Modelling smallholder farmers' preferences for soil fertility management technologies in Benin: A stated preference approach.

The decline of soil fertility is a major constraint which results in lower levels of crop productivity, agricultural development and food security in Sub-Saharan Africa. This study is the first to perform a focalized investigation on the most interesting technological profiles to offer to each category of producers in Benin agricultural development hubs (ADHs) using the stated preference method, more precisely, the improved choice experiment method. The investigation focused on 1047 sampled plots from 962 randomly selected producers in villages of the Smallholder Agricultural Productivity Enhancement Program in Sub-Saharan Africa of the ADHs. An analysis of the experimental choice data with the endogenous attribute attendance and the latent class models was carried out to account for the attribute non-attendance phenomenon and the heterogeneity of the producers' preferences. However, three classes of producer with different socio-economic, demographic, and soil physicochemical characteristics were identified. Thus, the heterogeneity of preferences was correlated with the attributes linked to the cost, sustainability, and frequency of plot maintenance. All producers, regardless of the ADHs, had a strong attachment to accessibility of technologies with short time restoration of soil fertility, and the ability to obtain additional benefits. These latest attributes, added to that relating to cost, tended to have a low probability of rejection in the decision-making process. These results have implications for local decision-makers facing the complex problem of resolving land degradation and local economic development challenges. The generalizability of these findings provides useful insight and direction for future studies in Sub-Saharan Africa.

Introduction

Agriculture is the key to ensuring food security [1]. However, dean class="Chemical">pan clne">ass="Chemical">span>ite its imass="Chemical">pan>ortance, agriculture has not yet reached its potential in emerging economies because it is subject to several land, social, and economic constraints. It is therefore exposed to high <pan class="Chemical">span class="Disease">rainfall depn>endencen>an>, low level of use of technological innovations, and low level of education among its stakeholders [1]. In addition, one of the reasons often attributed to the lack of agricultural research performance in Benin, and Sub-Saharan Africa in general, is related to land and eco<sppan>an class="Gene">system degradation [2-4].

Africa loses 8 million metric tons of soil nutrients each year and 95 million ha of land has been degraded to the an class="Chemical">point of significantly reducing productivity [3,5]. This situation is explained by high population growth and the intensification of land use with the direct consequence of reducing the fallow period [6], decreasing land productivity, and increasing charges to maintain crop production levels. Steiner [7] showed that the direct consequence is the threat to food supply. A research carried out in Benin shows that the loss of major nutrients from the soil is very high and far exceeds the inputs (mineral fertilizers in particular), resulting in a negative balance of <span class="Chemical">N, P, and K minerals in particular [6,8]. These losses are mainly due to exportation through crops, <span class="Chemical">water erosion and leaching, and mining practices [9].

In rean class="Chemical">pan clne">ass="Chemical">span>onse to this situation and the continuing need to increase agricultural ass="Chemical">pan>roductiopan class="Chemical">n, a set of strategies and technologies for soil fertility improve<pan class="Chemical">span class="Spn>ecies">menn>an>t in general and agroforestry in particular, which could help producers to improve crop yields and limit pressure on forest resources, have been developed through agricultural research and disseminated over the years. In Benin, several strategies have been developed by farmers over several centuries to maintain and restore soil fertility [10]. However, the existing body of literature support that other strategies (agronomic, biological, zootechnical, agricultural land use planning, and combined practices) have been introduced by research and develop<sppan>an class="Species">ment structures [4,11,12]. Researchers have strongly argued that despn>ite the actions of the govern<span class="Species">ment and technical and financial partners to support the dissemination and adoption of these technologies, the problem of declining soil fertility persists [4,12]. Besides mineral fertilizers, these new technologies have been poorly implemented. Several previous studies have highlighted the low level of implementation of these technologies by producers despite their technical performance [2,4]. This situation may be linked to the incompatibility of the characteristics of these new technologies with those desired by farmers [4]. Moreover, these previous studies do not reveal the heterogeneity of producers’ preferences for soil fertility management technologies.

The an class="Chemical">producers’ willingness to accept new technologies will increase through access to better technologies that fit their individual characteristics [13-16]. These characteristics are related to the socioeconomic, demographic, farm, and soil characteristics of the producers. Thus, this study contributes to the literature by including the physicochemical and biological characteristics of producers’ soil in the sources of heterogeneity of preferences by considering producers’ plots as an experi<span class="Species">mental observation unit.

This study offered the an class="Chemical">persistent effort to address the gaan class="Chemical">ps and limitations identified from the previous studies. To better involve stakeholders in the field and enable them to adaass="Chemical">pan>t their ass="Chemical">pan>ractices, this study assessed the technological profiles appropriate for producers’ plots, using the stated preference (SP) assess<span class="Species">ment method. The SP method is usually based on random utility theory and relies on assumptions of economic rationality and utility maximization [17]. To operationalize this approach, several data collection methods have been developed. These are the pairwise comparison approach, traditional full profile approach, adaptive or hybrid joint approach, and experimental choice (EC) or discrete choice approach [18,19]. The first three approaches are difficult and require respondents to have high intellectual skills unlike the EC approach, which is much easier [19,20]. In addition, the experimental choice method is very popular and provides a solid basis for stated preference studies by providing a measure of the benefits sought by different categories of individuals [19,21,22]. This method has been used in consumer choice studies to understand how consumers make their purchasing decisions, predict their behaviour, and determine the value and importance of different characteristics of goods to offer them products that correspond well to their requirements [23,24]. The experimental choice method proposes to the respondent to make a choice from a set of several alternatives or profiles. It is closer to the real purchase or adoption process of the economic agent, and therefore, better represents real market situations [21]. The latter approach is the one used in the present study.

Our study is an class="Chemical">part of a process to further develop all of the above-<pan class="Chemical">span class="Species">menpan>tioned work using improved econometric models, the latent class logit model (LCM), allowing for the heterogeneity of preferences to be taken into account [25,26]. This choice of modelling highlights the expectations of different classes of producers and their willingness to adopt for attributes. The attribute non-attendance (ANA) was particularly taken into account by using the endogenous attribute attendance econometric model (<ass="Chemical">san class="Chemical">ppan>an class="Chemical">EAA), which constitutes progress compared with previous experimentation studies on choice [27,28]. In addition, the present study was carried out following the new conformation of the agricultural sector. To this end, it provides updated data in line with current agricultural policy, which involves agricultural development hubs (ADHs).

This study should guide the actions of an class="Chemical">public and private decision-makers, as well as those of NGOs, in the develop<pan class="Chemical">span class="Species">menpan>t or dissemination of appropriate technologies per <ass="Chemical">san class="Chemical">ppan>an class="Gene">ADH. This manuscript is structured according to the following plan. The following section presents the materials and experi<span class="Species">mental selection methods as well as the empirical model used. The third section presents the estimated results, which are then discussed in Section 4. The final section presents the conclusions and implications for the development and a widespread adoption of soil fertility management technologies in Benin.

Materials and methods

Study zones

The study was conducted in 2017 in Benian class="Chemical">pan clne">ass="Chemical">n,an> West Africa, more ass="Chemical">pan>an class="Chemical">specifically, in 18 villages of the intervention zones of the Smallholder Agricultural Productivity Enhance<n>an class="Chemical">span class="Species">ment Program (SAPEP) in Sub-Saharan Africa. Villages of the intervention zones of the SAPEP, were randomly selected during a baseline survey, taking into account the agricultural production areas and the presence and levels of production in the promising sectors targeted by the project. Subsequently, these villages were divided into <spn>n>an>an class="Chemical">ADHs to spatially characterise the demand in terms of soil fertility management technology. The villages explored for the present study are illustrated in Fig 1.

Fig 1

Study zone.

Selection of producers and sampling of plots

The study was conducted not only with the heads of an class="Chemical">producers’ households that were previously randpan class="Gene">omly selected during the baseline study conducted by the project but also with an individual of the oass="Chemical">pan>ass="Chemical">pan>osite sex to the head of the household to cpan class="Gene">ompare intra-household perceptions. Thus, 96pan class="Gene">2 producers were surveyed (51.61% male, 48.39% female).

The an class="Chemical">producers’ an class="Chemical">plots were the observation units for soil sampling and experi<span class="Species">mental technology selection. Soil samples were taken from the producers’ plots to assess the physicochemical and biological characteristics of the soil. The study was carried out on private land. We confirm that, the owner of the land gave permission to conduct the study on this site, and no specific permissions were required for these locations/activities. The agents of the research centers, the town <span class="Chemical">hall, the village chiefs, and the resource <span class="Species">persons facilitated the interviews with the producers in the villages. The identification of these characteristics facilitates the determination of the preferences of producers’ seg<span class="Species">ments with particular soil characteristics. Given the large size of each producer’s plots, the total number of plots sampled (N) was determined according to the following formula of Glele-Kakai and Sinsin [29]: where N is the total number of plots to be explored; t21 − α/2 (α = 5%) is the critical value of the Student distribution t that converges to the normal distribution for large samples (N > 30) and is equal to 1.96; Cv is the coefficient of variation of the number of producers’ plots in the villages considered. It is equal to 51.26% (baseline survey); and d is the margin of error set at 5%. Thus, a total of 1047 parcels were investigated. It should be noted that the sample plots were randomly selected from those of producers while ensuring diversity in integrated soil fertility management practices. Note that soil samples were collected on the sampled plots producers to assess the physico-chemical and biological characteristics of soils. These composite soil samples weighing around 500 grams were taken 20 cm deep.

Assessment of soil parameters to determine their level of limitation

The can class="Chemical">pan clne">ass="Gene">omal">pan>osite soil samples taken were sent to the INRAB <span class="Chemical">Water and Env<span class="Chemical">ironment Soil Science Laboratory for analysis.

<an class="Chemical">pan clne">ass="Chemical">span>an class="Gene">Sysass="Chemical">pan>an> et al. [30] agreed five degrees of intensity of limitations for soil parameters:

Degree I: no limitatioan class="Chemical">pan clne">ass="Chemical">n,an> the characteristic of the soil considered is oass="Chemical">pan>timal;

Degree II: slight limitations, referring to situations that could slightly decrease yields without however iman class="Chemical">posing pan class="Chemical">special cultivation techniques;

Level III: moderate limitations, referring to situations which cause a greater reduction in yields or the iman class="Chemical">ple<an class="Chemical">pan class="Chemical">span class="Sass="Chemical">pan>ecies">menass="Chemical">pan>an>tation of span class="Chemical">pan>ecial cultivation techniques. These limitations do not affect profitability;

Level IV: severe limitations, referring to situations which cause a reduction in yields or the iman class="Chemical">ple<an class="Chemical">pan class="Chemical">span class="Sass="Chemicne">al">pan>ecies">menass="Chemical">pan>an>tation of cultivation techniques which could jeopardize profitability;

Degree V: very severe limitations, referring to situations that no longer allow the use of land for the an class="Chemical">pan clne">ass="Chemical">span>ecific ass="Chemical">pan>urpose.

Thus, the degrees of intensity of the limitations of the chemical characteristics, for each of the saman class="Chemical">pled plots of the producers were determined by following the parameters defined by <pan class="Chemical">span class="Gene">Syspan> et al. [30]. Table 1 summarizes the criteria for evaluating the degrees of limitation of soil chemical parameters, defined by <sppan>an class="Gene">Sys et al. [30].

Table 1

Criteria for assessing the degrees of limitation of soil chemical parameters.

