The state of soil fertility directly correlates with agricultural productivity (Bastin et al., 2022). However, loss of soil fertility in most smallholder farms is widely recognized as the primary biophysical factor contributing to the decrease in per capita food production across Africa (Terefe, 2023). In the absence of adequate nutrient replenishment and soil fertility management practices, soil degradation accelerates, limiting the land’s productive capacity over time (Saikia and Saikia, 2022). In response to this decline, farmers are compelled to extend cultivation into less productive or marginal lands as a means of sustaining their livelihoods (Mogaka et al., 2025). This further exacerbates land degradation, a problem widely acknowledged as a critical constraint to sustainable food production in the semi-arid tropics of Eastern Africa (Ziadat et al., 2022).
       
In the drylands of Lower Eastern Kenya particularly in Makueni County, declining soil fertility has emerged as a critical barrier to agricultural productivity and a persistent threat to household food security. The problem is largely attributed to nutrient depletion, soil erosion and unsustainable agronomic practices (Kichamu-Wachira et al., 2024). The concern for soil nutrient depletion and declining soil fertility has led to the promotion of various technologies aimed at restoring soil productivity in the County (Fue et al., 2025). However, these technologies have not been adopted to any appreciable extent by farmers. For instance, a study conducted across five maize-producing sub-counties in Makueni County revealed that adoption of modern irrigation technologies remains low, with only 23% of surveyed smallholder farmers using drip or sprinkler systems (Muluki et al., 2022). Furthermore, adoption of key soil and water conservation practices was also limited, with only 17% of respondents utilizing zai pits and 33% implementing terracing indicating constrained uptake of these technologies despite their proven agronomic benefits (Kuria et al., 2024). Several studies have explored the reasons behind the low uptake of SFETs, yet no conclusive or consistent explanation has emerged to fully account for this trend (Spurk et al., 2020).
       
Nevertheless, acceptability and uptake of new technologies depend not only on their biophysical performance and economic returns, but also on a comprehensive understanding of various factors such as how users perceive the core issue, their attitudes, beliefs and existing practices in relation to the target technological solutions (Yeo and Keske, 2024). In addition, Farmers’ attitudes and perceptions toward technologies play a crucial role in shaping their decisions and behaviors related to adoption of new agricultural innovations (Nyairo et al., 2022). The attitude toward using a range of technologies combined with the perception of tangible benefits and costs and the belief that one has the capacity and support to implement them has been found to positively influence farmers’ intentions to adopt new agricultural innovations (Mahata and Sharma, 2021). Despite the recognized importance of attitudes and perceptions in technology adoption, there remains limited research specifically examining how these factors influence the uptake of soil fertility enhancement technologies. Given the critical role of such technologies in improving agricultural productivity and restoring soil fertility, the present study aims to examine the attitudes, knowledge and perceptions influencing the adoption of soil fertility enhancement technologies among farming households in Makueni County, Kenya.

Study area
 
The study was conducted in Makueni County, located in the drylands of Lower Eastern Kenya, a region classified as arid and semi-arid land (ASAL). The county was purposively chosen owing to its prolonged exposure to soil fertility enhancement practices and the persistent challenges it faces, including severe soil degradation, erratic rainfall patterns and low agricultural productivity. These conditions render it a strategic location for investigating strategies aimed at improving soil fertility in dryland conditions. Makueni covers approximately 8,176.7 km² and is geographically positioned between latitudes 1^o352 and 3^o002 South and longitudes 37^o10′ and 38^o30′ East (KNBS, 2019). It is administratively divided into six sub-counties: Makueni, Mbooni, Kibwezi East, Kibwezi West, Kaiti and Kilome. The area receives an average annual rainfall of about 600 mm, varying from approximately 300-400 mm in the lowlands to about 800-1200 mm in higher elevations such as the Mbooni Hills. Mean annual temperatures range between about 22^oC and 24^oC, with maximum temperatures reaching up to 35^oC, while annual evaporation averages around 2000 mm, frequently exceeding rainfall and contributing to soil moisture deficits. The dominant soils are Ferralsols and Luvisols with sandy to sandy-loam textures and generally low organic matter content, which contributes to declining soil fertility. The county falls within the Lower Midland agro-ecological zones LM5 and LM6, characterized by low moisture retention and marginal growing conditions. Farming systems in these zones are largely based on drought-tolerant crops such as maize, beans, green grams, cowpeas, mangoes and vegetables. Primary data for this study were obtained during the 2025 cropping season as part of academic research at the University of Embu.
 
