Cogent Social Sciences ISSN: 2331-1886 (Online) Journal homepage: www.tandfonline.com/journals/oass20 Motivations and incentives for mechanization in Zambia: a mixed-methods analysis Kelvin Mulungu, Lushomo Molly Ngoma , Rumbidzai Mhembere , Mark Manyanga , Esau Simutowe , Christian Lutz Thierfelder , Md. Abdul Matin & Hambulo Ngoma To cite this article: Kelvin Mulungu, Lushomo Molly Ngoma , Rumbidzai Mhembere , Mark Manyanga , Esau Simutowe , Christian Lutz Thierfelder , Md. Abdul Matin & Hambulo Ngoma (2025) Motivations and incentives for mechanization in Zambia: a mixed-methods analysis, Cogent Social Sciences, 11:1, 2588017, DOI: 10.1080/23311886.2025.2588017 To link to this article: https://doi.org/10.1080/23311886.2025.2588017 © 2025 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group Published online: 02 Dec 2025. Submit your article to this journal Article views: 26 View related articles View Crossmark data Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=oass20 https://www.tandfonline.com/journals/oass20?src=pdf https://www.tandfonline.com/action/showCitFormats?doi=10.1080/23311886.2025.2588017 https://doi.org/10.1080/23311886.2025.2588017 https://www.tandfonline.com/action/authorSubmission?journalCode=oass20&show=instructions&src=pdf https://www.tandfonline.com/action/authorSubmission?journalCode=oass20&show=instructions&src=pdf https://www.tandfonline.com/doi/mlt/10.1080/23311886.2025.2588017?src=pdf https://www.tandfonline.com/doi/mlt/10.1080/23311886.2025.2588017?src=pdf http://crossmark.crossref.org/dialog/?doi=10.1080/23311886.2025.2588017&domain=pdf&date_stamp=02%20Dec%202025 http://crossmark.crossref.org/dialog/?doi=10.1080/23311886.2025.2588017&domain=pdf&date_stamp=02%20Dec%202025 https://www.tandfonline.com/action/journalInformation?journalCode=oass20 African Studies  | R esearch Article Cogent Social Sciences 2025, VOL. 11, NO. 1, 2588017 Motivations and incentives for mechanization in Zambia: a mixed-methods analysis Kelvin Mulungua , Lushomo Molly Ngomaa, Rumbidzai Mhembereb, Mark Manyangab, Esau Simutowea, Christian Lutz Thierfelderb, Md. Abdul Matinb and Hambulo Ngomab aCIMMYT Zambia, C/O IITA Campus, Lusaka, Zambia; bCIMMYT Zimbabwe, Harare, Zimbabwe ABSTRACT Smallholder farmers in Zambia face low agricultural productivity, and while tractor mechanization offers a solution, adoption rates remain low. The contextual factors driving this low uptake, including farmers’ preferences for different ownership and financing models, are not well understood. This study uses a mixed-methods approach, including surveys with 208 farmers, 18 focus group discussions (FGDs) and 28 key informant interviews (KIIs) across three districts in Zambia to examine these issues. The analysis shows that the motivations for tractor ownership are centred around both productivity enhancement and income generation through service provision, with farmers valuing tractors’ ability to improve operational timeliness given narrowing planting windows. The results reveal a clear divergence in ownership preferences. Individual ownership is favoured by male farmers and those in more mechanized districts seeking operational autonomy, while group ownership is preferred by female farmers and those in less-mechanized areas for its cost-sharing and risk-management benefits. In terms of incentives for ownership, risk-contingent credit (RCC), particularly when bundled with repair insurance, emerges as the most desirable incentive for encouraging tractor ownership. This highlights the need to de-risk mechanization investments. Key barriers to adoption include high maintenance costs, poor access to spare parts and the financial strains due to social obligations. These results demonstrate that a ‘one-size-fits-all’ approach to promoting mechanization is ineffective. Instead, successful interventions must be flexible, gender-responsive and tailored to the local context. Such an approach would likely increase ownership and improve livelihoods. 1.  Introduction Agricultural productivity among smallholder farmers in Zambia, as in many African countries, remains persistently low, presenting a significant challenge to rural development and food security (Adu-Baffour et  al., 2019). This low productivity has gained renewed urgency given the continent’s growing population pressure, land scarcity and labour constraints in the agricultural sector (Hamilton et  al., 2022; Kirui et  al., 2023). Recent studies document growing labour constraints in African agriculture, with climate change contributing significantly to the reduction in agricultural productivity growth in sub-Saharan Africa since 1961, driving unprecedented rural–urban migration and creating labour shortages during peak farming periods (Tietjen et  al., 2023). Rising rural wages and seasonal labour shortages further compound these challenges (Kotu et  al., 2017; Simutowe et  al., 2023). Against this backdrop, agricultural mechanization has emerged as a promising pathway for transform- ing smallholder farming systems, enhancing efficiency and improving rural livelihoods (Daum & Birner, 2020). Agricultural mechanisation encompasses a comprehensive spectrum of farming technologies, from basic implements to sophisticated motorised equipment that harness human, animal or mechanical power sources (Sims & Kienzle, 2016). This range of technologies offers farmers varying levels of acces- sibility based on their specific contexts and needs, with the potential to address multiple challenges at © 2025 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group CONTACT Kelvin Mulungu k.mulungu@cgiar.org CIMMYT Zambia, C/O IITA Campus, Ngwerere Road, Lusaka, Zambia https://doi.org/10.1080/23311886.2025.2588017 This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The terms on which this article has been published allow the posting of the Accepted Manuscript in a repository by the author(s) or with their consent. ARTICLE HISTORY Received 13 March 2025 Revised 15 October 2025 Accepted 7 November 2025 KEYWORDS Agricultural mechanization; ownership models; gender dynamics; risk management; smallholder farmers; Zambia SUBJECTS African Studies; Gender Studies; Rural Development http://orcid.org/0000-0002-4904-4122 http://orcid.org/0000-0002-7050-9625 mailto:k.mulungu@cgiar.org https://doi.org/10.1080/23311886.2025.2588017 http://creativecommons.org/licenses/by/4.0/ http://crossmark.crossref.org/dialog/?doi=10.1080/23311886.2025.2588017&domain=pdf&date_stamp=2025-12-1 2 K. MULUNGU ET AL. once: reducing the physical burden of farming, improving the timeliness of agricultural operations, expanding cultivated areas and enhancing overall farm profitability (Kirui & von Braun, 2018). Tractors represent a particularly transformative technology for smallholder agriculture, offering the potential to significantly expand land cultivation capacity, reduce the physical demands of farming and improve the timing of agricultural operations (Simutowe et  al., 2023). This timing aspect is particularly important in rain-fed agricultural systems, where the window for optimal planting and other field oper- ations can be narrow. The ability to perform timely pre-rain cultivation and respond quickly to weather patterns can substantially improve yields and reduce crop losses. Additionally, tractors offer versatility for multiple farming operations and the potential for income generation through service provision to other farmers, creating dual benefits for ownership. Therefore, in this study, we focus on tractors as proxies for mechanisation. However, despite these clear benefits, tractor adoption rates among smallholder farmers in Zambia remain relatively low, creating persistent barriers to agricultural development (Middelberg, 2017). This contradiction between the evident advantages of mechanisation and its limited uptake reflects the com- plex nature of technology adoption decisions in smallholder farming contexts. A recent behavioural anal- ysis by Omulo et  al. (2024) highlights that Zambian farmers’ attitudes significantly influence their willingness to adopt mechanised farming practices, suggesting that adoption barriers extend beyond simple economic constraints to encompass social, cultural and institutional factors. Previous research on agricultural mechanisation has often focused on quantifying the willingness to pay for tractors, either at the individual (Ngoma et  al., 2023) or group level (Kotu et  al., 2023), without fully exploring the underlying preferences for different ownership models or the complex motivations driving these preferences (Takeshima et  al., 2013). This gap is particularly pronounced in understanding when and why different ownership models succeed or fail. While group ownership is often recommended for smallholder mechanisation in sub-Saharan Africa (Sims & Kienzle, 2016), evidence suggests that indi- vidual approaches may sometimes be more effective (Abebaw & Haile, 2013). Studies have estimated farmers’ willingness to pay for tractor services and purchase tractors using various methodologies, includ- ing contingent valuation, choice experiments and experimental auctions (Diao et  al., 2014; Takele & Selassie, 2018). However, these approaches, while valuable for pricing and market assessment, often overlook the fundamental question of whether farmers prefer individual or collective ownership models and, more importantly, why they hold these preferences (Hamilton et  al., 2022). This gap in understand- ing baseline preferences and motivations is non-trivial as they can significantly influence the success or failure of mechanisation initiatives, regardless of pricing or financing mechanisms (Lu et  al., 2022; Omulo et  al., 2024). In addition to ownership models, understanding of incentives that can encourage ownership whether at the group or individual level is not well studied. While incentive design is well-studied for divisible, low-cost technologies like seeds or fertiliser (Carter et  