Soil chemical parameters	Degrees of limitation
	Degree I (Without limitations)	Degree II (Weak limitations)	Degree II (Weak limitations)	Degree: IV (Severe Limitations)	Degree: V (Very severe limitations)
Organic matter	> 2	2–1.5	1.5–1	1–0.5	< 0.5
Total nitrogen	> 0.08	0.08–0.06	0.06–0.045	0.045–0.03	< 0.03
P ppm (Bray 1)	> 20	20–15	15–10	10–5	< 5
K (meq/100 g of soil)	> 0.4	0.4–0.3	0.3–0.2	0.2–0.1	< 0.1
Sum of exchangeable bases (meq/100 g of soil)	> 10	10–7.5	7.5–5	5–2	< 2
Base Saturation (V)	> 60	60–50	50–30	30–15	< 15
CEC (meq/100 g of soil)	> 25	25–15	15–10	10–5	< 5
pH	6.5–6.0	6.0–5.5	5.5–5.3	5.3–5.2	< 5.2
	6.5–7.8	6.5–7.8	7.8–8.3	8.3–8.5	> 8.5

Experimental choice method

The EC method is usually based on randan class="Chemical">pan clne">ass="Gene">oman> utility theory and relies on assumass="Chemical">pan>tions of econpan class="Gene">omic rationality and utility maximization [17,31]. In randpan class="Gene">om utility models, the probability of observing a n>an class="Chemical">specific realization of a choice to be modelled is determined according to a decision rule formulated in terms of latent or unobservable variables that are associated with it [32]. Thus, the indirect utility (U) that the producer (i) receives from the given attributes for an alternative (j) takes the following linear form: where Vij is the deterministic (non-stochastic) part of the utility and ε is the random (or stochastic part consisting of the error terms of the model) that takes into account uncertainty. Depending on the assumptions adopted to represent the distribution of the random portion, different discrete choice models can be distinguished. However, mathematically, discrete choice models are generally based on the assumption that the choice probabilities relating to the utility function can be estimated by the multinomial logit model (MLM). However, this model has limitations regarding to the assumption (Gumbel’s law) of identically distributed independence (IDI) of error terms between alternatives and observations, and therefore assumes homogeneity of preferences [33,34]. Another limitation of the MLM is related to the assumption of independence of irrelevant alternatives (IIA). It is the capital limit of multinomial logit [35]. To overcome these limitations, several other alternative models are available. These are the nested logit, crossed nested logit, latent class model (LCM), polytomic probit model, mixed logit, and generalized multinomial logit [34].

The nested logit first an class="Chemical">proan class="Chemical">posed by [35], which is also part of the same family of generalized extreme values as the multinomial logit, does not allow us to completely avoid the IDI and IIA hypotheses. More flexible than previous models, the polytomic probit is not constrained by the three previously developed limits. However, the estimation of this model generates too heavy econometric calculations. The mixed logit model (MXL) is not constrained by any of the above limitations and can detect possible heterogeneity not observed in preferences [34]. The latent class logit model (LCL) also does not violate the IIA hypothesis and differs from the MXL model in that it allows the distribution of coefficients to be discrete rather than continuous. It uses a statistical methodology based on the concept of likelihood to identify sources of heterogeneity at the seg<span class="Species">ment level rather than at the individual level as does mixed logit [26]. It can be considered to be an improvement of the MXL. This model has been increasingly applied in recent segmentation studies and has produced promising results [25,26]. Several authors have compared the two approaches (MXL and LCL). Some concluded that each approach has its own merits and even that the LCL is more efficient in terms of estimation [36]. Others, such as Scarpa et al. [37], explain that the LCL has the advantage of being based on a joint estimate and that it allows a more intuitive interpretation, facilitating communication with decision-makers.

In additioan class="Chemical">pan clne">ass="Chemical">n,an> more recent studies have revealed that reass="Chemical">pan>an class="Chemical">spondents participn>ating in discrete choice expn>eriences often ignore certain attributes in their decision-making process (called ANA). Resulting biases in modelling occur when these an>an class="Chemical">spects are not taken into account [28,38,39]. Two approaches were proposed to account for the ANA. These include stated ANA and inferred ANA. Stated ANA is an experi<span class="Species">mental approach, which consists of asking the respondent to answer spn>n>an>ecific questions about the attribute ignored during decision-making [24]. Inferred ANA is an econometric approach that provides a better fit for the model, while the stated ANA is not consistent[27,28]. First, the fact that respondents attribute low importance to certain attributes, which could be ignored at the beginning of their choices, leads to overestimation [28]. In addition, the answers to questions relating to the ignored attributes can be a source of potential problems regarding endogeneity bias [27]. <span class="Chemical">LCL and EAA are econometric models widely used to account for inferred ANA models [38,40]. For these reasons, LCL and EAA were applied in the present study for data analysis.

Latent class logit model

The basic assuman class="Chemical">ption of the <an class="Chemical">pan class="Chemical">span class="Chemical">LCLass="Chemicne">al">pan>an> model is that an individual belongs to a ass="Chemical">san class="Chemical">pan class="Chemical">pan>ecific seg<span class="Species">ment but that the members of each segment are unobservable. As a result, respondents from different segments will have different preferences. The LCL simultaneously estimates the probability that a producer will select a given technology from the set of choices and belongs to a specific segment [25]. Let us consider producer i who selects alternative j from K technology alternatives in the set of choices. Furthermore, assuming that he/she belongs to segment s, where s ∈ S, the indirect utility function for his/her preferred technology profile j is written as follows: where X is a vector representing the attributes concerning to the K alternatives and β is the parameter vector of the segment s associated with the vector X and ε as error terms. Assuming that the error terms are independently and identically distributed (IID) and follow the Gumbel distribution, the probability that i will select an alternative among the K alternatives while belonging to a given segment is:

Let us now consider a function of the an class="Chemical">probability (an class="Chemical">P) belonging to a seg<pan class="Chemical">span class="Species">menass="Chemical">pan>an>t s among the S unobservable seg<ass="Chemical">san class="Chemical">pan class="Chemical">pan>an class="Species">ments where Z represents individual characteristics of the producer and λ_s (s = 1, 2, 3…S) represents the parameters to be estimated ass="Chemical">span>ecific to each seg<ass="Chemical">span>an class="Species">ment. Assuming that the term of error ξ is independently and identically distributed between producers and seg<span class="Species">ments and follows a Gumbel distribution, then the probability that he/she belongs to segment s can be expressed as follows:

By can class="Chemical">pan clne">ass="Gene">oman>bining Eqs (4) and (5), we obtain the following exass="Chemical">pan>ression which represents the probability that producer i belongs to seg<span class="Species">ment s and selects technology profile j:

The log likelihood function to obtain the an class="Chemical">parne">ameters λ and βs is given by the expression: where I is an indicator variable of the observed choice.

Endogenous attribute attendance model (EAA)

In the <an class="Chemical">pan clne">ass="Chemical">span>an class="Chemical">EAAass="Chemical">pan>an> model, each choice is considered a two-step process in which the decision maker first decides which attributes to take into account when comparing profiles. Then, he selects the profile with the best characteristics, taking into account his individual characteristics [28,38]. Thus, the formulation of the basic logit of <span class="Chemical">pan>an clne">ass="Chemical">EAAan> is presented as follows: where represents the individual (i) who selects the modality of the attribute (k) relative to the profile (j) of the subset of choice attributes C, in the choice situation (t); and denotes the ass="Chemical">span>ecific coefficient for the attribute (k).

According to [41], the an class="Chemical">probability that individual i tne">akes into account attribute k is an class="Chemical">pan class="Chemical">specified by , where z is a vector of individual characteristics and γ is a vector of ass="Chemical">pan>arameters to be estimated [28,42]. Assuming that these ass="Chemical">pan>robabilities are independent of attributes, the probability of choosing a profile designated as a subset of attributes (C) is given by:

The an class="Chemical">probability that individual i selects an class="Chemical">profile j from the set of choices (C) in a given situation (t) can be written as follows: where Y takes the value 1 when option (j) is choseal">pan>an class="Chemical">n, and 0 otherwise; f(β|θ) denotes the density for β in which θ is the distribution parameter.

Experimental design

Since the EC technique is characterized by a statistical design of hyan class="Chemical">pothetical alternne">atives [43,44], soil fertility manne">age<an class="Chemical">pan class="Chemical">span class="Sass="Chemical">pan>ecies">menass="Chemical">pan>an>t technology profiles were composed based on the main choice attributes reported by users during the exploratory phase. This first phase of the survey took place in 2017 through focus groups and using an interview guide developed on the basis of previous studies. Thus, analysis of the data from this phase using the Kendall method made it possible to prioritize an exhaustive list of the main attributes of the selection of practices. Consequently, six main attributes were selected, each with two levels, except for the cost level; cost was the monetary attribute.

Restoration time is the duration of soil remediatioan class="Chemical">pan clne">ass="Chemical">n,an> which can be fast or slow deass="Chemical">pan>ending on the technology used. It has been demonstrated previously that the misuse of chemical fertilizers is due to its rapid effect on crop growth. Shrub legumes have a much longer cycle and therefore assume slow soil restoratiopan class="Chemical">n, unlike herbaceous legumes (e.g. Mucuna) or synthetic chemical fertilizers [3,45].

Accessibility correan class="Chemical">pan clne">ass="Chemical">span>onds to the availability of the basic materials involved in the realization of technology. The unavailability of chemical fertilizers has always been raised as a constraint to imass="Chemical">pan>le<pan class="Chemical">span class="Species">menpan>tation [46]. Consequently, the massive use of technologies would be owing to their high availability and ease of access.

Regarding the an class="Chemical">possibility of obtaining additionne">al benefits, san class="Chemical">pan class="Gene">ome technologies (regeneration based oass="Chemical">pan>an class="Chemical">n, e.g., soya, <span class="Species">cowpea, and <span class="Species">cassava) are appreciated for their ability to restore soil fertility (biomass supply) and increase yield but also facilitate the production of edible or marketable products contributing to increasing, e.g. income, household food security, and livestock feeding [47]. Others (regeneration with, e.g. Mucuna) are less adopted because they do not facilitate the production of additional benefits. Soil fertility retention time corresponds to the effectiveness of the technology over one or more production campaigns following application. The sustainability of innovation is a continuous process of perceiving, which enables business organizations to have new markets, improved products and services [48].

The control exan class="Chemical">presses the frequency of maintenne">ance of the an class="Chemical">plot. The attributes that were used during the choice experi<pan class="Chemical">span class="Species">menpan>t are presented in Table 2 with the associated attribute levels.

Table 2

Attributes and associated attribute levels.

