Traits preferences heterogeneity and biofortified cassava varietal adoption among farming households in Southwest, Nigeria

Abstract

Promoting improved crop varietal adoption will prevent a waste of scarce resources deployed to breeding programs for sustained agricultural development in Nigeria. Using household-level survey data, this study aimed to identify farmers’ trait preferences and their influence on the adoption of cassava varieties in Southwest, Nigeria. Specifically, it profiled the preferences based on the perspectives of the cassava farming households and determined the factors that influenced the rate and intensity of adoption. The results are in two folds: First, for reasons bordering on the realization of cassava farming households’ sustenance and economic goals, early maturing and bigger root size were given preferences among others. Second, education of household head, farm size, extension contacts, and market-related traits were the major drivers of adoption and intensity of adoption of biofortified cassava varieties. Thus, poor market demand was key to the low adoption of biofortified cassava varieties. We suggest that crop development efforts should be tailored to address farmers’ concerns. This will undoubtedly increase the adoption rate and also lower the advocacy cost.

Introduction

For most people living in the lowland and sub-humid tropical regions of West and Central Africa, cassava is the primary source of dietary energy^1,2. Cassava is a food security crop. Cultivated in the tropical regions, it is regarded as the most significant crop in terms of both production value and weight, yielding around 200 million metric tons yearly. With an annual production of about 59.4 million tons, Nigeria is currently the world’s top producer of cassava^3,4. However, smallholder farmers, who are mostly found in the country’s central and southern regions, control over 90% of their production^5,6,7. Cassava is now a cash and industrial crop, having transitioned from a smallholder subsistence crop over time. Therefore, the development of novel cassava applications and markets highlights the numerous policy and research endeavors in cassava varietal improvement, production, and processing.

Previous studies on cassava development gave greater significance to productivity and less importance to nutritional value. However, studies like⁸ found that the majority of Nigerian diets are made up of staple food crops, which are often low in nutritional content. Cassava is the most widely consumed of these staple food crops⁹. Thus, vitamin A deficiency affects more than 30% of Nigerian children under five, 13% of mothers, and 20% of pregnant women^10,11,12. This illness causes several issues, including eye strain, diminished physical ability, increased medical expenses, and decreased economic production, as well as an increased risk and severity of infection.

The federal government of Nigeria has been working for almost 20 years to address the issue of vitamin A deficiency in the diet, which is estimated to cost the country’s GDP over US$1.5 billion^13,14. This work has included fortifying staple foods like sugar, wheat flour, vegetable oils, and salt with vitamin A and supplementing the population’s diet on a regular basis or through vaccinations, especially for children. Nevertheless, it appears that these intervention programs were ineffective because the proportion of children under five who lack vitamin A continues to range from 29.5% to 48%^15,16. Therefore, a different, more potent approach must be taken in order to reach the poor and rural households, which make up around 70% of the population. Thus, biofortification, the development and delivery of staple crop types with high pro-vitamin A content, was examined by^17,18,19 as a reliable substitute for fortification, supplementation, and diet diversification. As the most widely grown and consumed crop in all societal strata, cassava holds a special place in people’s diets. As such, it seems to be the most suitable and affordable vehicle crop for distributing the necessary vitamins^20,21. This resulted in the International Institute of Tropical Agriculture (IITA) and the National Roots Crops Research Institute (NRCRI) funding research projects through Harvestplus that developed the biofortified cassava varieties²². However, conventional breeding programs are still typically conceptualized by scientists rather than being solely end-user-oriented in many research resources. The main drawback of these efforts to improve crops is that the qualities that consumers and farmers desire are not taken into account when evaluating new varieties. Thus, the farmers’ low adoption rates, dis-adoption, and non-adoption of some enhanced crop varieties are indications of the outcome^23,24,25. Thus, employing the model village idea for cultivation, processing, and consumption, biofortified cassava stems were distributed free of cost to around 100,000 households in roughly 260 communities in the states of Akwa Ibom, Imo, Oyo, and Benue in 2011^26,27,28.

This study addresses a critical gap in the literature by integrating farmer preferences, socioeconomic factors, and gender dynamics into the analysis of technology adoption, specifically for biofortified cassava, a crop enhanced with micronutrients like provitamin A to combat malnutrition. First, the emphasis on heterogeneity in trait preferences is a relatively novel angle. While previous studies^{26,27,28,29,30,31} on cassava adoption often focus on broad determinants such as yield, disease resistance, or access to inputs, this research examines how diverse preferences for specific cassava traits (e.g., root size, cooking quality, maturity period, or nutritional content) influence adoption decisions. This approach acknowledges that farmers are not a monolithic group and that individual or household-specific preferences can significantly shape technology uptake. This is innovative because it moves beyond the “one-size-fits-all” assumptions often embedded in earlier adoption models. Second, biofortified crops represent a newer frontier in agricultural technology, blending agronomic improvements with nutritional outcomes. Unlike traditional improved varieties focused solely on productivity, biofortified cassava (e.g., vitamin A-enriched varieties) targets both food security and public health. The novelty here lies in analyzing adoption through the lens of a technology that serves dual purposes, which introduces additional layers of complexity, such as consumer acceptance of biofortified products (e.g., yellow cassava) and market dynamics, beyond what is typically considered in studies of conventional crop varieties. Third, focusing on Southwest Nigeria, the study contributes to a geographically nuanced understanding of adoption patterns. Nigeria, the world’s largest cassava producer, exhibits significant regional variation in farming practices, gender roles, and market access. Southwest Nigeria, with its proximity to urban markets like Lagos and distinct cultural preferences (e.g., for white gari over yellow), provides a unique setting to explore how these factors interplay with biofortified cassava adoption, potentially differing from findings in other regions like the Southeast or North Central zones. This research’s contributions lie in the fact that earlier studies, such as²⁵ identified factors like access to improved cuttings, education, and extension services as key drivers of adoption. While foundational, these studies often treat adoption as a binary outcome (adopt or not) without delving into the heterogeneity of farmer preferences. The current research builds on this by unpacking why certain traits matter to different households, offering a more granular understanding of adoption dynamics. In addition, studies like^9,25 on biofortified cassava dissemination emphasized consumer willingness to pay and nutritional benefits. However, these often focus on post-adoption acceptance rather than pre-adoption trait preferences among farmers. The current research bridges this gap by linking farmer preferences to adoption, potentially revealing how agronomic and nutritional traits jointly influence uptake, a contribution that complements consumer-centric studies. Furthermore³², showed how gender-responsive breeding increased adoption by aligning varieties with end-user needs. While insightful, it spans multiple regions and lacks the Southwest-specific focus of this study. The current research contributes by offering a deeper dive into Southwest Nigeria, where market proximity and cultural preferences (e.g., for white gari) might moderate biofortified cassava adoption differently than in Benue or Imo states. Lastly, this study employs advanced methods like double hurdle model³⁰, which could quantify preference heterogeneity more robustly than traditional regression models used in earlier works³³. This methodological rigor enhances the precision of its findings, setting it apart from prior studies that relied on simpler descriptive or econometric approaches. The results of this study on farmers’ preferences and choices of cassava varieties will therefore guide breeders and researchers in their breeding efforts and provide farmers with better-performing varieties in their localities.