al., 2021; Liverpool-Tasie, 2014), limited research exists on incentive preferences for capital-intensive mechanisation investments, particularly incentives that account for risk and the impact of social obligations on investment capacity, factors that are rarely considered in standard mechanisation financing models despite their documented importance in rural African contexts (Haggblade et  al., 2004). Empirically, this study addresses an important gap by shifting the focus from demand estimation (i.e. willingness to pay) to a foundational analysis of ownership model preferences (individual versus group) and the underlying motivations that shape them. Our primary contribution is the explicit examination of how these preferences are influenced by gender, the local mechanisation context and socioeconomic factors within Zambia, dimensions that have been largely overlooked in previous mechanisation research. We frame the choice between individual and group ownership not merely as a financial decision but as a calculated trade-off between the desire for operational autonomy and the benefits of shared costs and collective-risk management. Furthermore, by analysing preferences for incentives such as risk-contingent credit (RCC), our study applies a behavioural economics lens, showing that investment decisions are driven as much by farmers’ perceptions of risk as by their access to resources. This framing allows for a more nuanced understanding of adoption barriers than is possible based purely on economic models. This study focuses on three distinct districts in Zambia (Luano, Kalomo and Monze), selected for their varying levels of agricultural development and tractor adoption rates. Luano, characterized by its pre- dominantly rural setting, exhibits high tractor ownership levels, while Monze demonstrates lower rates Cogent Social Sciences 3 of tractor adoption, providing valuable comparative context. Kalomo represents the middle ground, with large farms and medium tractor ownership rates. This geographical diversity enables a more comprehen- sive understanding of the factors influencing tractor ownership and adoption patterns across different agricultural contexts in Zambia. This study aims to address several important questions surrounding mechanization in Zambian agricul- ture: (i) What motivates smallholder farmers to invest in tractor ownership? (ii) What factors influence farm- ers’ preferences between individual and group ownership models? (iii) Which incentives (traditional loans, risk-contingency loans, risk-contingent loans and repair services, repair services, assurance of customers, or combinations of these) can effectively promote tractor adoption among smallholder farmers? (iv) What are the primary barriers to tractor ownership? By employing a mixed-methods approach, this study seeks to provide nuanced insights into these questions, contributing to our understanding of mechanisation adop- tion in smallholder farming contexts. The findings will be particularly relevant for stakeholders working to promote sustainable agricultural intensification and improve smallholder farmers’ access to mechanical power and machinery, including efforts to reduce drudgery and enhance farm productivity under rain-fed agricultural conditions in sub-Saharan Africa. 2.  Literature review 2.1.  Mechanization adoption in developing countries Agricultural mechanisation is widely regarded as a critical pathway for transforming smallholder farming systems in developing countries (Lu et  al., 2022). Its adoption is consistently linked to increased agricul- tural productivity, improved rural livelihoods and enhanced food security outcomes (Baudron et  al., 2015; Cele, 2021; Lu et  al., 2022; Van Loon et  al., 2020). Supporting this, studies have shown that countries that invest significantly in mechanisation experience marked improvements in agricultural output and effi- ciency, allowing farmers to increase both the area of land they cultivate and their overall productivity (Pingali, 2007). However, this optimistic view is countered by a significant paradox: despite clear benefits, adoption rates in many parts of Africa, including Zambia, remain persistently low (Ngoma, Marenya, et  al., 2023; Ngoma, Simutowe, et  al., 2023). A case study in Zambia by Adu-Baffour et  al. (2019) raises concerns that mechanisation may cause rural unemployment, highlighting contradictory outcomes that depend on the local context. This contradiction reflects the complex reality that mechanisation benefits are not auto- matic but depend on the implementation approaches and local conditions. The level of mechanisation adoption varies considerably across different contexts, influenced by access to technology, local agricultural conditions, economic circumstances and farmers’ technical capacities. Research specific to the Zambian context by Ngoma, Marenya, et  al. (2023) and Ngoma, Simutowe, et  al. (2023) on scaling smallholder mechanisation in Zambia, Zimbabwe and Malawi identifies key barriers, including a lack of finance for purchasing and hiring equipment, inadequate knowledge about mechani- sation benefits, and limited access to equipment. These findings are particularly relevant to understand- ing the Zambian mechanisation landscape, where Middelberg (2017) notes that few private tractors are used for ploughing and haulage among smallholders, with the majority of farmers relying on hired oxen or manual labour for land preparation. The adoption of agricultural technologies by smallholder farmers is shaped by economic, social and institutional factors (Onomu & Aliber, 2021). Economic factors, particularly household income and access to credit, often serve as the primary determinants of adoption capacity. However, social factors such as education level, farming experience and membership in farmer organisations also play crucial roles in the adoption of technology (Tesema et  al., 2023). Gender dynamics significantly influence technology adoption. This gender dimension is particularly important in the Zambian context, where Kirui (2019) suggests increasing mechanisation initiatives and training that specifically targets women and youth. Despite these factors, mechanisation adoption remains limited across Africa, with only approximately 18% of farmers using motor-powered machinery compared to light machinery and animal-powered equipment (Kirui, 2019). This pattern is evident in Zambia, where traditional farming methods predomi- nate among smallholders, creating opportunities for targeted mechanisation. 4 K. MULUNGU ET AL. 2.2.  Group versus individual ownership models in agricultural technology The debate between group and individual ownership models (Daum, 2023; Tufa et  al., 2024) for agricultural technology reveals conflicting evidence about optimal approaches. Farmer groups have been increasingly recognised as potential catalysts for technology adoption, particularly in areas where individual resource constraints limit access to agricultural technologies (Wossen et  al., 2017). Sims and Kienzle (2017) specifi- cally recommended group ownership and custom hire service provision as appropriate models for making mechanisation accessible to smallholder farmers in sub-Saharan Africa. For example, Fischer et  al. (2018) found that the benefits of mechanisation depend on gendered patterns of labour allocation, and advo- cated for more flexible finance schemes to encourage women and youth to engage in mechanisation and ensure equitable use. These collective arrangements can facilitate improved agricultural productivity, pro- mote commercialisation and enhance market linkages through collective marketing efforts. However, group-based approaches face significant challenges in their implementation. Mulungu et  al. (2017) found that actual membership in farmer groups usually burdens one group (mostly the minority) over the other, suggesting barriers to collective action that, if unaddressed, could undermine group own- ership success. This concern is supported by Abebaw and Haile (2013), who challenge conventional wis- dom about group benefits by finding that individual adoption rates for improved agricultural technologies sometimes exceed those of group-based approaches. These conflicting findings suggest that the success of an ownership model is not inherent but highly dependent on institutional design and local context. As Bernard and Spielman (2009) argue, the perfor- mance of farmer organisations often hinges on factors such as internal governance, trust among mem- bers and the tangible economic benefits they provide. The mere presence of a group does not guarantee success. This reveals a critical gap in the literature: most studies focus on adoption rates rather than farmers’ underlying preferences for ownership models and the motivations that drive them. However, it is unclear when or why a particular model succeeds. Our study navigates this tension by directly inves- tigating how gender, local mechanisation levels and social structures influence farmer preferences between individual and group ownership in Zambia. 2.3.  Role of incentives in technology adoption Incentives are crucial for shaping technology adoption. Financial tools, such as credit, subsidies (Bopp et  al., 2019; Mohammed et  al., 2023), and other risk-sharing mechanisms, are often recommended to make mechanisation affordable for smallholders. However, adoption is not driven by financial calculations alone (Haggblade et  al., 2004; Ngoma et  al., 2020). As Duflo et  al. (2011) demonstrate in their work on Kenyan agriculture, even when technologies are profitable, behavioural biases such as procrastination and risk aversion can significantly impede adoption. This highlights the importance of non-monetary factors, such as social recognition and knowledge sharing (Shikuku & Melesse, 2020), which can also be powerful incentives for adopting new agricultural practices. The literature confirms the potential of mechanisation but highlights complex, context-specific adop- tion barriers. The review above reveals conflicting evidence regarding optimal ownership models and a limited understanding of how social and financial lives intertwine in rural mechanisation decisions. These gaps are particularly important given Zambia’s low mechanisation levels among smallholders (Middelberg, 2017) and the government’s interest in promoting agricultural mechanisation. Understanding farmers’ preferences for ownership models and incentive mechanisms before implementing interventions is essen- tial for designing mechanisation programmes appropriate to local social and economic contexts. 