Attributes	Attributes levels
Restoration time	1 = Short2 = Long
Accessibility	1 = Difficult2 = Easy
Possibility of obtaining additional benefits	1 = Impossible2 = Possible
Soil fertility retention time	1 = Temporary (1 production campaign)2 = Long (more than one campaign)
Regular control (frequency of maintenance of the plot)	1 = Less control2 = Regular control
Purchase cost CFAF per hectare	0; 70,000; 100,000; 150,000; 220,000

Considering the number of attributes and the associated attribute levels, 25 × 51 = 160 an class="Chemical">possible can class="Chemical">pan class="Gene">ombinations or theoretical ass="Chemical">pan>rofiles were constructed. It would be very difficult for reass="Chemical">pan>an class="Chemical">spondents to objectively consider and judge 160 profiles before making a precise choice. To facilitate their choice, the discrete choice sets were restricted to 16 realistic profiles divided into four groupn>s of four profiles each (4 × 4) with the efficiency index Dz estimated at 98.81% [44]. This result was possible in the SAS software pan>an class="Gene">ckage using the experi<span class="Species">mental design of Street and Burgess [49], which is based on the criterion of optimality of efficient design. Among the profiles thus created, 10 existing practices could be identified by their characteristics and put into play in the choice sets with the other six hypothetical profiles. A reference situation (status quo) was added to each set of choices. The different standard technology profiles were presented to each producer in the form of a game. For each set of choices, the respondent was asked to select a technology profile for each of their sampled plots or to select the reference situation that correspn>n>an>onds to their current practice. Each respondent revealed the fictitious situation that provided the most useful information, thus expressing their interest in the attributes according to the profile. To make the exercise easier to understand for the respondents, the different scenarios were illustrated with photos or pictograms accompanied by a brief caption. Examples of the maps proposed to the producers during interviews are shown in Fig 2. The various actors were assured that their virtual choice in the experi<span class="Species">ment would not have any real immediate consequences on their activities. It was clarified that the results would be used more generally to determine the appropriate technology model for soil fertility manage<an class="Chemical">span clne">ass="Species">menan>t. Data were collected in 2017 using an application installed on tablets that contained a digital version of the survey guide. It should be noted that soil samples were taken from the sampled plots of each producer to assess the physicochemical and biological characteristics of the soil.

Fig 2

Example of a set of cards proposed during the interview.

Empirical model

The dean class="Chemical">pendent variable Y correan class="Chemical">pan class="Chemical">sponds to the choice of a ass="Chemical">pan>rofile ass="Chemical">pan>referred by the producer for each of his plots in each set of choices. It takes the value Y = 1 if he/she selects a profile and Y = 0 for the others that are not chosen. In model specification, the utility that the individual derives from a profile model depends on the main attributes of the technology and some individual characteristics of the producer [13,50,51]. Apart frpan class="Gene">om the main attributes of the technology, the individual characteristics of the producer/household (socio-econn>an class="Gene">omic, plot, physicochemical, and biological characteristics), and the production area that can be used for segmentation, are those that could affect the choice or abandon<span class="Species">ment of fertility management practices [52]. According to the studies carried out by the researchers, potential characteristics are related to gender, formal education, number of active agricultural members, access to credit, number of plots and area planted, fertility level, hubs and sub-hubs of development. Not all of the explanatory variables were included in the LCL estimate. A correlation matrix was used to eliminate highly correlated variables and those that did not exhibit variability within the alternatives. Variables such as income were not included in the model because the correlation threshold between the variable and the area planted was very high (p <0.01). As a result, producers with high incomes were those who planted large areas. The current reference situation or practice (status quo) was used in the model. The negative sign of the coefficient of this variable would generally imply the motivation of producers to adopt new soil fertility management technologies. Table 3 provides details of the explanatory variables inserted in the model.

Table 3

Variables used in econometric models.

Variable	Modality
Status quo	1 = yes and 0 = otherwise
Attributes
Cost	Continuous variable
High restoration speed	0 = Slow; 1 = Quick
Accessibility	0 = Difficult; 1 = Easy
Possibility of obtaining additional benefits	1 = yes and 0 = otherwise
Long conservation life	0 = Temporary (one production campaign); 1 = Long (more than one campaign)
Maintenance frequency (regular)	0 = Less control; 1 = Regular control
Maintenance frequency (regular) × Cost
Accessibility × Cost
Physicochemical characteristics of the soil
Organic matter rate (MO)	Continuous variable
N	Continuous variable
P	Continuous variable
K	Continuous variable
Soil pH level	Continuous variable
Fertility level	Continuous variable (0 = weak; 1 = average; 2 = high)
Socio-economic and demographic characteristics
Gender	0 = Woman; 1 = Man;
Formal education	1 = yes and 0 = otherwise
Number of active agricultural members	Continuous variable
Access to credit	1 = yes and 0 = otherwise
Acreage	Continuous variable
Duration of fallow period	Continuous variable

an class="Chemical">pan clne">ass="Chemical">Pan>arameters estimated frass="Chemical">pan>an class="Gene">om the latent class logit can also be used to calculate the willingness-to-adopn>t (WTA) for each attribute, which helpn>s to understand the ren>an class="Chemical">spondents’ motivation and quantify their levels of preference for the attributes. Suppose for attribute X1 that the WTA1 wants to be estimated. Estimated parameter β1 of attribute X1 can be interpreted as the marginal utility of this attribute. In addition, let us note by δ the parameter estimated for the monetary attribute i.e. "the cost", which represents the marginal utility of the currency [41]. The WTA associated with attribute X1 is given by the following formula:

Results

Socio-economic and demographic characteristics of respondents

Table 4 an class="Chemical">presents the socio-econan class="Chemical">pan class="Gene">omic and demograass="Chemicne">al">pan>an class="Gene">phic characteristics of the respondents, showing that the proportion of men (51.79%) was significantly higher than that of <span class="Species">women in any hub (p <0.01). With regard to the level of class reached, all producers had an average of seven (06) years. Significant differences were observed at the levels of the producers of ADH2 and ADH3, which totalled 5 years of class reached. On average, seven people made up a household in the total population. The significantly differences show that the household size of producers of ADH2 (12 people), ADH3 (11 people) is significantly higher than those of ADH4, ADH5, and ADH6 (8 people). Less than 14% of all producers had access to agricultural credit. The variation observed within the hubs showed that the level of access to credit for producers of ADH3 (25.45%) and ADH5 (18.83%) was significantly higher than that of ADH2 (13.14%), ADH4 (13.68%), and ADH6 (5.17%) (p <0.01). Likewise, the area planted with those of ADH2 (6.55 Ha) and ADH3 (8.81 Ha) was significantly larger than those of ADH6 (2.33 Ha), ADH5 (1.83 Ha), and ADH4 (2.33 Ha) (p <0.01).

Table 4

Socio-economic and demographic characteristics.

ADH	2	3	4	5	6	All ADH	Statistic test
Gender (1 = man; 0 = woman) (%)	52.19	50.91	51.67	51.12	51.79	51.61	67.36 ***
Class level reached (in years)	4.67 (3.44)	5.00 (2.44)	6.15 (2.56)	5.76 (3.00)	5.96 (1.82)	6.18 (2.80)	-3.53 ***
Number of active agricultural members	12.03 (9.14)	11.43 (5.75)	7.56 (4.44)	8.41 (5.49)	8.10 (4.66)	6.74 (5.69)	2.78 ***
Access to credit (%)	13.14	25.45	13.68	18.83	5.17	13.81	13.82 **
Acreage (in Hectare)	6.55 (4.96)	8.81 (6.27)	2.96 (1.55)	1.83 (1.28)	2.33 (1.03)	3.8 (1.53)	23.15 ***

ADH = agricultural development hub.

<an class="Chemical">pan clne">ass="Chemical">span>an class="Gene">ADHal">pan>an> = agricultural develop<span class="Chemical">pan>an clne">ass="Species">menan>t hub.

Assessment of the nutritional status of soil in ADHs

The data in table A1 in aan class="Chemical">pan class="Chemical">pendix present the average of the nutrient contents of the soils (soil parameter), and the degrees of intensity of the associated limitations, according to the PDAs (Fig 3). Following Fig 3, the parameters which induce severe limitations in the soils of producers of <pan class="Chemical">span class="Gene">PDA2pan>, PDA4, and PDA6 are mainly related to the <sppan>an class="Chemical">phosphorus content, the sum of the cations, and the capacity cation exchange. On the other hand, at the level of PDA3 and PDA5, the limiting factors are linked to the cation exchange capacity for PDA3, and the phosphorus for PDA5.

Fig 3

Degree of intensity of associated limitations, according to ADHs.

Results of econometric estimates of experimental choices

Results of latent class logit

20,940 observations fran class="Chemical">pan clne">ass="Gene">oman> the reass="Chemical">pan>an class="Chemical">sponses of 96n>an class="Gene">2 respondents for 1,047 plots were analysed using the LCL model as part of the technology choice experi<span class="Species">ment. Respondents’ individual characteristics were assumed to affect their profile choices indirectly through their effect on class membership. The determination of the optimal number of classes requires the statistics reported in Table 5 to be used, mainly the Akaike, Bayesian, and Consistent information criteria (AIC, BIC, and CAIC, respectively), which must be minimal [41,53]. The results showed that AIC and BIC were respectively minimized to five and six segments, while the CAIC criterion was minimized to four segments. Five or six segments were already too high. In addition, the BIC criterion was the most reliable [26,41], and also allowed for a more robust explanation of the heterogeneity of preferences. Consequently, three segments were selected.

Table 5

Calculation of Akaike (AIC), Bayesian (BIC), and Consistent (CAIC) information criteria.

Number of segments	Likelihood log	AIC	BIC	CAIC
2	−4658.60	9347.21	9398.16	9362.79
3	−4607.61	9421.52	9165.34*	9401.79
4	−4551.67	9436.52	9318.90	9155.17*
5	−4545.79	9261.23*	9349.90	9387.99
6	−4530.58	9375.16	9169.59	9434.99

* indicates the lowest values of AIC, BIC, and CAIC.

* indicates the lowest values of <an class="Chemical">pan clne">ass="Chemical">span>an class="Disease">AICal">pan>an>, <span class="Chemical">pan>an clne">ass="Chemical">BICan>, and C<al">span>an class="Disease">AIC.

The results of the estimation of the logit model with three latent classes are an class="Chemical">presented in Table 6. As shown in Table 6, the first level presented the utility coefficients associated with the attributes of soil fertility technologies, while the second seg<pan class="Chemical">span class="Species">menass="Chemical">pan>an>t presented those of individual producer characteristics. The coefficients associated with the individual characteristics of the third seg<ass="Chemical">san class="Chemical">pan class="Chemical">pan>an class="Species">ment producers were reduced as zero, i.e. the S3 seg<ass="Chemical">span>an class="Species">ment was considered as a reference. The pseudo R2 of the model was 0.69, which suggests a very good fit. The predicted probability that a producer would belong to segments S1, S2, and S3 was 33%, 31%, and 36%, respectively. The distribution of the sample among the three classes finally appeared to be fairly balanced: 32.09% were assigned to S1, 30.95% to S2, and 36.96% to S3.

Table 6

Estimation of the model 3 latent classes.

Attribute	Segment 1	Segment 2	Segment 3
Status quo	−0.14 (0.08) **	−3.10 (0.29) ***	−3.44 (0.32) ***
Cost	0.04 (0.02) **	−0.05 (0.01) **	−0.34 (0.08) ***
Short restoration time	1.17 (0.07) ***	1.89 (0.12) ***	3.83 (0.49) ***
Accessibility	0.33 (0.06) ***	0.65 (0.10) ***	0.98 (0.10) ***
Possibility of obtaining additional benefits	0.67 (0.12) ***	0.99 (0.22) ***	3.81 (0.49) ***
Long conservation time	0.14 (0.09)	−1.68 (0.13) ***	0.81 (0.17) ***
Maintenance frequency (regular)	0.10 (0.07) *	0.03 (0.01) **	−0.11 (0.30)
Maintenance frequency (regular) × Cost	0.28 (0.35)	−1.51 (0.98) ***	−0.82 (0.01) **
Accessibility × Cost	1.65 (1.19) **	−0.14 (0.05)	−4.81 (2.66) **
Physicochemical characteristics of the soil
Organic Matter (OM) Rate	−0.14 (0.24)	1.18 (0.40) ***
N rate	12.88 (4.31) ***	−1.85 (6.06)
P rate	0.01 (0.01)*	−0.64 (0.01)*
K rate	−1.73 (0.03) **	−0.19 (0.05)*
Soil pH level	0.25 (0.09)	−0.55 (0.12)
Fertility level	−0.46 (0.01) **	−0.28 (0.06) *
Duration Fallow period	−0.05 (0.15)	−0.36 (0.08) **
Individual characteristics
Gender (1 = man; 0 = woman)	0.42 (0.30)	0.62 (0.05) **
Formal education	−1.58 (0.35)	−0.53 (0.35)
Number of active agricultural members	0.67 (0.18) ***	0.84 (0.26) ***
Access to credit	1.42 (0.36) ***	−0.18 (0.06) *
Acreage	0.15 (0.19)	1.03 (0.20) ***
ADH3	−0.57 (1.08)*	1.27 (0.83)
ADH4	−17.40 (27.38)	6.79 (1.67) ***
ADH5	2.13 (0.69) ***	0.74 (0.84) *
ADH6	19.19 (1.40)	10.17 (8.68) ***
Constant	−2.21 (0.71) ***	0.73 (0.11) ***
Probability of belonging to each class	0.33	0.31	0.35
Number of individuals (%)	32.09	30.95	36.96
Number of observations	20,940 = 1047*5*4
Number of respondents = 962; Number of plots = 1047
Likelihood log	−4322.48
R2	0.69
Test Wald Chi2(60)	13587.93***

***, **,* mean, respectively, that the coefficients are significant at the 1%, 5%, and 10% threshold; numbers in parentheses represent standard errors.