Research design, sample size and sampling technique
 
The study utilized a cross-sectional survey approach and used a multistage stratified random sampling method to determine sample size for data collection. The sample size was estimated to a minimum of 384 farming households using Cochran formulae. To enhance representation and adequately address the research questions, the study increased the sample to 414 households. In first stage, Makueni County was purposively selected due to its classification as an ASAL region that is highly vulnerable to climate change and variability, yet the population heavily depend on rain fed agriculture as the main source of livelihood.  In second stage, two sub-counties; Kibwezi West and Makueni were randomly selected from the three sub-counties located in the dryland zones of the county. In third stage, two wards were randomly chosen from each selected sub-county, giving a total of four wards. The fourth stage involved randomly selecting two locations from each ward, resulting in eight locations and in the fifth stage, one sub-location was randomly picked from each location, giving a total of eight sub-locations. A probability proportionate to size sampling procedure was used to calculate the number of farming households to be interviewed in each sub-location, based on a sample frame obtained from the Ward Agricultural Offices. Finally, the proportional-to-size approach was used by taking the number of farming households in each selected sub-location, dividing it by the total households across all eight sub-locations and then multiplying by the overall sample size to determine how many households to sample from each sub-location as shown below:

 
Where, 
M= Total number of farming households to be interviewed in each Sub location. 
n=  Total number of farming households in each location.
N=  Total number of farming households from the eight locations.
 
Instruments of data collection
 
This study employed a semi-structured questionnaire programmed in Kobo Toolbox to collect information from farming households. The questionnaire centered on the adoption of SFETs and examined farmers’ knowledge, attitudes and perceptions influencing the intensity of adoption. The technologies assessed included compost manure, farmyard manure, inorganic fertilizer, micro-nutrient supplementation, crop rotation, use of cover crops, mulching, irrigation, zero tillage, soil testing and mapping, legume intercropping, crop residue management, zaipits and terracing.

Model specification
 
Heckman two-stage selection model was employed to evaluate farmers’ KAPs on adoption intensity of soil fertility enhancement technologies. This model addresses the issue of selection bias by acknowledging the potential interdependence of two parts of the model such as probability to utilize a technology and intensity of adoption (Degaga et al., 2020). In the first stage, probit model was run to determine the decision to adopt technologies as expressed in equation below:
 

                             Y_i=aX_i+…+aX_n+ ε                          ...2

 
Where 
Y_i = Decision of a farmer to adopt SFET technologies. 
X_i = Vector of independent variables.
a = Vector of the parameter estimates hypothesized to effect chances of i^th farmer choosing the technologies.
ε = Error term.
       
Ordinary least squares (OLS) consistently estimated parameters in the second stage by including the Inverse Mills Ratio from the probit model as an additional explanatory variable as expressed below:
 

                                   Y_i= a₀+ a_iX_i + μ_iλ_i + v_i                              ...3

 
Where 
Y_i= Adoption intensity. 
X_i = Explanatory variables.
a₀ = A constant term. 
a_i = Parameter to be estimated. 
λ_i = Inverse mills ratio. 
μ_i = Correlation of first and second stage  error and  is error term in second stage.
       
In Heckman outcome model, the dependent variable was continuous and below is its specification:
 

                                      Y_i= ϕX_i+ … + ϕX_n + μ_i                            ...4

 
Where
Y_i= Land area under SFET technologies.
X_i … Xn = Independent variables.
ϕ = Vector of parameter estimates of independent variables.
μ_i = Error terms.

Distribution of farmers by land size (Acres)
 
Table 1 shows the distribution of farmers by the size of land they owned. Majority of farmers (47.3%) owned 2 acres or less, while 44.4% owned between 2.1 and 4 acres. A smaller proportion of farmers owned larger farms, with 6.8% holding 4.1-6 acres and only 1.4% holding more than 6 acres. These results indicate that the majority of farmers in the study area operated on small landholdings.


