Farmers’ perception of their level of participation in extension in Ethiopia: Policy implications

This study examined farmers’ perception of their level of participation in Public Agricultural Extension Service (PAES) in Soddo-zuria Woreda in Southern Ethiopia vis-à-vis seven selected farmers’ characteristics; namely, sex, age, educational status, wealth status, farming experience, experience with extension and frequency of contact with extension agents. For this study, 225 farmers were randomly selected and interviewed with a semi-structured questionnaire. The study showed that farmers in Soddozuria Woreda perceived their level of participation to be low, and had significant correlation with sex, educational status, wealth status and frequency of contact with extension agents. Female, illiterate and poor farmers’ perception of participation in the PAES was found to be lower than their male, literate and resource-rich counterparts. In a regression analysis, sex, educational status and wealth status explained 42.2% of the variance in farmers’ perceived level of participation in the PAES, with educational status alone contributing about 35%. To enhance farmers’ participation in the PAES, the Soddo-zuria Woreda Bureau of Agriculture and Rural Development need to properly mainstream gender, combine pro-poor development strategies and integrate literacy programmes into the routine extension activities.


INTRODUCTION
The importance of farmers" participation in agricultural extension services both as a means to an end and as an end by itself is a widely understood phenomenon among development practitioners.In extension, farmers are key stakeholders at grassroots level.Their participation in all stages of extension programme development and delivery processes enhances efficiency and effectiveness of the planned changes as participation facilitates mutual learning among stakeholders, develops ownership of the change programme, and brings about long lasting and sustainable change both on the farm and in the behaviours of farmers.Owing to this effect, the success of an extension programme largely depends on the roles played by farmers in the programme.Out of several successful outcomes of farmers" participation in development endeavours, the following empirical findings are just some examples to elaborate the issue.
In Ethiopia, Mendesil et al. (2007) and Shamebo and Belhu (1999) reported added values gained in incorporating farmers" indigenous knowledge through their participation in developing effective mechanisms to control sorghum storage pest and in developing sweet potato varieties that address farmers" criteria of valuing, respectively; in India, Hedge (2005) documented farmers" contribution in selecting and developing superior genotypes of crops; in Burkinafaso, Van Asten et al. (2004) were able to identify productive and unproductive spots of farm land with participation of farmerson those *Corresponding author.E-mail: nberhanu2001@yahoo.com.Fax: +251(0)462205421.spots, later on, soil tests were carried out and farmers" assessments were confirmed to be right; in Indonesia, the pioneering work of Integrated Pest Management (IPM) that was basically a result of farmers" cultural practices is of another interesting example [Environmentally and Socially Sustainable Development Network (ESSDN), 1995).Those aforementioned arguments indicate that successful technological development and meaningful change among clienteles require input from the clienteles themselves as from the development practitioners.
In Ethiopia, public agricultural extension service (PAES) has a history of more than sixty years, and is entrusted to bring about meaningful change to farmers.Currently, the PAES has adopted a "participatory demonstration and training extension system" (PADETES) approach to extension.According to Belay and Abebaw (2004:164), a presumed importance of the PADETES approach is "the promotion of active participation of rural communities in problem identification, analysis, planning, implementation and evaluation".However, there is little information on how farmers perceive their level of participation, in which stages of the extension process they participate, and the differences in participation levels among different groups of farmers.This study was, therefore, undertaken to get feedbacks on those issues; and the work was part of a wider scale study conducted to assess effectiveness of the PAES in programme development and delivery.

Objective of the study
The general objective of this study was to understand farmers" perception of their level of participation in the public agricultural extension service to provide feedback for policy directions.The specific objectives were to: (1) Determine male and female farmers" perceived level of participation in the different stages of extension; (2) Test if there was a difference in perceived levels of participation between male and female farmers; (3) Examine relationships between selected characteristics of farmers" and perceived levels of participation; and (4) Determine the best predictors of farmers" perceived level of participation in the PAES.