***, **,* meaan class="Chemical">pan clne">ass="Chemical">n,an> reass="Chemical">pan>an class="Chemical">spectively, that the coefficients are significant at the 1%, 5%, and 10% threshold; numbers in parentheses represent standard errors.

The estimated an class="Chemical">parameters of the utility function showed that the constne">ants (stne">atus quo) an class="Chemical">pan class="Chemical">specific to S1, S2, and S3, were highly and negatively significant in the model. Consequently, these classes of ass="Chemical">pan>roducers were ass="Chemical">pan>redipan class="Chemical">sposed to change their current situation (status quo) to adopt a new form of technology in view of their soil with a low fertility level (p <0.01).

an class="Chemical">pan clne">ass="Chemical">Pan>references were hass="Chemical">pan>an class="Gene">omogeneous across all seg<span class="Species">ments for technology with a high restoration speed (subsequent rapid growth of crops and yields) and would favour the production of additional benefits (p <0.01), taking into account the positive sign of the coefficient associated with these variables. Also, all producers revealed a very strong preference for a technology that will be available at all times and easily accessible.

The three seg<an class="Chemical">pan clne">ass="Chemical">span class="Sal">pan>ecies">menpan>ts identified differ in their preference for attributes about cost, sustainability of soil conservation, frequency of plot maintenance after application of technology, and individual characteristics. In addition, Wald’s test was able to show that the seg<sppan>an class="Species">menan>ts differed from each other mainly with respect to these attributes and individual characteristics because the value of the statistical test of the likelihood ratio estimated from the model (13,587.93) exceeded the critical value of 100.42 for a distribution at 60 degrees of freedom (p <0.01). Therefore, the null hypothesis of all parameters and interaction terms jointly equal to zero was rejected. The variable related to the level of education and soil pH had a non-significant coefficient for all classes. This implies that producers who have the same level of education and soil pH are randomly assigned to the three seg<span class="Species">ments. This result justifies the relatively stable soil pH level at the level of the plots of most producers.

Concerning class 1, the membershian class="Chemical">p coefficients for this clne">ass of an class="Chemical">producers show that they were <pan class="Chemical">span class="Species">menass="Chemicne">al">pan>an> from <ass="Chemical">san class="Chemical">pan class="Chemical">pan>an class="Gene">ADH2, <ass="Chemicne">al">span>an class="Gene">ADH3, ADH4, ADH5, and ADH6 who sowed vast areas and practiced short fallow periods. In addition, given their ease of access to campaign credit and the high proportion of active farm members in their households, they revealed preferences for expensive technologies even if they require regular maintenance or monitoring of the plot after application.

The coefficients an class="Chemical">pertne">aining to the terms of interactions with maintenne">ance cost and frequency, and accessibility were an class="Chemical">positive, indicating these producers’ preferences for expensive accessible technologies even if they require regular maintenance. The coefficients of the variables associated with fertility level and soil chemical components (organic matter content: OM; K) were negative, and those of the variables associated with <span class="Chemical">nitrogen and <span class="Chemical">phosphorus content were positive. This result indicates that the fertility status of their soil was practically low with high levels of N and P and low levels of OM and K.

Seg<an class="Chemical">pan clne">ass="Chemical">span>an class="Sass="Chemical">pan>ecies">menpan>t S2 of the model was significantly characterized by <sppan>an class="Species">menan> from all <span class="Chemical">ADHs who planted vast areas and practiced a short fallow period as in S1. The heterogeneity of preference between S1 and S2 appeared at the attribute levels specifying the duration of soil fertility conservation and cost. In addition, the variable related to the number of the active members and that of the attribute specifying the maintenance frequency were positive and showed that the probability of adopting technologies requiring high-frequency maintenance increases proportionally with the proportion of active agricultural members in the household of producers in S1 and S2. Conversely, the coefficient related to the term of interaction with the cost and the accessibility was negative, indicating these producers’ preferences for cheaper technologies when they are available. Similarly, the coefficient associated with the interaction term, cost and maintenance frequency was negative. This result indicates that producers in this class were also oriented towards high-cost technologies that require less maintenance or work. The coefficients of the variables associated with fertility level and soil chemical components (<span class="Chemical">N, P, and K) were negative, and those of the variables associated with OM levels were positive. This finding indicates that the fertility status of the soil was practically low with a low rate of <span class="Chemical">N, P, K and a high rate of OM.

In relation to the individual an class="Chemical">producers’ characteristics in seg<an class="Chemical">pan class="Chemical">span class="Sass="Chemical">pan>ecies">menass="Chemical">pan>an>t S3, which is considered as a reference seg<span class="Chemical">pan>an clne">ass="Species">menan>t, the parameter coefficients may be interpreted as regards the coefficients of the other two seg<ass="Chemical">span>an class="Species">ments provided that these coefficients are all of the same sign and significance [26]. However, the S3 class was significantly characterized by <span class="Species">men and women from ADH2, ADH3, ADH4, and ADH6, who practiced fallowing over a long period of time with preferences for less-expensive technologies. The fertility level of their soil was relatively low with a very high K level. The sign of the utility coefficients of the segment variables (S3) revealed that producers of this class shared almost the same preferences with those in S2. Indeed, preferences were heterogeneous with regard to the attribute relating to the duration of conservation and maintenance frequency after the application of technology. All producers in S3 revealed a determining position for all attributes except for the one with regard to maintenance frequency, which was not significant. This result indicates that these producers did not also attach particular importance to the frequency of maintenance after the technology was applied. Nevertheless, the sign of the coefficient of the variable shows that they had a preference for technologies that promote soil fertility over a long period of time and require less maintenance.

Results of endogenous attribute attendance model. The results fran class="Chemical">pan clne">ass="Gene">oman> the <ass="Chemical">pan>an class="Chemical">span class="Chemical">EAAn>an> model are presented in Table 7 with the probabilities of ANA. In model 1, the attributes sppan>ecifying the possibility of obtaining an additional benefits had the lowest significant ANA probabilities, which indicates that the probability that the attribute linked to the possibility of obtaining an additional benefit, is ignored in a choice situation is 13%. Lower probabilities were obtained for attributes pertaining to restoration sppan>eed (7%). The probabilities regarding to cost and accessibility were not significant, indicating that these attributes do not explain the probability of ANA. The attribute most often ignored was sustainability (86%) followed by maintenance frequency (52%), which was then excluded from model 2.

Table 7

Results of endogenous attribute attendance model.

Attribute	Model 1		Model 2
	Coefficient	p ANA	Coefficient	p ANA
Status quo	−3.28 (0.40) ***	0.69 (0.03) ***	−2.74 (0.2) ***	0.66 (0.03) ***
Cost	−0.10 (0.01) ***	0.01 (0.03)	−0.09 (0.01) ***
Restoration time	2.13 (0.09) ***	0.07 (0.01) ***	2.11 (0.10) ***	0.08 (0.02) ***
Accessibility	0.66 (0.13) ***	0.10 (0.16)	0.56 (0.03) ***	0.02 (0.01)
Possibility of obtaining additional benefits	1.92 (0.20) ***	0.13 (0.05) **	1.77 (0.21) ***	0.13 (0.05) **
Soil fertility retention time	2.95 (0.32) ***	0.86 (0.02) ***	1.19 (0.04) ***
Frequency of maintenance of the plot	−0.15 (0.06) **	0.52 (0.13) **	−0.086 (0.05)
p excluded attribute				0.57 (0.10) ***
Number of observations	20940		20940
Likelihood log	−4675.32		−4710.56
Wald Chi2(8)	806.48 ***		1109.68***
AIC	9376.65		9445.13
BIC	9479.99		9540.52

***, **,* mean, respectively, that the coefficients are significant at the 1%, 5% and 10% threshold; p: probability of ANA.

***, **,* meaan class="Chemical">pan clne">ass="Chemical">n,an> reass="Chemical">pan>an class="Chemical">spectively, that the coefficients are significant at the 1%, 5% and 10% threshold; p: probability of ANA.

Model an class="Chemical">pan clne">ass="Gene">2 an>ass="Chemical">pan>resents estimates jointly with model 1, according to the attributes excluded frpan class="Gene">om the choice in model 1. As a result, model pan class="Gene">2 involves only the attributes with the lowest probability of rejection in model 1. In model 2, the attributes n>an class="Chemical">specifying the sustainability and the frequency of maintenance were excluded. These variables were jointly ignored at 57% when they were in competition with the other attributes (cost, short restoration time, accessibility, and possibility of obtaining additional benefits). These latter attributes tend to have a low probability of rejection (<13%) when confronted together in a choice process. This indicates that these attributes are likely to play an essential role in the decision-making process. The likelihood of rejecting the status quo in both models was high, implying that respondents pay more attention to alternative situations during choices. Therefore, they are interested in new recom<span class="Species">mendations regarding the manage<span class="Species">ment of soil fertility.

Estimate of the willingness to adopt

The an class="Chemical">parameters estimated fran class="Chemical">pan class="Gene">om the <ass="Chemical">pan>an class="Chemical">span class="Chemical">LCL model were used to calculate the WTA for each attribute at the level of each seg<sass="Chemical">pan>pan>an class="Species">ment (Fig 4) from the derivative of the cost-related monetary attribute. This makes it possible to understand the motivation of the different classes of stakeholders and quantify their levels of preference. The results presented in Fig 4 show that the producers of seg<span class="Species">ment S1 (CFAF 133,512.83 or 220.05 dollars US, at a fixed exchange rate of 1 dollar US = CFAF 606,73), S2 (CFAF 33,357.59), and S3 (CFAF 55,987.00) were more interested in the attribute linked to the time to restore soil fertility compared with other attributes.

Fig 4

Estimate of the willingness-to-adopt from the three latent classes model.

With the excean class="Chemical">ption of fran class="Chemical">pan class="Gene">om the restoration time, attributes linked to the ass="Chemical">pan>ossibility of obtaining additional benefits and accessibility were also valued. Reass="Chemical">pan>resented in order of importance, the attributes regarding to short restoration time, possibility of obtaining additional benefits, and accessibility were more popular with producers of S1 and S3. Those of S2 prioritized, in order of importance, short restoration time, accessibility, and possibility of obtaining additional benefits. Controlling for the attribute linked to long shelf life, it was found that this attribute was important to those in S1 (CFAF 16,305.89) and S3 (CFAF 11,968.21) but not those in S2. This result indicates that those of S2 did not attach any particular importance to the duration of restoration of soil fertility. By considering the attribute specifying the frequency of regular maintenance, it appears that the producers of S1 and S2 considered this attribute positive unlike those of S3.