Conceptual framework

The study is anchored on a two-stage conceptual framework that explains both the decision to adopt biofortified cassava and the intensity of its adoption. Drawing on³⁴ double-hurdle model, the framework recognises that the choice to adopt is distinct from the extent of adoption, allowing different factors to influence each stage. Figure 1 of the study illustrates three interrelated blocks of determinants. Household and demographic characteristics including age, education, household size, farming experience, farm size and off-farm income, are hypothesised to shape both the probability of adoption and the proportion of land allocated to biofortified varieties. Prior studies have consistently shown that better-educated farmers and those with larger or more diversified households adopt improved technologies more readily^26,35. Institutional and market factors constitute the second block. Access to extension services, membership of farmer organisations, credit availability, and market signals such as price premiums and consumer demand are expected to facilitate adoption and to raise the scale of planting^36,37. For example, regular contact with extension agents can reduce information asymmetry about the nutritional and agronomic benefits of yellow cassava⁹, while participation in farmer groups lowers transaction costs and improves access to planting material.

The third block captures the technology’s trait attributes, which are central to the novelty of this study. Farmers’ preferences for specific varietal traits like planting-material availability, yield potential, maturity period, stem height, resistance to pests and diseases, dry-matter and starch content, colour of gari, and pounding quality are explicitly incorporated as drivers of adoption^25,38. Biofortified cassava adoption is therefore conceived not merely as a response to socioeconomic endowments but as a function of how well varietal characteristics align with farmer and consumer priorities. For instance, high dry-matter content and desirable gari colour are known to command market premiums that feed back positively into adoption decisions⁹. The framework also allows for reinforcing feedback loops: successful market performance and consumer acceptance of yellow gari increase farmers’ incentives to expand the area planted, which in turn raises visibility and stimulates further demand.

This conceptualisation guides the empirical strategy in two important ways. First, the separation of the adoption decision from adoption intensity is operationalised through a probit model for the first hurdle and a truncated regression for the second, consistent with³⁴. Second, inclusion of both socio-institutional variables and detailed trait-level indicators ensures that the econometric model captures the multidimensional nature of technology uptake. Heterogeneity across locations such as differences between Oyo and Ogun States in market proximity, prior awareness campaigns, and consumer preferences is naturally accommodated within this structure. By combining socioeconomic, institutional and trait-specific drivers within a double-hurdle framework, the study advances understanding of how biofortified cassava can achieve scale in Nigeria’s food systems and provides a basis for designing policies and extension strategies that simultaneously enhance nutritional outcomes and farmer livelihoods^9,36.

Fig. 1

Source: Author modification.

Conceptual framework.

Methods

Study area

The study was conducted in Oyo and Ogun States (Fig. 2) in the Southwestern zone of Nigeria. Agriculture and allied activities constitute the main occupation of the people of this zone, and this serves the primary role of providing food, shelter, employment and industrial raw materials. The vegetation and climatic condition of the zone encourages the cultivation of crops like rice, yam, maize, and cassava, among other arable crops, and was purposively chosen because of its potential and present position among cassava-producing zones in the country. The zone consists of Oyo, Ogun, Osun, Lagos, Ondo, and Ekiti States. According to⁷, of the total national aggregate of cassava production within the Southwestern zone of Nigeria, Oyo and Ogun States recorded the highest cassava production figures^7,25. Some of the cassava varieties cultivated in the zone include odongbo, oko-iyawo, otherwise called toko-taya in some places, IITA, Texaco, Mojaporoofo, isunikankoniyan, idileruwa, and biofortified cassava varieties among others. Oyo State is located between latitudes 2⁰ 38¹ and 4⁰ 35¹ east of the Greenwich meridian. It is bounded in the north by Kwara State, in the east by Osun State, in the south by Ogun State and in the west partly by Ogun State and the Republic of Benin. The state has a population figure of 5,580,894 million consisting of 50.2 per cent male³⁹ and lies on a lowland, predominantly agrarian, and covers an approximate area of 28,454 km². The state has two distinct agro-ecological zones- the rain forest to the south and the derived savannah to the north and divided into four agricultural zones namely, Ibadan/Ibarapa, Oyo, Ogbomoso and Saki zones. The climate is equatorial, notably with dry and wet seasons and relatively high humidity. The wet season spans from April through October while the dry season lasts from November to March. The mean daily temperature throughout the year ranges between 25 °C and 35 °C. The climatic condition of the state favours the production of crops like maize, yam, millet, rice, plantain and cassava. Other crops that thrive in the state are cashew, citrus, tobacco and fruits. The state has two notable research institutes (IITA and IAR&T) that have national mandates to develop the production of cassava.

Ogun State is bounded in the North by Oyo State, Lagos State in the south, Ondo State in the East and the Republic of Benin in the West. The state lies within latitude 6⁰ 20’ and 7⁰ 58^’ N and longitude 2⁰ 40^’ to 4⁰ 35’ E of the Greenwich Meridian. It has a total land area of about 16,409.26km² with an estimated population of 3.8 million comprising 49.7 per cent male. The state has two dominant agro-ecologies- savannah and rain forest with four agricultural zones namely, Abeokuta, Ilaro, Ikenne and Ijebu-Ode. The major arable crops cultivated include rice, maize, cassava, fruits and vegetables while tree crops grown include oil palm, rubber, cocoa and kola-nut.

Fig. 2

Source: Authors modification using MATLAB software.

Map of Nigeria showing the southwest.

Sample selection procedure

The study population consists of cassava producing households in the study area. The sampling design was such that all households in the villages that cultivated and processed cassava within the last 12 months before the survey had equal chance of being selected. A multi-stage sampling procedure was employed to select respondents for the study (Table 1). The first stage involved a random selection of three agricultural zones (AZs) in each state to allow for spread and adequate representation. In the second stage, two local government areas (LGAs) that are known for the cultivation of biofortified cassava varieties were purposively selected from each of the three zones. In the third stage, purposive sampling technique was employed to select three villages that are known for high production of biofortified cassava and an equal number of villages that are not known for biofortified cassava production. In the last stage, stratified sampling technique was used to select nine cassava farming households from each village, to give a total of 648 household heads. The sample size of 648 households is justified as representative of farming households in Southwest Nigeria for studying biofortified cassava adoption and trait preference heterogeneity. It correlates with the statistical minimum for a 95% confidence level and 5% margin of error, accounts for population heterogeneity and potential clustering effects and aligns with the study’s analytical demands and regional scope. Compared to similar studies, it provides sufficient power and coverage, ensuring reliable inferences about the diverse factors influencing adoption in this context.

