understand complex natural and social systems, moreover, it is an appropriate methodology when there are system uncertainties or when decision stakes are high (Funtowicz and Ravetz, 1991; Tognetti, 1999:691; Funtowicz and Ravetz 2003:2; Ravetz, 2004:347). PNS brings ‘facts’ and ‘values’ into a unified conception of problem solving, and its principle of the plurality of legitimate perspectives leads to a focus on dialogue, mutual respect and learning (Funtowicz and Ravetz, 2003:4).
Rodela, Cundill and Wals (2012:18) asserted that a PNS approach tends to be issue-driven, policy relevant, trans-disciplinary and emphasises issue improvement. In all that, the researcher recognises the value of different ways of knowing and different types of knowledge hence is engaged in boundary crossing and collaboration (ibid). The approach embodies the precautionary principle, and depend on public debate and engagement of extended peer community (Ravetz, 2004:347; Rodela, Cundill and Wals, 2012:18). As alluded to earlier, the focus of this study is climate change adaptation in the agricultural sector. In that context, Saloranta (2001:400) suggests that the climate change issue fulfils the attributes needed to belong to the domain of PNS. Saloranta claims that the enhanced problem solving in the climate change issue has been facilitated by use of such approaches, for example, in the Second Assessment Report of the IPCC.
Similarly, Etkin and Ho (2007:627) gave four reasons why climate change is an example of a problem that fits into PNS paradigm. Firstly, the linkages between science and society are profound. Secondly, the problem must be viewed holistically, with consideration of feedbacks between the climate system, the human system and ecosystems. Thirdly, there is large uncertainty and a plurality of legitimate perspectives with respect to risk. Fourthly, the climate change issue is complex, difficult or impossible to fit into a traditional linear problem-solving model. In addition, Westra (1997:238-239) postulates that the PNS comprises of at least three compatible aspects.
• The increased relevance of values, ethics and social aspects, hence the need for public discourse and debate.
• The switch from the expectations of the Newtonian scientific paradigm to the complex systems theory.
• Recognising that the ecological/ biological point of view may foster several possible developmental paths that may result in different configurations for ecosystems (ibid).
The discussion above highlight that the PNS paradigm is relevant to this study. The main thrust of the thesis is to assess climate change awareness and adaptation amongst smallholder farmers, and how ICTs contribute to this. Undoubtedly, there is a lot of uncertainty about climate change itself, which is compounded by the uncertainty in terms of how climate change will affect smallholder farmers such as those in Seke and Murewa districts and how they will respond. How people perceive, understand and adapt to climate change can be interpreted differently by different people. Furthermore, different methods can result in varied findings and outcomes thus the relevancy of the PNS paradigm.
Another relevant paradigm to the study is pragmatism. Morgan (2007:48) highlights that much of the recent discussion in social science research methods has focused on the distinction between qualitative research and quantitative research. Likewise, Clark and Creswell (2008:7-11) observe that paradigm wars between qualitative (constructivist) and quantitative (positivist) researchers led to the emergence of pragamatism, which considers the importance of what works. Pragmatism research paradigm argues that the only way we can acquire knowledge is through a combination of action and reflection (Tashakkori and 78 Teddie, 2010:112). Whereas the connection of theory to data for qualitative approaches is inductive reasoning and for quantitive approaches is deductive reasoning, contrariwise, the pragmatic approach uses abductive reasoning which moves back and forth between induction and deduction (Morgan, 2007:71; Clark and Creswell, 2008:58-59). Morgan (2007:71) highlighted the main differences between quantitative, qualitative, and pragmatic approaches as shown in Table 3-1.
A key issue that can be drawn from the above discussion is that pragramatism emphasises flexibility, due to need to focus on what works. This study is multidisciplinary, hence, the relevancy of pragmatism. It combines aspects from many disciplines which demands a combination of diverse methods. This requires the need to move back and forth between qualitative and quantitative methods, combining methods that are complementary and most appropriate. Having discussed pragmatism the next section focuses on the mixed methods approach.