Discussion

The an class="Chemical">producers were an class="Chemical">predisposed to change their current situation and follow other recom<span class="Species">mendations. The main attributes prioritized during the choices relate to cost, short recovery time, accessibility, and the possibility of obtaining additional benefits. The assess<span class="Species">ment of the heterogeneity of producer preferences made it possible to identify three classes of producers, appearing in different proportions (S1: 32.09%; S2: 30.95%; and S3: 36.96% in S3). The heterogeneity of preferences was observed at the level of attributes concerning to cost, duration of soil conservation, and frequency of maintenance of the plot after application of the technology. On the other hand, preferences were homogeneous across all segments, with respect to technology whose speed of recovery is fast (subsequently rapid growth of crops and yields), and would be able obtain additional benefits. In addition, technologies favouring the obtaining of additional benefits promote not only soil fertility, but also contribute to the income and food security of farmers through the provision of edible or marketable products. These results confirm those of Yabi et al. [45] who found that the technology about to the use of Mucuna, Aeschinomene Hytrix is less used because it does not facilitate the obtaining of an edible food.

All the an class="Chemical">producers show a very strong preference for a technology that will be available at all times, and easily accessible. This result agrees with that of Assogba et al. [46], and Maliki [47] which shows that supply difficulties and lack of funding limit the use of technologies. Katengeza [54] finds that the unavailability of the technology subsequently leads to non-compliance with recom<span class="Species">mended doses. Likewise, Adekambi et al. [55] assert that the adopters of aqueous botanical extracts abandoned this practice after having experienced it themselves at least once on their respective plots because of the problems of unavailability of the leaves of these extracts.

The heterogeneity of an class="Chemical">preferences shows that clne">ass S1 brings together <an class="Chemical">pan class="Chemical">span class="Sass="Chemical">pan>ecies">menass="Chemical">pan>an> from <span class="Chemical">pan>an clne">ass="Gene">ADH2an>, <ass="Chemical">span>an class="Gene">ADH3, <span class="Gene">ADH4, ADH5, and ADH6, who cultivate large areas, and who belong to a group. In addition, given their ease of accessing seasonal credit, and the high proportion of active agricultural members in their household, they reveal preferences for expensive technologies even if they require regular maintenance or control of the plot. after application. This result can also be explained by the fact that they consider the price to be a good quality indicator. Also, credit for them can be an for the poor to invest. In addition, the income situation of producers leaves little room for self-financing. IFS [56], and Yabi et al. [45] show that agricultural financing is important for the viability of the soil fertility management action plan in Benin. Also, they had an attraction for expensive accessible technologies, even if they require regular maintenance. Given the high cost of the technology, they want to perform regular maintenance in order to benefit from its economic performance over several production campaigns. These observations confirm the research results of Maliki [47], and of Baco et al. [57], who show that technologies that require regular field control also promote weed control, which may make the soil fertility conservation technology in use more effective. The fertility status of their soil was practically low, with high nitrogen and phosphorus content, and low in organic matter and potassium. In view of their preference, they were named “segment of men who are major producers of ADH 2, 3, 4, 5, and 6, given the size of their large area, having access to credit with infertile soils (especially with low OM and K rate), users of a costly practice in terms of purchase and use (maintenance), easily accessible, favouring acceleration, sustainability of soil fertility, and obtaining additional benefits. Indeed, the characteristics of the technology desired by these producers were similar to the use of crop residues. As a result, the use of crop residues would promote the delivery of nutrients rich in organic elements to the soils, and mineral fertilizers promote the delivery of the mineral elements. This result agrees with that of RAMR [58] who shows that the practice concerning the use of crop residues cannot be done over a large area, because the storage of residues requires a very large labor of plot.

Seg<an class="Chemical">pan clne">ass="Chemical">span>an class="Sass="Chemical">pan>ecies">menpan>t (S2) was characterized by less educated producers of <sppan>an class="Gene">ADHan>3, <span class="Gene">ADH4, ADH5, and <span class="Gene">ADH6 who practiced a short fallow like those in S1. Given the difficulties in accessing seasonal credit, they reveal preferences for technologies that promote soil conservation over a short period (one production campaign), and less expensive. This result supports the findings of Abbasi et al. [59,60] who show that the real economic activity grows when the price of electricity decreases and the electricity demand rises.

an class="Chemical">pan clne">ass="Chemical">Pan>overty is, after ignorance, one of the factors limiting the adopan>tion of technological approaches, and invest<span class="Species">ment in the regeneration of land fertility. Under these conditions, the granting of well-studied seasonal credits is more appropriate [3,57].

In additioan class="Chemical">pan clne">ass="Chemical">n,an> the ass="Chemical">pan>robability of adoption of technologies requiring a high maintenance frequency increases proportionally with the proportion of members of agricultural labor in the household of producers in S1 and S2. The preferences of the producers in the S2 and S3 seg<span class="Species">ment were also oriented towards technologies that are expensive, but require less maintenance. They believe that new technologies should help simplify work. The position of this segment of producers is particularly evident with regard to technologies requiring regular maintenance, involving the use of important production factors that require significant costs. For the most part, in order to minimize costs and achieve profit margins, they prefer less expensive technologies, which do not incur significant production costs. The fertility status of their soil was practically low with a low level of N, P, K, and a high organic content. They were qualified as a segment of producers who do not have access to credit with infertile soils (low rate of N; P; K and high rate of OM), users of a practice that is less expensive in terms of purchase, easily accessible and requiring regular control of the field, obtaining additional benefits favouring the acceleration and conservation of soil fertility over a production campaign. Indeed, the characteristics of the technology desired by these producers were similar to those of the use of mineral fertilizers (potash fertilizers) given the low level of N; P; K, and a high level of organic matter. These results are similar to the findings of Krah et al. [61] in level of smallholders of Malawi.

Class S3 has been characterized by <an class="Chemical">pan clne">ass="Chemical">span>an class="Sass="Chemical">pan>ecies">menpan>, and <sppan>an class="Species">womenan> from <span class="Gene">ADH2, ADH3, <span class="Gene">ADH4, and ADH6, who are not in groupings, and who have an attraction for less expensive technologies. The fertility of their soil was marked only by a high potassium element content. They were attracted to technologies that favour soil fertility over a long period of time, and which require less maintenance.

The an class="Chemical">position of an class="Chemical">producers in this seg<pan class="Chemical">span class="Species">menass="Chemicne">al">pan>an>t is particularly evident in view of the small size of the active members of the household. Indeed, the characteristics of the technology preferred by these producers were similar to those of the practice of <ass="Chemical">san class="Chemical">pan class="Chemical">pan>an class="Disease">crop rotation. This technique is preferred in order to regenerate, and guarantee the conservation of the soil structure over a long period. Sanginga and Woomer [62], reveal that rotations integrating legumes directly contribute to the constitution of soil organic matter, which plays multiple functions in improving the physicochemical and biological characteristics of soils. They were characterized by <ass="Chemicne">al">span>an class="Species">men and women of ADH2, ADH3, ADH4, and ADH6 having soils rich in potassium element, users of less expensive practice in terms of purchase and use (maintenance), favouring acceleration and sustainability of soil fertility, and obtaining additional benefits.

Conclusion

The main concern of small farmers in Africa remains the manage<an class="Chemical">pan clne">ass="Chemical">span>an class="Sass="Chemical">pan>ecies">menpan>t and conservation of soil fertility from one season to another to ensure food supplies and improve well-being. This work is important to highlight the serious fertility problems in Benin. This precise research study offers an original method to determine and examine the heterogeneity of preferences by identifying different seg<sppan>an class="Species">menan>ts of producers with particular preferences for the attributes of soil fertility manage<span class="Species">ment technologies at the ADH level. The results of the experi<span class="Species">mental choices of technology profiles for the plots show that the majority of producers are inclined to opt for novel recommendations given the state of their relatively infertile soil. All producers, regardless of gender or ADH, tended to prioritize accessible technologies, favouring improvements in soil fertility and the ability to produce additional benefits. These attributes, added to that relating to cost, tended to have a low probability of rejection during the selection process. Preferences were heterogeneous for attributes specifying cost, sustainability of soil conservation, frequency of plot maintenance after technology was applied, and individual characteristics. However, three classes of producers were identified. Class S1 was referred to as: segment of large male producers of ADH2, ADH3, ADH4, ADH5, and ADH6, with access to credit and infertile soils (especially low levels of OM and K) and users of practices that are expensive in terms of purchase and use (maintenance), easily accessible, promote the improvement and sustainability of soil fertility, and allow for producing additional benefits. Indeed, the typical technology profile desired by S1 women was similar to that for the use of crop residues. Thus, except fr mineral fertilizers, this practice was the most preferred by this class of producers (S1) unlike those of the other segments (S2 and S3) given the relatively low fertility level of their soil with high levels of N and P and low levels of OM and K. In addition, the rotation technique was preferred to guarantee the conservation of soil structure over a long period of time. Those in class S2 were qualified as large producers who do not have access to credit but do have infertile soils (low rate of N, P, and K and high rate of OM) and users of practices that are less costly in terms of purchase, easily accessible, and require regular field monitoring, and allow for the production of additional benefits that promote the improvement and conservation of soil fertility over a production campaign. In addition, the characteristics of the technology preferred by these producers were similar to those for mineral fertilizer use. This technology was preferred given the low level of soil fertility with a low rate of N, P, and K and a high rate of OM. The characteristics of the technology desired by those in class S3 were similar to those of the crop rotation practice. From their assessment, we can see that they were characterized as men and women from ADH2, ADH3, ADH4, and ADH6 who have less fertile soils and are users of practices that are less expensive in terms of purchase and use (maintenance), easily accessible, promote the acceleration and sustainability of soil fertility, and allow for the production of additional benefits. In short, these results show that all producers are not driven by the same expectations. The critical contribution of this research study, from a theoretical viewpoint, is the analysis of the respondents of ADHs, and their understanding of the soil fertility technological profiles, which, in turn, could help business firms to attain sustainable performance.

Policy recommendations

The results of this research will enable local decision-makers to measure an class="Chemical">possible interest in an class="Chemical">pan class="Chemical">spatially discriminating soil fertility manage<ass="Chemical">pan>an class="Chemical">span class="Species">ment measures by <sass="Chemical">pan>pan>an class="Gene">ADH. The definition of these measures at the technical level must be based on the frequent support of extension and technical agents towards producers, the technical performance of technologies, and the installation of school plots. At the economic level, strategies aimed at financial support (e.g. subsidies and grants) and facilitating access to agricultural credit must be implemented. The institutional and political measures relate to the ease of access in villages through the construction of infrastructure, definitions of accompanying measures, and access and availability of technologies. Promoting the advantages of the technologies sought by <span class="Gene">ADH through channels responsible for the dissemination of information or awareness and popularization campaigns would encourage the implementation of technologies. Furthermore, the development of a sectoral policy that takes into account the monitoring of the specificity of each ADH would also encourage the implementation of technologies. The objective is to create an adequate environment in which farmers can acquire the technologies they need to meet their demands for sustainable supply and support sectors. Moreover, it will be necessary to strengthen the cooperation between the public and private sectors for developing technologies to encourage the use of technologies by end users and facilitate access to markets. Technologies should be applied by users with sustainability criteria for rational biomass management. An interesting perspective would be to explore the preferences of producers of other ADHs (1 and 7) that are not associated with the SAPEP intervention areas.

