Factors related to the adoption of organic farming in Spanish olive orchards

The area given over to the cultivation of organic olive trees in Spain has greatly increased in the last ten years. The diffusion of organic farming is desirable since its principles are in line with the aim of the current EU Common Agricultural Policy to harmonise agrarian production and environmental protection, and to encourage the production of quality agricultural goods. This paper characterises organic olive growers and farms in the south of Spain and compares them to their conventional counterparts. Information was gathered in a survey of 322 olive growers. The organic growers differed from their conventional counterparts in that their orchards were less productive and in the more part-time nature of their dedication to agriculture. They were also younger, more involved in management and administration of the holding, attended more courses, were more commonly members of agricultural associations, received more information via the Andalusian Committee for Organic Agriculture (a control, certification and training organization), had a more negative opinion regarding the use of chemicals, and believed that organic agriculture required more time and effort but provided greater returns. Additional key words: adoption factors, diffusion of innovations, Olea europaea L., organic practices, technological innovations.


Introduction
The impressive growth experienced by organic olive farming in Spain stands out within the general increase in organic food production seen over the last ten years.The area given over to organic olive farming increased from 1,068 ha in 1991 to 91,209 ha in 2003; in 2003 some 13% of the country's total organic production land was occupied by olive orchards (Figure 1).Moreover, olive cultivation is the second most important organic crop in Spain after cereals and leguminous plants1 (MAPA, 1992(MAPA, -2004)), at least as far as cultivated area is concerned.The expansion of
Despite this increase in organic olive production, most Spanish olive groves are still cultivated according to conventional practices.Just 3.6% of olive land in Spain was cultivated according to organic techniques in 2001 (MAPA, 1992(MAPA, -2004;;2000).Indeed, in Andalusia (southern Spain), the foremost olivegrowing region of the world, only 2.2% of olive land was cultivated in this way (MAPA, 1992(MAPA, -2004;;Junta de Andalucía, 1999) despite the region being the most important in the country in terms of the area given over to organic olive production in 2003.Though for now organic production is relatively restricted in Andalusia, the region has great potential for development in the coming years.
Although studies on organic olive farming and olive oil are beginning to become abundant, few have dealt with the process of diffusion/adoption organic farming, even though this type of information is in great demand in the olive sector.Among those available, Lampkin and Padel (1994) analysed the characteristics, motives and barriers to organic farming adoption from an international perspective, while Diebel et al. (1993) analysed potential economic barriers to the adoption of low-input agriculture using mathematical non-linear programming.In Spain, Chinchilla Fernández (1999) used sociological techniques to analyse the role of the different parties involved in the diffusion of organic practices (from the research level to the transfer of technology to farmers) among Andalusian olive growers.The authors' previous work (Parra López and Calatrava Requena, 2002a;2002b), which dealt with the same items but in less depth, is a direct forerunner of the present paper.In addition , Parra López (2003) and Parra López and Calatrava Requena (2004) analysed the diffusion over time of organic farming as a technological innovation in Andalusian olive cultivation in three scenarios according to the yields of different olive orchards.Organic and conventional farming practices -and in some cases their environmental impacts-have also been compared, generally from an interdisciplinary perspective (Ardenclarke and Hodges, 1987;1988;Mawampanga and Debertin, 1996;Berentsen et al., 1998;Clark et al., 1998;Conacher and Conacher, 1998;DEPA, 1999;Sánchez Jiménez, 1999;Stolze et al., 2000;Wells et al., 2000;Hansen et al., 2001;Rigby et al., 2001;Pacini et al., 2003;2004;Parra López et al., 2004;and Rasul and Thapa, 2004).
1995).The latter identifies factors associated with the adoption of innovations, in this case organic agricultural system.Grower characteristics (age, possession of a formal or informal education and training, dedication to agriculture, access to information channels, opinions and attitudes toward organic farming, etc.) and farm characteristics (size, yield, olive varieties, irrigation, slope, etc.) were analysed to see which are related to the adoption of organic practices.It is important to understand that the factors analysed are statistically related to the adoption of organic agriculture, although it is impossible to determine whether they are causes or effects.However, the model constructed attempts to include only those that would appear to be causes (e.g., sources of knowledge concerning organic agriculture), and to avoid those that are clearly effects (e.g., no-use of chemicals).Causality analysis would require analyses over a long period of time, which this is beyond the scope of this paper.