Primary data were employed for this study. The data were collected with the aid of a pre-tested structured questionnaire (supplementary files). Information elicited from the households focused on their socioeconomic characteristics, improved cassava varieties introduced to them and their preferences, and outstanding traits of their preferences. The final attributes in the questionnaire presented to the farmers were based on the initial pre-survey carried out by the researchers and supported by literature. Data were also collected on the rate and intensity of adoption of biofortified cassava varieties as well as on farmers’ income. All methods were carried out in accordance with relevant guidelines and regulations. The study received formal approval from the postgraduate committee of the Department of Agricultural Economics, Obafemi Awolowo University, Ile-Ife with approval no-AGP15/16/h/0856. Informed consent was obtained from all respondents.

The study employed various analytical techniques to address the stated objectives of the study. Descriptive statistics was used to describe and rank respondents’ cassava varietal and trait preferences as well as adoption decisions of biofortified cassava varieties. The double hurdle model was used to determine the factors that influence the rate and intensity of adoption of biofortified cassava varieties. Descriptive statistics involving frequencies, percentages, means and standard deviations were used to describe and classify different cassava traits based on perceived relative importance attached by the farmers. This relativity of trait’s importance ranged from “very important”, “important”, “just important” to “not important”. Trait performance of biofortified cassava varieties was then assessed respectively vis-à-vis the relativity of importance by the adopters as “very good”, “good”, “fair” and “poor”. Trait performance was considered to be satisfactory when the responses were “very good” and or “good”; minimum when it was “fair” and unsatisfactory when the response was “poor”.

Descriptive statistics was also used to describe the socioeconomic characteristics of the farmers such as age, sex, education, farming experience, access to credit and extension services. The t-test statistics was employed to compare the socioeconomic characteristics of adopters and non-adopters of biofortified cassava varieties in the study area. The test compares the actual difference between two means in relation to the variation in the data expressed as the standard deviation of the difference between the means⁴⁰.

Table 1 Sampling procedure and sample size.

Econometric procedure

The double-hurdle model

The double-hurdle model was used to analyse the determinants of both the rate and intensity of adoption of biofortified cassava varieties. The double-hurdle model is motivated for this topic because it effectively captures the two-stage nature of biofortified cassava adoption, deciding to adopt and determining the extent, while handling zero-heavy data and preference heterogeneity among Southwest Nigerian farming households. Its flexibility, precedent in agricultural adoption research, and ability to address distinct decision drivers (e.g., awareness vs. resources) makes it superior to simpler models like Tobit or probit alone. By providing nuanced insights into both adoption barriers and intensity factors, it supports the study’s aim to understand trait preferences and inform biofortification strategies.

This model was chosen because it assumes that farmers make two sequential but functionally independent decisions with regard to adoption of technologies such as biofortified cassava varieties. Each hurdle is conditioned by the farmer’s socio-economic characteristics, biotic and abiotic factors, and traits of biofortified cassava varieties. In estimating the double-hurdle model, a probit regression (using all observations) is followed by a truncated regression on the non-zero observations [35]. The double-hurdle model is designed to analyse instances of an event which may not take place and also, if it takes place. In the case of a farmer’s adoption of biofortified cassava varieties, a decision on adopting each improved variety is made first, and then the decision on the intensity of use follows. A different latent variable is used to model each decision process, with a probit model to determine the probability that a household will adopt a variant of biofortified cassava varieties and a truncated regression model to determine the intensity of its adoption. Thus, the model is specified in two separate parts as.

$$y_{{i1}}^{*} = w_{i}^{/} \alpha + v_{i} \left( {{\text{decision to adopt biofortified cassava variety}}} \right)$$

(1)

$$y_{{i2}}^{*} = x_{i}^{/} \beta + \mu _{i} \left( {{\text{intensity of use for}}\;y_{{i2}}^{*} } \right)$$

$${\text{Truncated regression model}}:\;y_{i}^{/} \beta + \mu _{i} \;{\text{if}}\;y_{{i1}}^{*} \;{\text{> }}\;{\text{0 and }}\;y_{{i2}}^{*} {\text{> }}\;{\text{0}}$$

(2)

where

$y_{i1}^{*}$ = latent variable describing the farmer’s decision to adopt biofortified cassava varieties. $y_{i2}^{*}$ = latent variable describing the intensity of adoption. $y_{i}^{{}}$ = actual percentage of cultivated land under biofortified cassava varieties. w_i = vectors of variables explaining the adoption. $x_{i}$ = vectors of variables explaining the intensity. α and β are coefficients of parameters to be estimated. µ_i and v_i = error terms assumed to be independent and normally distributed as $v{}_{i}$ ~N(0,1); and µ_i ~ N(0,δ²).

The expected value of the latent variable is

$Ey_{i}^{*} = x_{i}^{/} \beta$.

Observation i is only made if $y_{i}^{*}$ is above a certain known threshold, which is zero in this case. The probability density function (pdf) of the observed truncated variable $y_{i}^{{}}$ is therefore, the pdf of the latent variable conditional on it being observed.

$$f(y_{i} |x_{i} ) = f(y_{i}^{ * } |y_{i}^{ * }> 0,x_{i} ) = \frac{{f(y_{i}^{ * } |x_{i} )}}{{p(y_{i}^{ * }> 0|x_{i} )}} = \frac{{\sigma^{ - 1} \varphi (\frac{{y_{i} - x_{i}^{{}} \beta }}{\sigma })}}{{1 - \Phi (\frac{{0 - x_{i}^{{}} \beta }}{\sigma })}} = \frac{{1\varphi (\frac{{x_{i}^{{}} \beta - y_{i} }}{\sigma })}}{{\sigma \Phi (\frac{{x_{i}^{{}} \beta }}{\sigma })}}$$

(3)

When estimating the truncated regression model, the simple linear regression of the observed variable $y_{i}$ on $x_{i}$ would yield biased estimates of beta, as the error term would be correlated with the independent variables (Heteroskedasticity). According to^41,42, when heteroscedasticity and a non-normal error structure are allowed, the truncated regression model will be estimated using maximum likelihood of the form:

$$L(\alpha ,\beta ,\hbar ,\theta ) = \Pi_{o} \left[ {1 - \varphi (w_{i} \alpha )\varphi (\tfrac{{x_{i}^{/} \beta }}{{\delta_{i} }})} \right]\tfrac{{}}{{}}X\tfrac{{}}{{}}\Pi_{1} \left[ {(1 + \delta^{2} y_{i}^{2} )^{{ - {\raise0.5ex\hbox{$\scriptstyle 1$} \kern-0.1em/\kern-0.15em \lower0.25ex\hbox{$\scriptstyle 2$}}}} \varphi (w_{i}^{/} \alpha )\delta_{i}^{ - 1} \phi (\tfrac{{T(\theta y_{i} ) - x_{i}^{/} \beta }}{{\delta_{i} }})} \right]$$

(4)

To assess the impact of the regressors on the extent of adoption, it is necessary to analyse the marginal effects of the selected variables. According to Jensen and Yen (1996), and Yen and Jones (1997), the intensity of adoption conditional on adoption is computed as

$$E(y_{i} |y_{i}> 0) = \varphi \left( {\frac{{x_{i}^{/} \beta }}{{\delta_{i} }}} \right)^{ - 1} \frac{{}}{{}}\int_{o}^{\infty } {\left( {\frac{{y_{i} }}{{\sqrt[\delta ]{{1 + \delta^{2} y_{i}^{2} }}}}\Phi \left( {\frac{{T(\theta y_{i} ) - x_{i}^{/} \beta }}{{\delta_{i} }}} \right)} \right)\partial y_{i} }$$

(5)

Where Ø and ф, respectively are the probability density function and the cumulative density function of the standard normally distributed random variable.

