Smallholder farmers’ access to improved groundnut production and value addition technologies in Eastern Uganda

Groundnuts are a key crop in Uganda, both as a source of nutrition and income. However, lack of knowledge and information on progressive practices along the groundnut value chain is a key contributor to the poor performance of the sub-sector. This study envisages establishing baseline knowledge on key aspects of groundnut production, processing and marketing with a view to identifying areas and gaps for capacity building interventions. A total of 155 farmers were randomly selected and primary data collected in early 2013 from three districts in Eastern Uganda namely; Bukedea, Mbale and Tororo. Results indicate that farmers were aware of most recommended pre and post harvest technologies/ practices including knowledge on improved groundnut varieties, superior agronomic practices, proper drying and cleaning, long-life storage, grading, sorting and packaging. However, even though most farmers had knowledge of the progressive technologies; quite a substantial number of them did not use these technologies in groundnut production and value addition. This finding points to presence of constraining factors that may hinder access and application of known technologies. Therefore, capacity building efforts to increase both access and utilization of value-enhancing groundnut technologies will eliminate the gap between awareness and use of these technologies for improved livelihoods.


INTRODUCTION
Uganda like many Sub-Saharan African (SSA) countries' economies depends largely on agriculture from which the majority of the rural populations derive their livelihoods and incomes.The agricultural sector in Uganda employs 73% of the population (aged 10 years and above), the majority of whom (over 80%) dwell in rural areas.It contributes over 20% to the national Gross Domestic Product (GDP) and supplies agro-based industries with raw materials (UBOS, 2008;MAAIF, 2010;NEMA, 2010).Due to the importance of the agricultural sector, the Government of Uganda has put renewed emphasis on this sector.This is supported by many policy documents including the Development Strategy and Investment Plan (DSIP) (MAAIF, 2010).However, agriculture remains largely at a subsistence level (Dijkstra, 2001;UBOS and MAAIF, 2011), is characterized by low agricultural productivity, partly due to high land degradation, nutrient mining and soil fertility depletion (Nkonya et al., 2005;Isabirye et al., 2007;NEMA, 2009), limited investment in soil improving technologies, lack of appropriate information and low adoption of available technologies due to inadequate incentives.Soil fertility has continued to decline to levels that are currently prohibitive to profitable agriculture.
Originally fertile lands have been degraded with cereal crop yields of less than 1 ton ha -1 becoming common (Sanchez and Leakey, 1997;Sanchez, 2002).Similar decline in yields have been observed in groundnut production.Okello et al. (2010) reported that while yields of over 2.5 t/ha have been achieved in experimental plots, farm level data show averages of 0.75 t/ha.
Increasing land productivity and being able to contribute to profitable agriculture requires among other things, addressing the major biophysical factors on land through technologies on soil enhancement, prevention of nutrient loss, and conserving available nutrients either individually or in technology combinations.In light of this, the Uganda National Development Plan (NDP) also seeks to facilitate the availability and access to critical production inputs, especially in agriculture and industry, as well as the promotion of science, technology and innovation (UBOS and MAAIF, 2011).
In the case of groundnuts, research and innovation has led to a number of technologies that have been disseminated to farmers.These technologies include: Resistant groundnut varieties such as Serenut 2 and 4, optimal spacing, motorized shelling machines, peanut grinding machines among others.Some of these technologies are low cost and thus expected to be affordable by a majority of small-scale farmers.However, this requires that the intended user, the promoter and the suppliers be aware and knowledgeable about the benefits of each technology.This is because knowledge and awareness is the critical first step in using a technology.Awareness helps to create perceptions, attitudes and beliefs about a technology that lead to an agent's decision to adopt or not adopt the technology (Rogers, 1995).Unfortunately there is limited knowledge on what farmers currently know about the technology, their sources of information about the technologies, the method of information gathering and the factors that determine technology awareness.Knowledge of this is important for a number of reasons: First, it helps researchers to understand the potential usage of the disseminated technologies and this information acts as a check on impact of the researchers' effort.Second, it helps to compare what they know and what they utilize which is important to understand the constraints in employing technologies that they know but do not use.Thirdly, knowing what farmers know about a technology is important in order to design ways of improving knowledge transfer to intended users.It is therefore important to know the level of farmers' awareness of land improvement technologies, the networks that farmers join to access that information and the barriers to technology access and utilization (Duff et al., 1992;Tucker and Napier, 2002).
The major thrust of this paper is to analyze the determinants of awareness of soil fertility improvement technologies in the three districts of Bukedea, Mbale and Tororo.The specific objectives include: Understanding the level of awareness of integrated soil fertility management (ISFM) technologies in the area, identifying challenges faced by groundnut farmers in accessing soil fertility improvement technologies and to determine the factors that drive awareness of ISFM technologies by groundnut farmers.