Mixed methods approach
A mixed method approach involves the use and combination of various quantitative and qualitative data collection and analytical methods in a single study in which the data are collected concurrently or sequentially and involves the integration of the data at one or more stages in the process of research (Clark and Creswell, 2008:165). If researchers focus on one approach, there is a possibility of failing to capture the bigger picture (Knox, 2004:123). Hence, the need to employ a mixed methods approach. There is a close link between pragramatism and the mixed methods approach. Pragmatism is generally regarded as a philosophical base for mixed methods research (e.g. Johnson and Onwuegbuzie, 2004:16; Denscombe, 2008:273; Feilzer, 2009:14; Tashakkori and Teddie, 2010:96).
The general characteristic of mixed methods approach is the methodological eclecticism (Tashakkori and Teddie, 2010:8). This involves selecting and synergistically integrating the most appropriate techniques from qualitative and quantitative methods to investigate thoroughly, a phenomenon of interest (ibid). Feilzer (2009:6) posited that mixed methods research “has been hailed as a response to the long-lasting, circular, and remarkably unproductive debates discussing the advantages and disadvantages of quantitative versus qualitative research”. The goal of mixed methods research is not to replace either of these approaches but to exploit the strengths and minimize the weaknesses of both (Johnson and Onwuegbuzie, 2004:14; Brewer and Hunter, 2006:63). Thus, it combines the rigor and precision of quantitative methods with deep understanding provided by qualitative methods (Rudestam and Newton, 2001:45).
The mixed methods approach is used for many reasons, which include the need to improve the accuracy of the data, to produce a more complete picture, and to avoid biases intrinsic to single-method approaches (Denscombe, 2008:272). It is more advantageous as each approach brings special strengths, and each compensates for the weaknesses of the other (Baxter and Babbie, 2004:65). The approach enables triangulation, which involves analysing something from multiple viewpoints (combination of two or more data collection methods and multiple sources of information to obtain data) which improves accuracy and validity (du Plooy, 2009:40; Neuman, 2006:149; Yeasmin and Rahman, 2012:156). Despite the relatively high cost of triangulation in terms of time, money and energy, it leads to synergy, enrichment and complementarity of data (Munyua and Stilwell, 2010:15). As a result, this allows for newer or deeper dimensions to emerge which enrich our understanding (Clark and Creswell, 2008:107). Brewer and Hunter (2006:39) argue that the mixed methods approach is more than just triangulation; it is a perspective that permeates all stages of the research process.
It is important to note that both triangulation and the mixed methods approach have some drawbacks. For instance, Tashakkori and Teddie (2010:9) asserts that even though triangulation has generally been associated with convergence of results, it is important to realise that results from different sources can also diverge. Furthermore, Yeasmin and Rahman (2012:160) maintain that triangulation demands creativity and ingenuity in the collection and interpretation of data for it to produce a satisfactory outcome. In addition, Denscombe (2008:280) criticise the mixed methods approach, arguing that the combination of quantitative and qualitative methodologies is liable to be fragmented and inconsistent due to lack of an overarching philosophy. Besides the criticisms the mixed methods is of relevance to this study, what matters is the appropriate combination of the methods. Tognetti (1999:701) emphasised that, “what is important is not which perspective dominates, but how diverse kinds of knowledge can all contribute not only to the decision process, but also towards a new shared understanding”.
In sum, three complementary paradigms were discussed. The PNS provides the overall guidance on understanding issues related to perceptions, information and knowledge of climate change. Pragmatism is closely related to the mixed methods approach – they emphasise flexibility and a combination of various methods that work best to achieve the desired outcome. The actual mixed methods used in the study are explained in the next section.
DESIGN OF THE STUDY
The data are only as good as the instruments that were used to collect them and the research framework that guided their collection (Pallant, 2011:3). As a result, measures were taken to ensure that the study and instrument design would produce data of highest quality to meet the objectives of the study. To recap, the main objective of the study was to analyse the contribution of ICTs in addressing climate change amongst communal farmers in Seke and Murewa districts of Zimbabwe. In order to meet the research objectives, it is important to state the research approach.
Generally, there are four common and useful approaches to research, which are exploration, description, causal/functional explanation, and understanding (Baxter and Babbie, 2004:30-31). From these approaches, the most appropriate approach for this study was exploratory. Baxter and Babbie advised that an exploratory approach is suitable when the researcher examines a new interest or when the subject itself is relatively new. This study analyses the contribution of ICTs in climate change adaptation in the agriculture sector, which is a relatively new field of enquiry (as highlighted in section 1.5 in Chapter 1). Now it is 81 important to discuss the design of the study, methods and instruments. The steps undertaken in designing the study are shown in Figure 3-1.
