This stage involved two types of research: firstly, discourse analysis of primary sources of farmer information; and secondly, content analysis of 200 magazine articles from the New Zealand Dairy Exporter.
Justification: This segment of the research aimed to obtain a picture of cultural factors that influence the decisions of dairy farmers, particularly the images and
The discourse analysis was intended to convey the richness of detail about the messages that farmers receive from different parts of the dairy industry. It allows interpretive analysis, but cannot ‘prove’ that messages discerned by the analyst are the same as those picked up by the farmer. The content analysis was intended to assess the extent to which the popular dairy farmer media addresses production issues compared with environmental matters.
A further justification for both the discourse and content analyses was to provide a test for the theoretical proposition that farmers ignore or oppose protection or retention of native forest remnants because they are influenced by cultural institutions and values that militate against alternative management values. The discourse and content analyses applied to documentary sources that are typical of those which cross the farm gate daily or weekly. If I were to discover marked differences between the values, attitudes and farm management philosophy expressed by farmers with the images conveyed by farm media, it would suggest that cultural ‘constructivist’ interpretations were perhaps not as significant as the theoretical propositions suggest. If the media express messages that are strongly productivist and weakly environmental, and farmers likewise indicate attitudes that are strongly productivist and weakly environmental, it would reinforce the cultural constructivist interpretations.
The use of multiple methods, triangulation and robustness
Use of multiple methods and different sources of information provides for robustness of research that allows the inferences from each method to be tested against the others and, if necessary, modified. Some of my assumptions at Stages 1 and 2 of the research had to be modified in the light of the Stage 3 survey findings. For example, I was surprised by the proportion of respondents to the survey who indicated they had native bush on their farm, and by the proportion of farmers who said they would have liked native bush on their farm.
Use of multiple methods is known as ‘triangulation’ (Sarantakos 1993, 155). It is often practised to increase the validity and reliability of the research; if the same finding is discovered by different methods, it is considered a reliable finding, and if a conclusion is in accord with a theoretical proposition or hypothesis, it is considered to be valid. Sarantakos cites Lamnek (1988) to the effect that triangulation is inadequate if the methods are based on wrong conditions and
views and interests; if the methods cannot be replicated; or if the methods are not relevant to the subject under study (Sarantakos 1993, 156).
In the case of this study the use of multiple methods was intended to add robustness to the research. Trying to find answers to the two research questions was like fishing for a number of elusive fish in murky waters. The different methods are like nets of different mesh size. Even if a number of ‘fish’ have escaped the investigative net, the use of multiple methods enabled a more representative ‘catch’ than would dependence on just one or two methods.
ETHNOGRAPHY AND PARTICIPANT OBSERVATION
Ethnography and participant observation share similar objectives and methodological philosophies. Both involve participation by the researcher in the life-worlds of the people being researched, and both assume that meaning is an important dimension of human action. Participant observation covers the range of research methods, from full participation in the lives of the people being researched, to detached observation at the other. Ethnography involves the former and can run for months or even years.
The field research for this thesis is better described as participant observation because I did not live with the farmers I studied. However, it involved some aspects of ethnography in that it was spread over 9 months, involved repeated visits to the same families, and participation in their family and farm activities (e.g. family meals and social events as well as milking and feeding out hay to the cows). Because it involved repeated visits over many months, I came to recognise and understand aspects of farm life that would not have been evident from one-off visits or discrete observations. In addition, the aims of ethnography as articulated by anthropologist Clifford Geertz were central to my views of the researcher’s task.
To Geertz (1973), ethnography is ‘thick description’. By that term he means a process of coming to understand, translate and record the details of other peoples' lives in ways that convey the sense and logic of their way of seeing and doing. It is a way of rendering the meaning of other people’s actions, and recording that meaning in terms that readers of the ethnographic record can understand. Geertz notes three distinct aspects of the ethnographic process:
interpretation (i.e. identifying the social frame); and fixing the discourse into a timeless frame by lifting it from its temporal moment (i.e. recording the discourse as a fixed record outside its temporal context).
