Previous PageTable Of ContentsNext Page

Different pebbles, same pond: ‘Farming styles’ in the Loddon catchment of Victoria

Don Thomson

Sinclair Knight Merz, PO Box 952, Bendigo VIC 3552 Australia,
phone +61 3 5444 1861 e-mail
dmthomson@skm.com.au

Abstract

It is acknowledged in the recent literature on the adoption of recommended farming practices that divergent types or ‘styles’ of farmer will have different propensities to adopt technologies and to participate in learning activities, industry development and planning processes. The same literature laments the lack of tools available to allow policy developers, resource managers and extension providers to predict adoption rates among these ‘styles’ or monitor programs to evaluate their success in attracting a wide range of farmers. This paper describes a ‘farming styles’ tool that has the potential to address this current situation.

The ‘farming styles’ tool described here was developed during PhD research by the author in the Loddon catchment of Victoria. Respondents to a mail survey (366) were assigned to ten groups by K-Means clustering, using their responses to 31 attitude and goal statements about farming. Significant behavioural differences were found between these groups in terms of their participation in natural resource management programs, their adoption of information technology and quality assurance, and their participation in learning activities.

This paper provides a background to ‘farming styles’ research both in Australia and internationally. It outlines the theoretical and methodological basis of the approach developed by the author, and presents a summary of the key results. The paper highlights that many of the assumptions made in other methods of classifying farmers have severely limited our ability to understand the drivers of change in agricultural practice.

I draw the analogy between farming styles and pebbles being thrown into a pond because it is evident from the results of this research that some farmers have more influence on the definition of, ‘good’ farming. Some farmers make a big splash and their influence spreads widely, others make smaller impacts in their local area only. Particular ‘styles’ of farmer play a disproportionate role in defining what ‘good’ farming is and are able to influence the practices of other farmers, often through industry and government programs: their ripples are stronger and can overpower other ideas about ‘good’ farming.

Introduction

The notion of ‘farming styles’ has received some attention in Australia because of its potential application in understanding the diversity of approaches to farming. However, until recently, research into farming styles has been unable to provide resource managers, policy makers and extension agencies with the tools necessary to apply the theory (Thomson 2000, Barr and Cary 2000). While research in Australia has described styles of farming in various regions, it has been unable to accurately classify farmers into styles, and has concluded that styles exist as ‘heuristic parables’, not tangible entities (Howden and Vanclay 2000, Howden et al. 1998, Vanclay et al. 1998).

I developed an alternative farming styles methodology in an attempt to provide an efficient tool for resource managers and policy makers and to address some theoretical and methodological challenges remaining from the Australian and international research. One of the keys to unlocking the potential of the farming styles approaches was to quantify the existence of styles and measure their attitudinal and behavioural differences.

‘Farming styles’ research in Australia

Market segmentation has become a popular part of designing extension programs and targeting key messages to particular audiences. Several market segmentation exercises that seek to classify farmers into types or styles have been attempted in the wool industry (e.g. O’Keeffe and Fletcher 1998), the dairy industry (e.g. D. McKinna et al Pty Ltd 1994) and the adopters of perennial pastures (e.g. Barr et al 2000, Barr and Cary 2000). However, these tend to be a-theoretical (Barr and Cary 2000) and are often limited in their scope by focusing on particular industries, practices, geographic locations and/or focus on structural and demographic variables (Thomson 2000).

In Australia, in-depth research into ‘farming styles’ has followed in the vein of van der Ploeg (of the Wageningen Agricultural College in The Netherlands) and has been predominantly undertaken by Frank Vanclay, Peter Howden, Luciano Mesiti and Scott Glyde of Charles Sturt University, Wagga Wagga (see van der Ploeg 1994, Vanclay et al 1998). The Australian research has concluded that, despite some methodological problems, farming styles are useful in understanding the diversity of approaches to farming and could therefore aid the targeting of extension programs and communications. However, their research has not drawn any firm conclusions about the classification of farmers into styles. Although the application of ‘farming styles’ theory may not require the classification of farmers into styles, a more complete understanding of the nature and extent of farming styles is reliant upon a more holistic classification method. As Vanclay et al (1998:102) observe “analysis of differences between styles still rests on an accurate classification of farmers into styles.”