3.  Methodology 3.1.  Study area This study was conducted in three districts, Luano, Kalomo and Monze in Zambia. These were selected through a stratified sampling approach based on their varying levels of mechanisation adoption and par- ticipation in the CGIAR Diversification in East and Southern Africa Initative; Ukama Ustawi project. Ukama Ustawi was a CGIAR initiative aimed at address food and nutrition risks that arise from overreliance on Cogent Social Sciences 5 maize, through a climate-resilient, water-secure and socially inclusive approach. In Zambia, the initiative promoted mechanisation in various districts, including Monze, which is included in this study’s sample. In the Luano District, data were collected from four agricultural camps: Lunsemfwa, Chikupili, Matuka and Kaundula. In Kalomo District, research was conducted in two camps: Nantale and Nahila, whereas in Monze District, three camps were sampled: Chisuwo, Malende and Namakube. The selection of these sites followed a maximum variation sampling strategy (Patton, 2014), enabling the capture of diverse mechanisation con- texts within Zambia’s agricultural landscape. On average, the majority of farmers in all the districts are smallholder farmers, cultivating about a hectare of land, as shown in Table 1. The sampling framework prioritises areas with established cooperative structures and savings groups. The goal was to sample the most common groups in the community and understand whether there were differences in the way they perceived mechanisation ownership models. This sampling strategy was relevant given the study’s focus on comparing individual- and group-based ownership models. Participants for the key informant interviews (KIIs) were identified using purposive snowball sampling. Initial informants were identified with the help of local extension agents, and these individuals recom- mended other tractor owners in their communities. This method was effective in obtaining a represen- tative sample of local service providers. For the focus group discussions (FGDs), two FGDs were conducted in each camp: one with cooperative members and one with savings group members, with 10–12 partic- ipants each. Participant selection was performed by the extension officer, who was informed ahead of the schedule. The criteria were to select a good representation of women and men youth, and different scales of production. Overall, approximately 30% of the participants were women. Each FGD was led by a trained moderator using a semi-structured guide, with a separate note-taker recording the discussion and observing group dynamics. 3.2.  Study design and data collection This study employed a sequential mixed-methods design (Creswell, 2014), combining qualitative and quantitative approaches to provide a comprehensive understanding of mechanisation preferences and incentives. This methodological triangulation approach has been widely used in agricultural adoption studies to enhance the validity and reliability of findings. Data collection involved three primary instruments, all implemented simultaneously at around the same time. Integration during the analysis was achieved using qualitative tools for deeper contextual analysis and understanding of the quantitative results. Two instruments were used for qualitative data collection and one for quantitative data collection. 3.2.1.  Qualitative instruments i. KIIs: Semi-structured interviews were conducted with tractor owners at the study agricultural camps. A guide was used to implement the KIIs. These interviews focused on motivations for owning and Table 1.  Characteristics of the sampled households (participants in the FGDs) and number of KIIs. Luano (94) Monze (n  =  69) Kalomo (n  =  45) Overall (n  =  208) Mean SD Mean SD Mean SD Mean SD Prefers group model (1/0) 0.32 0.47 0.72 0.45 0.62 0.49 0.51 0.50 Age (years) 44.86 10.62 51.41 13.70 49.11 14.15 47.95 12.79 Gender (1= Female) 0.19 0.40 0.43 0.50 0.22 0.42 0.28 0.45 Education level (years of schooling) 8.03 3.08 8.35 2.62 6.64 2.74 7.84 2.92 Household size (number) 8.39 4.35 6.86 2.59 7.51 3.14 7.69 3.64 Primary occupation as a farmer? (1/0) 0.87 0.34 0.94 0.24 1.00 0.00 0.92 0.27 Land owned (ha) 4.01 6.70 1.41 3.11 1.46 2.20 2.60 5.28 Rented a tractor before (1/0) 0.24 0.43 0.13 0.34 0.20 0.40 0.20 0.40 Number of tractors in the community 2.06 2.26 0.86 0.97 1.42 1.74 1.52 1.88 Number of cattle owned 7.32 12.83 6.96 9.24 6.87 6.32 7.10 10.51 Number of oxen owned 3.04 4.44 2.12 2.01 3.02 2.61 2.73 3.44 Experience in farming (years) 17.39 9.28 24.52 12.55 23.07 12.53 20.99 11.61 FGDs 8 6 4 18 KIIs (number) 17 4 8 28 SD is standard deviation. 6 K. MULUNGU ET AL. understanding operational dynamics, challenges and opportunities in tractor service provision. The goal of the KIIs was to help understand the mechanisation motivations of those who already own a tractor, the challenges they face in using the tractors, and their observed benefits and perceptions. ii. FGDs: Two FGDs were conducted in each camp, one with cooperative members and another with savings group members – with 10–12 participants per group. The FGD protocol was designed fol- lowing Krueger (1988) guidelines for focus group methodology, with particular attention to gender representation, especially in savings groups, where female membership typically predominates. Here, the discussion focuses on ownership models, ranking of incentives and understanding tractor own- ership and the hiring landscape. To understand the preferences for different financial incentives that could promote tractor ownership, this study presented farmers with five options during FGDs and structured individual interviews: (i) traditional loan: a standard loan of ZMW 50,000 (~ USD 2000)1 with 15% annual interest, repayable over five years, regardless of harvest outcome; (ii) RCC: similar to traditional loans, risk-contingent credit includes insur- ance coverage that waives repayment obligations during poor harvest years for an annual premium of ZMW 2000; (iii) insurance for repairs: coverage of all repair costs for an annual premium of ZMW 2000; (iv) customer assurance: the tractor owner is guaranteed by customers who hire the mechanisation ser- vice; (v) service at your doorstep: free service at farm location with owners paying only for parts. These incentives were identified through consultations with key stakeholders including the Ministry of Agriculture, Conservation Farming Unit and Agricultural Leasing Company. Each of the five incen- tive options was written on a flipchart, and farmers ranked these incentives by discussing what they preferred the most during FGDs to determine preferences and discuss potential combinations that would be the most effective. 3.2.2.  Quantitative data For the quantitative data, structured questionnaires were administered to FGD participants from different households to collect demographic, socioeconomic and farm characteristics, as well as preferred owner- ship models and incentives (similar to those administered in the FGDs). The survey instrument was adapted from validated tools used in similar mechanisation studies (Takele & Selassie, 2018). Structured questionnaires were administered to FGD participants to collect individual household-level data that could not be effectively gathered through group discussions. While FGDs captured collective perspec- tives on mechanisation preferences, motivations, and barriers, the questionnaires collected quantitative data on demographic characteristics, asset ownership, farm size and socioeconomic indicators for each participating household. Furthermore, we asked the individual members which of the five incentives they preferred the most (stated in the qualitative section). This approach enabled us to (1) analyse how indi- vidual household characteristics influence preferences expressed during group discussions, (2) conduct regression analysis linking household factors to ownership model preferences and preferences for certain incentives and (3) triangulate qualitative insights from FGDs with quantitative household data. 3.3.  