METHODOLOGY
The research was a survey with a correlation design aimed at describing farmers" perceived level of participation in the PAES and determining relationships between some characteristics of farmers and their level of participation.The study was conducted in Soddozuria Woreda (equivalent to a district) of Wolaita Zone, Southern Ethiopia.The research subjects were selected with a multi-stage selection process.First, 9 Kebeles (lowest administrative unit below a district level) were randomly selected out of 31 Kebeles of the Woreda.In each of the nine Kebele offices, lists of all farmers who had been participating in the PAES, stratified by sex and wealth status, were obtained as a sampling frame.Accordingly, based on the proportions of male and female headed households and their wealth statuses, 225 farmers were included in the study.The instrument used to collect data was a questionnaire with two parts.The first part was on farmers" characteristics listed above and the second part was on farmers" level of participation in the different stages of the PAES.Prior to the development of the questionnaire, a thorough discussion was made with staff members of the Woreda Bureau of Agriculture and Rural Development to identify comprehensive and non-overlapping stages in extension programme development and delivery process.
Accordingly, seven different stages were identified and the questionnaire was developed in line with those stages.For face validity of the questionnaire, the researchers made all the necessary precautions.For content validity, a panel of three experts drawn from Hawassa University, Southern Agricultural Research Institute, and Bureau of Agriculture and Rural Development of the Southern Region of Ethiopia evaluated the questionnaire.The questionnaire was field-tested by taking 30 farmers in Damot-wyde, a neighbouring Woreda where it was believed that the respondents exhibit more or less similar characteristics to those included in the study.The second part of the questionnaire, on farmers" level of participation, was measured on a Likert type scale.To assess the consistency of the items identified to measure the construct "level of participation" a Cronbach"s alpha coefficient of reliability analysis was made and found out to be 0.80.

Characteristics of farmers
The selected characteristics of farmers included sex, age, educational and wealth statuses, farming experience, experience with the PAES and frequency of contact with extension agents.The findings are summarized in Table 1.

Sex composition
The total number of farmers included in the study was 225 with 177 male and 48 female.At a glance, the number of female farmers seems small; that is because of the small size of female headed households in the Woreda.The total households in the Woreda are 27,768, out of which 24,527 are male-headed and 3,241 femaleheaded (Soddo-zuria, 2001).This shows that in the Woreda, the female headed households are only about 11.7%.Thus, quite an adequate number of female farmers (21.3%) were included in the study.

Age structure
The majority of both male (92 or 52%) and female (26 or 54%) farmers fall in the 40 to 50 years age category.Mean ages for male and female farmers were 42.31 and 40.77years.The study did not indicate variations in level of participation in extension as a result of age variations.Belloncle (1989) argues that young farmers better

Educational status
Out of the five categories of educational status, the majority of the male farmers, 62 or 35%, fall in the "grade 5 to 8" followed by other 45 or 25.4% in "unable to read and write" category.Regarding with the female farmers, the majority, 31 or 64.6%, were in the category "unable to read and write".If the data are aggregated, the majority of the farmers fall in the category "unable read and write" because of the highest number of the female farmers in that category.The 35% of the male farmers in grade 5 to 8 is, even, specifically true only for that area in the country.Berhanu (2008), for instance, in a study conducted in North-Central part of Ethiopia indicated much lower numbers of farmers who reach to grade level educational status.Therefore, a high level of illiteracy among farming communities, and particularly in female farmers is a characteristic feature in Ethiopia.Several authors (Apantaku et al., 2003;Belloncle, 1989;FAO and World Bank, 2000;Hegde, 2005;UNESCO, 2005;Weir and Knight, 2000) reported the importance of education among farmers for effectiveness of extension.For instance, Apantaku et al. (2003) argues that low educational status of farmers is a limiting factor to make better use of extension services.In this respect, women farmers in the study area are in a more disadvantageous position.