For this study, at the first stage the dependent variable $y_{1}$ = adoption of biofortified cassava variety (1 if adopted, 0 if otherwise), while at the second stage the dependent variable $y_{i2}^{*}$ = actual percentage of cultivated land under biofortified cassava varieties and the independent variables are defined as: $x_{1}$ = Age of household head (years); $x_{2}$ = Education of the household head (years); $x_{3}$ = Farming experience (years); $x_{4}$ = Household size (number); $x_{5}$ = Farm size (ha); $x_{6}$ = Off-farm income (N); $x_{7}$ = Access to credit (1 if yes. 0 if otherwise); $x_{8}$ = Frequency of extension service contacts (♯); $x_{9}$ = Planting material availability (1 if adequate, 0 if not); $x_{10}$ = Yield trait (categorical); $x_{11} =$ Stem height (categorical); $x_{1`2}$ = Pest resistant trait (categorical); $x_{13}$ = Starch content (categorical); $x_{14}$ = Market demand (categorical); $x_{15}$ = Product price (categorical); $x_{16}$ = Product colour (categorical); e = error term; βs are coefficients of parameters to be estimated.

Results and discussion

Descriptive characteristics of cassava farming households

This section presents the background of cassava farming households in the study area. A household, usually represented by the household head, consists of one (or more) people who live in the same dwelling and share meals. The socioeconomic characteristics of the households include age, sex, household size, education, farm size and farming experience. Also reported are other variables such as access to extension services and membership of cooperative societies.

Age effect on improved agricultural technology adoption is indeterminate. While some studies revealed positive influence, some revealed negative effects. The age distribution of the respondents in the study area is shown in Table 2 where the mean age of the non-adopters of biofortified cassava varieties was 49.11 ± 13 and 44 ± 11.32 for the adopters. The results suggest that an average cassava farmer in the study area was in the active age bracket that is physically active and fit to carry on the production of cassava. The test of difference however shows that there is a significant difference (p < 0.05) in the mean age of the adopters and non-adopters. The results show that the mean age of the non-adopters of biofortified cassava varieties was higher than the mean age of the adopters. Further analysis of the household’s different age brackets revealed that the mean number of female that were less than 5 years of age among the non-adopters was 2 ± 1.24 while the mean number of their male counterparts was 2 ± 0.86. While the test of difference suggests that there was no significant difference between the age bracket categories of the non-adopters and adopters of biofortified cassava varieties, the result revealed the preponderance of individuals under 5 years of age among the respondents. Furthermore, the sex of an individual can influence the type and quality of work carried out by the individual. The frequency distribution of household heads based on sex is presented in Table 2. According to the table, about 77 per cent of adopters of biofortified cassava varieties were males. This result was also similar for the non-adopters as majority (85.3%) of them were also males. For all samples, the results showed that majority (81.94%) of the respondents were male. This implies that cassava production in the area was male dominant.

Education is an important determinant of adoption decisions as well as an item of human capital development^43,44. To a large extent, it indicates the ability to read and write, which will in turn afford the respondents the opportunity to carefully analyse the advantages and disadvantages of various farming options and decisions. The number of years of formal education determines the literary level of the household heads. It avails rural farmers the opportunity to access and assess agricultural innovations. In addition, it affords farmers the opportunity to clearly weigh and compare the advantages and disadvantages of various innovations or technology introduced in order to make a rational decision for adoption. The distribution of the years of education of the household heads in the study area is presented in Table 2. The results showed that the mean years of education of household heads that are non-adopters of biofortified cassava varieties was 3 ± 0.811 while 9 ± 4.33 was for the adopters. The t-test result shows a significant difference (p ≤ 0.05) in the mean year of education of household heads between the adopters and non-adopters of biofortified cassava varieties.

It has been established that land is an economic resource that has alternative uses. Land is considered inseparable to the people of the study area, being a traditional source of identity, social and political relationships, sustenance and wealth-creating. Land is the most valuable form of property in agrarian societies because of its economic, political and rural importance^45,46. Land availability has been seen as a means of survival for vast majority of rural populations including the study area. According to⁴⁷, the amount of land available to a household may determine the amount of agricultural production, and hence, its economic welfare. The amount of the resource available to the farmer and the number of enterprises of interest to him will therefore determine the amount allotted to various uses. Land therefore remains the limiting factor of production in the study area because it constitutes the largest share of agricultural inputs. As shown in Table 2, the study revealed that the average farm size among the non-adopters of biofortified cassava varieties was 2.7 ± 0.84 hectares while it was 2.5 ± 0.55 hectares for the adopters. This implies that the farm size in the study areas is small. This could be traced to the land tenure factors associated with land holding in the southwest Nigeria rural areas which encourage farm land fragmentation. Table 2 also showed the results of ownership of land among the adopters and non-adopters of biofortified cassava varieties. According to the table, about 89 per cent of non-adopters indicated that they owned the land cultivated while 76.89 per cent of the adopters indicated that they owned the land cultivated. This implies that majority of the cassava farmers in the study area owned the land cultivated.

In addition, as shown in Table 2, the mean years of farming experience among the adopters of biofortified cassava varieties was 20.31 ± 10.41 years while the mean year of experience among the adopters was 14.26 ± 9.9 years. This implies that most of the cassava producers started farming at a young age. The test of difference results showed that the mean years of experience between the non-adopters and adopters was significant (p ≤ 0.05). Extension visits make available to farmers, information about new technologies. The greater the frequency of contacts of farmers with extension agents, the better informed and knowledgeable they become, at least in their areas of concern⁴⁴. Table 2 showed the extension access information of households in the study area. According to the table, 48.4 per cent of non-adopters indicated access to extension services while 54.7 per cent of the adopters indicated access to extension services. The low rate of access to extension services among the non-adopters might be responsible for the non-adoption of biofortified cassava varieties in the study area.