MATERIALS AND METHODS
The study was carried out as part of a project in view to upscale ISFM technologies in East and Central Africa.In Uganda, purposive selection of survey respondents was conducted using a multi-stage sampling procedure.First, the districts of Bukedea, Tororo and Mbale were selected based on past involvement under Phase 1 of the ISFM project (for Mbale) and also based on the importance of the project enterprises in the districts (Tororo and Bukedea for Groundnuts).The second stage of sampling involved sampling of sub-counties based on their share of production of groundnuts in the district.The third stage of sampling involved random selection of farmers from the selected sub-counties.Lists of farmers obtained from the sub counties were used as sampling frames and a total of 50 farmers were randomly sampled from each district, making a total 150 groundnut farmers in the three districts.
Data collection was by use of direct face-to-face interviews with the aid of questionnaires.The questionnaire captured data on socio-demographic characteristics of the respondents, production data, knowledge and awareness of technologies used in production, post-harvest handling and value addition in groundnuts in the selected districts and sub-counties.Data analysis was performed using SPSS v16 and STATA version 12.0.Descriptive statistics of selected socio-economic characteristics (means, standard deviations, frequencies and percentages) were generated and are presented in Table 1 along with the sampled areas.
The present study uses descriptive statistics (frequencies, coefficient of variation, means and standard deviation) to analyze the farmers' level of awareness of integrated soil fertility management (ISFM) technologies and the challenges faced by groundnut farmers in accessing soil fertility improvement technologies.The ordered logit model is used to analyze the determinants of farmers' awareness of ISFM groundnut technologies using extension staff as their source of information.

The ordered logit model
The Ordered Logit Model is used when the response variable is categorical and has some meaningful ordering.In the present study, farmers were asked to rate the use of extension personnel as their source of information for awareness about ISFM groundnut technologies.Farmers ranked the source as secondary, important and major.In this context; the dependent variable is ordered from a scale of 1 to 3; with y=1 (secondary) being the lowest rating and y=3 (major) being the highest rating.In order to model the use of extension personnel as a function of a set explanatory variable, we used the generalization of the binary-choice framework known as the ordered logit estimation technique (http://fmwww.bc.edu/GStat/docs/StataIntro.pdf).

Consider an index model for a single latent variable
where; (i) The probability that observation will select alternative is: The ordered choice model generalizes the notion of multiple thresholds.In a sense, we cannot observe directly but only the range with which it falls.The observed choice might reveal only an individual's relative preference.
For the ordered logit, is the logistic cumulative density function given as: (iv) The parameters to be estimated are a set of coefficients

EXTENSION
Rating the use of extension personnel as a source of agricultural information (1= secondary, 2=important and 3=Major).

Explanatory variable
Where, the marginal effects of each variable on the different alternatives sum up to zero.The interpretation of the marginal effects is such that each unit increase in the dependent variable increases/decreases the probability of selecting alternative by the marginal effect expressed as a percent relative to the base category.
The reduced form equation that was estimated in this paper is given as: The definition of the variables used in the reduced form equation and their measurement is given in Table 2.