Wealth status
In Ethiopia, as stated under the method, at present, all farmers have been categorized into poor, medium and better off in terms of their wealth status.In this study (Table 1), the majority of the male farmers fall in the "medium" (82 or 46.3%) and "poor" (77 or 43.5%) categories.On the other hand, a great proportion of the female farmers [30 (62.5%)], fall in the "poor" category.This indicates that female farmers are poorer than male farmers although on aggregation the majority of the farmers in the study area were also poor.Several other authors have also indicated that women farmers are on the average poorer than male farmers (Anandajayasekeram et al., 2008;Belloncle, 1989;Saito and Weidemann, 1990).

Farming experience
Regarding with farming experience, the majority of both the male and female farmers fall in the categories of 11 to 20 and 21 to 30 years.Mean farming experiences were 22.7 for male and 17.8 years for female farmers.This indicates that in the study area, female farmers have shorter farming experiences than the male farmers.

Experience with the PAES
The mean male farmers" experience with the PAES was 10.7 years with standard deviations of 6.9; and that of the female farmers" was 7.6 with standard deviations of 4.7.There were no female farmers who had more than 20 years of experience with the PAES.But, there were male farmers who had experience with the PAES up to 35 years.Both the range and the mean values indicate that female farmers had shorter years of experience with the PAES.This was, in fact, in agreement with their shorter years of farming experience presented earlier.

Frequency of contact with extension agents
The study showed that mean frequencies of contacts with extension agents were 2.50 for male and 2.04 for female farmers.Male farmers" mean frequency of contact was somehow close to the "high" category; whereas, female farmers" frequency of contact was in the "low" category.This shows that male farmers had better frequency of contact with extension agents than the female farmers.Lahai et al. (2000) have found a direct relationship between farmers" frequency of contact with extension agents and their levels of participation in extension.In their view, frequent contact of farmers with extension agents helps them to internalize well the extension education they receive as issues can be clarified whenever the contact occurs.Different authors have also argued that farmers" frequency of contact with extension agents has a direct relationship with effectiveness of extension -the more the frequency of contact of farmers with extension agents the better the effectiveness of the extension service (Aphunu and Otoikhian, 2008;Sarker and Itohara, 2009).

Farmers' perceived level of participation in the PAES
The results showed that the overall mean participation levels of male and female farmers in the PAES were 1.99 and 1.63, respectively.Both these values fall between "low" and "very low" of the scale.An independent t-test, however, showed that the mean participation level difference was statistically significant (p < 0.05) implying that female farmers" participation in the PAES is adversely very low (Table 2).With the exception of "selecting appropriate methods for extension" where percentages of male and female farmers participated in it were almost the same, in the other stages, the proportion of female farmers was lower than that of the male farmers.Regarding with participation levels, male farmers" means were above 2.0 for "need assessment and problem identification", "identifying alternative courses of actions for extension" and "monitoring implemented extension programmes".For the female farmers", a value of above 2.0 was observed only for "need assessment and problem identification"; this was, even, with wide dispersion as a high standard deviation of 0.83 shows.M = male, F = female, M + F = male and female.Scale: 1 = Very low participation, 2 = low participation, 3 = high participation, 4 = very high participation.
The low participation level of farmers in the PAES in this study is, in contrary to what is stipulated by the federal and regional bureaus of Agriculture and Rural Development (GFDRE, 2002;MoARD, 2006;BoARD, 2007) that claim participatory processes in the extension system.This shows existence of mismatches between farmers" own perception of their level of participation and the agricultural development practitioners.