Table 2 also showed the cooperative membership of respondents in the study area. Cooperative is a self-help approach where individuals pool money together from where they can access loan and other inputs at affordable rates. According to the results, about 88% of non-adopters indicated that they were members of cooperative society while 76% of the adopters indicated membership of cooperative society. The higher proportion of non-adopters who were members of the cooperative society did not translate to technology adoption in the study area. The results further revealed that higher proportion (93.14%) of adopters of biofortified cassava varieties belong to other social groups. While the findings suggest that access to cooperative loan by cassava producers in the study area may not be a determinant of adoption of improved cassava varieties, adoption of biofortified cassava varieties could have necessitated the formation of other social or interest groups on the one hand or be traced to interactions afforded by membership of these existing social groups on the other hand. About 38.29% of the respondents adopted biofortified cassava while 61.74% did not which indicates a low level of adoption of biofortified cassava in the study region.

Table 2 Socioeconomic characteristics of cassava farming households.

Traits preferences of cassava varieties

Ceteris paribus, cassava varietal choice of two farmers in the same location may be significantly differentiated by their traits preferences and goals of production. The production, processing and marketing traits preferences for the respondents are shown in Table 3. The production traits include resistance to drought, pests and diseases, yield, maturity and stem height; the processing traits are pounding, dry matter content, starch and fibre content; while marketing traits are market demand, good quality of gari, texture, price command, root size and colour. The results are presented for the entire sample of respondents and according to State. The field performance of a cassava variety is consequentially and directly related to how well the variety can perform relative to drought, pests and disease resistance abilities. Results from the pooled estimates showed that 69.75 per cent, 77.47 per cent and 76.08 per cent of the respondents respectively indicated that drought resistance, pest resistance and disease resistance traits are very important in their decision to adopt any cassava variety. On the average, about three-quarters (74.43%) of the respondents considered these three traits as “very important” cassava characteristics, 20.73 per cent and 4.84 per cent of the respondents considered the traits as “important” and “just important” respectively. The results across the States revealed that 74.90 per cent and 73.97 per cent of the respondents in Oyo and Ogun States respectively indicated that drought resistance is vital in their decision to adopt cassava variety.

The adoption of a specific cassava variety is a derived adoption. Cassava varieties are adopted based on their perceived abilities to timely meet the farmer’s needs. Hence, maturity and yield traits (root size and number) are two crucial production traits that play important role in the achievement of farmer’s production objectives. These objectives often have to do with the expected higher returns, household sustenance and livelihood improvement^48,49. Maturity trait explains how long a variety of cassava takes to mature for harvest. Expectedly therefore, farmers tend to be more favourably disposed to cultivating cassava varieties that mature early. This is because of the burden to keep up with family’s food and financial demands rests mainly on the farm enterprise. About 34 per cent and 67 per cent of the respondents expressed respectively that early maturing and yield traits are very important in their adoption decision of cassava varieties. While 37.04 per cent and 15.43 per cent of the respondents respectively indicated that early maturing traits as “important” and “just important”, 25.77 per cent and 7.25 per cent of the farmers expressed yield traits as “important” and “just important” respectively. Taking early maturing and yield traits as complements, half (50.31%) of the respondents considered the traits as very important. While about one-third (31.40%) of the respondents indicated that the traits are “important”, 11.34 per cent of the farmers expressed that the traits are “just important”. Thus, regardless of the intrinsic benefits of a new cassava variety, this finding implies that cassava varieties with perceived long maturity period may suffer adoption setback or dis-adoption when early maturing alternatives are available. The results however showed that yield traits are more important to farmers than maturity traits as all the respondents attached varying weights of importance to the former and 13.89 per cent expressed that the latter trait is not important. This finding is in support of²⁵ who reported that yield traits of new varieties are part of the major determining factors that farmers in the Southwestern Nigeria consider in their adoption decisions. Stem height trait is valued differently by the respondents. While some prefer tall stem height because it allows weeding and other field operations relatively unhindered, some others prefer short stem height for the presumed ability to grow more branches with canopy that hinders weed growth and, majority attached no importance to the trait. Unlike the other production traits therefore, about two-thirds (68.67%) of the respondents attached no importance to stem height trait of cassava variety in their adoption decisions of cassava variety.

The results also gave a clear picture of the respondents’ valuation of the various processing traits of what they considered a good cassava variety. Table 3 also shows that less 16.24 per cent of the respondents did not agree to poundability of cassava root as a trait preference. This suggest that cassava is not sought primarily for its pounding trait as there exists another staple crop in the study area, yam, which is considered a preferred substitute with better pounding characteristics. However, about 21 per cent of the respondents claimed that pounding trait plays very important role in their adoption decision of cassava variety.

The quantity of products, usually gari and lafun, from cassava roots during or after processing is a preferred trait to cassava farming households. The quantity of products derivable from a given quantity of roots is directly proportional to the dry matter content and inversely proportional to the water content of the roots^50,51. The trait probably becomes more prominent among others because of its direct impact on the volume of output that will be available to the market. Expectedly therefore, high dry matter content trait is one of the major features in cassava variety to which the respondents attached considerable weights. As shown in Tables 3 and 57.25 per cent of the respondents considered high dry matter content as a very important characteristic to look out for in cassava varietal selection. While no single respondent considered dry matter content of no importance, the rest (22.38% and 20.37%) of the respondents attached varying degrees of weights to the dry matter contents. The results suggest that other economic uses of cassava are not being considered or exploited. The results further shows that dry matter content trait of cassava is the most important processing trait sought after by cassava farmers and would have positive influence on adoption of new varieties.

The starch content trait of cassava is not primarily sought for starch production by farmers but for the binding role it plays in gaari and lafun production. While the pooled result shows that about 26 per cent of the respondents would consider high starch content trait of cassava variety in their adoption decisions, 32.56 per cent claimed that the fibre content trait is very important in cassava varietal adoption decision. Across the States, respondents in Ogun State attached greater importance to starch content than their counterparts in Oyo State as 35.80 per cent of the former against 21.91 per cent of the latter claimed the trait is very important in their varietal adoption decision.

The likelihood of the producer to achieve his set production goals and objectives depend partly on the ability of his produce to favourably compete with other similar produce in the market. Table 3 shows how the marketing traits of the products from the roots of a cassava variety would influence respondents’ adoption decisions. The pooled result shows that about 46 per cent and 44 per cent of the respondents claimed that the ability of the products derived from roots of a cassava variety to attract market demand is “very important” and “important” respectively, and would influence their adoption decision. Similarly, consideration for cassava varieties whose products command good price showed that 58.95 per cent, 30.40 per cent and 10.65 per cent of the respondents claimed that good price command trait of cassava products is “very important”, “important” and “just important” respectively. This finding supports the submission of³⁸ who posited that the ability of products from a cassava variety to command good price in the market will positively influence varietal adoption decisions of farmers.

