4. Regression Modelling
4.2 Model One
The first regression model Figure 4.17 examines differences between the 328 Committers and the 406 Non-Committers. Non Committers are those producers that have not always committed to one company in the last five years. This includes some producers that have committed once or twice, but for some reason are not full committers. These “Sometimes” committers are analysed separately in more detail in the following section on reasons for committing, and differences in influences of decisions and behaviour. However for simplicity and to allow binary regression modelling, they are included in the Non Committers category in this section. This allows the comparison of what makes consistent Committers different from others.
The largest cause of differentiation is based on the producers score for the Conscientiousness component (Figure 4.17). Producers are drafted initially by whether or not they have a Conscientiousness score below -0.06583. Initially from the 734 producers, 333 have low Conscientiousness scores and are split to the left. Of these producers 23 per cent are committers. To the right are 401, of which 63 per cent are committers. Those in the left-hand branch are then grouped based on whether or not producers are located in the North or South Island. Similarly those with a high Conscientiousness score are secondarily grouped based on their location. At the left-hand side second branch, there are two final nodes. If a producer has a low Conscientiousness score, and is located in the East Coast, they end up at node A and there is a probability of 0.13 that they commit. If a producer has a low Conscientiousness score, and is located in Otago-Southland, they end up at node B and there is a probability of 0.38 that they commit. For each group it is possible to compare the likelihood that they will commit given they are in a particular Island and with a high or low Conscientiousness score. There is more chance of being a Committer at both high and low Conscientiousness if a producer is in Otago-Southland.
Table 20: Comparison of commitment
Probability will commit East Coast Otago-Southland
High Conscientiousness 45% 74%
Low Conscientiousness 13% 38%
Producers with a high Conscientiousness score are also secondly split by island, with those in the East Coast entering the left-hand branch and those in Otago-Southland in the right-hand branch. Those in Otago-Southland are then at final node C and this group has a 74 per cent chance of being a Committer. The East Coast producers are then split again on their level of Conscientiousness. Those with a very high Conscientiousness score are partitioned and enter final node D where they have a 71 per cent chance of being a Committer. Producers with medium Conscientiousness scores, they enter the final left-hand branch and are partitioned based on their score for the Influencer component. Those with a high Influencer score are partitioned to the left and have a 19 per cent chance of being a committer. Those with a low Influencer score are partitioned to the right and have a 52 per cent chance of being a Committer.
Consc< -0.06583 Location Location Consc< 1.711 Influencer>=0.154 0.13 n=196 0.38 n=137 0.19 n=58 0.52 n=56 0.71 n=42 0.74 n=245
A
B
C
D
E
F
0.23 n=333 0.63 n=401 0.45 n=156 0.35 n=114 0.45 n=734Low Conscientiousness High Conscientiousness
Otago-
Southland Otago-Southland
East Coast East Coast
Med Cons
Very high Cons
High Influencer Low Influencer
Given that there is a 45 per cent chance from the total sample that a producer will be a Committer, it is useful to analyse situations where producers are much more or less likely to commit. In cases where a producer is from Otago-Southland, and has a high Conscientiousness, they are allocated to nodes that have a higher chance of being a committer. Therefore from this we can conclude that those producers that commit are more conscientious in their selling decisions and more likely from the Otago-Southland. Those with a high Influencer score are less likely to commit and measures the effect of how high bargaining power and leadership status reduces the chance of committing. The Influencer score also measured the willingness to try new marketing supply options. Either this variable is overpowered by leadership status or bargaining power, or while producers may have high willingness to try new programmes it may not translate into increased commitment of lambs.
Given the importance of the Conscientiousness and Location it is useful to look at the differences in numbers of producers in each group. There is a higher proportion of producers in Otago-Southland with a high Conscientiousness score (64%) compared with the East Coast (44%).
Table 21: Island and Conscientiousness comparison
Proportion of respondents in each
group
East Coast Otago-Southland Total
High Conscientiousness 156 (44%) 244 (64%) 400 (55%) Low Conscientiousness 196 (56%) 137 (36%) 333 (45%) Total 352 (100%) 381 (100%) 733 (100%)
The level of Conscientiousness correlates with the principal components Market and Active which measures how market focused the producer is and how active they are in their selling behaviour. This may mean that the importance of Conscientiousness has overpowered the Market and Active components, and while Market and Active may also be important, their importance is incorporated through the Conscientiousness component.
Caution needs to be taken, as it could be that producers that commit need to be more conscientious to fulfil their commitments, rather than the conscientiousness tendencies leading to a desire to commit. Nevertheless the model is designed to be exploratory in nature rather than predictive and therefore it is useful to understand the strength of the relationship between Conscientiousness and commitment. The average scores for the Conscientiousness component by Island and whether or not producers commit are outlined in Table 22. On average Otago-Southland producers have higher Conscientiousness scores than East Coast producers in both cases of committing and not committing (i.e. a lower negative score when not committing). Therefore overall Otago-Southland sample participants are more conscientious. There is a greater level of deviation in scores of Non Committers.
Table 22: Conscientiousness score on Island and Commitment Mean Conscientiousness score
(Standard Deviation)
East Coast Otago-Southland
Commit 0.816 (1.50) 0.863 (1.57)