Figure 3-1 shows the various steps undertaken in designing the study. Based on the research objectives and research questions, the steps included deciding which information to collect (variables), where was it going to be collected from (study area), from who (the target population), how was it going to be collected (methods), and how was it going to be analysed (data analysis). This study is a cross-sectional one, as it investigates the state of affairs in a population at a certain point in time (Adèr and Mellenbergh, 1999:110). As mentioned before, a mixed methods approach was adopted, particularly, the dominant-less dominant mixed methods approach (as explained by Rudestam and Newton, 2001:45). For this study, the quantitative method based on a general household survey targeting individual farmers was the dominant method, while the qualitative method based on key informant interviews was the less dominant method. The next section talks about the variables used and the design of the questionnaires.
VARIABLES USED AND QUESTIONNAIRE DESIGN
Recommendations and guidelines from various books and manuals were followed to properly structure and design the survey questions for the study. Some of the guidelines that were followed include use of simple language, mutually exclusive and exhaustive response categories, avoiding using vague and emotionally loaded words and avoiding double-barrelled, complex, incomplete, and ambiguous questions (du Plooy, 2009:126-161; Pallant, 2011:10; Stopher, 2012:188-197). The choice of variables and the design of the questionnaire were guided by conceptual framework outlined in section 2.11 (some of the variables are further discussed in detail in section 3.6.2). An important step was to operationalise the constructs in the study in order to accurately measure the desired aspects (du Plooy, 2009:71). Operational definitions describe how the concept will be measured (Black, 1999:35; du Plooy, 2009:65).
The main purpose of the questionnaires in this study was to measure perceptions, views, and opinions of respondents (Black, 1999:215). In that regard, the questions and statements were formulated in an easy-to-understand way for both enumerators and respondents. Various types of questions were used in the questionnaires. These included binary questions (which required yes/no responses); rating scales; Likert scales; and open-ended questions. Likert scale questions involved asking respondents to rate a particular statement by selecting one of these responses: strongly agree, agree, neutral, disagree, or strongly disagree (du Plooy, 2009:142). The study used two questionnaires, one for the general household survey and the other for key informants (attached in Appendix 1 and 2). The next subsections describe how the ICT and climate change variables were formulated.
ICT use and access variables
The ITU (2003:4) highlights the importance of understanding who has access, where and how people use ICTs. Access to ICTs can be measured at different levels; individual, household, and the community (ITU, 2003; Alampay, 2006:2). In this study, access to ICTs was mainly 83 evaluated at two levels namely the individual level and the household level. The study analysed ownership and access to both old and new ICTs, namely, radio, television, video cassette recorder (VCR), digital video disc (DVD) player, fixed telephone, mobile phone, satellite decoder, computer, and internet. The really old ICTs were also included (outlined in Section 2.8.1 in Chapter 2). These are the traditional print media namely, newspapers, farming/environmental magazines, business magazines, entertainment magazines, church magazines, and posters. It is important to note that the term ICT largely refers to electronically based technologies, nonetheless, in this particular study, it was deemed necessary to include even those in print format. Various questions were formulated to explore ways in which ICTs were contributing to climate change adaptation amongst rural households. The questions included those related to basic knowledge of ICTs, ownership of ICTs in terms of who owned and the number owned per household, access to and use of ICTs within the household, issues to do with sources of energy power for the ICTs, perceptions related to maintenance and accessibility costs, perception of importance of various ICTs to livelihoods, climate information, and climate change awareness and adaptation.
Climate change related variables
The main issues that were evaluated include respondents’ perceptions on rainfall and temperature trends, climate change awareness and adaptation, as well as rainfall forecasting and early warning systems. Elicitations of views on climate change which rely on past events can serve as analogues for future changes and as a means of stimulating critical consideration of impacts (Lorenzoni and Pidgeon, 2006:81). In addition, the perceptions are important in developing an effective strategy to climate change adaptation (Ruddell et al. 2012:581). Moreover, studying the use of forecasts provides a more detailed view of the forecast process, which helps to anticipate use patterns and increase the effectiveness of forecast dissemination (Pfaff et al. 1999:649; Letson et al. 2001:59). Furthermore, there is need to identify key farming decisions that would be sensitive to climate and weather information (Stone and Meinke, 2006:18).