The effort and process of understanding (i.e. interpreting) the meaning of a particular action can only take place through analysis of the cultural context within which that action occurs. Actions derive their social meaning and purpose as part of a system of shared conventions, expectations and the beliefs held by the actors. For example, within the context of New Zealand culture, the person who casually greets an acquaintance with, ‘Hello! How are you?’ does not expect the other to go into a long description of his or her aches and pains. However, if the exchange is between a doctor and a patient, or if the person greeted is a close friend, then the greeter might expect a detailed reply. For a New Zealander, further differences of meaning (and the social identity or status relationships, of the people involved), can be inferred by the use of different terms: 'Good morning', 'G'day', 'Kia ora', 'Morena', 'Tena koe', as well as by non- verbal gestures such as shaking hands, nodding, eye contact etc. The significance of ‘How are you?’ can only be assessed by reference to the context of the occasion and in the knowledge of a code of conventions and mutual expectations between the people involved. That knowledge comes from the gradual accumulation of knowledge about the discourses that surround greetings, and the way they vary from one social context to another.
Ethnography allows cues to be picked up in ways that are not possible from one or a series of interviews. For example, it was soon evident that milk hygiene is an important issue for dairy farmers, and that bacterial cell counts in milk are a major concern for the average farmer because they may be penalised for milk with a high cell count. But full understanding of the complex exchange of information between a dairy company and a farmer only came after several months of observing the behaviour of farmers when the daily record of information on milk production and milk quality for the previous fortnight arrived. In that respect, the field research was ethnographic because it enabled a build-up of knowledge and understanding about the meaning of events experienced by dairy farmers.
farmers of experiential learning discussed in Chapter 5. As they showed me their animals, paddocks and elements of the landscape, the farmers demonstrated how acute were their perceptions and judgements related to management decisions. On the basis of lived experience they could assess the amount of grass in a paddock and how much of the area they would need to fence off each day to provide their cows with enough grass. If they gave the cows too much then precious feed would be wasted; if too little then the cows would suffer and produce less milk. Farmers constantly have to make fine judgements about the welfare of their cows and the state of their land. Some do so on the basis of recommended formulae – so many kilograms of dry matter per cow per day; so much fertiliser per hectare – but the variability of cows, weather, soils, grass, and other factors, mean that no formula can account for every circumstance and most farmers learn to make their judgements through experience.
Ethnography as participant observation can engender an effective learning atmosphere because it involves the person as a learning ‘instrument’. Because it is situational and interactive it tends to involve most or all of the sensory perceptions of the learning person and this, in turn, helps enrich and reinforce memory. For example, I was able to remember incidents of behaviour and interaction on the farm more vividly than I could remember conversations during interviews. This was because remembered incidents gave access to diverse cues when searching memory to write down field notes. Ethnographic observation provided a form of physical, sensory and emotional immersion that reinforced the learning response. It also involves redundancy and repeated messages in different contexts. If I failed to pick up a cue the first time, I could pick it up again the second or third time. The different contexts in which such cues did or did not occur helped to indicate more about the significance and context of particular cues.
The ethnographic method is an iterative process of asking questions, collecting data, analysis, interpretation, making an ethnographic record, hypothesising/asking questions on the basis of the record, collecting further data, analysing, interpreting and recording. It serves to develop an accumulated understanding of actions and their meanings.