The methodology applied in the Australian research involves, in essence, the identification of types of farmers using focus groups of farmers and extension officers, the preparation of written portraits of each ‘style’, and the ratification of these styles quantitatively (e.g. Mesiti and Vanclay 1997) or by case studies (Howden 2000, Howden and Vanclay 1998, 2000). Howden and Mesiti both found that farmers had difficulty choosing a style from the portraits presented to them, and extension officers had trouble in nominating farmers into particular styles. Howden and Vanclay’s (2000) conclusion is that styles can be presented as important dimensions of farming, rather than real management strategies. Styles exist as part of farming discourse and people involved in farming, and farmers themselves, have images of different types of farmer.

I argue that the underlying assumption of the existing farming styles research that farmers are universally aware of different types of farmers, as evidenced by their use of metaphors for these types of farmers, is problematic. This approach assumes, firstly, that people set goals to be ‘like’ ‘farmer X’ or ‘style X’; secondly, that farmers are consistent in their perception of what constitutes ‘style X’; and thirdly, that farmers are free to choose to farm in certain ways. Research so far has also been confined within agricultural commodities and/or regions: there is little account of the mix of industries that often occur within and around a geographic location (Thomson 2000).

Howden and Vanclay observed in their later work that it was the methodology adopted and the perception of farming styles as a classification scheme that had limited the Australian research considerably:

Being governed by theory and steeped in the findings of van der Ploeg, Leeuwis and Mesiti and Vanclay, we were blinded by our conviction that styles were tangible… As a consequence we were perhaps too ready to dismiss the diverse meanings that emerged from the focus groups and too ready to accept the focus group results as being ‘real’ farming styles (Howden and Vanclay, 2000:308, emphasis added).

But the passage quoted above contains a clue to a potential way out of the farming styles dilemma, in its recognition of the diverse perceptions of focus group participants. The psychological basis of differences in perception by individuals, itself a result of socialisation and cultural discourses, applies equally to individual perceptions of other people (or farmers), the landscape (their farms) and every other aspect of social life. It is this diversity of perceptions and understandings that I believe is at the heart of resolving the ‘farming styles’ dilemma.

While the notion of farming styles has considerable practical and theoretical interest, the published Australian research has not made significant progress towards achieving its potential. Existing methods have been unable to quantify the attitudinal and behavioural differences between styles and are unable to account for spatial and temporal variation within and between styles. In order to unravel the farming styles mystery, it is necessary to develop a truly derived classification scheme so that attitudinal and behavioural differences between styles can be measured.

A dialectical approach

My approach to resolving some of the theoretical shortcomings of farming styles was to re-conceptualise farming styles from a dialectical perspective. Dialectics is the study of flows, fluxes and processes. The debates in sociology about the relative importance of ‘structure’ and ‘agency’ on individual behaviour and social institutions have entrenched the notion that these constructs are dichotomous (see Hays 1994). A dialectical approach rejects this view and considers the formation and modification of social institutions as a process. Individual actions are constrained (and enabled) by a number of social, cultural, economic, physical and political influences. However, ‘institutions’ and structures are also shaped by the actions of individuals. Over time, and between different localities, diversity emerges in both social institutions and actors as a result of this process.

I envisaged farming styles as groups of farmers who have a similar pattern of responses to social, cultural, political, economic, historical and farm management ‘forces’. Drawing on Kelly’s (1955) ‘Personal Construct Theory’, I conceptualised that farmers’ ideas about what constitutes ‘good’ farming are a function of their personal constructs, which are developed through their life experiences and interactions with other people, including other farmers, extension officers and agribusiness. Each farmer has a different behavioural environment, and is therefore potentially exposed to a different range of influences. In addition to influencing perception, personal constructs influence the type of people we relate to (Kelly, 1955). Personal constructs also effect the way information is interpreted. When exposed to information, people tend to accept the information that conforms to their personal constructs, and shed the rest (Abel et al, 1998, citing Salmon, 1981).