In evaluating climate change awareness and knowledge, the aim of this study was not to ask respondents, scientific definitions or technical aspects of climate change, but to evaluate their basic knowledge on various climate change aspects. Myers (2005:13) notes that research that measures knowledge, attitude, behaviour and practice (KABP) is useful for finding out what 84 your target audience already knows and practices. Similarly, Grothmann and Reusswig (2006:118) emphasise the importance of studying people’s own perceptions and assessments of these aspects. However, the challenge with asking people about climate change is that most of them would not have thought about the issue at any length and thus cannot give meaningful responses (Kempton, 1997:13). Thus, extra effort was taken in order to evaluate climate change awareness.
Questions were presented as statements designed to elicit attitudes and opinions. Marshall et al. (2010:4) used the same approach in assessing awareness and attitudes of environmental conditions and climate change issues of dive tourists and dive operators in the Egyptian Red Sea. Marshall et al. (2013:31) also examined the influence of climate change awareness on the capacity to adapt to climate risks of peanut growers in Queensland, Australia. Likewise, Stamm, Clark and Eblacas (2000:223) measured the breadth of knowledge by counting the number of items each respondent had heard about global warming, and depth of concern by counting the number of the causes or solutions considered important or helpful by the respondents. Following the procedures used in the studies, the next paragraph outlines what was done in this study.
First, respondents were asked a yes/no question on whether they were aware of climate change or had some knowledge about it. This initial question separated those who were aware and those who were not aware of climate change. For the respondents who had indicated that they were aware of climate change, they were asked to rate how strongly they agreed with each of the 11 statements that covered various aspects of climate change including causes, effects, adaptation and mitigation. The statements were based on a five-point Likert scale (1= strongly agree, 2= agree, 3= undecided, 4= disagree, 5= strongly disagree). In addition to evaluating climate change awareness, other questions focused on sources of the climate change information, adaptation and access to rainfall forecasting information. The next section explains the data collection process and methods employed in the study.
DATA COLLECTION AND METHODS
The choice of data collection methods is directly related to the research topic and available resources (Fowler, 2009:69). As discussed before, a household survey and key informant interviews were used in the study. On one hand, the household survey was undertaken to get quantitative data from individual farmers, and on the other hand, the key informant interviews were used to map the institutional context in which farmers made their decisions and to cross-reference the information provided. The next sub-section discusses the study area, target population and the sampling methods used in the study.
The study area, Population and Sampling
As highlighted before the study was conducted in Seke and Murewa districts situated in Mashonaland East Province of Zimbabwe. The general description, map and more details of the study area have been presented in section 1.6 in Chapter 1. According to the census of year 2002, Zimbabwe had a total population of about 11 634 663 people and Mashonaland East province had a total population of about 1 127 413 people and 309 198 households (CSO, 2004:3). Most of the population in the province is rural, accounting for about 76 percent of the total population (ibid). The target population were rural households whose livelihoods were mainly agricultural dependant. The next section discusses the sampling approach employed in the study.
A multi-stage sampling was used for the survey. It is a sampling approach that involves selecting a sample of clusters and, within each cluster, further selecting a sample of study units (Patel et al. 2004:162). The multi-stage sampling approach falls under qausi-probability sampling which is similar to a probable sample except that the procedure used to draw the sample differs (du Plooy, 2009:118). This approach is appropriate when drawing a true probable sample is unfeasible, when there is no adequate list of the individuals in the population and there is no way to get to the populations directly (Fowler, 1993:18; Black, 1999:118; Fowler, 2009:28, du Plooy, 2009:120). It allows the researcher to avoid the need to create an expensive sampling frame when there is no sampling frame (Stopher, 2012:305). It is important to note that for this study, the sampling frames were unavailable at the district, provincial and national level, they were only available at the village level. Admittedly, Stopher (2012:266) observes that sampling frames for human populations are usually difficult to obtain in most countries due to non-existence of the lists or their unavailability to researchers.
The sampling process undertaken in the study is as follows. The multi-stage sampling involved a combination of purposive and cluster sampling approaches. Purposive sampling involves handpicking subjects based on specific characteristics, while, cluster sampling involves random samples of successive clusters (Black, 1999:118). Purposive sampling was used to select the province and the subsequent selection of the two districts. One of the desired characteristics was to have two rural districts, one of which was supposed to be closer to a major urban area while the other district was a bit far. The districts’ proximity to urban areas was important, as there was a need to evaluate how such differences could influence access to ICTs and climate change awareness.
