Farmers’ Preferred Genotype Traits and Socio-Economic Factors Inﬂuencing the Adoption of Improved Cowpea Varieties in South-Central Niger

: Cowpea, Vigna unguiculata (L.) (Walpers, 1842), is an important legume for food and nutrition security, and income generation. Despite decade-long efforts to disseminate improved varieties, cowpea productivity remains low in Niger. This is due, in part, to the limited adoption of improved cowpea varieties among farmers. Increasing the adoption of improved cowpea varieties requires a better understanding of farmers’ preferred genotype traits and socio-economic factors that inﬂuence their decision. We interviewed 634 farmers from the south-central regions (Maradi and Zinder) of Niger to assess factors that inﬂuence their decision to adopt improved cowpea varieties. The average age of the respondent was 41 years with 29 years of farming experience. Eight improved cowpea varieties were grown by these farmers with average adoption rates ranging from 3.3 to 38.0%. Genotype traits that inﬂuenced farmers’ decision to adopt improved cowpea varieties included early maturing (86.9%), high yielding (73.9%), and high market value (50.5%). Socio-economic factors that signiﬁcantly inﬂuenced adoption were age, gender, membership in a farmers’ organization, and contact with the extension services. Adoption is constrained by the limited availability of cowpea varieties with farmers’ preferred genotypes traits. Farmers’ preferences for genotype traits must be considered in the early stages of breeding programs and the release of new varieties to increase adoption. Stakeholders involved in the cowpea value chain can use this information to improve cowpea adoption and productivity.


Introduction
Cowpea, Vigna unguiculata (L.) is the most important legume grown in the semi-arid tropics and one of the world's major legumes consumed in West Africa [1,2].It is adapted to the Sahelian climate which is characterized by low rainfall and poor soil fertility [3].Worldwide, Niger is the second largest cowpea producer after Nigeria [4].In terms of area and production, cowpea is the second most important crop after millet in Niger [5].Cowpea plays a key role in food and nutrition security, animal feed (fodder), and a cash crop for smallholder farmers [6,7].
Cowpea production suffers from important abiotic and biotic constraints including pests, diseases, and the effects of climate change [8,9].Breeding programs have developed cowpea varieties that are tolerant or resistant to these constraints [2,10].Several cowpea varieties have been released to farmers in Niger including those developed for pests, diseases, and/or drought tolerance or resistance [11][12][13].Many of these varieties were released to farmers through large-scale extension activities by various projects [11,14].
Despite crop improvement and dissemination of new varieties, cowpea productivity remains low in Niger.Poor cowpea productivity is due, in part, to the limited use of im-proved varieties among farmers [15,16].Adoption of improved seed can be constrained by several factors including seed types (technologies), the lack of awareness of or information on the availability of seed, or risk mitigation by farmers [17].Smallholder farmers may be hesitant to invest in improved seed because of risks associated with crop failure (e.g., drought, pests, or diseases).Understanding smallholder farmers' behavior particularly their perceived preferences for genotype traits may help increase the adoption of improved cowpea varieties.
Prior studies on the adoption of improved cowpea varieties in Niger have focused on (i) the effect of socio-economic factors on the adoption of improved cowpea or millet production technologies (e.g., improved genotypes, biopesticides, biological control) disseminated under different extension approaches (i.e., farmers' field schools or demonstration plots) in the Maradi and/or Zinder regions [13,18], and (iii) modeling farmers' decision to adopt and increase the use of improved cowpea varieties in the Zinder and Tillabery regions [19].These studies looked at the adoption of improved cowpea varieties in general (without a specific focus on a particular genotype) or a limited number of varieties, and focused on project areas/effects of extension approaches that influence the adoption of improved cowpea varieties.There is a need to assess whether farmers' preferred genotype traits and socio-economic variables may influence their decision to adopt improved cowpea varieties regardless of interventions (i.e., project areas and extension approaches).
The objective of this study was to identify improved cowpea varieties grown, assess their level of uptake, and measure factors affecting the preference of genotypes among smallholder farmers in the regions of Maradi and Zinder.We hypothesize that farmers' preferred genotype traits (e.g., insect pest and disease resistance, high yield, precocity), in addition to socio-economic characteristics, can influence the adoption of improved cowpea varieties.The results will be useful to development partners interested in the dissemination of cowpea varieties to increase crop productivity and improve the income of smallholder farmers.
This study, unlike others in Niger [13,18], makes three major contributions: first, in addition to socioeconomic characteristics, this study takes into account genotype traits (e.g., early maturity, insect pest resistance, disease resistance, drought resistance, yield, fodder production, and market value) in assessing the adoption of improved cowpea varieties.Second, the study assesses smallholder farmers' preferred genotype traits which is critical in understanding gaps in the adoption of improved cowpea varieties.Third, the study focus is wider than all prior publications.It identifies factors that influence the adoption of improved cowpea varieties disseminated in both the Maradi and Zinder regions regardless of intervention approaches used by development partners (e.g., NGOs, government, or others).