Although people have different behavioural environments, elements of commonality draw people with similar personal constructs ‘together’ (that is, they understand and can relate to each other). This potentially provides the basis for reinforcing beliefs and attitudes about farming, which manifest in consistent patterns of behaviour within these ‘groups’. At any one place or time, farmers with similar patterns of belief and behaviour could be identified (see Figure 1) and grouped together as a ‘style’ of farming (the dashed circles encompassing groups of farmers). Farmers (small circles in Figure 1) respond (the small arrows extending from the ‘farmers’) in different ways to a range of external pressures (or the field of forces illustrated in Figure 1 by the box). The relative position of individuals (farmers) and groups (farming styles) within this field of forces is subject to constant change. However, because farmers of a particular style have similar personal constructs, they are likely to respond to change in a similar way, and thus move within this ‘field of forces’ together. However, farmers also observe the practices of other farmers, without necessarily being able to discern their ‘style’ or personal constructs. So, farming styles are partly defined with reference to other styles of farmer, represented in Figure 1 by the arrows linking the styles. Under this approach, it is possible for individuals to change their style over time, should they be exposed to significant emotional events or other circumstances that alter their personal constructs and perceptions.

Figure 1. Theoretical ‘field of forces’ within which farming styles emerge as regular patterns of response to external influences on farming.

I argue that diversity in agricultural practice is a product of this ‘balancing act’ between external influences on farming and internal cognising about farming and the variation in people’s behavioural space. My PhD research was structured around testing this approach.

Research aims

One of the primary aims of the research that is reported in this paper was to develop and test a method of grouping farmers according to their pattern of beliefs and attitudes about farming (or personal constructs) and then to quantify the behavioural differences between these groups.

Methodology

The methodology involved both qualitative and quantitative components. The qualitative component of the research is not discussed here (see Thomson 2000), but was concerned with understanding the role of the landscape in farmers’ interpretation of farming practice and the ways in which they differentiate landscape type. The Loddon catchment was selected as a case-study catchment because of its diversity of land uses and landscape types. It covers an area of approximately 1.5 million hectares and falls within the Murray Darling Basin.

The quantitative component of the research is based on a mailed survey of 1000 landholders from all industries within the Loddon catchment where the capital improved value of the holding was over $50,000. The survey was in four parts. Part One gathered data on the structural characteristics of the farm enterprise (size, industry). Part Two contained gathered data on the participation of respondents in natural resource management initiatives, their adoption of recommended practice and their information preferences. Part Three contained the 31 attitude/goal statements about farming. A five-point Likert-type scale was used for each item. Part Four included questions pertaining to demographic and income characteristics. A response rate of forty-two percent was achieved after one reminder. Three hundred and sixty-six cases were available to be assigned into farming styles.

Cases (farmers) were assigned to groups (farming styles) using K-Means clustering – a non-hierarchical technique that, in effect, groups respondents according to their pattern of responses to any number of input variables. The input variables used here were 31 attitude/goal statements derived from qualitative research into farming styles by Mesiti (Mesiti and Vanclay 1997, pers. comm. 1999) and from research into different styles of farming in Scotland (Austin et al. 1996). An extensive statistical exploration of the data was undertaken to determine the number of ‘natural’ groups in the data. The aim of this exercise was to identify the number of times the data could be ‘split’ (i.e. the number of groups selected) where the within-group similarity and between-group difference was maximised. These criteria were measured using a range of tests1. Based on this assessment it was decided to split the data into ten groups. The resulting group populations are provided in Table 1.

Table 1. Group populations

Farming Style

n

One

37

Two

44

Three

43

Four

53

Five

29

Six

48

Seven

30

Eight

8

Nine

59

Ten

15

Total

366

I did not name the emergent farming styles for several important reasons. Firstly, Howden (Howden et al 1998; Howden and Vanclay, 2000; pers. comm. 2000) observes that the metaphors or names of farming styles identified in their research are influenced by the language of extension officers; they are not necessarily ethnotaxonomic. Secondly, metaphors may not be universally used, or have the same meaning in one region as another. Thirdly, any names applied are likely to influence the interpretation of the characteristics of the style because individuals have their own idea about what constitutes a 'progressive' farmer or a 'Collins Street' farmer, for example.