(DOCX)

Clian class="Chemical">pan clne">ass="Gene">ckan> here for additional data file.

(DOCX)

Clian class="Chemical">pan clne">ass="Gene">ckan> here for additional data file.

(DOCX)

Clian class="Chemical">pan clne">ass="Gene">ckan> here for additional data file.

Estimate of the willingness-to-adopt from the three latent classes.

(DOCX)

Clian class="Chemical">pan clne">ass="Gene">ckan> here for additional data file.

(DOCX)

Clian class="Chemical">pan clne">ass="Gene">ckan> here for additional data file.

(DOCX)

Clian class="Chemical">pan clne">ass="Gene">ckan> here for additional data file.

(DOCX)

Clian class="Chemical">pan clne">ass="Gene">ckan> here for additional data file.

(DOCX)

Clian class="Chemical">pan clne">ass="Gene">ckan> here for additional data file.

(DOCX)

Clian class="Chemical">pan clne">ass="Gene">ckan> here for additional data file.

(DOCX)

Clian class="Chemical">pan clne">ass="Gene">ckan> here for additional data file.

(DOCX)

Clian class="Chemical">pan clne">ass="Gene">ckan> here for additional data file.

Average nutrient content of soils, and the degree of intensity of associated limitations, according to ADHs.

(DOCX)

Clian class="Chemical">pan clne">ass="Gene">ckan> here for additional data file.

Survey guide.

(DOCX)

Clian class="Chemical">pan clne">ass="Gene">ckan> here for additional data file.

3 May 2021

<an class="Chemical">pan clne">ass="Chemical">span>an class="Chemical">PONE-D-21-07803ass="Chemical">pan>an>

Modelling Smallholder Farmers' an class="Chemical">pan clne">ass="Chemical">Pan>references for Soil Fertility Manage<ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menpan>t Technologies in Benin: A Stated Preference Approach

an class="Chemical">PLOS ONE

Dear Dr. Segla Roch Cedrique Zossou,

Thank you for submitting your manuscrian class="Chemical">pt to an class="Chemical">PLOS ONE. After careful consideratiopan class="Chemical">n, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscriass="Chemicne">al">pan>t that addresses the ass="Chemical">pan>oints raised during the review process.

an class="Chemical">pan clne">ass="Chemical">Pan>lease submit your revised manuscriass="Chemical">pan>t by May 7, 2021. If you will need more time than this to cpan class="Gene">omplete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revisiopan class="Chemical">n, log on to httpn>s://www.editorialmanager.cn>an class="Gene">om/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

an class="Chemical">pan clne">ass="Chemical">Pan>lease include the following items when submitting your revised manuscriass="Chemical">pan>t:

A rebuttal letter that rean class="Chemical">pan clne">ass="Chemical">span>onds to each ass="Chemical">pan>oint raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Repan class="Chemical">sponse to Reviewers'.

A marked-uan class="Chemical">p coan class="Chemical">py of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Trapan class="Gene">ck Changes'.

An unmarked version of your revised an class="Chemical">paper without trapan class="Gene">cked changes. You should upload this as a seass="Chemical">pan>arate file labeled 'Manuscriass="Chemical">pan>t'.

If you would like to make changes to your financial disclosure, an class="Chemical">please include your uan class="Chemical">pdated state<pan class="Chemical">span class="Species">menass="Chemicne">al">pan>an>t in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer com<ass="Chemical">san class="Chemical">pan class="Chemical">pan>an class="Species">ments at the end of this letter.

If aan class="Chemical">pplicable, we recom<span class="Species">mend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscrian class="Chemical">pt.

Kind regards,

Carlos Alberto Zúniga-González, an class="Chemical">pan clne">ass="Gene">Phan>.D

Academic Editor

an class="Chemical">PLOS ONE

Additional Editor Can class="Chemical">pan clne">ass="Gene">oman><ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menpan>ts:

No can class="Chemical">pan clne">ass="Gene">oman><ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menpan>t

Journal Require<an class="Chemical">pan clne">ass="Chemical">span class="Sal">pan>ecies">menpan>ts:

an class="Chemical">pan clne">ass="Chemical">Pan>lease review your reference list to ensure that it is cass="Chemical">pan>an class="Gene">omplete and correct. If you have cited papn>ers that have been retracted, please include the rationale for doing so in the manuscripn>t text, or remove these references and replace them with relevant current references. Any changes to the reference list should be <n>an class="Chemical">span class="Species">mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

When submitting your revisioan class="Chemical">pan clne">ass="Chemical">n,an> we need you to address these additional require<ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menpan>ts.

1. an class="Chemical">pan clne">ass="Chemical">Pan>lease ensure that your manuscriass="Chemical">pan>t meets pan class="Chemical">PLOS ONE's style require<pan class="Chemical">span class="Spn>ecies">menn>an>ts, including those for file naming. The PLOS ONE style templates can be found at

and

httan class="Chemical">ps://journne">als.an class="Chemical">plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. an class="Chemical">pan clne">ass="Chemical">Please include additional information regarding the survey or interview guide used in the study and ensure that you have al">pan>rovided sufficient details that others could replicate the analyses. For instance, if you developed a questionnaire or interview guide as part of this study and it is not under a copyright more restrictive than CC-BY, please include a copy, in both the original language and English, as Supporting Information.

3. We note that the grant information you an class="Chemical">provided in the ‘Funding Information’ and ‘Finne">ancial Disclosure’ sections do not match.

When you resubmit, an class="Chemical">plene">ase ensure that you provide the correct grant numbers for the awards you received for your study in the ‘Funding Information’ section.

4. We note that Figure an class="Chemical">pan clne">ass="Gene">2 an>includes an image of a<ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">participn>antpan> in the study.

As an class="Chemical">per the an class="Chemical">PLOS ONE policy (http://journals.plos.org/plosone/s/submission-guidelines#loc-<span class="Species">human-subjects-research) on papers that include identifying, or potentially identifying, information, the individual(s) or parent(s)/guardian(s) must be informed of the terms of the PLOS open-access (CC-BY) license and provide specific permission for publication of these details under the terms of this license. Please download the Consent Form for Publication in a PLOS Journal (http://journals.plos.org/plosone/s/file?id=8ce6/plos-consent-form-english.pdf). The signed consent form should not be submitted with the manuscript, but should be securely filed in the individual's case notes. Please a<span class="Species">mend the methods section and ethics statement of the manuscript to explicitly state that the patient/participant has provided consent for publication: “The individual in this manuscript has given written informed consent (as outlined in PLOS consent form) to publish these case details”.

If you are unable to obtain consent fran class="Chemical">pan clne">ass="Gene">oman> the subject of the ass="Chemical">pan>an class="Gene">photogran>an class="Gene">ph, you will need to remove the figure and any other textual identifying information or case descriptions for this individual.

[Note: HTML markuan class="Chemical">p is below. an class="Chemical">Please do not edit.]

Reviewers' can class="Chemical">pan clne">ass="Gene">oman><ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menpan>ts:

Reviewer's Rean class="Chemical">pan clne">ass="Chemical">span>onses to Questions

Can class="Chemical">pan clne">ass="Gene">oman><ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menpan>ts to the Author

1. Is the manuscrian class="Chemical">pt technically sound, and do the dne">ata suan class="Chemical">pport the conclusions?

The manuscrian class="Chemical">pt must describe a technically sound piece of scientific research with data that supports the conclusions. Experi<pan class="Chemical">span class="Species">menpan>ts must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

2. Has the statistical analysis been an class="Chemical">performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscrian class="Chemical">pt fully avne">ailne">able?

The an class="Chemical">pan clne">ass="Chemical">Pan>LOS Data ass="Chemical">pan>olicy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability State<span class="Species">menpan>t in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. <span class="Species">participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscrian class="Chemical">pt presented in an intelligible fashion and written in standard English?

an class="Chemical">pan clne">ass="Chemical">Pan>LOS ONE does not copan>yedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Can class="Chemical">pan clne">ass="Gene">oman><ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menpan>ts to the Author

an class="Chemical">pan clne">ass="Chemical">Pan>lease use the ass="Chemical">pan>an class="Chemical">space provided to expn>lain your answers to the questions above. You may also include additional com<span class="Species">ments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attach<span class="Species">ment if it exceeds 20,000 characters)

Reviewer #1: I found this study informative, which an class="Chemical">presents a new idea entitled, " Modelling Smallholder Farmers' an class="Chemical">Preferences for Soil Fertility Manage<pan class="Chemical">span class="Species">menass="Chemical">pan>an>t Technologies in Benin: A Stated Preference Approach."

This article states that the decline of soil fertility is a major constraint which results in lower levels of croan class="Chemical">p an class="Chemical">productivity, and agricultural develop<pan class="Chemical">span class="Species">menpan>t and food security in Sub-Saharan Africa. This study analyses the most interesting technological profiles to offer to each category of producer in Benin agricultural develop<ass="Chemicne">al">san class="Chemical">ppan>an class="Species">ment hubs (ADHs) using the stated preference method, more precisely, the improved choice experiment method. The investigation focused on 1047 sampled plots from 962 randomly selected producers in villages of the Smallholder Agricultural Productivity Enhancement Program in Sub-Saharan Africa of the ADHs.

Abstract and Introduction iman class="Chemical">prove<an class="Chemical">pan class="Chemical">span class="Sass="Chemicne">al">pan>ecies">menass="Chemical">pan>an>t:

I am glad to assess this informative study. In my oan class="Chemical">pinioan class="Chemical">pan class="Chemical">n, I have sass="Chemicne">al">pan>an class="Gene">ome guidelines for the authors to enhance the study quality before endorsing it for publication. As the Abstract is the main door or "FACE" of the manuscript, it should briefly present high-quality English with new information. I am recom<span class="Species">mending the authors of this study to expand Abstract, as it is too short. The Abstract should be around 250 words. I have suggested some studies to check the abstracts and improve yours and cite them in the introduction and build your study objectives like these studies.

Hussaian class="Chemical">pan clne">ass="Chemical">n,an> T., Abbas, J., Wei, Z., & Nurunnabi, M. (2019). The Effect of Sustainable Urban Planning and Slum Disa<ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menn>an>ity on The Value of Neighboring Residential Property: Application of The Hedonic Pricing Model in Rent P<sppan>an class="Species">rice Appraisal. Sustainability, 11(4). doi:10.3390/su11041144

Abbas, J., Raza, S., Nurunnabi, M., Minai, M. S., & Bano, S. (2019). The Iman class="Chemical">pact of Entrepreneurial Business Networks on Firms’ Performance Through a Mediating Role of Dynamic Capabilities. Sustainability, 11(11). https://doi.org/10.3390/<pan class="Chemical">span class="Chemical">su11113006pan>

Hussaian class="Chemical">pan clne">ass="Chemical">n,an> T., Abbas, J., Wei, Z., Ahmad, S., Xuehao, B., & Gaoli, Z. (2021). Imass="Chemical">pan>act of Urban Village Disa<pan class="Chemical">span class="Species">menpan>ity on Neighboring Residential Properties: Empirical Evidence from Nanjing through Hedonic Pricing Model Appraisal. Journal of Urban Planning and Develop<sppan>an class="Species">ment, 147(1), 04020055. https://doi.org/10.1061/(asce)up.1943-5444.0000645

Literature section

It an class="Chemical">presents a good summary of the literature. I suggest authors add the literature as recan class="Chemical">pan class="Gene">om<ass="Chemical">pan>an class="Chemical">span class="Species">mended below to improve the manuscript. Overall, the authors have creatively linked variables. It reflects an innovative model of the study. I am pleased to read this article. However, I have some suggestions for the authors to enhance the quality of the literature section. The authors can add few lines about technological innovations and env<sass="Chemical">pan>pan>an class="Chemical">iron<span class="Species">mental respn>onsibility practices. Please see the suggested studies and cite them to enhance the literature section.