Study Area
The study was implemented in the south-central agricultural zone of Niger; in the regions of Maradi and Zinder (Figure 1).This is a Sahelian zone with a rainfall of 300 to 600 mm per year.Rainfall is poorly distributed and characterized by episodes of drought [20].The rainy season begins in June and ends in October.The soils are sandy and sandy clay and generally poor.Agriculture is the main economic activity with more than 80% of the active population.Millet, sorghum, cowpea, and peanuts are the main crops.

Data Collection
This survey was implemented in July 2020 by interviewing 634 farmers in four departments in the Maradi (Mayahi and Guidan Roumdji) and Zinder (Mirriah and Magaria) regions.The regions and departments were selected based on their importance in cowpea production (i.e., quantity).These two regions produced 42.0% of the total national cowpea production in 2020 [5].Villages in each department were selected based on cowpea production (i.e., quantity) and accessibility (road conditions due to flooding and security challenges) by the enumerators.Fifteen villages were randomly selected per region based on a list provided by local extension agents.In each village, 15 to 25 representatives of households were randomly selected from a list of cowpea producers who were available in the village during the study.Villages and producers' names written on slips of paper were drawn from a basket until the desired number was reached.The research unit was a cowpea producer who resided in any of the selected villages.

Data Collection
This survey was implemented in July 2020 by interviewing 634 farmers in four departments in the Maradi (Mayahi and Guidan Roumdji) and Zinder (Mirriah and Magaria) regions.The regions and departments were selected based on their importance in cowpea production (i.e., quantity).These two regions produced 42.0% of the total national cowpea production in 2020 [5].Villages in each department were selected based on cowpea production (i.e., quantity) and accessibility (road conditions due to flooding and security challenges) by the enumerators.Fifteen villages were randomly selected per region based on a list provided by local extension agents.In each village, 15 to 25 representatives of households were randomly selected from a list of cowpea producers who were available in the village during the study.Villages and producers' names written on slips of paper were drawn from a basket until the desired number was reached.The research unit was a cowpea producer who resided in any of the selected villages.
Six enumerators were recruited and trained on how to implement the survey.For the identification of varieties, enumerators were trained on the characteristics of the 14 varieties registered in Niger.To facilitate the identification of improved cowpea varieties during the survey, we developed an illustration guide with pictures and packaged small samples of the 14 cowpea varieties registered in Niger.The illustration guide and the 14 samples of cowpea seed were given to each enumerator.The verification of the cowpea variety was carried out by cross-checking several pieces of information including the physical characteristics of the variety and also the visual observation of the seed sample by each respondent.In addition, producers brought their seeds to compare them with the samples Six enumerators were recruited and trained on how to implement the survey.For the identification of varieties, enumerators were trained on the characteristics of the 14 varieties registered in Niger.To facilitate the identification of improved cowpea varieties during the survey, we developed an illustration guide with pictures and packaged small samples of the 14 cowpea varieties registered in Niger.The illustration guide and the 14 samples of cowpea seed were given to each enumerator.The verification of the cowpea variety was carried out by cross-checking several pieces of information including the physical characteristics of the variety and also the visual observation of the seed sample by each respondent.In addition, producers brought their seeds to compare them with the samples provided to the enumerators.The shape, color, size, and appearance of the seed make it possible to distinguish cowpea varieties.