Eight indices were constructed from the 31 attitude/goals statements used in the clustering process to assist in the interpretation of similarities and differences between styles in terms of their attitudes and motives about farming. These indices are: finance, farming practice – business, farming practice – tradition (or a ‘lifestyle’ dimension), knowledge, labour, land (environment), planning/risk and technology/innovation. Index scores were calculated by summing the component item scores and then standardised.

Having split the data into ten groups, the behavioural, structural and demographic data from other parts of the survey were then used to profile each group, and explore the differences between each ‘style’.

Results

Attitudinal characteristics of ‘farming styles’

The characteristics of each farming style in terms of their attitudes and beliefs about farming are compared in Figure 2. This graph highlights the variance within each index and the level of significance of the differences between styles on each attitudinal index. For example, it is clear that Style One farmers' attitudes towards farming as a business are stronger than all other styles, and that they also believe strongly that upholding the traditional values of farming is important. While Style One farmers see farming as a business, they are not necessarily more willing to plan ahead and manage risk than other farmers.

Figure 2. Mean index scores (standardised) for selected ‘farming styles’ in the Loddon catchment.

A strength of the approach adopted here is that it acknowledges the possibility of 'multiple realities' because it searches for patterns in the data and does not attempt to classify cases (farmers) into predetermined classes. It allows for the possibility, for example, that some people may feel strongly that farming is a business and a way of life.

The most significant finding of this research however is that it allows a holistic understanding the underlying attitudes and motives of farmers. This has significant implications in terms of identifying indicators or predictors of behaviour. This is clearly demonstrated by grouping the survey respondents on structural, demographic and behavioural variables and comparing the resultant index scores (particularly the degree of variance on each index) to the farming styles method. Table 2 provides a summary of ANOVA tests for the mean index scores across the eight attitudinal indices for farming style, age, income and Landcare membership. Table 2 demonstrates the significant variance across all eight attitudinal indices for the farming styles groups. On the other hand, other classifications reveal significant differences on only a few indices. For example, classifications based on income provide some understanding of attitudes towards farming as a business (farmers in the higher income categories tend to have a stronger belief that farming is a business), but it provides little insight into their thoughts about training or the environment. Landcare membership appears to be more of an indicator of innovativeness and acceptance of new technologies than an indicator of orientation towards the environment.

Table 2. Variance in index scores for farming styles compared with structural, demographic and behavioural classifications.

Index

Farming Style
(9 df)

Age
(9 f)

Income (total cash receipts) (6 df)

Landcare Membership
(1 df)

F

P

F

p

F

p

F

p

Finance

76.00

***

2.34

*

6.91

***

0.04

n.s.

Farming Practice – Business

45.01

***

0.63

n.s.

8.55

***

1.06

n.s.

Farming Practice – Tradition

12.94

***

2.78

**

1.08

n.s.

0.66

n.s.

Labour

8.41

***

2.165

*

0.41

n.s.

2.29

n.s.

Knowledge

13.51

***

3.91

***

0.68

n.s.

0.35

n.s.

Land (environment)

13.49

***

1.609

n.s.

1.78

n.s.

8.21

**

Planning/Risk

33.98

***

5.04

***

3.29

**

9.28

**

Technology/Innovation

31.02

***

2.96

**

3.90

***

21.09

***

Key to p values: *** = >99%, ** = 99 to 95%, * <90%, n=366

A 'taxonomic key' to farming styles in the Loddon catchment is presented (Figure 3) to illustrate the similarities and differences between some styles. Styles are differentiated according to the attitude indices that exhibit a significant difference between styles (at the 95% confidence level as a minimum requirement).

Figure 3. A Taxonomic key of farming styles in the Loddon Catchment.

While this section has illustrated that the farming styles identified here have significantly different patterns of attitude and belief about farming, a better test of these attitudes is to see to what extent they manifest in patterns of behaviour that are different from farmers in other styles.

Behavioural characteristics of farming styles

The real test of the potential of this farming styles method as a tool for resource managers and policy makers is its ability to explain and predict patterns of behaviour. This section highlights that there are significant differences between the ten styles in average behaviours. But more importantly, the pattern of behaviours is coherent and reflects the underlying attitudes of each style. It is this coherence that has value in terms of predicting behaviours.