Farmers' characteristics and their participation in the PAES
With the exception of ages of farmers, all the key characteristics of farmers assessed in the study showed a direct relationship with the level of farmers" participation in the PAES (Table 3).Based on Davis (1971) convention, the association between ages of farmers, farming experience and experience with the PAES with farmers" level of participation in the PAES can be described as "negligible".On the other hand, sex of farmer, wealth status and frequency of contact with extension agents had "moderate" associations with farmers" level of participation in the PAES.Educational status had a "substantial" association with the level of participation in the PAES.In general, all the farmers" characteristics that had "moderate" and "substantial" associations with farmers" level of participation were also statistically significant.A stepwise regression analysis was also made in order to determine the best predictors of farmers" level of participation (Table 4).The fitness of the regression model to the data was checked and found fit because the significance value of the F-statistic in the ANOVA was below 0.05 (95% confidence interval).Therefore, the dependent variable, farmers" perceived level of participation (Y) can be explained with the model.Out of the four farmers" characteristics including sex, educational status, wealth status and frequency of contact with extension agents that were significantly correlated with farmers" perceived level of participation, the first three became important predictors of the dependent variable.If we take the adjusted R-square, the three independent variables together explained 42.2% of the variance in the dependent variable.
The contribution of each of the three independent variables can be seen in the equation: Where α is the intercept; β 1 to β 3 are the regression coefficients for the three independent variables; and e is an error term which points out the proportion of the variance in the dependent variable, farmers" perceived level of participation in the PAES, that is not explained by the regression equation.In other words, the error term is the residue obtained when we compute 1-R 2 .This is a measure of the unexplained variance in the dependent variable which may arise as a result of some combinations of the influence of other variables, measurement errors and random chances.If the model is appropriate for the data, the distribution of the residuals will show a normal curve (Bryman and Cramer, 2005).Substituting the values for α, and β-weights we get: Y = 1.098 + 0.145X 1 + 0.171X 2 + 0.1389X 3 + e As we used the standardized β-weights, the relative contribution of the three independent variables to the change on farmers" perceived level of participation can be easily seen from the regression coefficients.With the issue of sex, in the present study, male farmers" participation level was significantly higher than female farmers" participation.So, the explanation is that as the situation becomes more favourable for female farmers", their participation can also increase.Thus, proper mainstreaming of gender into the routine extension activities is by far essential for this change to occur.

Conclusions
In many cases, farmers" participation level in extension is reported to the high side when data are collected and analyzed only from the responses of agricultural professionals.This indicates the need for triangulating source of information by including farmers" own responses for valid conclusions.On matters that have to do directly on farmers" life such as extension, their direct response is particularly important.In this study, as it was found out from farmers" response, their level of participation in the PAES was low, and contrary to the naming "participatory demonstration and training extension system" and the guidelines on participation that the country is presumed to follow.Significant differences were also observed in levels of participation in the PAES owing to sex, educational status, and wealth status of the farmers.Keeping other things constant, female farmers, poor farmers and illiterate farmers had lower levels of participation in the PAES.

Policy implications
Based on the findings, the following recommendations can be made with the aim to enhance farmers" participation.
(i) Mainstream gender properly into the routine extension activities including establishing women extension groups that facilitate peer education and free interaction for exchange of knowledge and information; and recruiting more female extension agents who can better understand women farmers" extension problems and with whom the female farmers would be at ease to communicate and enhance their participation; (ii) Design contextually relevant and pro-poor strategies of development such as, for instance, interventions that enable poor farmers to participate more in the extension activities by using their family labour rather than mere focus on capital intensive extension packages; and (iii) Integrate functional literacy into the extension programmes.

Table 1 .
Some selected characteristics of the respondent farmers.

Experience with the PAES (in year)
participate in extension than older farmers owing to their versatile nature.On the other hand, Nwachukwu(2005)reported low level of participation of young farmers in extension owing to resource limitations.

Table 2 .
Farmers" participation level in the different stages of the PAES disaggregated by sex.

Table 3 .
Relationship between the independent variables and the dependent variable, level of participation in the PAES (N = 225).

Table 4 .
Stepwise multiple regression on farmers" characteristics and their perceived level of participation in the PAES.