The pooled result shows that 88.58 per cent of the respondents attached importance to vitamin trait to cassava variety. This result further shows that about two-third (67.07%) of the 37.96 per cent of the respondents that claimed vitamin trait will play a “very important” role in their cassava varietal adoption decisions were from Oyo State. The revealed high preference for vitamin trait by respondents from Oyo State might not be unconnected to the effect of series of advocacies and enlightenment campaigns that had taken place in the State.

Colour of cassava products is often determined by the experience of the processor on the one hand and the level of technology employed in processing on the other hand. Across Oyo and Ogun States, 20.06 per cent and 35.35 per cent of the respondents respectively claimed that the trait is very important in their cassava adoption decision. This result suggests higher colour sensitivity among the respondents in Ogun State than in Oyo State. Thus, colour trait may play a very important role in their adoption decision. The pooled result however shows that about 29 per cent of the respondents claimed that colour of cassava products has a very important role to play in varietal adoption decision.

In the same way that different products derived from the roots are offered for sale so are the roots offered in the emerging produce markets to independent commercial processors. Apart from the weight factor, the size of the roots tend to affect the price they attract because root size determines the relative ease of peeling and time required for processing. The result therefore shows that while about 51 per cent and 21 per cent of the respondents respectively attached biggest and least weights to root size trait, no respondent indicated preparedness to cultivate varieties of cassava with smaller roots. The information elicited on root size trait therefore affirmed the departure from sustenance as the main production goal of cassava farmers and affirmed the findings of^24,25,51,52 who reported the preference of processors for cassava varieties with bigger roots for their manageability and higher yield.

Table 3 Cassava traits preferences for respondents.

Biofortified cassava varieties traits performance valuation

Crop’s variety trait performance assessment by farmer, to a large extent would influence the rate and intensity of its adoption. Farmer’s assessment of a crop is usually idiosyncratic and is often reflexively done over a period of time, relative to another variant of the same crop that is considered the best or the most preferred by the farmer. The “very good” and “good” valuation indexes adopted in this study are considered to be favourable and satisfactory assessments of the varieties by the farmers. Therefore, juxtaposing these valuations with the respondents’ earlier stated trait preferences above, which was carried out in the “absence” of biofortified varieties under examination, Table 4 shows the adopters of biofortified cassava varieties’ assessment of the various production, processing and marketing traits.

For the drought resistance component of production traits, the result shows that 33.16 per cent and 44.60 per cent of the adopters in Oyo and Ogun State respectively claimed that biofortified cassava varieties performance was “very good”. The pooled result thus shows that 80.50 per cent of the adopters claimed that biofortified cassava varieties’ drought resistance performance is satisfactory. While 29.70 per cent and 30.69 per cent of the respondents in Oyo State claimed that biofortified cassava varieties’ pest and disease resistance traits were “very good”, 42.67 per cent and 40.00 per cent respectively of their Ogun State counterparts gave the same assessment. Adopters’ assessment of biofortified cassava varieties in terms of how long it takes the varieties to mature can be said to be above average. This is revealed as 71.11 per cent of the adopters gave favourable valuation. The same proportion (71.11%) of the adopters of biofortified cassava varieties also expressed of the yield trait as satisfactory. The result however shows across the States that a higher proportion of the adopters in Ogun State (42.67%, 40.00% and 34.67%) as against 29.70%, 30.69% and 23.76% in Oyo State respectively claimed that pest, disease and maturity traits of biofortified cassava varieties are “very good”. The much observed variance expressed in the result across the States might be due to the disproportionate difference between the ratios of adopters to non-adopters in the two States. The pooled result shows that 92.06 per cent of the adopters’ assessment of biofortified cassava varieties’ stem height was satisfactory.

The processing traits of cassava include those that affect the quality and quantity of products derivable (i.e. dry matter content) and texture (i.e. fibre and starch content) of the root. They also include other intrinsic features that determine the final economic returns. The result (Table 4) shows that less than half (48.38%) of the adopters in the study area claimed that the pounding trait is satisfactory. This finding lends credence to the challenge of low pounding trait expressed varieties by⁵³. This may negatively impact on future adoption campaigns when the observed trait deficiencies are addressed in the future releases of variants of biofortified cassava varieties. While in Oyo State, 9.41 per cent and 18.32 per cent of the adopters claimed that the varieties’ dry matter content trait is “very good” and “good” respectively, 8.00 per cent and 20 per cent of their counterparts in Ogun State accorded the varieties dry matter content trait “very good” and “good” valuations respectively. The pooled result shows that less than one-third (27.8%) of the respondents claimed that biofortified cassava varieties dry matter trait is satisfactory. This result affirmed the findings of⁵³ that biofortified cassava varieties have low dry matter content. The favourable respective “very good” and “good” assessments of 9.03 per cent and 18.77 per cent of the adopters may then have been influenced by the information on the nutritional benefits of the cassava varieties available to them during the adoption process. This result further portends a negative implication on the adoption advocacy and efforts for future releases of improved biofortified cassava varieties in the study area.

In addition, vis-à-vis the adopters’ expectation of starch trait earlier expressed in Table 3, the result (Table 4) shows that 13.86 per cent and 32.18 per cent of the adopters in Oyo State and 18.67 per cent and 24.00 per cent of adopters in Ogun State claimed that biofortified cassava varieties starch trait is “very good” and “good” respectively. The pooled result thus shows that more than half (54.88%) of the adopters claimed that the starch trait of the varieties falls below their expectation as 35.02 per cent and 19.86 per cent assessed the trait as “fair” and “poor” respectively. In addition, this result suggests that the low adoption level and the dwindling intensity of use of the varieties in the study area might be due to their unsatisfactory trait expression. Furthermore, 29.60 per cent and 37.91 per cent of the adopters as shown in the pooled result claimed that the fibre trait of biofortified cassava varieties is “very good” and “good” respectively.

The pooled result shows that about 36 per cent and 46 per cent of the adopters gave “very good” and “good” assessments of the products of biofortified cassava roots respectively. However, in spite of this, the result reveals a generally low market demand for the adjudged good products as 15.84 per cent and 23.27 per cent of adopters in Oyo State and 14.67 per cent and 22.67 per cent of their counterparts in Ogun State respectively claimed that biofortified cassava products had “very good” and “good” market demand. The low market demand for biofortified cassava products may be attributed to lack of information to the buyers about the nutritional benefits of the products. The study however considers the seemingly skewed positive valuation of the products by the farmers on the other hand to the gamut of information available to them on the products.

