Kiwifruit Results and Sector Comparisons

Get Complete Project Material File(s) Now! »

Modifications to the list of factors

The 36 factors used in the 2005 kiwifruit study (see Table 1) were the starting point for the factors used in the 2008 study. These original factors were derived from 14 open-ended interviews with a variety of farmers near Lincoln University. The challenge of moving to a more participant-driven list involves the potentially large variety of terms individuals might employ to refer to very similar elements of management systems. There is a trade-off between number of factors and efficiency of research design: more factors would increase the chance that each farmer would have all the key factors they would need but additional factors would make the sorting task more difficult. Accordingly, some attention was given to the number of factors used but with a concern not to significantly increase the total number.

Analysis of Q-sort data

The ordering of items in a Q-sort array shows which items are important or salient in some way and provides the quantitative data used for factor analysis. The columns in the Q-sort are routinely assigned numbers with zero for the middle column and, in this case, ranging from -4 for the left hand or unimportant column, to 4 in the right hand or very important column. This numbering is somewhat arbitrary and is deployed merely to provide numbers for the subsequent correlation between all Q-sorts and to impose (somewhat artificially) an ordering of the data that has a mean of zero and is not skewed. In effect, the columns in the middle of the Q-sort, enumerated as zero or close to zero in the data analysis, can be taken to mean some importance since the orchardists put most factors into the important and some importance piles at the beginning of their Q sort.

Group map data for all 30 orchardists

When the data for each of the 30 cases had been entered into individual Excel spreadsheets as a data matrix it was possible to create an equivalent data matrix for the group map by calculating the average score for each cell in the group matrix. These average scores then formed the basis of further calculations. The complete matrix for the group map data showed that for the average group map there were a total of 195 separate connections between factors, considerably short of the theoretical maximum of 41 times 40 or 1,640 connections, (or double that if consideration is taken that connections can go both ways). This is still rather too many to represent easily on a single map (see later). In this section of the report the data are presented by first focusing on the group map data and then focusing on the group map generated by these data.

Group map

The centrality scores show which factors are important but they do not show, in detail, how all the factors are linked. To show linkages, we need to use the average data to generate a causal map based on strength of causal connections. However, the full group map has linkages between many factors and is difficult both to present and to interpret. To simplify the group map it is useful to include only those factors with centrality scores of 3 or more. The derived group map is shown in Figure 5. The map was created by taking the three top factors and placing them in triangular fashion in the centre of the map. The top three factors are shown in the largest font and with a lightly shaded background. Then the next four factors were placed around these so as to minimise the overlapping of arrows. These four factors are shaded slightly darker. Finally, the remaining factors were positioned in close proximity to the factors with which they connected while keeping the lower rated factors on the perimeter. The figure shows only one arrow with double arrowheads and two numbers on the line. The number nearest to the arrowhead applies to that arrowhead.

Data from questions asked after the mapping

The main purpose of the analysis of orchardists’ response to the questions asked after the mapping was to assess if there was any obvious change in their situation. To that end, two questions were given attention: that asking about their goals and that inquiring about significant changes having occurred since 2005. These data are collated in Appendix 1 and show that for most of the orchardists their reported goals were unchanged between 2005 and 2008. There were ten (36 per cent) for which we assessed there was no change in their goals and there were 13 (46 per cent) for which we assessed there was little change in their goals. However, one orchardist had moderate change (goals in 2005 related to production only, in 2008 related to family and the threat of rezoning), and four orchardists had significant change (now land banking rather than maximising production, moving off orchard, subdivision pressures, drop in orchard gate returns – selling property). These data show that while the majority of orchards are in a similar if not identical position, a minority were in a different situation in 2008. This observation means that the 2008 causal mapping results cannot be expected to be identical to the 2005 results.

Group map data for each panel

The assessment of differences between the group maps created for each of the ARGOS management panels involved the identification of significant differences for centrality scores and map characteristics among the panels. These data were examined using unbalanced one-way ANOVA analyses and the significant results are shown in Table 5. Superscript letters are used to denote a statistically significant difference between scores, and bolding is used to show the highest centrality for each factor. At first glance, seven of the eight significant differences are at the five per cent level for a two tailed test. However, for the factor, vine health, our earlier research (Fairweather et al., 2006; Hunt et al., 2005) suggests a hypothesis that Organic orchards would have a higher score a hypothesis that Organic orchards would have a higher score for several reasons.

Contents :

• List of Tables
• List of Figures
• Acknowledgements
• Summary
• Chapter 1 Introduction: Background and Research Objectives
  • 1.1 Background
  • 1.2 Research objectives
  • 1.3 Outline of report
• Chapter 2 Method: Q-methodology and Causal Mapping
  • 2.1 Introduction
  • 2.2 Modifications to the list of factors
  • 2.3 Outline of Q methodology
  • 2.4 Interview procedure
  • 2.5 Analysis of Q-sort data
  • 2.6 Analysis of Map Data
  • 2.7 Conclusion
• Chapter 3 Kiwifruit Results and Sector Comparisons
  • 3.1 Introduction
  • 3.2 Group map data for all 30 orchardists
  • 3.3 Group map data for each panel
  • 3.4 Group data for each Q-sort type
  • 3.5 Comparisons between sectors for conventional and organic panels
  • 3.6 Comparisons between management systems in the different sectors
  • 3.7 Comparisons between sectors for Q-sort types
  • 3.8 Comparisons across management systems that are common to Q-sort types – Q-sort effects
  • 3.9 Overall sector comparisons
  • 3.10 Conclusion
• Chapter 4 Key Findings and Discussion
  • 4.1 Introduction
  • 4.2 Summary of approach
  • 4.3 Summary of results
  • 4.4 Discussion and Interpretation of Results
  • 4.5 Conclusion
• References
  • Appendix 1 Assessment of change in goals from 2005 to

GET THE COMPLETE PROJECT
Kiwifruit causal mapping in 2008: Comparisons to 2005 and to other sectors