Empirical and Conceptual Models
The dependent variable "adoption" is dichotomous.It takes the value one (1) if the farmer adopts an improved cowpea variety, and zero (0), otherwise.The adoption of improved cowpea seeds is influenced by their genotype attributes and socioeconomic factors.Based on these attributes various factors (explanatory variables) are likely to influence the adoption of the technologies (Table 1).Several factors interact in an individual's decision whether or not to adopt a technology [21].These are the socio-economic characteristics of the individual, the technology, the environment, and the institutional factors.Including farmers' perceptions of technology-specific attributes may help understand factors conditioning adoption choices [22].The LOGIT model was used to assess the adoption of cowpea varieties [18].In the literature, three types of models are mainly used to analyze the decision of producers to adopt an agricultural technology: linear probability models, LOGIT, and PROBIT.The first model has drawbacks because the probability can often exceed 1.The last two models are the most commonly used to specify the relationships between the probability of choice and the determining variables of choice.In the context of this study, a cowpea producer can adopt several improved varieties at the same time.However, we were interested to understand the adoption rate of each improved variety and the genotype traits and socio-economic factors that influence a farmer's decision.Thus, the binary logit model was considered more appropriate than the multinomial model.Unlike multinomial models where there are alternatives, here the choice is binary.LOGIT has the advantage of facilitating the interpretation of the β parameters associated with the explanatory variables X i [23].
The decision to adopt an innovation only occurs when the combined effect of factors reaches a value from which the decision maker agrees to use or adopt an innovation.Considering the hypothesis that the effect is measured by an unobservable index I d for the decision maker d and I 0d the critical value of the index from which he adopts the technology, two scenarios can arise: If I d is greater than or equal to I 0d , then he adopts the technology, and the adoption variable Y takes the value 1.The greater the index I d is above the critical value, the higher the probability that the producer will adopt.If I d is less than I 0d , it rejects the innovation and Y is equal to 0. This is presented in Equation (1).
For individual d, the index I d can be a linear combination of variables X i that determine adoption and coefficients β i to be estimated.Its expression is then mathematically given by: less than I 0d , it rejects the innovation and Y is equal to 0. This is presented in Equation (2).
With X id the ith independent variable explaining the adoption of the technology by individual d and β i its corresponding parameter to be estimated.The probability Pd for individual d to adopt the innovation is then (Equation ( 3)): As the index I 0d is a random variable, if we denote by F its function of cumulative probability or distribution function, it comes to (Equation ( 4)): The functional form of F is determined by that of the probability density function of the random variable I d .For the logit model, it is a logistic function (Equation ( 5)):

Econometric Model
The empirical equation resulting from the model theory can be expressed as (Equation ( 6)): with: where β 0 is the constant term; β i the coefficients to be estimated and e i error terms.

Demographics and Other Characteristics of Respondents
The average age was about 41 years, 74% of respondents were men, 94% were married, and there were about five adults in each household (Table 2).Agriculture was the most important activity for all respondents, with an average experience of about 30 years.The average size of a field per respondent was 3 ha, and about half of it was used for cowpea production.A little less than half (45%) of the respondents had contact with the extension service and only 38.3% were members of farmers' organizations.

Adoption of Improved Cowpea Varieties
Adoption of improved cowpea varieties by region ranged from 3.3 to 38% (Table 3).Improved cowpea variety grown by the majority of farmers was IT90K 372-1-2 (38%) followed by UAM-09-1055-6 (11.4%).The use of local varieties by farmers was a little over three times higher in the Zinder than in the Maradi region.The adoption of improved varieties was comparable between the two regions, except for IT97K-499-35, which was higher in the Maradi region, and IT07K 292-10 (1), which was higher in the Zinder region.

Factors Affecting the Adoption of Improved Cowpea Varieties
On average, most improved cowpea varieties were preferred by farmers due to early maturity (86.9%), high yield (73.9%), and high market value 50.5% (Table 4).Varietal attributes such as drought, disease, striga or witchweed (S. gesnesrioides), and insect pest resistance had low preferences among the respondents (Table 4).All genotype attributes were similar among all improved cowpea varieties except for high yield.