Data analysis This study utilised a concurrent mixed analysis approach (Onwuegbuzie and Combs, 2011) to integrate the qualitative and quantitative findings. Qualitative data from FGDs and KIIs were transcribed verbatim and analysed using a thematic analysis framework (Braun & Clarke, 2006). The qualitative analysis employed a hybrid approach combining large language model (LLM) assistance with traditional researcher-led analyses. We used Claude 3.5 Sonnet (Anthropic, 2024) for initial thematic coding of tran- scripts from FGDs and KIIs. The LLM coding process involved: (1) developing structured prompts that instructed the model to identify themes related to mechanisation motivations, ownership preferences, barriers, and incentive preferences while maintaining sensitivity to local context and agricultural termi- nology; (2) processing transcripts in batches while maintaining participant anonymity; and (3) generating initial codes and preliminary theme categorisation. The LLM outputs were systematically reviewed by the research team for contextual accuracy and cultural appropriateness in the Zambian agricultural context. This hybrid approach enhanced theme extraction while ensuring that nuanced local meanings and cul- tural contexts were preserved through human oversight and interpretation. Similar approaches have Cogent Social Sciences 7 been used in literature with the rise of LLMs (Dengel et  al., 2023; Jalali & Akhavan, 2024; Mathis et  al., 2024; Tai et  al., 2024; Xiao et  al., 2023). For quantitative analysis, descriptive statistics were computed to characterise the demographic and socioeconomic profiles of the sample population. To identify factors influencing incentive own- ership model preferences, we employed a linear probability model following the approach of Takeshima et al., (2013): Y i k i i = + +β β 0 X ε (1) where Y i represents the binary outcome variable (group ownership preference), X i is a vector of house- hold characteristics and ε i is the error term. The list of variables included in the model is summarised in Table 1, with their definitions, mean and standard deviation. These variables were selected based on the literature and availability based on our questionnaire. We also estimated a similar model to understand preferences regarding the type of financial incentive. For robustness and given the challenges of the linear probability model (LPM), where probabilities less than zero or greater than one may be estimated, we also use the probit model and present the marginal effects (Wooldridge, 2010). Overall, we find that the results of the two approaches are comparable; hence, we maintain the linear probability models. Stata version 18 (StataCorp, College Station, TX) was used for all analyses. In all models, we interpret the marginal effects as correlations and not as causal, as we cannot eliminate endogeneity. We used the Huber–White robust standard errors to account for any potential heteroscedasticity. In the checks for multicollinearity, no two variables were highly multicollinear. The integration of qualitative and quantitative findings followed a parallel mixed data analysis approach (Guest, 2013), enabling triangulation of findings and development of more nuanced insights into mechani- sation preferences and adoption patterns. This mixed-methods approach allows for both generalisation of findings through quantitative analysis and deep contextual understanding provided by qualitative insights (Greene, 2007). The combination of these methods strengthens the validity of the findings through meth- odological triangulation, while providing rich contextual data to inform policy recommendations. 3.4.  Ethical clearance The study was approved by the ethics committee prior to commencement. Ethical clearance was approved by the International Maize and Wheat Improvement Centre (CIMMYT) Internal Research Ethics Committee (IREC), application number 2024.015. In addition, because the study involved human par- ticipants, informed verbal consent for participation in the study was obtained before the commence- ment of the survey. Verbal consent is one of the two main ways to obtain consent from the participants and is deemed more culturally appropriate and ethically sound than written consent for several rea- sons: (1) many participants had limited formal education and were unaccustomed to signing formal documents, which could create anxiety and undermine trust-building essential for free participation in the FGDs (Abay et  al., 2016; Noë et  al., 2025); (2) in anthropological and social contexts, formal legal-looking documents may intimidate participants and be inconsistent with local communication patterns that emphasise oral agreements (The American Anthropological Association, 2012); and (3) written consent forms can present barriers to meaningful participation among communities where lit- eracy levels are low and signing documents may raise concerns about potential future use of signa- tures (Molyneux & Geissler, 2008). 4.  Results and discussion 4.1.  Sample characteristics The study sample comprised 208 smallholder farmers across three districts, with Luano District contrib- uting the largest share (45%, n  =  94), followed by Monze (33%, n  =  69) and Kalomo (22%, n  =  45). More respondents preferred the group ownership model in Monze (72%) followed by Kalomo (62%) and lastly in Luano. The average participant was 48  years old with nearly 21  years of farming experience, indicating a sample of experienced agricultural practitioners. The sample was predominantly male (72%), with 8 K. MULUNGU ET AL. notable variation across districts: Monze had the highest proportion of female participants (43%), com- pared to Luano (19%) and Kalomo (22%). The average education level was 7.8  years of schooling, with participants in Luano and Monze having slightly higher educational attainment than those in Kalomo. Farming was the primary occupation for 92% of the participants, with all participants in Kalomo being full-time farmers compared to 87% in Luano and 94% in Monze. Significant differences in asset ownership and farm characteristics were observed across the districts, reflecting varying levels of agricultural development and mechanisation. Luano District farmers owned con- siderably larger pieces of land on average (4.01 hectares) compared to Monze (1.41 hectares) and Kalomo (1.46 hectares), consistent with Luano’s higher mechanization levels. The number of tractors in communities varied substantially, with Luano averaging 2.06 tractors per community compared to 0.86 in Monze and 1.42 in Kalomo, supporting the district’s characterisation as having the highest level of mechanisation. Livestock ownership was relatively consistent across the districts, with farmers owning an average of 7.1 cattle and 2.7 oxen. Previous tractor rental experience was most common in Luano (24%), compared to Monze (13%) and Kalomo (20%), suggesting greater familiarity with mechanisation services in the more mechanised dis- trict. These characteristics reflect the study’s successful sampling strategy for capturing diverse mechanisa- tion contexts, from high adoption (Luano) to moderate (Kalomo) and low adoption (Monze) areas. In total, this sample came from 18 FGDs, with Luano having 8, Monze 6 and Kalomo 4. For the KIIs, which were farmers who owned tractors, there were 17 interviews in Luano, 4 in Monze and 6 in Kalomo. Unfortunately, we did not collect demographic data on the key informants. 4.2.  Motivations for and tractor ownership patterns Among the 208 participants, 24 (11.5%) owned tractors, with ownership heavily concentrated in Luano District (87.5%, n  =  21) compared to Monze (8.3%, n  =  2) and Kalomo (4.2%, n  =  1). Tractor owners were predominantly male (87.5%, n  =  21) with a mean age of 47  years. Two-wheel tractors existed exclusively in Luano (n  =  8), while four-wheel tractors showed broader distribution (Luano: n  =  13, Monze: n  =  2, Kalomo: n  =  1). Among the non-owners, 41 farmers (19.7%) hired tractors, with four-wheel tractors strongly preferred (39 farmers) over two-wheel tractors (two farmers) (see Table 1). As a Lunsemfwa camp FGD participants explained: The four-wheeled tractor is the most efficient but there aren’t many around here… it saves time and makes it possible for early sowing so as to be right on time with the rains”. The primary motivation for tractor ownership centres on operational efficiency in response to climate variability. A Namakube tractor owner stated: What motivated me to purchase the tractor and the other equipment I mentioned was that work was difficult to finish, but after I got the tractor and the other equipment, it has become easier. This efficiency enables scale expansion; for example, a Choma owner doubled his cultivation from 20 to 40 hectares within 2  years of acquisition. Farmers consistently empha- sized timing advantages: “by the time you finish ripping with cattle rains would barely still be there, on the other hand with a tractor you would rip and plant a 50 kg bag in just a day. Disease-related livestock losses are a critical push factor. Multiple farmers transitioned to tractors after losing draft animals: “What made me buy a tractor is being a maize farmer, I had challenges because of the low rain fall we were experiencing… all my cattle had died so I decided to buy a tractor. This risk mitigation aspect distinguishes Zambian mechanisation from conventional intensification. Tractors serve dual purposes as both farming tools and as income-generating assets. One owner explained: “I use it for my own farm activities with ease, and when it is booked, I make money”. Service provision proves particularly profitable with appropriate equipment: “when it is time for shelling and people have harvested very well, we have a lot of business because in 45 minutes the sheller can shell about 40 x 50 kg bags [2 metric tons] of maize”. However, equipment gaps limit income potential; four of five Kalomo tractor owners lacked shellers despite high demand. These findings reveal that mechanisation adoption in Zambia operates through a dual-purpose invest- ment logic, where tractors serve simultaneously as productivity tools and business assets, explaining why timing efficiency rather than yield gains drives adoption. The prominence of livestock disease as a trigger suggests that mechanisation functions as risk diversification rather than simple intensification. Group ownership preferences correlate inversely with local mechanisation levels, indicating that it serves as a transitional mechanism rather than a permanent preference. Cogent Social Sciences 9 4.3.  