Challenges in using improved soil fertility enhancing technologies by smallholder farmers
All the respondents (100%) reported that the high cost of chemical fertilizer was the major reason inhibiting its usage.Other reasons included lack of knowledge about the technology.About 50% of the respondents believed that fertilizers destroy the land and 48% of the sampled respondents reported that fertilizers were unavailable (Table 3).About 40% of the farmers did not use chemical fertilizers because these particular farmers perceived their farms to be fertile hence without need for application of chemical fertilizers.
The results in Table 3 show that the biggest challenges to adoption of chemical fertilizer included high cost of fertilizer (expensive), lack of technical knowledge regarding use of the technology and unavailability of fertilizers.This implies that if increased use of this technology is to be realized, these three factors must be considered.
As shown in Table 4, farmers were examined for their knowledge on the average yield of crops with and without chemical fertilizer use.The average yields of maize, groundnuts, rice, millet and beans in kilogram per hectare were 1013, 484, 962, 527 and 600 kg/ha respectively without fertilizer use.However, the coefficient of variation for the crop yields in the three districts was high due to a number of reasons.First, yield differences may be due to differences in soil fertility with some farmers realizing better yields because of good soils while others realize low yields due to infertile soils.Another reason is attributed to the differences in the types of crop varieties planted and differences in managerial abilities of the different farmers in the three districts.In addition, a given crop variety may yield differently in different geographical locations.
The results in Table 4 indicate that with application of chemical fertilizers the average yields for maize, groundnuts, millet, banana, cassava, rice and sweet potato were; 1089,1106,1503,6740,1607,1950 and 1393 kg/ha, respectively.Comparing the results of average yields with and without chemical fertilizers, it is observed that that with exception of maize yield, there is higher margin between average yields without chemical fertilizers and those with chemical fertilizers implying that farmers can realize higher crop yields with fertilizer use.The coefficients of variation for average yields with chemical fertilizers are more consistent compared to those without application of chemical fertilizers (Table 4).The finding that farmers realized higher yields with the use of chemical fertilizers compared to non-use is an indication that use of improved technologies enhances agricultural productivity.This implies that efforts taken to address the challenges in the use of improved technologies shall be instrumental in changing the status quo; from subsistent to commercialized agriculture.Such efforts should focus on improving accessibility and affordability of the superior technologies (such as improved seed and fertilizers) and farmers' technical knowledge.

Awareness and use of soil improvement technologies
The results in Table 5 show levels of awareness of soil improvement technologies.In order of decreasing importance, awareness was highest for crop rotation (96%) intercropping (95%) and farm yard manure (90%) whereas seventy two percent of all the farmers were aware of chemical fertilizers and optimal plant populations.Technologies that were least known as soil enhancing technologies included green manure, soil bands, contouring, terracing, agro-forestry and minimum tillage.Terracing and contouring are technologies most applicable to hilly terrain.The terrain in most of the study area (with the exception of Mbale) was relatively flat and hence these two technologies would be least expected to be practiced.
In the overall sample 67% of the farmers had adopted farm yard manure as a soil fertility improvement technology.It is important to note that the 48 farmers that used chemicals applied them on multiple crops.From the findings, the fact that 67% of the farmers applied farm yard manure in agricultural production implies efforts by farmers to intensify agricultural production.In addition, 31% of farmers used fertilizer as an agricultural input in their production, which corresponds to 43% of those who were aware of chemical fertilizers4 .The fact that 43% of those who had heard about chemical fertilizer actually used them is an indication of high adoption potential of ISFM technologies.As depicted in Table 5, about 54% of the farmers were aware about soil fertility improvement technologies using green manure and out of these, 64% had used green manure.This gives a difference of 36% who had not adopted green manure and yet they were aware about it.The reason underlying this behavior may be due to famers' perceptions about green manure as a technology that is bulky to apply hence labor intensive; therefore some farmers may not adopt it.
Close to 96% of the respondents were aware about crop rotation and 94% of all farmers had adopted the technology.Crop rotation plays a major role in conserving soil fertility in that it reduces the depletion of soil nutrients.Additionally, crop rotation is important in the control of crop pests by breaking the cycle of pest and disease-causing organisms thereby minimizing the risk of yield losses.Therefore the high adoption of crop rotation is important for crop production in Eastern Uganda.A number of reasons can be given for its wide adoption.First, farmers in Eastern Uganda (with exception of Mbale) have relatively large acreages which can allow for rotation.Secondly, crop rotation is not a capital intensive technology, that is, farmers do not need very high investment to adopt it.
The results in Table 5 also show that intercropping is widely known and used by farmers.As with crop rotation, intercropping does not require much capital to be used by the farmers.It is an efficient land resource utilization cropping system for the smallholder farmers who lack access to land resources.This is therefore the reason as to why 95 and 89% of the farmers knew about it and used it respectively.About 43% of the farmers were not aware of soil bands while close to 57% of the farmers had knowledge about it.Overall, 56% of the farmers did not adopt (most of whom did not know about the technology) while about 78% of the respondents who knew about the technology were using it.
5 Soil bands were mostly used on maize, beans and sorghum in that order.