For the reasons that have just been discussed, Mashonaland East province was selected. The province has Seke (rural) district, which is closer (adjacent) to the major urban centres namely, Chitungwiza town and Harare the capital city. Murewa district is relatively far from these urban centres (about 100 kilometres). Having selected the province and the two districts, three wards were purposively selected in each district based on the criterion of geographical spread across each district (using district maps). Cluster sampling was then used to select five villages (clusters) in each of the three wards. Bluman (2004:11-13) suggests that cluster sampling is used when the population is large or when it involves subjects residing in a large geographic area. At the village level, the sampling units were selected systematically according to the sampling frames that had complete list of all households in the village. In terms of key informant interviews, a snowballing sampling approach was used. This involved identifying and interviewing initial key informants, who also identified other key informants. The next section discusses the sample size for the survey.
TABLE OF CONTENTS
TABLE OF CONTENTS
LIST OF TABLES
LIST OF FIGURES
LIST OF TEXTBOXES
LIST OF EQUATIONS
LIST OF ACRONYMS
CHAPTER 1: INTRODUCTION
1.1 INTRODUCTION AND BACKGROUND
1.2 RESEARCH PROBLEM
1.3 RESEARCH QUESTIONS
1.6 STUDY AREA
1.7 LIMITATIONS OF THE STUDY
1.8 OUTLINE OF THE THESIS
CHAPTER 2: LITERATURE REVIEW
2.2 THEORATICAL FOUNDATION
2.3 The Agricultural Knowledge and Information System, and the Agricultural Innovation System
2.4 The Cynefin Framework
2.5 Information and Communication Technology for Development (ICT4D) approach
2.6 EMPIRICAL LITERATURE REVIEW SECTION
2.7 CLIMATE CHANGE LITERATURE
2.8 LITERATURE ON ICTs
2.9 THE CONTRIBUTION OF ICTs TO CLIMATE CHANGE ADAPTATION IN THE AGRICULTURAL SECTOR
2.10 ZIMBABWE’S ICT POLICY AND E-AGRICULTURE
2.11 THE CONCEPTUAL FRAMEWORK
CHAPTER 3: RESEARCH METHODOLOGY
3.2 THE RESEARCH PARADIGM
3.3 DESIGN OF THE STUDY
3.4 VARIABLES USED AND QUESTIONNAIRE DESIGN
3.5 DATA COLLECTION AND METHODS
3.6 DATA ANALYSIS
3.7 RELIABILITY AND VALIDITY ISSUES
CHAPTER 4: DESCRIPTIVES OF FARMING SYSTEM CHARACTERISTICS AND ACCESS TO ICTs
4.2 Household demography
4.3 Marital status
4.5 Land ownership
4.6 Source of income and agricultural production
4.7 Participation in organisations and positions of authority
4.8 Farmer-to-farmer input exchange and interactions
4.9 Agricultural extension
4.10 Farmer training
4.11 ICT ownership, access and usage
4.12 Access to print media
4.13 Access to ICT services from other community members and through shared facilities
4.14 Languages in which the various ICT services are accessed
4.15 Energy sources for the ICTs
4.16 Affordability of ICTs
4.17 Conclusion on access to ICTs
CHAPTER 5: SOCIO-ECONOMIC DETERMINANTS OF CLIMATE CHANGE AWARENESS
5.2 Assessment of climate change awareness
5.3 Calculation of the climate change awareness index
5.4 Perceptions on how other areas were experiencing climate change
5.5 Respondents who had talked/discussed about climate change
5.6 Important sources of climate change information
5.7 Regression results on the determinants of climate change awareness
CHAPTER 6: ACCESS TO EARLY WARNING AND WEATHER FORECASTING INFORMATION
6.2 Perceptions on climate variables and extremes
6.3 Early warning on violent storms, floods and droughts
6.4 Access to and use of forecasting information
6.5 Main sources of agricultural information
CHAPTER 7: SUMMARY OF FINDINGS, CONCLUSIONS AND SUGGESTIONS
7.2 The objectives of the study
7.3 Summary of key findings
7.5 Contributions of this study
7.7 Areas of further research
GET THE COMPLETE PROJECT