Ownership preferences Ownership preferences exhibit strong geographic and demographic patterns (Figure 1). In highly mech- anised Luano, farmers overwhelmingly preferred individual ownership for operational autonomy: “I prefer individual [ownership]. Group setups are tricky when it comes to time management and use patterns may collide… the tractor may not be readily accessible at any time”. Conversely, Southern Province farm- ers favored group ownership – 72% in Monze (n  =  50) and 62% in Kalomo (n  =  28) (see Figure 1), viewing it as a stepping stone to individual ownership: “It would be easier as a group because we can put our little monies together…then afterwards we can start buying individually”. Table 2 quantifies these patterns. Farmers in Monze are 33.0 pp more likely to prefer group ownership than those in Luano (p  <  0.01), with Kalomo showing a 25.3 pp higher preference (p  <  0.01). These substantial geographic effects persisted even after controlling for individual characteristics, suggesting that local mechanisation levels and social networks strongly influence preferences beyond personal attributes. Female farmers are 19.6 pp more likely to prefer group ownership than males (p  <  0.05), reflecting both resource constraints and risk management strategies. Female participants emphasized: “Group own- ership allowed us to share costs of purchase, maintenance and repairs and repayment of loans making it more affordable than individual ownership”. Male farmers prioritized autonomy and profit: “freedom to use the tractor at my own time” and “would have more profit since I wouldn’t need to share it. This 12-percentage point gender gap (Figure 2) suggests that mechanisation programmes should offer flexi- ble models rather than assuming fixed gender preferences. Individual ownership advocates acknowl- edged maintenance challenges as one individual may not have enough resources if the repair costs are high: Maintenance may be costly as an individual…[in the end] the same tractor maybe hired out to raise funds for upkeep and maintenance”. Education and assets negatively predict preference for group ownership witheach additional year of schooling reducing preference for group ownership by 3 pp (p  <  0.01), while each hectare of land owned reduces it by 0.1 pp (p  <  0.05). Similarly, farming experience reduces preference for grouwp ownership by 0.8 pp per year (p  <    0.1). These findings align with those of Takele and Selassie (2018), who found that endowment positively influenced individual tractor use. Group ownership faces substantial operational challenges. Timing conflicts threaten productivity as mentioned in one FGD: “Maybe rains will stop and the ground will dry while others are still using the tractor before it’s my turn”. The likelihood of maintenance disputes among group members creating fric- tion also reduces the preference for the group ownership model: “If it develops a fault while I am using it and I tell the group members that we should buy another tyre, they can say that the tyre was dam- aged when I was using the tractor so I should buy a new one”. Figure 1.  Preference for the group or individual ownership models by district. The graph shows that in districts with higher mechanisation the individual model was more preferred while in Monze with less mechanisation the group ownership model was more preferred. The districts were selected based on levels. 10 K. MULUNGU ET AL. Successful groups require a detailed governance framework to reduce conflicts, and have better scheduling approach. Participants proposed: When it rains, we need to discuss as a body [group] to see who may use the tractor first until the entire group is catered for… The tractor is very powerful; there- fore, it would be unreasonable for one person to use it for an entire day. It needs to rotate among us in hours”. There were also equity concerns among the groups that proved important with one member stating that equal contributions should be made instead of unequal as those who contributed more would feel powerful and assume more user rights: “Equal [contributions]. If one contributed more than others, they may assume more rights over it [tractor]’. Table 2.  Regression of preference for group model with social and farm characteristics. Variables LPM Probit (marginal effects) Age (years) 0.005 0.005* (0.003) (0.003) Female farmer 0.196** 0.172** (0.075) (0.071) Education level (years of schooling) −0.030*** −0.029** (0.011) (0.011) Household size −0.009 −0.012 (0.010) (0.011) Primary occupation as a farmer? (1/0) −0.210 −0.157 (0.129) (0.101) Land owned (ha) −0.001** −0.003*** (0.001) (0.001) Rented a tractor before (1/0) −0.077 −0.076 (0.088) (0.079) Number of tractors in the community −0.017 −0.012 (0.016) (0.016) Number of cattle owned −0.001 −0.001 (0.004) (0.005) Number of oxen owned −0.002 −0.006 (0.013) (0.015) Experience in farming (years) −0.008* −0.008** (0.004) (0.004) Monze district 0.330*** 0.310*** (0.081) (0.078) Kalomo district 0.253*** 0.246*** (0.090) (0.082) Constant 0.827*** (0.200) Observations 208 208 R-squared 0.229 Robust standard errors in parentheses. The base category for the dependent variable is individual own- ership model preference. The reference district is Luano. Districts are coded as 0/1 for observations from that district. The R-squared is not reported for probit. ***p  <  0.01; **p  <  0.05; *p  <  0.1. Figure 2.  Preference of group vs individual ownership model by gender. The percentages are shown as a percent of the total within that specific gender. Cogent Social Sciences 11 How much each individual was willing to contribute (financial commitment) varies significantly accord- ing to ownership model preferences. Farmers favouring individual ownership would contribute ZMW 97,000 versus ZMW 21,000 for group advocates towards buying a tractor, a 4.6-fold difference (Figure 3). This gap, larger than that found by Kotu et  al. (2023) in similar contexts but for a different equipment, may reflect income disparities or free-riding intentions. The correlation between local mechanization lev- els and individual preferences suggests group ownership serves as a transitional mechanism towards individual ownership as mechanization exposure increases. This gender disparity in ownership preferences can be explained by several factors evident in the qualitative data. Female farmers’ preference for group ownership likely reflects risk-sharing strategies and resource-pooling benefits. These findings have important implications for policy design and implementation. They suggested that mechanisation programmes should offer flexible ownership models that accommodate gender-specific preferences and constraints rather than advocating for a single approach. Our findings suggest a mechanisation transition pathway unique to the Zambian context, in which individual ownership preferences correlate directly with mechanisation exposure levels. This pattern indi- cates that group ownership may serve as a stepping stone mechanism rather than a permanent prefer- ence, with farmers gravitating toward individual ownership as local mechanisation levels increase. The gender-based preference differences we observed reflect not just resource constraints but gendered risk management strategies; women’s preference for group ownership represents a rational response to exclusion from individual asset ownership rather than simply financial limitations. This suggests that mechanisation programmes that treat gender preferences as fixed may overlook the underlying struc- tural constraints that shape these preferences. 4.3.  Purchase models, incentives for tractor adoption and barriers The tractor market operates through distinct financing channels. Used tractors (costing between ZMW 47,000–250,000) typically require cash purchases, whereas new tractors (costing between ZMW 450,000– 600,000) necessitate loans. Farmers face trade-offs between capital preservation and equipment quality However, loan financing carries risks when equipment quality disappoints as noted by one key informant tractor owner: “You get a tractor on loan which does not do much [work] for you… the tractor I bought has been giving me problems. What makes it worse is that I got it on loan, and it keeps breaking down. I don’t make much money out of it, but I must still repay the loan. Figure 3. A mount willing to contribute towards a group ownership model by preference and district. The amount people are willing to contribute are lower as the levels of mechanization in the district reduces. Farmers who prefer the group model are willing to pay less – something which is a result of low income (21,000 versus 97,000 annual income) or attempting to free ride. The spikes show the 95% confidence interval. 12 K. MULUNGU ET AL. The interviews revealed nuanced perspectives on the advantages and disadvantages of the two financing approaches. Although cash purchases were seen as advantageous for avoiding interest pay- ments and ongoing financial pressure, loans were valued for allowing farmers to retain working capital for other farming operations. A farmer in Chifula explained, “With a loan, you can get a new machine, however, with cash I got a second-hand machine”. Another farmer noted, “[A] loan is better because you will remain with something [your own money] that you can use for other farm needs. Cash payment is tricky because you will need to use all your capital on one equipment and the rest of the farm activities will suffer”. Another emerging aspect after talking to tractor owners is that there is a lack of knowledge; once this is overcome, it seems that those who can ultimately do buy tractors. For example, three tractor owners interviewed in Choma bought tractors in the same year, and they all attributed their purchase decisions to the knowledge they received from the Chief, who had been conducting mechanisation campaigns in the village. One owner said, “I think there is lack of knowledge, because we also thought working with cattle is sufficient”. There is a lack of knowledge about what a tractor can do and how it can help with farming. Once this knowledge barrier is overcome, it seems to lead to more people who can afford to buy tractors. One owner gave an example of how most farmers own big pickup trucks ‘that are not helpful at the farm except for transport’ but not tractors, and the cost is almost the same. This study examined farmers’ preferences for five financial incentives to encourage tractor ownership (Figure 4). RCC was overwhelmingly preferred (Figure 4) because it addresses farmers’ primary concern about weather-related repayment risks. Similar to the farmer who got a loan and the tractor keeps breaking down and hence cant repay, farmers are scared of getting a loan and the weather turns out bad such that they cannot manage to make a profit and repay. Farmers valued protection from RCC, with one explaining, “This one [RCC] is even more favourable than traditional loans…we live in a time that is prone to droughts. When droughts hit, insurance steps in to repay the loan. This preference was rein- forced by recent (2023/2024 agricultural season) experiences in which tractor loan companies suspended repayments due to El-nino induced drought, highlighting the need to formalise such arrangements. Quantitative analysis (Table 3) supports this preference for comprehensive approaches: education level significantly increases the preference for the combined (RCC and insurance for breakdowns) package (3.1% points per additional year of education), while women show a significantly higher preference for repair insurance alone (9.8% points), likely reflecting greater risk aversion and limited access to repair services. Other factors singifcantly influencing preference for incentive type are age, with older farmer not prefrring either loan(traditional or RCC). Female farmers do not prefer repair at service at the door. Figure 4.  