Awareness of other technologies
Table 6 shows farmers' awareness about the key practices/activities used in crop enterprise pre and post harvest handling.Results show that the majority of the farmers were aware about the key activities/practices that were asked by the research team.Over 90% of the farmers were aware about value addition, improved planting materials, better agronomic practices, timely harvesting, proper drying/cleaning, proper storage, and grading/sorting; while over 80% of the farmers expressed their awareness about packaging or bagging, promotion and measuring/weighting.Over 50% of the farmers were aware about processing and preservation.On the other hand, the majority of the farmers were unaware about branding, labeling and display/weighting.

Sources of information on ISFM technologies
Farmers gave various sources of information for various technologies.With regard to chemical fertilizer, about 72% of the entire sample of farmers (30% of all responses) reported that extension staff/NAADS were the major source of information on chemical fertilizer while 63, 36 and 34% of the entire sample (or 26, 15 and 14% of all responses) said that fellow farmers, agro-dealers and media were the major sources of information for chemical fertilizer use.NGOs, development partners or research projects were insignificant sources of information on chemical fertilizer usage (Table 7).The finding that extension staff are the major source of farmers' agricultural information implies that interventions that are focused towards strengthening and enhancing the extension system shall be better placed in enhancing farmers access to information on ISFM technologies and ultimately enhance the adoption of superior agricultural technologies.Farmers were told to rank the importance of extension staff as a choice of information.The responses in Table 8 show that about 66% of respondents rank extension as a major source of information while 22 and 12% said that extension staff was an important and secondary source of information respectively.This indicates that agricultural research and extension policies should foster interventions that can enhance the development of an effective and efficient agricultural technology transfer policy for information dissemination to farmers.

Importance of other information sources for farmers in Tororo, Mbale and Bukedea districts
About 48% of the farmers reported that agro-dealers were an important source of information while 27 and 25% ranked agro-dealers as a major and secondary source of information respectively.In addition, a large number of farmers (82 out of the entire sample) ranked fellow farmers as a source of information.Of these, 37% ranked fellow farmers as a major source of information while 40% and 23% of the farmers ranked fellow farmers as an important source and secondary source of information respectively.This implies that the farmer-to-farmer linkage formed a strong platform for information sharing amongst groundnut producers.From such informal platforms, farmers actively gather information from fellow farmers to enhance their knowledge characterized by pooling information or observing the behavior of others and imitating it (Katungi et al., 2008).
Therefore, policies should target and encourage informal sources as important sources of agricultural information for smallholder farmers.Ordered Logit Regression results (Table 9) show the determinants of extension as an important source of agricultural information by farmers.From these findings, sex of respondent is positively (coeff.= 2.37 P > |z| = 0.011) related to choosing extension as a choice of agricultural information source.Male farmers are more inclined to seek agricultural information from extension agents.This results has been arrived at by various researchers in different settings (Mbaga-Semgalawe and Folmer, 2000;Bayard et al., 2007;Tiwari et al., 2008;Mugisha et al., 2012).
In addition, the level of education was positively related to choosing extension agents while the farmer's experience was negatively related to choosing extension.This is because educated individuals tend to use formal channels of information access because they believe that the formal channels provide the right kind of information hence the reason as to why education is positively related to choosing extension.
This result is in line with findings of others like Sidibe (2005) and Mugonola et al. (2013).On the other hand, experience is negatively related to choice of extension as a major information source because experienced individuals tend to know much about their farming activities hence may not require or seek additional information from any formal sources.
Use of improved planting materials is negatively related to choice of extension (coeff.= -4.41P > |z| = 0.01).This may be attributed to the fact that most farmers access improved planting materials through farmer-to-farmer exchange rather than through extension agents therefore use of improved planting materials is likely to lead to choice of extension as a less important source of information by the farmers.
These results show that those who sought the services of extension service did so for purposes of obtaining other advice but not planting materials.

Model fitness
The test of whether the variables showing preferences for extension staff should be in the model (Table 9) showed that the included variables were significant at the 10% level (chi 2 (8) = 13.49,Prob > chi 2 = 0.0962) implying that without the included variables in the model, the categories of the dependent variable would not be significant.Marginal effects for the 3 outcome categories of the dependent variable presented in Table 9 are a test of the effect of a change of dummy variable from 0 to 1 for each category.The marginal effects in the two endgroups (Import_ext_staff =1 and Import_ext_staff =2) are highly significant.
By contrast, only two of the marginal effects in the last group are significant at the 10% significance level.This implies that, when a particular variable changes (say, age increases by 1 unit), the number of people moving from group 1 to group 2 is not more or less the same as the number of people moving out of group 2 into group 3, so that the probability of being in group 3 may significantly change.