Discussion
The results show that eight improved varieties were cultivated by smallholder farmers in both regions.These improved cowpea varieties are among the 14 genotypes registered in the Nigerien crop catalog [24].The IT90K 372-1-2 was the most widely used variety with an adoption rate of 38% across all regions.This genotype, which was disseminated in Niger about 25 years ago is still preferred and used by most smallholder farmers [11,14,19,25].The adoption rate of IT90K 372-1-2 in this study is a little higher than the 30.5% [8] and much lower than the 61.5% [13] previously reported.The differences may be explained by the data collection methods.Rabe et al., (2017) surveyed participants in farmers' field schools who were trained on improved seed while Baoua et al., (2021) results are based on a broad sample of cowpea farmers.Overall, there has been significant progress in the adoption of IT90K 372-1-2 compared to 20 years ago; when uptake was at about 17% [11].
Several socio-economic factors were found to explain differences in the adoption of improved cowpea varieties.In addition to the genotype trait (early maturity), contact with extension appeared to explain the high uptake of IT90K 372-1-2.Observed differences in adoption rates among varieties may be explained by the approaches used to disseminate improved cowpea varieties.For instance, some projects targeted different socioeconomic strata (e.g., gender, villages, and farmers' organizations) and included public extension services while others did not.The role of extension in increasing the adoption of improved seed has been demonstrated [15,17,19].A study in the Zinder region linked the low use of improved genotypes to inadequate communication, unavailability, and high cost of cowpea seed [28].Disseminating farmers' preferred genotypes using various extension approaches and targeting different socio-economic groups may help to accelerate the adoption of improved cowpea varieties.
The fact that IT90K 372-1-2 was the most adopted improved cowpea variety may be explained by its precocity (60-70 days) as well as large-scale dissemination [11,13].The IT90K 372-1-2 is also known to have yields that are higher or comparable to those of other existing improved cowpea varieties [8,19].In addition, the higher uptake of the IT90K 372-1-2 variety can be explained by the availability of its seeds.Seed availability plays an important role in technology adoption [17,29].About 71.5% and 17.3% of all certified cowpea seed stocks in both regions were made of IT90K 372-1-2 and the three remaining genotypes (UAM 09 1055-6, IT07K 292-10 (1), TN-121-80), respectively [30].Limited access (availability) to and low demand (preference) for these three improved genotypes (UAM 09 1055-6, IT07K 292-10 (1), TN-121-80) may explain the low adoption rate among farmers.Though drought resistance did not influence farmers' adoption of IT90K 372-1-2, this variety is known to have this attribute [26].Farmers in arid areas such as the Maradi and Zinder regions prefer varieties with multiple attributes that fit their production systems [19].Drought-resistant or tolerant short-cycle varieties with good yields will help farmers cope with the negative effects of climate change [6,10,31].Risk exposures are an integral part of farmers' decision-making to invest in agricultural inputs (e.g., fertilizer, seed) to increase crop productivity [32].Hence, breeding programs must consider farmers' preferred attributes in the early stage of genotype development and release to increase the potential for adoption.Omitting farmers' perceptions of technology-specific attributes would negatively influence their adoption decisions [22,33].

Conclusions
This study found that farmers in the Maradi and Zinder regions of Niger grew eight improved cowpea genotypes.IT90K 372-1-2, though introduced 25 years ago, was the most widely used and highest-performing variety compared to recently released genotypes.Early maturity was the most significant factor that explained the adoption of IT90K 372-1-2, followed by contact with extension agents.Beyond these genotype traits, the high use of IT90K 372-1-2 was also due to the availability of seeds.Observed adoption gaps among varieties can be explained, in part, by the scarcity of cowpea seeds with farmers' preferred genotypes traits.To scale up the adoption of improved cowpea varieties, there is a need to (i) include farmers' preferred genotype traits in the early stages of breeding programs and seed release; and (ii) improve the availability of high-performing (early maturing with good yields) cowpea genotypes; and (iii) increase awareness and training using various extension approaches.Given the geographic limitations of our study (focused on two out of seven regions that produce cowpea in Niger) and, lack of information on the quality and how farmers access seeds, we propose that future research efforts (i) conduct the same study in other regions to validate our findings; (ii) assess the best approach to improve access and increase the availability of farmers' preferred genotypes, (iii) evaluate genotype purity to improve the quality of seed being supplied to farmers.

Agronomy 2022 , 10 Figure 1 .
Figure 1.Map showing the area where the survey was conducted in both Maradi (left) and Zinder (right) regions, Niger.Each black dot represents a surveyed village.

Figure 1 .
Figure 1.Map showing the area where the survey was conducted in both Maradi (left) and Zinder (right) regions, Niger.Each black dot represents a surveyed village.

Table 1 .
Description of variables used in the LOGIT regression model.

Table 2 .
Socio-economic characteristics of respondents.

Table 3 .
Cowpea varieties adoption rates in Maradi and Zinder regions of Niger.

Table 5 .
Estimation model results on socioeconomics and genotype traits explaining the adoption of improved varieties.