Methodology
Based on the modern conception of the classical theory of the diffusion of innovations (Rogers, 1995), Andalusian olive growers were surveyed between October 2000 and June 2001 regarding the diffusion/ adoption of organic practices.The survey consisted of a structured questionnaire that was completed during face-to-face interviews of organic and conventional olive growers.A total of 322 growers were interviewed, one half of which followed organic cultivation practices, the other half of which followed conventional practices.The sample was stratified by zone (according the number of organic olive growers officially registered in each) according to the data of the Andalusian Committee for Organic Agriculture-Comité Andaluz de Agricultura Ecológica2 (CAAE, 2000).The same number of organic and conventional olive growers were interviewed in each zone to ensure the comparability of responses.
The questionnaire included several groups of questions dealing with different topics: -Farm characteristics (area cultivated, yield, year of plantation, varieties of olives, tenure, tree density, slope of the land, etc.).
-Dedication to agriculture (the importance of agriculture in the interviewee's income, number of years devoted to agriculture, type of work performed, training in agriculture, etc.).
-Degree of contact with sources of information (attendance at courses and conferences on agricultural topics, the amount of reading undertaken regarding agricultural topics, membership of agrarian and nonagrarian organisations, place of residence, trips outside the place of residence, etc.).
-Assessment of other factors from a grower's point of view (economic profits, avoidance of risk, personal prestige, environmental protection, production of healthy products, etc.).
-Steps leading to individual adoption or rejection of organic practices (knowledge of organic olive growing, sources of information through which the existence of organic cultivation was discovered, sources of further information, and the moment when organic practices were either adopted or rejected, etc.).
-Attitude towards organic agriculture as an innovation (image that the interviewee has of these practices, economic advantages offered, time and effort saved, social prestige, etc.).
-Personal information (age, sex, marital status, amount of formal education).
A binomial logit model was used to model growers' behaviour regarding the adoption of organic farming.The dependent variable was the main agricultural system used by the olive growers (organic or conventional) while the independent variables were the analysed factors.Prior to the construction of the model, the dependence between each of the independent variables (answers to the survey questions) and the dependent variable was individually tested (α > 0.05).The most suitable statistical tests were used in each case according to the nature of the answer3 .This bivariate analysis allows the selection of initial variables to be introduced into the model (Table 1).
The binomial logit model included only some of the initial variables identified.These were the factors that, jointly, were significantly associated with the adoption of organic farming.Thus, some of the factors shown to be individually related to the adoption of organic farming in bivariate analysis do not appear in the logit model, although all the factors in the logit model are bivariately related to the agricultural system used by growers.
Since the classic study on the diffusion of innovations by Ryan and Gross (1943) from the standpoint of rural sociology, which gave rise to the so called IOWA model, the most important areas concerning the diffusion of innovations, and the methodology for analysing them, have been defined.However, since the 1970s, some of the assumptions associated with the theory behind the diffusion of innovations have been questioned, leading to adjustments to the classical theory and the development of new approaches to its understanding.The extended-classical theory (Rogers, 1995) was used for the present work, which answers the main criticisms made of the classic model: a) The pro-innovation bias.It is very common to suppose that an innovation is always an improvement for all its potential adoptees (Rogers and Shoemaker, 1971), and that its rejection, abandonment or reinvention is attributable to «irrational» behaviour.In the present paper, it was not assumed that organic cultivation was superior to conventional practises.The following were therefore analysed to avoid bias: (1) the attitude of the interviewee towards organic agriculture (see Table 1), focusing on the subjective perception of growers regarding the advantages that organic cultivation might offer them (e.g., economic advantages, time and effort sparing, social prestige etc.), (2) the concordance of organic farming with the values, beliefs and needs of growers, and the assessment of factors such as care for the environment, the will to raise healthy foods and risk aversion, etc. (as suggested by Cimadevilla, 2000) (see Table 1), (3) the possibility of rejection and abandonment via the inclusion of specific questions in the questionnaire.Finally, a multicriteria comparison was performed of the two systems, including economic, technological, environmental and sociocultural issues; earlier work revealed organic olive cultivation to show an overall better performance than conventional cultivation (Parra López, 2003;Parra López et al., 2004).
b) The individual-blame bias.It is common to focus only on the features of individuals (usually economic or psychological in nature, i.e., intra-individuals variables) without regard for the influence upon them of other individuals (inter-individuals variables) or the Personal data AGE Age of the interviewee characteristics of the system to which they belong (system variables).In this paper a holistic approach was attempted.The authors' previous work shows the great importance of the geographic location of individuals in the year of adoption of organic practices, independent of their personal characteristics (Parra López, 2003).c) Origin of the innovation.Traditional diffusion research centres on certain phases of the diffusion process (adoption and diffusion) and fails to analyse others (the recognition of problems and needs, research, development, marketing and the analysis of the consequences of adoption) (Rogers, 1995).In our previous work we analysed the origin of organic agriculture from the perspective of the Theory of Induced Innovation (Ruttan and Hayami, 1984;1989), and showed that the development of organic practices in olive orchards as a technological innovation arises because of the autonomous development of science and technology and new demands made by society (Parra López, 2003).d) Existence of incentives.It is interesting -but not very common-to study the influence of positive and negative incentives (subsidies and taxes/punishments respectively) on the diffusion of innovations.The adoption rate of organic olive-growing would appear to be closely related to the subsidies available, especially in less productive regions (Parra López, 2003).
e) Consequences of adoption.Since the 1970s, following the disastrous consequences of the rash implantation of Green Revolution principles in countries different to those in which they were defined, studies of innovation diffusion have tried to analyse the foreseeable consequences of adopting innovations.Our earlier work shows organic olive growing to have greater social value than conventional cultivation (Parra López, 2003;Parra López et al., 2004), especially in less productive conditions (greater generation of employment, greater contribution to social justice, and better performance in disadvantaged regions).However, the diffusion of organic agriculture is slowed down by resistance to change.