 
Knowledge, attitude and perceptions on adoption of soil fertility enhancement technologies
 
Fig 1 and Fig 2 show the respondents’ knowledge, attitudes and perceptions concerning adoption of SFETs. The findings reveal a reasonable level of awareness, with 45% of respondents agreeing and 14% strongly agreeing that they are aware of SFETs. Furthermore, 50% of respondents indicated that they understood how SFETs worked. However, a sizable proportion remained neutral on this subject, indicating the presence of a knowledge gap in the research sample.  Respondents expressed optimism about the technologies’ acceptance. For example, 55% perceived SFETs as useful in fostering sustainable farming methods and 61% were willing to use them. Further, 58% of the participants expressed a willingness to invest in these technologies, with 27% strongly agreeing.




       
The function of agricultural extension services in promoting SFETs elicited conflicting opinions. While 11% of respondents stated that extension services actively promote SFETs, a greater proportion (26%) strongly disagreed. Similarly, although 45% agreed that information supplied through agricultural extension methods was beneficial in promoting these technologies, 24% were indifferent. This emphasizes the need to improve the quality and relevancy of extension communication. Participation in extension initiatives also found to affect adoption rates. Approximately 41% of respondents believed that increased engagement in extension services improved SFET adoption. However, 28% were neutral and 12% disagreed, demonstrating that the perceived advantages of engagement are not shared equally across the community.

Perceptions on the labor needs of SFETs were split. While 55% stated that SFET deployment necessitated extensive work, 45% disagreed. This implies that, while many people believe the technologies are labor-intensive, quite a few of respondents do not see labor demand as a big constraint especially if adequate support structures are in place.
       
A greater proportion (83%) of respondents acknowledged that SFETs have benefits over other techniques. This high degree of awareness has the potential to be very effective in stimulating adoption, especially if it translates into practical interest and behavioral change. However, the notion of high prices remains a key hurdle, with just 20% of respondents perceiving SFETs as cost-effective and 80% disagreeing. Land availability was also cited as a limiting factor, with 62% of respondents asserting that limited land hinders the use of SFETs. In comparison, 38% did not perceive land as an impediment. Concerning extension support, only 37% believed that such services encourage the adoption of SFETs, while 63% disagreed, underlining the current gap in the supply or accessibility of support services. While 92% of respondents felt that SFETs are typically accessible, this contradicts previous findings demonstrating restricted accessibility and insufficient extension support. The ambiguity implies that, while awareness is high, real availability and institutional support may be minimal to guarantee broad and successful adoption of SFETs in the region.
 
Determinants of the adoption of SFET
 
The results in Table 2 show the factors of SFET technology adoption as estimated by the probit regression model. The findings show that farmers’ perceptions on labor requirements had a significant positive impact on technology adoption (p=0.005). This favorable perception of labor needs increased the likelihood of adoption by approximately 83%. This indicates that when farmers perceive favourable returns to their labour investment they are more likely to use these technologies. The findings were consistent with research by Mwaura (2021), who discovered that availability to labor had significant effects on both the adoption and intensity of application of soil organic amendments.


       
Farmers’ willingness to commit resources was used to assess their attitudes about adoption. Although this variable was statistically significant (p=0.041), the association was negative, indicating that increasing willingness to devote resources leads to a 39% drop in adoption. This finding suggests that, while households exhibit a desire to deploy resources, this willingness may be accompanied with increased knowledge or anxiety about the possible financial burden and the uncertainty around the return on investment. Farming households may expect large initial expenses, delayed rewards, or danger of loss, which discourages real adoption behavior, despite their stated intention. This finding is consistent with that of Engelberts et al. (2021), who observed that the long-term advantages of soil fertility treatments might lower immediate adoption.
       
Adoption of SFETs was significantly positively correlated with attitudes towards access to agricultural extension services (p=0.001). A positive attitude regarding extension access was linked to an 86% increase in the chance of adoption. This highlights the critical role that extension services play in promoting the spread of agricultural technologies. Extension services provide a platform for capacity building through farmer training, timely information distribution and individualized assistance, all of which are essential in improving farmers’ decision-making processes around technology adoption. Similarly, Kiprotich et al. (2024) discovered that access to extension services had a substantial impact on the adoption of integrated soil fertility management strategies by sorghum farmers.
 