Most preferred financial incentives. Farmers were asked to rank these incentives to assist in buying a tractor including buying it without any financial assistance/incentive. Overall, risk contingent credit was the most preferred financial tool as most farmers felt it would give them relief in a bad year. The y-axis is the percent of farmers who prefer that given incentive. Cogent Social Sciences 13 However, an increase in number of tractors in the communities increases the likelihood of preferring traditional loans and reduces likelihood of preferring RCC, probably because others already bought suc- cessfully using traditional loans. 4.4.  Barriers to mechanization investment Social obligations emerge as important constraints on mechanisation investments. Farmers consistently reported funeral and wedding expenses diverting agricultural funds: “When I have a funeral, the money I kept for buying farming inputs will be diverted to funeral expenses thus reducing how much I will invest in my farming. This social embeddedness distinguishes the mechanisation challenges in Zambia from conventional models that assume individualistic decision-making. Information asymmetry compounds financial barriers. Farmers expressed uncertainty about accessing support on information about loans: “We just don’t know where we can do [access] these things because some of us really want to, we just don’t know where to insure our vehicles or fields”. Institutional intim- idation further limits access: “We are afraid of entering offices where we are supposed to get things but you who are used should inform us whenever there is new information. Knowledge barriers significantly constrain adoption. Farmers seem to have no knowledge of the adva- natges of using a tractor over cattle or other sources of power. Three Choma farmers who purchased tractors simultaneously attributed their decisions to mechanization campaigns by their Chief. One reflected: I think there is lack of knowledge, because we also thought working with cattle is sufficient. Table 3.  Factors associated with preference for certain financial incentives for tractor purchase. Variables No incentive Traditional loan Risk contingent credit (RCC) Insurance for breakdowns Repair service at your door RCC and insurance for breakdowns Age (years) 0.002 −0.005** −0.006* 0.003 0.001 0.004 (0.002) (0.002) (0.004) (0.002) (0.001) (0.004) Female farmer −0.048 0.001 0.007 0.098** −0.025* −0.060 (0.038) (0.063) (0.081) (0.041) (0.015) (0.077) Education level (years of schooling) −0.004 −0.015* −0.010 −0.003 −0.002 0.031*** (0.006) (0.008) (0.012) (0.007) (0.003) (0.011) Household size −0.002 −0.009 −0.003 −0.001 0.007 0.009 (0.004) (0.006) (0.010) (0.003) (0.005) (0.011) Primary occupation as a farmer? (1/0) −0.031 0.093 0.141 −0.015 0.023 −0.101 (0.060) (0.075) (0.152) (0.066) (0.022) (0.130) Land owned (ha) 0.000 −0.000 0.001 −0.000 −0.000 −0.000 (0.000) (0.000) (0.001) (0.000) (0.000) (0.001) Rented a tractor before (1/0) −0.017 0.011 −0.056 0.091** 0.013 −0.079 (0.029) (0.072) (0.090) (0.041) (0.033) (0.087) Number of tractors in the community 0.002 0.042** −0.037** 0.003 −0.004 0.005 (0.004) (0.017) (0.016) (0.009) (0.004) (0.016) Number of cattle owned 0.002 −0.001 0.005 −0.002 0.000 −0.005 (0.004) (0.002) (0.006) (0.001) (0.001) (0.005) Number of oxen owned −0.014* −0.002 −0.000 0.003 −0.002 0.004 (0.008) (0.007) (0.016) (0.005) (0.003) (0.015) Experience in farming (years) −0.000 0.003 0.001 −0.004* −0.002 0.004 (0.002) (0.003) (0.004) (0.002) (0.001) (0.004) Monze district 0.128*** −0.001 −0.188** −0.020 0.015 0.104 (0.044) (0.060) (0.090) (0.038) (0.025) (0.089) Kalomo district 0.046 0.125* −0.322*** −0.056** 0.034 0.181* (0.033) (0.074) (0.090) (0.022) (0.036) (0.092) Constant 0.012 0.337** 0.778*** −0.007 −0.040 −0.204 (0.129) (0.131) (0.213) (0.120) (0.058) (0.189) Observations 208 208 208 208 208 208 R-squared 0.109 0.118 0.119 0.127 0.061 0.079 Robust standard errors in parentheses. ***p  <  0.01; **p < 0.05; *p  <  0.1. The models are estimated separately using a linear probability model. The R-squared is reported for all models, but has little meaning in a linear probability model. Therefore, we do not use it to judge model fit. 14 K. MULUNGU ET AL. Another highlighted resource misallocation is that farmers own expensive pickup trucks “that are not helpful at the farm except for transport” while avoiding similar-cost tractor investment. This suggests that information provision could substantially increase adoption among those with sufficient resources. Despite individual investment decisions, farmers conceptualize mechanization as generating commu- nity benefits: “The community can also benefit because they will be hiring your tractor; and when you have enough diesel, you can help those who cannot afford to hire so that the whole community can benefit”. This social perspective supports hybrid financing with farmers asking NGOs and government to meet them halfway: “If 20 people come together and contribute what they can, then government or NGOs pay the deficit. These findings reveal that mechanisation adoption in Zambia operates through a dual-purpose invest- ment logic, where tractors serve simultaneously as productivity tools and business assets, explaining why timing efficiency rather than yield gains drives adoption. The prominence of livestock disease as a trigger suggests that mechanisation functions as risk diversification rather than simple intensification. Group ownership preferences correlate inversely with local mechanisation levels, indicating that it serves as a transitional mechanism rather than a permanent preference. The strong preference for RCC over tradi- tional loans, combined with social obligations constraining investment, suggests that successful mecha- nisation programmes must account for embedded social systems where individual investments serve collective insurance functions. 5.  Conclusion This study reveals that preferences for tractor ownership among Zambian smallholders are not uniform but are shaped by local mechanisation levels, gender, and socioeconomic factors of the farmers. Our mixed-methods analysis identifies a clear divergence: individual ownership is preferred in more mech- anised areas and by male farmers who prioritise operational autonomy, whereas group ownership is favoured by female farmers and those in less mechanised regions seeking to share costs and mitigate risks. Notably, female-headed households showed a significantly higher probability of preferring group ownership, reflecting distinct, gender-specific risk management strategies. When it comes to financing, RCC, especially when bundled with repair insurance, emerged as the most preferred incentive, high- lighting that farmers’ decisions are driven by a need for comprehensive risk management and not just access to capital. Finally, we found that social obligations, particularly funeral expenses, and persistent knowledge gaps are significant barriers that divert household funds and hinder investment in mechanisation. Our findings suggest that ‘one-size-fits-all’ mechanisation programmes are unlikely to succeed. Instead, interventions should be flexible and context-specific. The following recommendations are drawn from the results. (i) Offer diverse ownership models: Programmes should be designed to support both individual and group ownership pathways, recognizing that preferences are shaped by gender and the local con- text. Gender-responsive interventions are needed, acknowledging that women may benefit more from support for group-based models. (ii) De-risk financial products: Financial mechanisms must go beyond traditional loans. The strong preference for RCC indicates that products that protect farmers from climate-related shocks are essential for encouraging investments. (iii) Address social and practical barriers: To address the significant impact of social obligations, policymakers could explore innovative financial products. These may include flexible repayment windows that account for periods of high social expen- diture. The feasibility of more novel ideas, such as social obligation insurance, requires further explor- atory research but could offer potential solutions. Furthermore, support for maintenance, repairs, and access to spare parts must be integrated into any mechanisation programme to ensure its sustainability. (iv) Improve information dissemination: Knowledge gaps should be addressed by leveraging trusted