Results
Before presenting the results obtained with the logit model, a brief review of the answers regarding the most important variables of the questionnaire is helpful.
Table 2 shows that the mean farm productivity (PRODUCTIV) of the organic olive growers to be lower than that of conventional growers, that organic growers tend to be younger (AGE), and that they have spent fewer years dedicated to agriculture (DEDIC_AGR).
With respect to the categorical variables, it is interesting to note that 39.40% of the organic olive growers had contact with local agriculture extension agencies (EXTENS) while the same can be said of only 23.10% of conventional growers (Table 3).Organic olive growers were also more likely to attend courses and conferences on agrarian topics (COURSES) and be members of an association for integrated pest management (ATRIA)4 .Moreover, first contact with organic agriculture via a co-operative or «agricultural transformation society» [ATS] (KN_COOP), the Andalusian Committee of Organic Agriculture (KN_CAAE), the provincial headquarters of the official agriculture services (KN_HEADQ) or some other source (KN_OTHER) was significantly more common among the organic olive growers.
Table 4 shows the logit model.Different groups of survey questions were differentiated in order to discuss the most important results obtained with the model.Tables of relationships were produced for the categorical variables of the model with more than two levels (Tables 5 to 9).In these tables, the numbers of the rows and columns represent the codes for the responses to the survey questions.Reference categories are shown in the rows and the category with which these are compared in the columns.The level of signif icance and the direction of influence of the different values obtained are shown with respect to the reference categories.    1.The levels of the categorical variables are shown in Table 5 to Table 9.

Farm characteristics
Olive farm productivity (yield, kg of olives ha -1 ) (PRODUCTIV) was signif icantly (α > 0.01) and inversely related to the implementation of organic farming techniques by the olive growers (Table 4).High productivity implied an increased likelihood of the grower following conventional methods: the mean productivity of conventional olive growers was 2,126.92kg of olives ha -1 compared to 1,532.05 kg ha -1 for organic growers (Table 2).The reason may lie in the fact that organic olive orchards are more common in areas where the agricultural conditions are not optimal, or in an intrinsic drop in production associated with organic techniques.The latter is more likely since (1) the conventional and organic growers selected were in close proximity to avoid the distortion of results due to spatial reasons, and (2) neither the slope, land quality, orchard age, acreage, plantation density, irrigation techniques, the presence of crops among the trees, the orientation of production towards oil or olives nor the degree of erosion signif icantly differed between the two types of orchard in this survey.