Conclusions
The study used baseline survey data collected in the districts of Bukedea, Mbale and Tororo.Descriptive statistics were generated using SPSS v16 and detailed data analysis was performed using STATA v-12.This paper discusses the study results using percentages, means, frequencies and standard deviations presented in Tables 3 to 7. Results revealed that the biggest challenges include un-affordability, lack of technical knowledge regarding use of the technology and unavailability of fertilizers.This implies that if increased use of chemical fertilizers in groundnut production is to be realized, these three factors will have to be addressed.
In order of decreasing importance, awareness was highest for crop rotation (96%), intercropping (95%) and farm yard manure (90%).Seventy two percent of all the farmers were aware of chemical fertilizers and optimal plant population.Technologies that were least known as soil enhancing technologies included green manure, soil bands, contouring, terracing, agro-forestry and minimum tillage.For all ISFM technologies, there was a disparity between awareness and utilization implying that not all the farmers who knew about a given technology applied it.This suggests a potential for the project's interventions in increasing utilization of ISFM technologies.
Farmers had access to varied sources of information.Many farmers obtained farming advice from extension personnel and from fellow farmers.The least responses were for research projects, NGOs, the media and agrodealers as important information sources.The fact that extension service providers were the biggest contributors to information on chemical fertilizers for the sample implies that interventions in the area can leverage on this already existing establishment if they are to reach a wider coverage of farmers.In addition, the finding that farmers are a significant source of information for their fellow Mugonola et al. 255 farmers suggests that to improve the quality of the shared information, training of these farmers should be conducted so as to create a hub of (technical) information to support other farmers within a farmers' group.Another reason could be that formal sources of information (such as extension agents, NGOs and agro-dealers) may not be accessible such that farmers rely more on their peers than technical service providers.This latter case requires that efforts be put in place to make technical agricultural services more accessible.
corresponding to the explanatory factors in , as well as a set of threshold values corresponding to the alternatives.In the interpretation of the estimators in ordered logit, actual values of the response variable (in this case; rating the use of extension personnel) are not relevant.Larger values are taken to correspond to higher outcomes.If there are possible outcomes, a set of threshold coefficients or cut points { is defined where and .The ordered logit model with alternatives will have one set of coefficients with intercepts.You can recognize an ordered choice model by the multiple intercepts.In the interpretation of the coefficients, the sign the respondent spent in school EXP Number of years the farmer has spent accessing agricultural information from the service providers STAPLEINCOME Annual income received from staple foods CHEMICAL Duration of training on chemical usage in days IPM Whether farmers use IPM techniques (1=yes, 0=otherwise) HHSIZE Household size (total number of people living in the household by the time of the survey exercise) of the parameters shows whether the latent variable increases with the regressor.The ordered logit model with alternatives will have sets of marginal effects.The marginal effect of an increase in a regressor on the probability of selecting alternative is:

Table 1 .
Selected Socio-economic characteristics of respondents in Taroro, Mbale and Bukedea.

Table 2 .
Definition of variables used to fit the ordered logit model.

Table 3 .
Possible reasons hindering wide usage of Chemical Fertilizers in Tororo, Mbale and Bukedea districts.

Table 4 .
Average yield obtained in kg/ha with and without chemical fertilizer of various crops.

Table 5 .
Farmers' awareness and use of ISFM technologies in Eastern Uganda.

Table 6 .
Awareness of improved technologies by farmers in Tororo, Mbale and Bukedea districts (n=155).

Table 7 .
Sources of information on chemical fertilizer in Tororo, Mbale and Bukedea districts.

Table 8 .
Importance of extension staff as a source of information for farmers in Tororo, Mbale and Bukedea districts.

Table 9 .
Determinants of the use of extension staff as a source of information for ISFM technologies in Tororo, Mbale and Bukedea districts.

dx for extension as a major source of information (outcome 3)
**, ** and * 1, 5 and 10% level of significance, Figures in parentheses are standard errors. *