Table 3 compares farming styles with classifications based on structural, demographic and behavioural categories. While income could be interpreted from these results as a stronger predictor of a range of behaviours than farming styles, the relationship between income and adoption/participation is not necessarily linear. For some items, such as Landcare membership, the middle-income categories have a higher membership than the higher income groups. Also, income is not necessarily a determinant of these behaviours: the behaviours may be a determinant of income. The important difference between structural and demographic classifications and the farming styles approach described here is that the underlying motives and beliefs are more fully appreciated across a range of measures.

Table 3. Behavioural differences of farmers grouped by farming style, age, income and Landcare membership.

Index

Farming Style
(9 df)

Age
(9 df)

Income (total cash receipts) (6 df)

Landcare Membership
(1 df)

Chi2

p

Chi2

P

Chi2

p

Chi2

p

Participation in industry training

25.8

**

14.0

n.s.

53.2

***

25.2

***

Adoption of QA

18.3

*

(17.1)

(*)

28.6

***

10.2

**

Landcare membership

23.5

*

14.5

n.s.

14.0

*

   

Adoption of farm planning

47.8

***

22.4

**

32.4

***

2.9

n.s.

Adoption of conservation works

44.7

***

31.2

***

26.3

***

20.5

***

Use of funding incentives

19.6

*

16.3

n.s.

16.7

**

38.3

***

Key to p values: *** = >99%, ** = 99 to 95%, * <90%, n=366

Items in parentheses are suspect due to low frequencies in five or more cells in the contingency table.

The pattern of behaviours for each farming style illustrated in Table 4 reflects the attitudinal profiles of each style. That is, participation and adoption behaviour among farming styles is coherent and varies across different situations in accordance with their beliefs.

The consistency of behaviour over a range of variables for Styles Three, Five and Eight is extremely high. For some Styles, such as One, Four and Ten, the apparent inconsistency in behaviour is coherent with their attitudinal profiles. For example, Style One farmers have a high business and planning orientation, and report a moderate level of concern for the environment. This manifests in a high level of activity in planning and quality assurance, and moderate to low participation rates on items relating to natural resource management initiatives. The high involvement of Style One farmers in the adoption of QA programs seems contradictory to their level of participation in other initiatives. However, this reflects their very strong beliefs in farming as a business. This is also reflected in the relatively high adoption of farm planning among Style One farmers.

Table 4. Relative behavioural characteristics of each farming style

Variable

Proportion of respondents:

Farming Style

1
10%

2
12%

3
12%

4
15%

5
8%

6
13%

7
8%

8
2%

9
16%

10
4%

Participation (Industry training)

Mod

Low

High

Low

Low

Mod

Mod

Low

Mod

Mod

Adoption (QA)

High

Low

High

Low

Low

Mod

Mod

Mod

Mod

Mod

Landcare membership

Low

Low

High

Low

Low

Mod

Mod

Low

High

High

Adoption (farm planning)

High

Low

High

Low

Low

High

Mod

Low

Mod

High

Adoption (conservation works)

Low

Mod

High

Low

Low

High

High

Low

High

Mod

Use of funding incentives

Mod

Low

High

Low

Low

Mod

Low

Low

High

Mod

Awareness of RCS

Low

Mod

High

Mod

Low

High

High

Low

High

High

Read RCS

Mod

Low

High

Mod

Low

Low

High

Low

Mod

High

Involvement in RCS

Mod

Low

High

Mod

Low

Low

Low

Low

High

Mod

Read SMP

Low

Low

High

Mod

Low

Mod

Mod

Low

Mod

High

Involvement in SMP

Low

Mod

High

Mod

Low

Low

Mod

Low

Mod

Mod

Membership of CMA/IC

Low

Low

High

Mod

Low

Low

Mod

Low

High

Low

RCS = Regional conservation strategy, SMP = Salinity management plan, CMA = Catchment management authority, IC = Implementation Committee, QA = Quality Assurance Schemes (various). (n=366)

Conclusion

This research has confirmed that a ‘farming style’ can be defined as a pattern of beliefs, motives and attitudes about farming that is manifest in particular patterns of behaviour. By understanding the attitudes, motives and beliefs of farmers within particular styles, it is possible to predict behaviours, forecast the rates of adoption of practices, and target communications more effectively.