Dedication of the interviewee to agriculture
The importance of agriculture in the income of the interviewees (INCOM_AGR) was very significantly related (α > 0.01) to the agricultural system used.Table 5 shows that partial dependence on agriculture as a source of income (irrespective of whether mainly or secondarily) significantly increases the probability of a farmer following organic methods compared to those who either totally or marginally depend on agriculture.Thus, it was more common for organic olive growers to work part-time in this activity.
The greater number of years dedicated to agriculture (DEDIC_AGR), the lower the probability growers had of following organic methods (Table 4).Organic growers had usually worked for less time in agriculture.This is generally due to their being younger, as suggested by the bivariate analysis performed.The organic olive growers had worked in agriculture for a mean of 22.16 years compared to 26.50 years for conventional growers (Table 2).
With respect to the type of work performed directly on the farm (WORK), interviewees who performed management tasks, whether they also worked in the fields or not, were more likely to be organic growers than those who simply worked in the field (Table 6).The feature that seems to differentiate organic olive growers from their non-organic counterparts is their direct involvement in the administration of the holding, i.e., the rational control of the production process.
As for agricultural training, the most significant difference between organic and non-organic growers lay in the number of courses and conferences attended  etc. (COURSES).Conventional growers attended significantly fewer (α > 0.05, see Table 4).Moreover, 22.37% of the organic olive growers declared having been trained in agricultural subjects by attending courses and conferences compared to only 9.68% of the conventional growers (see Table 3).

Contact with sources of information
With respect to association membership, and taking into account the vast number of different possibilities (e.g., co-operatives, ATS, unions, ATRIA, etc.), only belonging to an ATRIA was related to adopting organic olive production: 41.72% of the organic growers belonged to an ATRIA compared to 16.56% of the conventional growers (see Table 3).

Assessment of different factors from the point of view of growers
The opinion of growers regarding the use of synthetic chemicals and their influence on human health (CHEMICAL) was also related to the agricultural system used.Growers with more extreme opinions (categories 1 and 2 on Table 7) behaved in significantly different ways to those with more moderate opinions.The former were more likely to be organic growers.

The process of diffusion: steps between attaining knowledge of the technique until its adoption or rejection
Having found out about organic practices through the CAAE (KN_CCAE) was the only factor that increased the likelihood of adopting such techniques according to the logit model.Some 11.11% of the present organic growers heard of the existence of organic agriculture through the CAAE, compared to 2.07% of the conventional growers (Table 3).

Interviewee attitude towards organic agriculture
With respect to opinions on economic prof itability (PROFIT), the only signif icant difference seen (according to the logit model) was between those who believed organic orchards to be more prof itable than the conventional type and all other possible opinions (even when growers confessed they had no idea).A clear, positive opinion regarding the prof itability of organic agriculture was far more commonly held by those who practised it (Table 8).
Significant differences were found between the two types of growers with respect to their opinion on the time and effort saved by organic agriculture (TIME_EFF).Organic growers believed that organic methods required more time and effort while conventional growers had the opposite opinion (Table 9).

Personal characteristics of the interviewees
None of the personal characteristics initially taken into account were significantly associated with the adoption of organic growing in the logit model.