SYMBOLIC INTERACTIONISM AND THE CONSTRUCTION OF SOCIAL REALITY
The theoretical relevance of symbolic interactionism was explained in Chapter 4. According to symbolic interactionism, individuals learn ways of viewing the world and realms of meaning from their interaction with others. Methodologically, symbolic interactionists ‘attempt to take the role of the subject and to interpret the context in which the behaviour takes place. . . . The point is to seek human truths as they are constructed and enacted by the subject’ (O’Brien and Kollock 1997, 19). I sought to follow this approach during participant observation and interviews by analyzing actions and discussion in the light of the circumstances. The approach has been outlined by symbolic interactionist, Herbert Blumer (1969, 40). He uses the term ‘exploration’ to describe ‘a flexible procedure in which the scholar shifts from one to another line of inquiry, adopts new points of observation as his study progresses, moves in new directions previously un- thought of, and changes his recognition of what are relevant data as he acquires more information and better understanding’. It is a stage of getting to know the worlds of people under study, before attempting to specify hypotheses or set out pre-designed research procedures.
Blumer (1969, 22-27) argued for an empirically based research methodology which he set in contrast to the conventions of the logico-positivist approaches that he saw as the ‘overwhelming bulk’ of sociological and social psychological research of his time. The shortcomings of conventional approaches, as he saw them, were an ‘almost universal failure to face the task of outlining the principles of how schemes, problems, data, connections, concepts, and interpretations are to be constructed in the light of the nature of the empirical world under study’ (Blumer 1969, 27). In his view, reality exists only in the empirical world and not in the methods used to study that world; it is to be discovered in the examination of that world and not in the analysis or elaboration of the methods used to study that world.
A criticism of the methodology suggested by Blumer is that it may lack theoretical rigour and fall prey to emergent social or political bias as a consequence of the ‘unstated assumptions of the researcher, the climate of opinion in the discipline, and the distribution of power in the interactive setting’ (Huber 1973b, 282, as cited by Denzin 1992, 49). This was a danger I appreciated in that I had to be careful not to impose my own preconceived notions on the data. I had assumed from the beginning that there would be significant value differences between
imposing preconceived notions onto individuals of one or the other ilk.
‘Grounded theory’ is a methodological procedure that developed from the views expressed by Blumer. It uses the research process to develop theory through systematic observation and induction (Flick 1998; Glaser and Strauss 1967; Strauss 1987; Strauss and Corbin 1990; 1997). Conceptually, the method aims to discover social organisation based on the assumption that it is the outcome of repeated interactions and meaningful relationships between people over time. Observation of such interactions will give rise to insights about the possible meanings and connections between observed patterns of behaviour, and these insights can be tested empirically by designing sampling procedures that will throw light on the behaviour in question. The theory is termed ‘grounded’ because it is said to come from and be ‘grounded’ in the research methods used.
The systematic procedures developed by Strauss and others have become highly elaborate, rigorous and detailed. They involve complex systems of coding, ‘memoing’ (i.e. recording of possible theoretical insights) and sampling designed to observe people’s behaviour under different, theoretically relevant, circumstances. I approached the participant observation studies of the farm families in the flexible manner suggested by Blumer and my data recording and analysis techniques for this element of the study were strongly influenced by the techniques proposed by grounded theory. However, I approached the observation of the five farm families with an open, but not an empty, mind. To the contrary; my mind was full of research assumptions and propositions from the outset. Empirical observation required change to some of these assumptions but gave the basis for new or revised theoretical propositions.
THE QUANTITATIVE SURVEY AND ITS JUSTIFICATION
Efficient and effective research requires all elements of a research design contribute to the objectives of the research (Bechhofer and Paterson 2000). This is especially so for methods, such as sample surveys, that can be demanding of time and resources.
To avoid the additional time and trouble of a sample survey, I explored sources of information that might provide a substitute for a survey. These included Ministry of Agriculture and Forestry publications and statistics, Livestock Improvement
Institute, the Ministry of Research, Science and Technology (MORST), and the Department of Statistics. Almost all the surveys examined did not draw a clear distinction between dairy farmers and other types of farmers. Although there is information available on land use and land cover, and different spheres of environmental management by dairy farmers (e.g. Fairweather and Keating 1994; Parminter and Perkins 1997; Parminter et al. 1998), there is little published information about the proportion of dairy farms that contain native vegetation, or the values, attitudes and management practices of farmers in relation to native vegetation (but see the references in Chapters 2 and 3 to Burns et al. 2000; Cruickshank and Peuckert 1989; Denyer 2000).