social networks and community leaders, rather than relying solely on formal extension services. Peer-to- peer learning and local demonstrations are likely to be more effective methods. While this study provides valuable insights, its limitations point toward avenues for future research. The findings are based on three districts in Zambia and may not be generalisable to all farming contexts in the country. The cross-sectional design captures a single point in time; longitudinal studies are needed to track how ownership preferences and mechanisation patterns evolve as farmers gain more experience. Cogent Social Sciences 15 Note 1. At the time of the survey, USD 1= ZMW 25. Acknowledgments We thank the households that participated in this study. We are grateful to the officials from the Ministry of Agriculture who worked with the field research enumeration teams. Author contributions CRediT: Kelvin Mulungu: Conceptualization, Data curation, Formal analysis, Methodology, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing; Lushomo Molly Ngoma: Data curation, Investigation, Writing – original draft, Writing – review & editing; Rumbidzai Mhembere: Investigation, Writing – original draft, Writing – review & editing; Mark Manyanga: Methodology, Software, Writing – original draft, Writing – review & editing; Esau Simutowe: Project administration; Christian Lutz Thierfelder: Funding acquisition, Project administration, Resources, Supervision, Validation; Md. Abdul Matin: Conceptualization, Investigation, Resources; Hambulo Ngoma: Funding acquisition, Resources, Supervision, Validation. Disclosure statement No potential conflict of interest was reported by the authors. Funding This work was supported by various donors through contributions to the CGIAR Trust Fund (https://www.cgiar.org/ funders), which funded the CGIAR Regional Initiative on Diversification in East and Southern Africa (https://www.cgiar. org/initiative/diversification-in-esa/), Ukama Ustawi. We acknowledge the additional support from the European Union: (Grant No 660 FED/2019/400-893) through the Sustainable Intensification of Smallholder Farming Systems in Zambia (SIFAZ) project. ORCID Kelvin Mulungu http://orcid.org/0000-0002-4904-4122 Hambulo Ngoma http://orcid.org/0000-0002-7050-9625 Data availability statement The data used in this study are available upon request. The request can be sent to the corresponding author, Dr. Kelvin Mulungu (email: k.mulungu@cgiar.org). References Abay, S., Addissie, A., Davey, G., Farsides, B., & Addissie, T. (2016). Rapid ethical assessment on informed consent content and procedure in Hintalo-Wajirat, Northern Ethiopia: A qualitative study. PLoS One, 11(6), e0157056. https://doi.org/10.1371/journal.pone.0157056 Abebaw, D., & Haile, M. G. (2013). The impact of cooperatives on agricultural technology adoption: Empirical evi- dence from Ethiopia. Food Policy, 38, 82–91. https://doi.org/10.1016/j.foodpol.2012.10.003 Adu-Baffour, F., Daum, T., & Birner, R. (2019). Can small farms benefit from big companies’ initiatives to promote mech- anization in Africa? A case study from Zambia. Food Policy, 84, 133–145. https://doi.org/10.1016/j.foodpol.2019. 03.007 Anthropic. (2024). Claude 3.5 Sonnet Baudron, F., Sims, B., Justice, S., Kahan, D. G., Rose, R., Mkomwa, S., Kaumbutho, P., Sariah, J., Nazare, R., Moges, G., & Gérard, B. (2015). Re-examining appropriate mechanization in Eastern and Southern Africa: Two-wheel tractors, conservation agriculture, and private sector involvement. Food Security, 7(4), 889–904. https://doi.org/10.1007/ s12571-015-0476-3 Bernard, T., & Spielman, D. J. (2009). Reaching the rural poor through rural producer organizations? A study of agricultural marketing cooperatives in Ethiopia. Food Policy, 34(1), 60–69. https://doi.org/10.1016/j.food- pol.2008.08.001 https://www.cgiar.org/funders https://www.cgiar.org/funders https://www.cgiar.org/initiative/diversification-in-esa/ https://www.cgiar.org/initiative/diversification-in-esa/ http://orcid.org/0000-0002-4904-4122 http://orcid.org/0000-0002-7050-9625 mailto:k.mulungu@cgiar.org https://doi.org/10.1371/journal.pone.0157056 https://doi.org/10.1016/j.foodpol.2012.10.003 https://doi.org/10.1016/j.foodpol.2019.03.007 https://doi.org/10.1016/j.foodpol.2019.03.007 https://doi.org/10.1007/s12571-015-0476-3 https://doi.org/10.1007/s12571-015-0476-3 https://doi.org/10.1016/j.foodpol.2008.08.001 https://doi.org/10.1016/j.foodpol.2008.08.001 16 K. MULUNGU ET AL. Bopp, C., Engler, A., Poortvliet, P. M., & Jara-Rojas, R. (2019). The role of farmers’ intrinsic motivation in the effective- ness of policy incentives to promote sustainable agricultural practices. Journal of Environmental Management, 244, 320–327. https://doi.org/10.1016/j.jenvman.2019.04.107 Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), 77–101. https://doi.org/10.1191/1478088706qp063oa Carter, M., Laajaj, R., & Yang, D. (2021). Subsidies and the African green revolution: Direct effects and social network spillovers of randomized input subsidies in Mozambique. American Economic Journal: Applied Economics, 13(2), 206–229. https://doi.org/10.1257/app.20190396 Cele, L. (2021). Empowerment and agricultural mechanisation: Perceptions and experiences of women tractor oper- ators in Ghana. Development in Practice, 31(8), 988–1001. https://doi.org/10.1080/09614524.2021.1937551 Creswell, J. W. (2014). Research design: Qualitative, quantitative and mixed methods approaches (4th ed.). SAGE Publications, Inc. Daum, T. (2023). Mechanization and sustainable agri-food system transformation in the Global South. A review. Agronomy for Sustainable Development, 43(1), 16. https://doi.org/10.1007/s13593-023-00868-x Daum, T., & Birner, R. (2020). Agricultural mechanization in Africa: Myths, realities and an emerging research agenda. Global Food Security, 26, 100393. https://doi.org/10.1016/j.gfs.2020.100393 Dengel, A., Gehrlein, R., Fernes, D., Görlich, S., Maurer, J., Pham, H. H., Großmann, G., & Eisermann, N. D. G (2023). Qualitative research methods for large language models: Conducting semi-structured interviews with ChatGPT and BARD on computer science education. Informatics, 10(4), 78. https://doi.org/10.3390/informatics10040078 Diao, X., Cossar, F., Houssou, N., & Kolavalli, S. (2014). Mechanization in Ghana: Emerging demand, and the search for alternative supply models. Food Policy, 48, 168–181. https://doi.org/10.1016/j.foodpol.2014.05.013 Duflo, E., Kremer, M., & Robinson, J. (2011). Nudging farmers to use fertilizer: Theory and experimental evidence from Kenya. American Economic Review, 101(6), 2350–2390. https://doi.org/10.1257/aer.101.6.2350 Fischer, G., Wittich, S., Malima, G., Sikumba, G., Lukuyu, B., Ngunga, D., & Rugalabam, J. (2018). Gender and mecha- nization: Exploring the sustainability of mechanized forage chopping in Tanzania. Journal of Rural Studies, 64, 112–122. https://doi.org/10.1016/j.jrurstud.2018.09.012 Greene, J. C. (2007). Mixed methods in social inquiry. John Wiley & Sons. Guest, G. (2013). Describing mixed methods research: An alternative to typologies. Journal of Mixed Methods Research, 7(2), 141–151. https://doi.org/10.1177/1558689812461179 Haggblade, S., Tembo, G., & Donovan, C. (Eds.) (2004). Household level financial incentives to adoption of conservation agricultural technologies in Africa. Food Security Collaborative Working Papers 54466, Michigan State University. https://doi.org/10.22004/ag.econ.54466 Hamilton, M., Salerno, J., & Fischer, A. P. (2022). Cognition of feedback loops in a fire-prone social-ecological system. Global Environmental Change, 74, 102519. https://doi.org/10.1016/j.gloenvcha.2022.102519 Jalali, M. S., & Akhavan, A. (2024). Integrating AI language models in qualitative research: Replicating interview data analysis with ChatGPT. System Dynamics Review, 40(3), e1772. https://doi.org/10.1002/sdr.1772 Kirui, O. (2019). The Agricultural mechanization in Africa: micro-level analysis of state drivers and effects. ZEF-Discussion Papers on Development Policy, (272). Kirui, O. K., Kornher, L., & Bekchanov, M. (2023). Productivity growth and the role of mechanisation in African agri- culture. Agrekon, 62(1), 80–97. https://doi.org/10.1080/03031853.2023.2176894 Kirui, O. K., & von Braun, J. (2018). Mechanization in African agriculture: A continental overview on patterns and dynam- ics (Center for Development Research (ZEF). Working Paper, University of Bonn. https://ideas.repec.org//p/ags/ ubonwp/273522.html Kotu, B. H., Alene, A., Manyong, V., Hoeschle-Zeledon, I., & Larbi, A. (2017). Adoption and impacts of sustainable in- tensification practices in Ghana. International Journal of Agricultural Sustainability, 15(5), 539–554. https://doi.org/1 0.1080/14735903.2017.1369619 Kotu, B. H., Manda, J., Mutungi, C., Fischer, G., & Gaspar, A. (2023). Farmers’ willingness to invest in mechanized maize shelling and potential financial benefits: Evidence from Tanzania. Agribusiness, 39(3), 854–874. https://doi. org/10.1002/agr.21801 Krueger, R. A. (1988). Focus groups: A practical guide for applied research (p. 197). Sage Publications, Inc. Liverpool-Tasie, L. S. O. (2014). Fertilizer subsidies and private market participation: The case of Kano State, Nigeria. Agricultural Economics, 45(6), 663–678. https://doi.org/10.1111/agec.12114 Lu, Q., Du, X., & Qiu, H. (2022). Adoption patterns and productivity impacts of agricultural mechanization services. Agricultural Economics, 53(5), 826–845. https://doi.org/10.1111/agec.12737 Mathis, W. S., Zhao, S., Pratt, N., Weleff, J., & De Paoli, S. (2024). Inductive thematic analysis of healthcare qualitative interviews using open-source large language models: How does it compare to traditional methods? Computer Methods and Programs in Biomedicine, 255, 108356. https://doi.org/10.1016/j.cmpb.2024.108356 Middelberg, S. L. (2017). Value chain financing: Evidence from Zambia on smallholder access to finance for mecha- nization. Enterprise Development & Microfinance, 28(1), 112–129. https://doi.org/10.3362/1755-1986.16-00027 Mohammed, K., Batung, E., Saaka, S. A., Kansanga, M. M., & Luginaah, I. (2023). Determinants of mechanized tech- nology adoption in smallholder agriculture: Implications for agricultural policy. Land Use Policy, 129, 106666. https://doi.org/10.1016/j.landusepol.2023.106666 https://doi.org/10.1016/j.jenvman.2019.04.107 https://doi.org/10.1191/1478088706qp063oa https://doi.org/10.1257/app.20190396 https://doi.org/10.1080/09614524.2021.1937551 https://doi.org/10.1007/s13593-023-00868-x https://doi.org/10.1016/j.gfs.2020.100393 https://doi.org/10.3390/informatics10040078 https://doi.org/10.1016/j.foodpol.2014.05.013 https://doi.org/10.1257/aer.101.6.2350 https://doi.org/10.1016/j.jrurstud.2018.09.012 https://doi.org/10.1177/1558689812461179 https://doi.org/10.22004/ag.econ.54466 https://doi.org/10.1016/j.gloenvcha.2022.102519 https://doi.org/10.1002/sdr.1772 https://doi.org/10.1080/03031853.2023.2176894 https://ideas.repec.org//p/ags/ubonwp/273522.html https://ideas.repec.org//p/ags/ubonwp/273522.html https://doi.org/10.1080/14735903.2017.1369619 https://doi.org/10.1080/14735903.2017.1369619 https://doi.org/10.1002/agr.21801 https://doi.org/10.1002/agr.21801 https://doi.org/10.1111/agec.12114 https://doi.org/10.1111/agec.12737 https://doi.org/10.1016/j.cmpb.2024.108356 https://doi.org/10.3362/1755-1986.16-00027 https://doi.org/10.1016/j.landusepol.2023.106666 Cogent Social Sciences 17 Molyneux, S., & Geissler, P. W. (2008). Ethics and the ethnography of medical research in Africa. Social Science & Medicine (1982), 67(5), 685–695. https://doi.org/10.1016/j.socscimed.2008.02.023 Mulungu, K., Madzudzo, E., Adjei-Nsiah, S., & Akatama, M. (2017). Platforms for institutional change: Assessing the potential of livelihood enhancement groups as community entry points in Zambia. Development in Practice, 27(6), 839–850. https://doi.org/10.1080/09614524.2017.1340434 Ngoma, H., Hailu, A. T., Kabwe, S., & Angelsen, A. (2020). Pay, talk or ‘whip’ to conserve forests: Framed field experi- ments in Zambia. World Development, 128, 104846. https://doi.org/10.1016/j.worlddev.2019.104846 Ngoma, H., Marenya, P., Tufa, A., Alene, A., Chipindu, L., Matin, M. A., Thierfelder, C., & Chikoye, D. (2023). Smallholder farmers’ willingness to pay for two-wheel tractor-based mechanisation services in Zambia and Zimbabwe. Journal of International Development, 35(7), 2107–2128. https://doi.org/10.1002/jid.3767 Ngoma, H., Simutowe, E., Matin, A., & Thierfelder, C. (2023). Challenges and opportunities to scaling smallholder mech- anization in Kenya, Malawi, Zambia, and Zimbabwe. https://cgspace.cgiar.org/bitstreams/0b903807-b9cf-4fac-b357- b577d9327dbb/download Noë, A., Vaillancourt, E., & Zawati, M. H. (2025). Verbal consent in biomedical resetowardsMoving toward a future standard practice? Frontiers in Genetics, 16, 1472655. https://doi.org/10.3389/fgene.2025.1472655 Omulo, G., Daum, T., Köller, K., & Birner, R. (2024). Unpacking the behavioral intentions of ‘emergent farmers’ towards mechanized conservation agriculture in Zambia. Land Use Policy, 136, 106979. https://doi.org/10.1016/j.landuse- pol.2023.106979 Onomu, A. R., & Aliber, M. (2021). Factors influencing smallholder farmers mechanization decisions in Nigeria: The case of tractor use in the fourth industrial revolution ERA. Asian Journal of Agriculture and Rural Development, 11(2), 199–209. https://doi.org/10.18488/journal.ajard.2021.112.199.209 Onwuegbuzie A. J., & Combs, J. P. 2011. Data analysis in mixed research: A primer. International Journal of Education. 3 (1) 1–25. Patton, M. Q. (2014). Qualitative research & evaluation methods: Integrating theory and practice. SAGE Publications. Pingali, P. (2007). Chapter 54 agricultural mechanization: Adoption patterns and economic impact. Handbook of agri- cultural economics (Vol. 3, pp. 2779–2805). Elsevier. https://doi.org/10.1016/S1574-0072(06)03054-4 Shikuku, K. M., & Melesse, M. B. (2020). Networks, incentives and technology adoption: Evidence from a randomised experiment in Uganda. European Review of Agricultural Economics, 47(5), 1740–1775. https://doi.org/10.1093/erae/ jbaa009 Sims, B., & Kienzle, J. (2016). Making mechanization accessible to smallholder farmers in Sub-Saharan Africa. Environments, 3(2), 11. https://doi.org/10.3390/environments3020011 Sims, B., & Kienzle, J. (2017). Sustainable agricultural mechanization for smallholders: What is it and how can we implement it? Agriculture, 7(6), 50. https://doi.org/10.3390/agriculture7060050 Simutowe, E., Chikwalila, E. M., Chipindu, L., Tufa, A. H., Alene, A. D., Matin, M. A., Chikoye, D., Thierfelder, C., Marenya, P. P., & Ngoma, H. (2023). Status of appropriate-scale mechanization in Zambia and Zimbabwe [Report]. CGIAR. https://hdl.handle.net/10568/138130 Tai, R. H., Bentley, L. R., Xia, X., Sitt, J. M., Fankhauser, S. C., Chicas-Mosier, A. M., & Monteith, B. G. (2024). An exam- ination of the use of large language models to aid analysis of textual data. International Journal of Qualitative Methods, 23, 16094069241231168. https://doi.org/10.1177/16094069241231168 Takele, A., & Selassie, Y. G. (2018). Socio-economic analysis of conditions for adoption of tractor hiring services among smallholder farmers, Northwestern Ethiopia. Cogent Food & Agriculture, 4(1), 1453978. https://doi.org/10.1080/2331 1932.2018.1453978 Takeshima, H., Nin—Pratt, A., & Diao, X. (2013). Mechanization and agricultural technology evolution, agricultural intensification in Sub-Saharan Africa: Typology of agricultural mechanization in Nigeria. American Journal of Agricultural Economics, 95(5), 1230–1236. https://doi.org/10.1093/ajae/aat045 Tesema, Y. M., Asrat, P., & Sisay, D. T. (2023). Factors affecting farmers’ hiring decisions on agricultural mechanization services: A case study in Ethiopia. Cogent Economics & Finance, 11(2), 2225328. https://doi.org/10.1080/23322039.2 023.2225328 The American Anthropological Association. (2012). AAA statement on ethics: Principles of professional responsibility table of contents. The American Anthropological Association. https://americananthro.org/about/policies/statement-on- ethics/ Tietjen, B., Jacobsen, K., & Hollander, J. (2023). Climate change and urban migration in Sub-Saharan African cities: Impacts and governance challenges. Journal of Climate Resilience and Justice, 1, 20–32. https://doi.org/10.1162/crc- j_a_00009 Tufa, A., Alene, A., Ngoma, H., Marenya, P., Manda, J., Matin, M. A., Thierfelder, C., & Chikoye, D. (2024). Willingness to pay for agricultural mechanization services by smallholder farmers in Malawi. Agribusiness, 40(1), 248–276. https://doi.org/10.1002/agr.21841 Van Loon, J., Woltering, L., Krupnik, T. J., Baudron, F., Boa, M., & Govaerts, B. (2020). Scaling agricultural mechanization services in smallholder farming systems: Case studies from sub-Saharan Africa, South Asia, and Latin America. Agricultural Systems, 180, 102792. https://doi.org/10.1016/j.agsy.2020.102792 Wooldridge, J. M. (2010). Econometric analysis of cross section and panel data (2nd ed). MIT Press. https://doi.org/10.1016/j.socscimed.2008.02.023 https://doi.org/10.1080/09614524.2017.1340434 https://doi.org/10.1016/j.worlddev.2019.104846 https://doi.org/10.1002/jid.3767 https://cgspace.cgiar.org/bitstreams/0b903807-b9cf-4fac-b357-b577d9327dbb/download https://cgspace.cgiar.org/bitstreams/0b903807-b9cf-4fac-b357-b577d9327dbb/download https://doi.org/10.3389/fgene.2025.1472655 https://doi.org/10.1016/j.landusepol.2023.106979 https://doi.org/10.1016/j.landusepol.2023.106979 https://doi.org/10.18488/journal.ajard.2021.112.199.209 https://doi.org/10.1016/S1574-0072(06)03054-4 https://doi.org/10.1093/erae/jbaa009 https://doi.org/10.1093/erae/jbaa009 https://doi.org/10.3390/environments3020011 https://doi.org/10.3390/agriculture7060050 https://hdl.handle.net/10568/138130 https://doi.org/10.1177/16094069241231168 https://doi.org/10.1080/23311932.2018.1453978 https://doi.org/10.1080/23311932.2018.1453978 https://doi.org/10.1093/ajae/aat045 https://doi.org/10.1080/23322039.2023.2225328 https://doi.org/10.1080/23322039.2023.2225328 https://americananthro.org/about/policies/statement-on-ethics/ https://americananthro.org/about/policies/statement-on-ethics/ https://doi.org/10.1162/crcj_a_00009 https://doi.org/10.1162/crcj_a_00009 https://doi.org/10.1002/agr.21841 https://doi.org/10.1016/j.agsy.2020.102792 18 K. MULUNGU ET AL. Wossen, T., Abdoulaye, T., Alene, A., Haile, M. G., Feleke, S., Olanrewaju, A., & Manyong, V. (2017). Impacts of extension access and cooperative membership on technology adoption and household welfare. Journal of Rural Studies, 54, 223–233. https://doi.org/10.1016/j.jrurstud.2017.06.022 Xiao, Z., Yuan, X., Liao, Q. V., Abdelghani, R., & Oudeyer, P.-Y. (2023). Supporting qualitative analysis with large lan- guage models: Combining codebook with GPT-3 for deductive coding. Companion Proceedings of the 28th International Conference on Intelligent User Interfaces (pp. 75–78). https://doi.org/10.1145/3581754.3584136 https://doi.org/10.1016/j.jrurstud.2017.06.022 https://doi.org/10.1145/3581754.3584136 Motivations and incentives for mechanization in Zambia: a mixed-methods analysis ABSTRACT 1. Introduction 2. Literature review 2.1. Mechanization adoption in developing countries 2.2. Group versus individual ownership models in agricultural technology 2.3. Role of incentives in technology adoption 3. Methodology 3.1. Study area 3.2. Study design and data collection 3.2.1. Qualitative instruments 3.2.2. Quantitative data 3.3. Data analysis 3.4. Ethical clearance 4. Results and discussion 4.1. Sample characteristics 4.2. Motivations for and tractor ownership patterns 4.3. Ownership preferences 4.3. Purchase models, incentives for tractor adoption and barriers 4.4. Barriers to mechanization investment 5. Conclusion Note Acknowledgments Author contributions Disclosure statement Funding ORCID Data availability statement References