In addition, the result reveals across the States that vitamin A trait is the driving factor of adoption for the varieties. This is as 43.56 per cent and 50.99 per cent of the adopters in Oyo State and 54.67 per cent and 38.67 per cent of the adopters in Ogun State respectively gave “very good” and “good” valuations for the trait. Even though they had no objective or scientific means of verifying their claim or belief, the study reasoned that the highly skewed adopters’ assessment of the trait was informed by the trust and belief in the information about the nutritional benefits of vitamin A that was made available to them during their adoption decision construct. The study reasons that adopters’ belief in the nutritional benefits of biofortified cassava varieties might have been strengthened by the source(s) from where they gathered their information. This result however affirmed the findings of⁵⁴ that adequate knowledge significantly affects adoption. The study also reasoned that the low adoption level of biofortified cassava varieties in the study area might be due to the unsatisfactory expression of some other “important” traits which to the respondents have more weight than the nutritional benefits. The result also shows that adopters’ opinion about colour trait of the products is divided along the States. This is revealed as 29.21 per cent of adopters in Oyo State and 9.33 per cent of their counterparts in Ogun State claimed that the colour of the biofortified cassava products is “very good”.

One very important parameter is the price that a product commands in the market. While it is established that biofortified cassava roots certainly do not yield as much products from an equal volume as some other improved cassava white roots do, it is expected that products from biofortified cassava roots will command higher prices in the market. Thus, the expected price differentials between the two classes of products are expected to buffer and compensate for the effects of low productivity of biofortified cassava roots at the processing phase. Though the result affirms that the products are good judging from the adopters’ perspectives, yet, the market demand was poor. The finding about the market performance of the products is in line with⁵¹ but contradicts the findings of⁹ that consumers are willing to buy and eat biofortified cassava products. Furthermore, the pooled result reveals that just about 15 per cent and 23 per cent of the adopters attested that the products performance in the market is “very good” and “good” respectively while 20.58 per cent and 26.71 per cent claimed that the products commanded “very good” and “good” prices. This finding corroborates³⁸ who found that the ability of products from a crop variety to command good price motivated farmers’ varietal selection. However, the unenthusiastic and unimpressive demand reported of biofortified products may be due to lack of sufficient relevant information to the majority of cassava products buyers in the market.

The pooled result shows that about half (53%) of the adopters were satisfied with the tuber size trait of the varieties as 22.47 per cent and 29.24 per cent assessed the trait as “very good” and “good” remarks. The findings are in agreement with²⁸ that production of biofortified cassava varieties is good as opposed to the processing and marketing of the products therefrom. However, the result contradicts that of⁹ who reported good market demand and processing-related attributes for biofortified cassava variety. Though, physical yield may seem not have been significantly compromised on farmers’ fields as declared by HarvestPlus⁹, the low dry matter content and high water content of the roots of biofortified cassava were reasons for the relatively low yield at the processing phase of production.

Table 4 Traits valuation of biofortified cassava varieties.

Determinants of biofortified cassava varieties adoption

The result in Table 5 shows that the log likelihood values between the dependent and independent variables were − 211.81, -111.23 and − 189.68 in Oyo State, Ogun State and the Southwestern Nigeria respectively. These values support the fitness of the model. The entire model was significant at 1 per cent level of probability. The marginal effect was also computed to show the magnitude of a unit change that independent variables had on the response variable.

The estimated probit model of the pooled result reveals that age, years of education of household head, farm size, off-farm income, extension visits, yield traits and dry matter content traits were significant and are more likely to influence the probability of adoption of biofortified cassava varieties. Also included among the variables that are likely to influence the probability of adoption were starch content trait, market demand for products and product price. There is disagreement on the direction of the effect on adoption of agricultural technology⁵⁵, however, the ability of a farmer to bear risk and be innovative has been reported to decrease with age. The pooled result shows an inverse relationship between age and adoption of biofortified cassava varieties in the study area. The result shows that an increase in the age of the household head by one year would reduce the probability of adoption of biofortified cassava varieties by 3.60 per cent. This implies that as the farmer grows in age, the probability of adoption of biofortified cassava decreases. The pooled result also shows that an increase in the year of education of the household head by a unit would raise the probability of adoption by 11 per cent. This implies that the more educated a household head is, the greater the likelihood that the household would adopt biofortified cassava variety. This result was in line with the findings of^25,51 who found that education and technology adoption are directly related. An increase in the size of farm cultivated by the household head by 1 hectare would raise the probability of adoption of biofortified cassava varieties by 22.2 per cent. This implies that as more land area is cultivated, the greater is the likelihood that the household would adopt biofortified cassava varieties. The positive sign is however contrary to the findings of⁵², that farm size and technology adoption are inversely related.

The result in Oyo State shows that an increase in farming experience of the household head by one year would raise the probability of adoption of biofortified cassava varieties by about 58 per cent. This result validates the higher adoption level in the State and implies that cassava farmers in the State are less risk averse than their counterparts in Ogun State. The result in Oyo State also shows that an increase in the volume of credit accessed by the household head would raise the probability of adoption by 21 per cent. This corroborates the findings of^56,57 who found a significant positive relationship between access to credit and adoption of agricultural technology. The result further reveals that there could have been more incentives and supports from the Institutes and other promoters of biofortified cassava varieties in the State which were not visibly present in Ogun State of the study area.

In addition, the pooled result shows that an increase in income from sources other than the farm by 1 would raise the probability of adoption by 10.10 per cent. This implies that as more and more income is made from sources other than farm, the greater will be the probability of a household head adopting biofortified cassava varieties. The positive sign observed is however not in agreement with²⁵, who reported that off-farm income and technology adoption are inversely related. This result therefore shows that off farm income might have provided farmers with liquid capital for timely purchase of productivity enhancing inputs such as planting materials and processing equipment.

Access to extension service was also positive and significant (p < 0.05) across the study area. This shows that an increase in the number of extension visit by a unit would raise the probability of adoption by 31 per cent. This implies that the higher the number of extension visit, the higher is the probability that household heads in the area would adopt biofortified cassava varieties. This is in agreement with the findings of^25,58.

The study shows that adoption of biofortified cassava was influenced not only by the socioeconomic characteristics of the farmers but also by some variety-specific traits. These traits include the yield, dry matter content, starch, market and product price. The result shows that an improvement in the yield trait of biofortified cassava varieties would increase the probability of adoption by 10.10 per cent. The result also reveals that improvement in the dry matter content and starch traits would increase the probability of adoption by 11.10 per cent and 2.10 per cent respectively. These imply that the higher the assessment or the valuation of these traits by the farmer, the higher is the probability that he would adopt biofortified cassava varieties. In addition, an increase in the product price by 1 would raise the probability of adoption of biofortified cassava varieties by 4.10 per cent. This implies that the higher the unit price of the products, the higher is the probability that farmers in the study area would adopt biofortified cassava varieties.

The results of probit analysis revealed that age has a negative influence on the probability of adoption of biofortified cassava varieties. The marginal effect of 0.031 implies that a unit increase in age of cassava farmers would decrease the probability of adoption of biofortified cassava by 3.1 per cent. This implies that the older a household is, the greater the likelihood that he would not adopt biofortified cassava.