I was unable to discover information in the standard Ministry of Agriculture and Forestry surveys or Department of Statistics agricultural surveys, or elsewhere about the nature and extent of native forest on dairy farms, and the values, attitudes and management practices of dairy farmers in relation to native habitat. Because dairy farmers are tied to a political economy of global dimensions, and subject to productivist pressures from the dairy industry, I concluded that a sample survey could be worthwhile. (See Appendix 3 for questionnaire.)
DISCOURSE ANALYSIS
Discourse analysis concerns the content and construction of the language and, latterly, visual communications media, used in everyday discourse to reveal structures of knowledge and social practice characteristic of different groups of people. A discourse is a communicative process within a realm of meaning that is shared by those who take part in it. The aim of discourse analysis is to uncover or ‘deconstruct’ the meanings and assumptions that underlie cultural forms of communication, often with the intention of exposing power relationships between and within different societal groups.
The focus of discourse analysis is the ‘text’ in its social context. ‘Text’ was originally used by linguistics to refer to linguistic texts, including any written or spoken product of language (Fairclough 1992). That meaning has been broadened to include virtually all forms of culturally meaningful production including economic, social and political institutions (Johnston et al. 1994, 621). Although the concepts of text and discourse have been broadened to include
analysis rests on language and language-related media.
An assumption underlying discourse analysis is that ‘linguistic practices [are] the primary medium through which social processes operate. Social and institutional diversity is established and perpetuated through diversity in linguistic usage, different ‘ways of speaking’’ (Lee 1992, x). Furthermore, linguistic practices reveal ‘frameworks that embrace particular combinations of narrative, concepts, ideologies and signifying practices, each relevant to a particular realm of social action’ (Barnes and Duncan 1992, as cited by Johnston et al. 1994, 136). Lee (1992), Fowler (1991), Fairclough (1992; 1995a; 1995b) and others have used the analysis of media texts (newspapers, TV, movies, advertisements) to analyse social and political structures and changes within society.
‘Critical discourse analysis’ is an approach which
. . . sets out to make visible through analysis, and to criticize, connections between properties of texts and social processes and relations (ideologies, power relations) which are generally not obvious to people who produce and interpret those texts, and whose effectiveness depends on this opacity (Fairclough 1995a, 97).
More succinctly, Wodak describes critical discourse analysis as a means to make ‘transparent structural relationships of dominance, discrimination, power and control as manifested in language’ (Wodak 1997, 173)
Fowler (1991, 222) provided a ground-breaking analysis of how language is used in newspapers to shape the ideas and beliefs of readers. He argued that,
News is not a natural phenomenon emerging straight from ‘reality’, but a product. It is produced by an industry, by the relations between media and other industries, and most importantly, by relations with government and with other political organisations. From a broader perspective, it reflects, and in return shapes, the prevailing values of a society in a particular historical context.
Fowler’s focus is the way in which media such as newspapers shape and reinforce the worldview of their audiences. Reader and journalist alike are involved in a discourse which contains values, assumptions, conventions and a world view that is the outcome of ongoing interaction. He notes that the construction of news is a process that involves both writers and readers in an
institutional context and independent of any specific individual.
The arguments put forward by the discourse analysts make sense, in particular that the media express as well as reinforce the values of their readers. Chapter 10 discusses the results of content analysis of articles from the New Zealand Dairy Exporter, a subscription magazine that relies heavily on its readership for subscription income and advertising. Given that it has survived competition from other farmer magazines (many of them distributed free) for decades, it is a reasonable assumption that it has survived because it reflects and reinforces the values, assumptions and concerns of New Zealand dairy farmers.
The linguist, Norman Fairclough, has had an important influence on the practice