Indicators of Adoption intensity of SFET
 
Table 3 presents the results on the influence of knowledge, attitudes and perceptions on the intensity of adoption of SFETs. The findings on perceptions of cost-effectiveness exhibited a statistically significant and positive effect on intensity of adoption of SFETs (p=0.027). Specifically, when farmers perceived SFETs to be economically viable, the likelihood of adoption intensity increased by 82%. This research emphasizes the importance of economic concerns in determining farmers’ technology adoption behavior. Farmers are more likely to devote resources and integrate a technology into their production systems if they feel it will provide considerable financial benefits, whether through lower input costs, higher yields, or better long-term profitability. This shows that, beyond knowledge, the perception of apparent economic gain is a powerful driver for increasing technology use. As a result, marketing SFETs should include both technical training and clear communication of economic benefits, such as evidence of input reductions, increased soil productivity and long-term return on investment. This finding affirms that of Wang (2022) which showed that farmers were far more inclined to accept and expand their usage of agricultural technology when they envisaged a profit gain of at least 5%, but cost-effectiveness uncertainty served as a barrier.


       
Attitude toward access to agricultural extension services was found to have a statistically significant and positive effect on extent of adoption of SFETs (p=0.012). Specifically, a favorable attitude toward extension access was associated with a 79% increase in adoption intensity. The provision of hands-on assistance and regular follow-up interactions enhances farmers’ engagement with the technologies and contributes to long-term adoption outcomes. These findings concurred with Mburu et al. (2024), who indicated that improved access to extension services significantly enhanced farmers’ learning and, consequently, increased both the uptake and intensity of utilization of soil and water conservation practices.
       
Perception of the benefits associated with SFETs showed a statistically significant influence on adoption intensity (p=0.064). The results indicate that a favorable perception of SFET benefits may increase adoption intensity by approximately 39%. It implies that farmers tend to increase the application of SFETs when they identify clear benefits, including better soil productivity, higher yields, or greater environmental sustainability. Adoption intensity indicates a more developed phase of technology integration, frequently necessitating a higher allocation of resources and ongoing dedication. This agrees with (Mutungi et al., 2025) who indicated that the perceptions held by farmers regarding the benefits of agricultural technologies played a crucial role in both the initial adoption and ongoing utilization of these technologies.

Adoption and intensity of SFETs were significantly influenced by farmers’ perceptions of labor requirements, willingness to invest, awareness of benefits and access to extension services. To improve uptake, it is vital to promote labor-efficient technologies, reduce cost barriers through financial support or demonstrations of long-term gains and enhance awareness through targeted education. Strengthening extension services is important for the provision of essential training and technical support to farming communities. Agricultural extension agencies and relevant stakeholders should intensify efforts to sensitize farmers on SFETs, shaping positive beliefs and attitudes while highlighting their benefits for soil fertility, crop productivity and climate resilience. Farmers should also be educated on SFETs that help regulate water regimes, such as mulching, irrigation, terracing, crop residue management and cover crops, since adequate soil moisture is critical for nutrient uptake and crop performance in drylands. Policymakers should integrate soil fertility into agricultural strategies, offer smart subsidies or affordable credit and improve extension delivery using digital platforms and private sector collaboration.

The present study did not receive any funding.
 
Disclaimers
 
The views and conclusions presented in this work are solely those of the authors and do not reflect the positions of their affiliated institutions. The authors take responsibility for the accuracy and completeness of the information provided. While the authors are dedicated to correcting any errors that may exist, they disclaim any liability for direct or indirect damages that may arise from use of this information.
 
Informed consent
 
Participation in the study was completely voluntary and respondents were under no pressure or obligation to take part.

The authors declare that they have no competing interests regarding this publication. The research was carried out independently, without financial assistance or sponsorship. No external funding played any role in shaping the study’s design, data gathering, analysis, publication decisions, or the writing of the manuscript.