Abbas, J., Zhang, Q., Hussaian class="Chemical">pan clne">ass="Chemical">n,an> I., Akram, S., Afaq, A., & Shad, M. A. (2020). Sustainable Innovation in Small Medium Enterass="Chemical">pan>rises: The Impact of Knowledge Manage<span class="Species">ment on Organizational Innovation through a Mediation Analysis by Using SEM Approach. Sustainability, 12(6). https://doi.org/10.3390/<span class="Chemical">su12062407

Methods and Results

The results section of the an class="Chemical">paper presents a good view of the study. This work presents a notable investigation on a selected topic. I suggest including some graphical presentations to improve the quality of this study. Please see the proposed studies and see the graphical representation. Improve your work like these studies and cite them in this section.

Abbasi, K. R., Abbas, J., & Tufail, M. (2021). Revisiting electricity consuman class="Chemical">ptio<an class="Chemical">pan class="Chemical">span class="Chemical">ass="Chemicne">al">pan>an class="Chemical">n, p<sass="Chemical">pan>pan>an class="Species">rice, and real <span class="Chemical">GDP: A modified sectoral level analysis from Pakistan. Energy Policy, 149, 112087. doi:10.1016/j.enpol.2020.112087

Abbas, J., Amaan class="Chemical">pan clne">ass="Chemical">n,an> J., Nurunnabi, M., & Bano, S. (2019). The Imass="Chemical">pan>act of Social Media on Learning Behavior for Sustainable Education: Evidence of Students frpan class="Gene">om Selected Universities in Pakistan. Sustainability, 11(6). https://doi.org/10.3390/<pan class="Chemical">span class="Chemical">su11061683n>an>

Abbasi, K. R., Hussaian class="Chemical">pan clne">ass="Chemical">n,an> K., Abbas, J., Adedoyiass="Chemical">pan>an class="Chemical">n, F. F., Shaikh, P. A., Yousaf, H., & Muhammad, F. (2021). Analyzing the role of industrial sector's electricity consumpn>tion, prices, and GDP: A modified empirical evidence from Pakistan [J]. AIMS Energy, 9(1), 29-49. doi:10.3934/energy.2021003

Abbas, J., Mahmood, S., Ali, H., Ali Raza, M., Ali, G., Amaan class="Chemical">pan clne">ass="Chemical">n,an> J., . . . Nurunnabi, M. (2019). The Effects of Corass="Chemical">pan>orate Social Repan class="Chemical">sponsibility Practices and Env<pan class="Chemical">span class="Chemical">ironn>an><sppan>an class="Species">mental Factors through a Moderating Role of Social Media Marketing on Sustainable Performance of Business Firms. Sustainability, 11(12), 3434.

Conclusion

I suggest you make a sean class="Chemical">parne">ate hene">ading of the conclusion and do not mix it with iman class="Chemical">plications.

an class="Chemical">pan clne">ass="Chemical">Pan>olicy Recass="Chemical">pan>an class="Gene">om<span class="Species">mendations

I again recan class="Chemical">pan clne">ass="Gene">oman><ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menn>an>d you to make a separate heading of the Policy Recom<sppan>an class="Species">mendations.

The conclusion section is accean class="Chemical">ptne">able. Overne">all, this an class="Chemical">presents a good piece of research work. I recom<span class="Species">mend that authors do a little more work and revise this article accordingly. I suggest the authors check English quality and fix some weak sentences. If you have already taken English editing service, ask them to recheck the quality to meet scientific merit for publication. I endorse this manuscript for publication after minor corrections, as suggested.

Reviewer #2: I am glad to review and assess this interesting article, entitled, Modeling Smallholder Farmers' an class="Chemical">pan clne">ass="Chemical">Pan>references for Soil Fertility Manage<ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menn>an>t Technologies in Benin: A Stated Preference Approach. This study analyses the most interesting technological profiles to offer to each category of producer in Benin agricultural develop<sppan>an class="Species">ment hubs (ADHs) using the stated preference method, more precisely, the improved choice experiment method. The organization of this article is good and satisfactory. The Introduction section and methodology portions are adequate. I suggest the authors improve the Materials and Methods section by adding some latest articles' citations to enhance the work quality and also concise this part. Also, Improve the Conclusion part as well.

Overall, the manuscrian class="Chemical">pt is a good piece of work. I recom<span class="Species">mend that authors do a little more work and add the latest literature to support the study, as suggested. The English level is good and smooth, e.g., the language standard, sass="Chemical">pan>ass="Chemical">pan>an>ecifically the grammar, of sufficient quality to meet scientific merit for publication. I accept this manuscript after minor revision, as I have recom<span class="Chemical">pan>an clne">ass="Species">menan>ded.

**********

6. an class="Chemical">pan clne">ass="Chemical">PLOS authors have the oal">pan>tion to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made an class="Chemical">public.

Do you want your identity to be an class="Chemical">public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer can class="Chemical">pan clne">ass="Gene">oman><ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menn>an>ts were submitted as an attach<sppan>an class="Species">ment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attach<span class="Species">ment files.]

While revising your submissio<an class="Chemical">pan clne">ass="Chemical">span>an class="Chemical">ass="Chemical">pan>an class="Chemical">n, pn>an>lease upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS require<sppan>an class="Species">ments. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

an class="Chemical">pan clne">ass="Gene">2 an>Jun 2021

Rean class="Chemical">ply to reviewers’ can class="Chemical">pan class="Gene">om<ass="Chemicne">al">pan>an class="Chemical">span class="Species">ments on <sass="Chemical">pan>pan>an class="Chemical">PONE-D-21-07803R1

“Modelling Smallholder Farmers' an class="Chemical">pan clne">ass="Chemical">Pan>references for Soil Fertility Manage<ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menpan>t Technologies in Benin: A Stated Preference Approach.”

Rean class="Chemical">pan clne">ass="Chemical">span>onse to academic editor

Rean class="Chemical">ply to an class="Chemical">pan class="Chemical">specific cal">pan>an class="Gene">om<span class="Species">ments:

1. We note the following text is only included in the Traan class="Chemical">pan clne">ass="Gene">ckan>ed Changes docu<ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menn>an>t: « The study was carried out on private land. We confirm that, the owner of the land gave permission to conduct the study on this site, and no sppan>ecific permissions were required for these locations/activities. The agents of the research centers, the town <sppan>an class="Chemical">hall, the village chiefs, and the resource <span class="Species">persons facilitated the interviews with the producers in the villages."

an class="Chemical">pan clne">ass="Chemical">Please confirm and include the above al">pan>ermissions text in the Methods section of your manuscript.

- Rean class="Chemical">pan clne">ass="Chemical">span>onse 1: The correction has been made as suggested. We confirm and include the following text concerning the ass="Chemical">pan>ermissions text in the Methods section of the manuscript. " The study was carried out on private land. We confirm that, the owner of the land gave permission to conduct the study on this site, and no specific permissions were required for these locations/activities. The agents of the research centers, the town <span class="Chemical">hall, the village chiefs, and the resource <span class="Species">persons facilitated the interviews with the producers in the villages."

\f

Rean class="Chemical">ply to reviewer 1:

First, we want to thank te reviewer for the time an class="Chemical">pan clne">ass="Chemical">span>ent going through this ass="Chemical">pan>aper for com<span class="Species">menpan>ts. We are convinced that these com<span class="Species">ments have substantially improved the paper.

Abstract and Introduction iman class="Chemical">prove<an class="Chemical">pan class="Chemical">span class="Sass="Chemicne">al">pan>ecies">menass="Chemical">pan>an>t:

I am glad to assess this informative study. In my oan class="Chemical">pinioan class="Chemical">pan class="Chemical">n, I have sass="Chemicne">al">pan>an class="Gene">ome guidelines for the authors to enhance the study quality before endorsing it for publication. As the Abstract is the main door or "FACE" of the manuscript, it should briefly present high-quality English with new information. I am recom<span class="Species">mending the authors of this study to expand Abstract, as it is too short. The Abstract should be around 250 words. I have suggested some studies to check the abstracts and improve yours and cite them in the introduction and build your study objectives like these studies.

Hussaian class="Chemical">pan clne">ass="Chemical">n,an> T., Abbas, J., Wei, Z., & Nurunnabi, M. (2019). The Effect of Sustainable Urban Planning and Slum Disa<ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menn>an>ity on The Value of Neighboring Residential Property: Application of The Hedonic Pricing Model in Rent P<sppan>an class="Species">rice Appraisal. Sustainability, 11(4). doi:10.3390/su11041144

Abbas, J., Raza, S., Nurunnabi, M., Minai, M. S., & Bano, S. (2019). The Iman class="Chemical">pact of Entrepreneurial Business Networks on Firms’ Performance Through a Mediating Role of Dynamic Capabilities. Sustainability, 11(11). https://doi.org/10.3390/<pan class="Chemical">span class="Chemical">su11113006pan>

Hussaian class="Chemical">pan clne">ass="Chemical">n,an> T., Abbas, J., Wei, Z., Ahmad, S., Xuehao, B., & Gaoli, Z. (2021). Imass="Chemical">pan>act of Urban Village Disa<pan class="Chemical">span class="Species">menpan>ity on Neighboring Residential Properties: Empirical Evidence from Nanjing through Hedonic Pricing Model Appraisal. Journal of Urban Planning and Develop<sppan>an class="Species">ment, 147(1), 04020055. https://doi.org/10.1061/(asce)up.1943-5444.0000645

- Rean class="Chemical">pan clne">ass="Chemical">span>onse : We thank the reviewer for this ass="Chemical">pan>roposal. We expand the Abstract as suggested. Span class="Gene">ome studies suggested by the reviewer have been exploited and cited in order to improve the abstract and the introduction and build our study objectives.

Literature section

It an class="Chemical">presents a good summary of the literature. I suggest authors add the literature as recan class="Chemical">pan class="Gene">om<ass="Chemical">pan>an class="Chemical">span class="Species">mended below to improve the manuscript. Overall, the authors have creatively linked variables. It reflects an innovative model of the study. I am pleased to read this article. However, I have some suggestions for the authors to enhance the quality of the literature section. The authors can add few lines about technological innovations and env<sass="Chemical">pan>pan>an class="Chemical">iron<span class="Species">mental respn>onsibility practices. Please see the suggested studies and cite them to enhance the literature section.

Abbas, J., Zhang, Q., Hussaian class="Chemical">pan clne">ass="Chemical">n,an> I., Akram, S., Afaq, A., & Shad, M. A. (2020). Sustainable Innovation in Small Medium Enterass="Chemical">pan>rises: The Impact of Knowledge Manage<span class="Species">ment on Organizational Innovation through a Mediation Analysis by Using SEM Approach. Sustainability, 12(6). https://doi.org/10.3390/<span class="Chemical">su12062407

- Rean class="Chemical">pan clne">ass="Chemical">span>onse : The correction has been made as suggested by the reviewer. We have added the literature as recass="Chemical">pan>an class="Gene">om<span class="Species">mended

The results section of the an class="Chemical">paper presents a good view of the study. This work presents a notable investigation on a selected topic. I suggest including some graphical presentations to improve the quality of this study. Please see the proposed studies and see the graphical representation. Improve your work like these studies and cite them in this section.