Possible ways of potentiating the diffusion of organic olive cultivation
Once the logit model had been constructed, it was used to design specific strategies for potentiating the diffusion of organic olive growing in Andalusia and similar regions.This required two debatable assumptions: (1) that the future behaviour of growers will be the same as that seen until now, and (2) that the specific variables selected actually cause grower behaviour and are not an effect (it has already been noted that the factors analysed could be causes or effects -or perhaps not even causally related to the adoption of organic techniques at all).The first assumption involves linearity in terms of grower response; this is very common in mathematical modelling.The second assumption is acceptable if the predictive variables are carefully chosen.Indeed, selected variables in this paper can be understood (with no conceptual doubt) as causes rather than consequences of grower orientation (conventional or organic).
For analytical purposes, the conventional olive growers were classified into two types: conventional growers with extreme answers (type I), i.e., conventional growers whose answers suggest they are unlikely to become organic growers, and modal conventional olive growers, i.e., those whose characteristics are more common among conventional olive growers.Once characterised, the probabilities of their being a priori organic growers, were calculated (Table 10).
Some of the variables of the model (taken as causes rather than effects of grower behaviour) could be easily managed by public institutions in a hypothetical strategy aimed at encouraging the adoption of organic olive growing, including: 1.The promotion of agricultural training and education through courses, conferences, etc.
2. Support for the development and creation of specialised public or private organisations dealing with the diffusion of environmentally friendly agricultural techniques (such as ATRIA etc.).
3. Support for training in organic farming techniques on the part of specialised organisations (such as the CAAE, etc.).
4. The incorporation of young people into agriculture.

The development of training programs related to agricultural management and business administration.
As a simulation exercise with our model, if variables (1), ( 2) and ( 3) are managed at the same time, conventional growers with less likelihood of following organic techniques (type I) would increase their probability of doing so from 0.10% to 39.92%.The same strategy applied to the modal conventional olive growers (type II) would increase this probability from 29.86% to 99.65%.This highlights the importance of adopting such strategies.

Conclusions
On the basis of the analysis performed, it can be concluded that: -The process of adopting organic techniques in olive tree cultivation is statistically related to a series of identified variables.However, the analysis made does not allow causality relationships to be established.
In this paper, the adoption phase of the diffusion process was focused upon.The authors' earlier work (Parra López, 2003;Parra López et al., 2004) facilitated a more holistic approach to the study of the diffusion of organic olive techniques.
-The adoption of organic techniques is inversely related to the productivity (yield) of olive tree cultivation.In all likelihood this is due to a reduction in productivity inherent to organic farming, particularly in the first steps of the conversion process.
-Part-time labour is more frequent among organic olive growers.-Organic growers have usually spent fewer years dedicated to agriculture than conventional growers.This can be explained, at least partially, in terms of their younger average age.
-Organic growers often spend more time than conventional growers personally tending to farm administration and managerial tasks.
-Organic olive growers attend training courses much more frequently than do traditional growers.
-Membership of an ATRIA (Association for Integrated Pest Management) is much more common among organic growers.
-Opinions regarding the use of synthetic chemicals are, logically, more restrictive among organic growers.
-Having found out about organic farming through the CAAE (Andalusian Committee for Organic Agriculture) was more common among those who became organic growers.
-Organic and conventional growers' opinions on the economic returns and labour requirements of organic farming differ.Organic growers believe that organic cultivation requires more time and effort than conventional growing, but generates greater returns.
-Working on a series of strategic variables can boost the adoption of organic techniques (assuming the linearity and causality of the studied variables).The following steps are suggested: (a) Agricultural training and education should be improved via sponsoring from specialised bodies (such as the ATRIA and the CAAE etc.); (b) Young growers should be incorporated into organic cultivation.Training programs on management and administration could indirectly create greater awareness and recognition of organic farming.

*
The value of continuous variables can not be categorized and its mean has been used to calculate probabilities of both type I and type II farmers.

Table 1 .
Initial variables individually related to the agricultural system used

Table 2 .
Description of the initial continuous variables VC: variation coefficient.

Table 3 .
Percentage of answers to some initial categorical variables CON: conventional.ORG: organic.

Table 5 .
Relationship between the different categories of «Importance of agricultural activity in final income»

Table 6 .
Relationship between the different categories of «type of work carried out directly on the farm» ** α > 0.01.* α > 0.05; ns: not significant.

Table 7 .
Relationship between the different categories of «opinion on the use of synthetic chemical inputs and their influence on human health» ** α > 0.01.* α > 0.05; ns: not significant.

Table 8 .
Relationship between the different categories of «opinion on the economic profitability of olive farming in the interviewee's zone»

Table 9 .
Relationship between the different categories of «opinion on the saving of time and effort in organic agriculture» ** α > 0.01.* α > 0.05; ns: not significant.