Determinants of intensity of adoption of biofortified cassava varieties

The second component of double hurdle model (truncated regression model) was employed to determine the factors that influenced the intensity of adoption of biofortified cassava varieties. The pooled result as revealed in Table 6 shows that the log likelihood was − 194.661. The diagnostics parameters support the fitness of the model and the entire model was significant at 1.0 per cent. The marginal effect was also computed to determine the magnitude that a unit change in the independent variables had on the dependent variable.

Years of education of farmers was hypothesized to have positive influence on the intensity of adoption of technology. However, while the result shows that education was positive and significant (p < 0.05) for adopters of biofortified cassava in Ogun State, it was not for their counterparts in Oyo State. This result is in agreement with⁵⁹ who reported a positive relationship between education and intensity of adoption of agricultural technology. In addition, in agreement with the apriori expectations, extension contacts and credit were also positive and significant (p < 0.05) for adopters in Ogun State. The positive influence of education, credit and extension contacts on the intensity of adoption of biofortified cassava in Ogun State could thus be traced to the fact that biofortified cassava varieties were relatively new in the State, and farmers’ adoption decisions were still being guided by factors other than their on-the-farm experiences with the biofortified cassava varieties. The result further shows that product colour was positive and significant (p < 0.05) in Oyo State. This result is in agreement with⁹ who reported that the yellow gari made from biofortified cassava variety was preferred in Oyo State where hitherto white gari is consumed.

The pooled result however shows that household size, farm size, off-farm income, extension contacts, pest resistant trait, market demand and product price were positive and significantly (p < 0.05) influenced the intensity of adoption of biofortified cassava varieties. Household size creates demand which in turn motivates increases in the intensity of adoption of new practices or technologies that would increase the farmers’ means of meeting household food and other economic demands. The result shows that increase in the household size by 1 member would raise the intensity of adopting biofortified cassava varieties by 1.9 per cent. This outcome is in line with the expectation as the sign could either be positive or negative³³.

The positive sign of farm size coefficient indicates that the larger the farm size, the more the likelihood of increase in the intensity of adoption of biofortified cassava varieties. The results therefore shows that an increase in the farm size by 1 hectare would raise the intensity of adoption by 3.1 per cent. These studies agrees with^25,52 that increase in farm size could stimulate and enhance better adoption of improved technologies.

The coefficient of off-farm income was positive and signicant (p < 0.05). The result shows that an increase in the off-farm income by 1 would raise the intensity of adoption of biofortified cassava varieties by 0.9 per cent. This suggests that an additional off farm income is channeled to biofortified cassava production. This implies that off-farm income and intensity of adoption of biofortified cassava varieties are directly related. This result is in agreement with^35,36,37 who established a positive direct relationship between off-farm income and agricultural technology adoption.

Extension contact was hypothesized to influence the adoption decision positively as it obviously facilitates the uptake of new technologies. The coefficient of extension contact was also positive and significant (p < 0.05). This implies that an increase in increase in the extension contacts by a unit would increase the intensity of adoption of biofortified cassava varieties by 13.3 per cent. This finding is in line with^36,55 who found that respondents who are frequently visited by extension agents have higher possibilities of adoption of new technology.

Similarly, the coefficients of product price and market demand for biofortified cassava products were positive and significant (p < 0.05). The result shows that an increase in the product price by 1 would raise the intensity of adoption by 10.3 per cent The result also shows that a unit increase in the market demand for biofortified cassava products would raise the intensity of adopting biofortified cassava by 8.10 per cent. This finding corroborates⁵⁹ who found that biofortified cassava products command premium prices in the market.

Table 5 Probit analysis of the determinants of adoption of biofortified cassava varieties.

Table 6 Truncated regression analysis of the intensity of biofortified cassava variety adoption.

Conclusions

This study aimed at identifying farmers’ trait preferences and their influence on the adoption of cassava varieties in Oyo and Ogun States, Nigeria. Specifically, it profiled the preferences based on the perspectives of the cassava farming households, determined the factors that influenced the rate and intensity of adoption. The conclusions are in two folds: First, for reasons bordering on the realization of cassava farming households’ sustenance and economic goals, early maturing and bigger root size were given preferences. Second, education of household head, farm size, extension contacts and market related traits were the major drivers of adoption and intensity of adoption of biofortified cassava varieties. Thus, poor market demand (low patronage for products) was key for the low adoption of biofortified cassava varieties. Farmers in the study area are smallholders who had few years of formal education which, perhaps accounted for the proportion of the respondents that adopted biofortified cassava varieties.

Policy recommendations

Promoting improved crop varietal adoption will prevent a waste of scarce resources deployed to breeding programmes for a sustained agricultural development in Nigeria. Based on the conclusions from the study, crop development efforts should be tailored to address farmers’ concerns. This will undoubtedly increase the adoption rate and also lower the advocacy cost. In addition, since farming is business and that market will continue to play a major role in influencing adoption decisions of rural farming households greater advocacy/awareness campaigns about specific interventions should be considered. This will create the needed market for the products and drive the farmers desire to increase the rate and extent of use of biofortified cassava varieties.

Policies on cassava value chain should be strengthened in order to stimulate and sustain cassava production in Nigeria and to encourage industrial uses of cassava root. This will assuredly in turn make cassava a competitive agricultural cash crop and enhance adoption of biofortified cassava varieties.

Limitations of the study

The three variants of biofortified cassava released in 2011 were UMUCASS 36, UMUCASS 37, and UMUCASS 38 commonly known in the research parlance as IITA-TMS-IBA011368, IITA-TMS-IBA1371 and IITA-TMS-IBA011412. These three varieties were taken as one because the respondents could not say which of them they cultivated and the researcher also could not afford the service of an agronomist for proper varietal identification that could allow for respondents’ disaggregation along the line of the variant(s) cultivated. The responses given therefore were taken as representative of all the variants. In addition, the study suffers from methodological shortcomings. First, the use of non-random sampling may limit the representativeness of the findings and introduce selection bias, thereby affecting the generalizability of the results. Second, the non-exhaustive attribute selection in the design phase means that some potentially important factors influencing preferences may have been omitted, which could narrow the explanatory power of the analysis. Lastly, the reliance on a non-experimental design to elicit preferences restricts the ability to infer causal relationships, as it cannot fully control for confounding factors or unobserved heterogeneity in respondents’ choices.

Suggestions for future research

Future research could address these limitations by employing probability-based random sampling techniques to enhance representativeness and reduce sampling bias. Future studies are also encouraged to expand the set of attributes and levels included in preference elicitation tasks through qualitative pre-studies, expert consultation, or participatory methods to ensure a more comprehensive capture of relevant factors. Additionally, adopting experimental or quasi-experimental designs, such as randomized controlled trials or choice experiments, could strengthen causal inference and provide deeper insights into the determinants of preferences. Furthermore, this study investigated the determinants of adoption of biofortified cassava varieties. Further studies could therefore be carried out to investigate the extent to which the adoption of biofortified cassava varieties has helped reduce vitamin A deficiencies in the population, particularly, among the rural farming households in Nigeria.