Abbasi, K. R., Abbas, J., & Tufail, M. (2021). Revisiting electricity consuman class="Chemical">ptio<an class="Chemical">pan class="Chemical">span class="Chemical">ass="Chemicne">al">pan>an class="Chemical">n, p<sass="Chemical">pan>pan>an class="Species">rice, and real <span class="Chemical">GDP: A modified sectoral level analysis from Pakistan. Energy Policy, 149, 112087. doi:10.1016/j.enpol.2020.112087

Abbas, J., Amaan class="Chemical">pan clne">ass="Chemical">n,an> J., Nurunnabi, M., & Bano, S. (2019). The Imass="Chemical">pan>act of Social Media on Learning Behavior for Sustainable Education: Evidence of Students frpan class="Gene">om Selected Universities in Pakistan. Sustainability, 11(6). https://doi.org/10.3390/<pan class="Chemical">span class="Chemical">su11061683n>an>

Abbasi, K. R., Hussaian class="Chemical">pan clne">ass="Chemical">n,an> K., Abbas, J., Adedoyiass="Chemical">pan>an class="Chemical">n, F. F., Shaikh, P. A., Yousaf, H., & Muhammad, F. (2021). Analyzing the role of industrial sector's electricity consumpn>tion, prices, and GDP: A modified empirical evidence from Pakistan [J]. AIMS Energy, 9(1), 29-49. doi:10.3934/energy.2021003

Abbas, J., Mahmood, S., Ali, H., Ali Raza, M., Ali, G., Amaan class="Chemical">pan clne">ass="Chemical">n,an> J., . . . Nurunnabi, M. (2019). The Effects of Corass="Chemical">pan>orate Social Repan class="Chemical">sponsibility Practices and Env<pan class="Chemical">span class="Chemical">ironn>an><sppan>an class="Species">mental Factors through a Moderating Role of Social Media Marketing on Sustainable Performance of Business Firms. Sustainability, 11(12), 3434.

- Rean class="Chemical">pan clne">ass="Chemical">span>onse : The observation are very relevant and has been taken into account as suggested by the reviewer. Sass="Chemical">pan>an class="Gene">ome gran>an class="Gene">phical presentations are now added to improve the quality of this study. Some studies suggested by the reviewer have been exploited and cited in the manuscript

Conclusion

I suggest you make a sean class="Chemical">parne">ate hene">ading of the conclusion and do not mix it with iman class="Chemical">plications.

- Rean class="Chemical">pan clne">ass="Chemical">span>onse : The correction has been made as suggested by the reviewer.

an class="Chemical">pan clne">ass="Chemical">Pan>olicy Recass="Chemical">pan>an class="Gene">om<span class="Species">mendations

I again recan class="Chemical">pan clne">ass="Gene">oman><ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menn>an>d you to make a separate heading of the Policy Recom<sppan>an class="Species">mendations.

The conclusion section is accean class="Chemical">ptne">able. Overne">all, this an class="Chemical">presents a good piece of research work. I recom<span class="Species">mend that authors do a little more work and revise this article accordingly. I suggest the authors check English quality and fix some weak sentences. If you have already taken English editing service, ask them to recheck the quality to meet scientific merit for publication. I endorse this manuscript for publication after minor corrections, as suggested.

- Rean class="Chemical">pan clne">ass="Chemical">span>onse : The correction is done as suggested by the reviewer. We have made a seass="Chemical">pan>arate heading of the Policy Recpan class="Gene">om<pan class="Chemical">span class="Spn>ecies">menn>an>dations. We have checked English quality and fix some weak sentences with English editing service

\f

Rean class="Chemical">pan clne">ass="Chemical">span>onse to Reviewer #ass="Chemical">pan>an class="Gene">2 Com<span class="Species">ments

I am glad to review and assess this interesting article, entitled, Modeling Smallholder Farmers' an class="Chemical">pan clne">ass="Chemical">Pan>references for Soil Fertility Manage<ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menn>an>t Technologies in Benin: A Stated Preference Approach. This study analyses the most interesting technological profiles to offer to each category of producer in Benin agricultural develop<sppan>an class="Species">ment hubs (ADHs) using the stated preference method, more precisely, the improved choice experiment method. The organization of this article is good and satisfactory. The Introduction section and methodology portions are adequate. I suggest the authors improve the Materials and Methods section by adding some latest articles' citations to enhance the work quality and also concise this part. Also, Improve the Conclusion part as well. Overall, the manuscript is a good piece of work. I recommend that authors do a little more work and add the latest literature to support the study, as suggested. The English level is good and smooth, e.g., the language standard, specifically the grammar, of sufficient quality to meet scientific merit for publication. I accept this manuscript after minor revision, as I have recommended.

Rean class="Chemical">ply to generne">al can class="Chemical">pan class="Gene">om<al">pan>an class="Chemical">span class="Species">ments:

- First, we exan class="Chemical">press our grne">atitude to the reviewer for devoting his time to this an class="Chemical">paper for the com<span class="Species">ments and for providing many relevant articles to the topic addressed in this paper. We are convinced that these com<spass="Chemicne">al">pan>an>an class="Species">ments and the review of suggested articles have substantially improved the paper. The correction has been made as suggested. We have improved the Materials and Methods, and Conclusion section, by adding some latest articles' citations to enhance the work quality and also concise this part.

Submitted filename: Rean class="Chemical">pan clne">ass="Chemical">span>onse to Reviewers.docx

Clian class="Chemical">pan clne">ass="Gene">ckan> here for additional data file.

7 Jun 2021

Modelling Smallholder Farmers' an class="Chemical">pan clne">ass="Chemical">Pan>references for Soil Fertility Manage<ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menpan>t Technologies in Benin: A Stated Preference Approach

<an class="Chemical">pan clne">ass="Chemical">span>an class="Chemical">PONE-D-21-07803ass="Chemical">pan>an>R1

Dear Dr. Segla Roch Cedrique Zossou,

We’re an class="Chemical">pleased to inform you thne">at your manuscrian class="Chemical">pt has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical require<pan class="Chemical">span class="Species">menpan>ts.

Within one week, you’ll receive an e-mail detailing the required a<an class="Chemical">pan clne">ass="Chemical">span>an class="Sass="Chemical">pan>ecies">menpan>d<sppan>an class="Species">menan>ts. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for an class="Chemical">pay<an class="Chemical">pan class="Chemical">span class="Sass="Chemical">pan>ecies">menass="Chemical">pan>an>t will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing depart<span class="Chemical">pan>an clne">ass="Species">menan>t directly at authorbilling@plos.org.

If your institution or institutions have a an class="Chemical">press office, an class="Chemical">please notify them about your upcoming paper to helass="Chemicne">al">pan> maximize its imass="Chemical">pan>act. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more informatio<span class="Chemical">n, please contact onepress@plos.org.

Kind regards,

Carlos Alberto Zúniga-González, an class="Chemical">pan clne">ass="Gene">Phan>.D

Academic Editor

an class="Chemical">PLOS ONE

Additional Editor Can class="Chemical">pan clne">ass="Gene">oman><ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menpan>ts (optional):

Reviewers' can class="Chemical">pan clne">ass="Gene">oman><ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menpan>ts:

Reviewer's Rean class="Chemical">pan clne">ass="Chemical">span>onses to Questions

Can class="Chemical">pan clne">ass="Gene">oman><ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menpan>ts to the Author

1. If the authors have adequately addressed your can class="Chemical">pan clne">ass="Gene">oman><ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menn>an>ts raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Com<sppan>an class="Species">ments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recom<span class="Species">mendation.

Reviewer #1: All can class="Chemical">pan clne">ass="Gene">oman><ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menpan>ts have been addressed

Reviewer #2: All can class="Chemical">pan clne">ass="Gene">oman><ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menpan>ts have been addressed

**********

2. Is the manuscrian class="Chemical">pt technically sound, and do the dne">ata suan class="Chemical">pport the conclusions?

The manuscrian class="Chemical">pt must describe a technically sound piece of scientific research with data that supports the conclusions. Experi<pan class="Chemical">span class="Species">menpan>ts must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been an class="Chemical">performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscrian class="Chemical">pt fully avne">ailne">able?

The an class="Chemical">pan clne">ass="Chemical">Pan>LOS Data ass="Chemical">pan>olicy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability State<span class="Species">menpan>t in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. <span class="Species">participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscrian class="Chemical">pt presented in an intelligible fashion and written in standard English?

an class="Chemical">pan clne">ass="Chemical">Pan>LOS ONE does not copan>yedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Can class="Chemical">pan clne">ass="Gene">oman><ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menpan>ts to the Author

an class="Chemical">pan clne">ass="Chemical">Pan>lease use the ass="Chemical">pan>an class="Chemical">space provided to expn>lain your answers to the questions above. You may also include additional com<span class="Species">ments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attach<span class="Species">ment if it exceeds 20,000 characters)

Reviewer #1: I am satisfied to evaluate the revised manuscrian class="Chemical">pt. I have found the revised version of this study effective and sne">atisfactory. The authors have made a good attempt and answered all my points to improve the quality of this study. I feel happy to avail myself of the opportunity to evaluate this informative study.

In my evaluatioan class="Chemical">pan clne">ass="Chemical">n,an> this version of the article entitled, "Modelling Smallholder Farmers' Preferences for Soil Fertility Manage<ass="Chemical">pan>an class="Chemical">span class="Spn>ecies">menpan>t Technologies in Benin: A Stated Preference Approach" has reach merit for publication. I believe that the authors have made an excellent revision to reach scientific merit for the publication of this study. The article is well structured, and the methodology is appropriate, well applied, and discussed. I accept and endorse this revised article in the current format, as the authors have made a satisfactory revision to achieve scientific merit for publication. Have a smooth publication procedure. Good Luck!

Reviewer #2: The authors have successfully addressed all my concerns in the revised manuscrian class="Chemical">pt. Hence I recan class="Chemical">pan class="Gene">om<ass="Chemicne">al">pan>an class="Chemical">span class="Species">mend the acceptance of this paper.

**********

7. an class="Chemical">pan clne">ass="Chemical">PLOS authors have the oal">pan>tion to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made an class="Chemical">public.

Do you want your identity to be an class="Chemical">public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: J. Abbas

Reviewer #2: No

14 Jun 2021

<an class="Chemical">pan clne">ass="Chemical">span>an class="Chemical">PONE-D-21-07803ass="Chemical">pan>an>R1

Modelling smallholder farmers' an class="Chemical">preferences for soil fertility manne">age<an class="Chemical">pan class="Chemical">span class="Sass="Chemical">pan>ecies">menass="Chemical">pan>an>t technologies in Benin: A stated preference approach

Dear Dr. Zossou:

I'm an class="Chemical">pleased to inform you thne">at your manuscrian class="Chemical">pt has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production depart<pan class="Chemical">span class="Species">menpan>t.

If your institution or institutions have a an class="Chemical">press office, an class="Chemical">please let them know about your upcpan class="Gene">oming paper now to helass="Chemicne">al">pan> maximize its imass="Chemical">pan>act. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until pan class="Gene">2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can helan class="Chemical">p with anything else, an class="Chemical">please email us at plosone@plos.org.

Thank you for submitting your work to an class="Chemical">pan clne">ass="Chemical">Pan>LOS ONE and supan>porting open access.

Kind regards,

an class="Chemical">PLOS ONE Editorial Office Staff

on be<an class="Chemical">pan clne">ass="Chemical">span>an class="Chemical">halass="Chemical">pan>an>f of

Dr. an class="Chemical">Prof. Carlos Alberto Zúniga-González

Academic Editor

an class="Chemical">PLOS ONE