The farming styles approach is more meaningful and useful than reductionist classifications based on structural, demographic or behavioural variables. Although structural and demographic characteristics sometimes correlate with the particular behaviours, there are too many exceptions to make them useful in terms of predicting behaviour. Structural classifications tend to be simplistic as they make too many assumptions about the motives of farmers and are therefore limited in their capacity to predict behaviour and responses to change.

The farming styles method described here has significant potential for organisations seeking to:

  • communicate with and engage diverse audiences
  • influence the behaviour of land managers
  • monitor the participation of key stakeholders/groups in a variety of forums and decision-making processes.

The importance of this research is the acknowledgment of ‘multiple realities’ and the rejection of the assumption of homogeneity in attitudes and motives of farmers. While the method described here is efficient and can be administered in a number of ways, it may not be necessary to undertake a classification exercise to identify farming styles in a particular region or commodity group. A rejection of the simplistic assumptions of reductionist classifications and a celebration of diversity may be all that is required in order to design more effective extension programs.

References

  1. Abel, N., Ross, H. and Walker, P. (1998) ‘Mental models in rangeland research, communication and management’, Rangeland Journal, 20(1): 77-91.
  2. Austin, E., Deary, I., Gibson, G., McGregor, M. and Dent, J. (1996) ‘Attitudes and values of Scottish farmers: yeoman and entrepreneur as factors, not distinct types’, Rural Sociology, 61(3): 464-474.
  3. Barr, N. and Cary, J. (2000) ‘Influencing improved natural resource management on farms’, Bureau of Rural Sciences, Kingston ACT.
  4. Barr, N., Ridges, S., Anderson, N., Gray, I., Crockett, J. Watson, B. and Hall, N. (2000) ‘Adjusting for catchment management’ Murray Darling Basin Commission, Canberra.
  5. D. McKinna et al. Pty Ltd (1994) ‘A marketing approach to developing and evaluating extension programs’, Department of Agriculture, Melbourne.
  6. Hays, S. (1994) ‘Structure and agency and the sticky problem of culture’, Sociological Theory, 12(1): 57-72.
  7. Howden, P. and Vanclay, F. (2000) ‘Mythologisation of farming styles in Australian broadacre cropping’, Rural Sociology, 65(2): 295-310.
  8. Howden, P. Vanclay, F., Lemerle, D. and Kent, J. (1998) ‘Working with the grain: farming styles amongst Australian broadacre croppers’, Rural Society, 8(2): 109-125.
  9. Kelly, G. (1955) ‘The psychology of personal constructs. Vol. 1 A theory of personality’, Routledge, London.
  10. Mesiti, L. and Vanclay (1997) ‘Identifying farming styles in Australian viticulture’, In F. Vanclay and
  11. L. Mesiti (Eds.) Sustainability in Social Research – Proceedings Conference of the Australian Association for Social Research, Charles Sturt University, Wagga Wagga, February 1997. Centre for Rural Social Research, Charles Sturt University, Wagga Wagga. pp. 257-287.
  12. O’Keefe, M. and Fletcher, M. (1998) ‘Productivity gains in the wool industry: towards a new perspective on adoption and extension’ Report prepared for the Bestwool 2010 Steering Committee, Rabo Australia Limited.
  13. Thomson, DM (2000) ‘As if the Landscape Matters: The social space of ‘farming styles’ in the Loddon catchment of Victoria. Unpublished PhD Thesis, School of Anthropology, Geography and Environmental Studies, University of Melbourne.
  14. Van der Ploeg, J.D. (1994) ‘Styles of farming: an introductory note on concepts and methodology’, in J.D. van der Ploeg and A. Long (Eds.) ‘Born From Within: Practice and Perspectives of Endogenous Rural Development’, Assen, Van Gorcum. pp. 7-30.
  15. Vanclay, F., Mesiti, L. and Howden, P. (1998) ‘Styles of farming and farming subcultures: appropriate concepts for Australian rural sociology?’, Rural Society, 8(2): 85-107.

1 These measures included: equality of group means (F-Ratio and Wilk’s Lambda), between-group sum of squares, within-group sum of squares and F-Ratio, and the percentage of cases correctly classified by discriminant analysis (jack-knifed and non-jack-knifed).

Previous PageTop Of PageNext Page