1 Introduction

Climate change affects global food security through rising temperatures, changes in rainfall patterns, and the increased frequency of extreme weather phenomena (Mbow et al. 2019; FAO 2018). This in turn negatively affects agriculture through food insecurity and changes in people’s livelihood; lower productivity in farming, aquaculture, and animal husbandry; and an increase in global greenhouse gas emissions (Dinesh et al. 2018; IPCC (Intergovernmental Panel on climate Change) 2014; FAO 2010).

Vietnam, while contributing significantly to global food security, is among those countries most affected by climate change (World Bank Group and the Asian Development Bank 2020; MONRE 2016; World Bank 2011). In recent decades, Vietnam has faced higher temperatures and a sharp rise in extreme weather events, such as droughts, floods, and tropical cyclones (Son et al. 2018; Trinh 2018; MONRE 2016; CCAFS-SEA (CGIAR Research Program on Climate Change, Agriculture and Food Security- Southeast Asia), 2016a; CCAFS-SEA 2016b; World Bank 2010). These have presented multiple risks to Vietnam’s agricultural sector (Vu et al. 2018). Past and ongoing studies confirm the strong negative impact of climate change on Vietnam’s agricultural production and efforts at poverty reduction, food security, employment, and export (Nguyen et al. 2022; Ferrer et al. 2022; Diallo et al. 2019; Huong et al. 2019; Nguyen and Nguyen 2019; Trinh 2018; Tivet and Boulakia 2017; Le et al. 2015a, 2015b; Arndt et al. 2015).

For example, saltwater intrusion during the winter–spring crop season in 2015/2016 adversely affected many agricultural zones across Vietnam (Le and Nguyen 2019; FAO 2016; UNDP 2016). Rice paddy production fell by 11.2% compared to the 2014/15 crop year (GSO (General Statistical Office) 2016). The marked changes in rainfall dynamics, with more frequent and intense precipitation events and droughts disrupting agricultural and other economic activities, presented new challenges for Vietnamese farmers. Farmers with traditional farming methods are unable to effectively respond to adverse events that threaten the economic viability of many farming communities. The situation is serious for Vietnam, whose agricultural sector is the backbone of the national economy, accounting for 24% of the country’s gross domestic product (GDP) and more than 20% of exports. Agriculture is also important for employment, with 70% of the country’s population depending on this sector for livelihood (Maitah et al. 2020).

Given the context, a change in agricultural practices is in order. Efforts have been made to develop, promote, and disseminate new farming systems, technologies, strategies, and measures that can help farmers, especially those in vulnerable areas, overcome climate change challenges (Bai et al 2019; Lipper and Zilberman 2018; FAO 2013, 2010).

Globally, the key response to climate change is the introduction of climate-smart agriculture (CSA). It is a systematic approach to developing technical, policy, and investment conditions that achieve sustainable agricultural development and food security under climate change (Lipper et al. 2014). The goal of CSA is to enable the sector to move toward more climate-resilient production systems and more sustainable livelihoods in the presence of climate change stressors and climate variability. The three pillars of CSA interventions are to: (i) sustainably increase agricultural productivity and income; (ii) adapt and increase resilience to climate change impacts; and, (iii) reduce and/or remove greenhouse gas emissions to the extent possible and appropriate (Asfaw and Maggio 2016; FAO 2013, 2010).

In Vietnam, the CSA is recognized as an important tool for managing the impact of climate change and climate-responsive policies were adopted (Simelton et al. 2017; Van et al. 2017; Pham et al. 2017; Tivet and Boulakia 2017). In 2014, climate-smart agricultural technologies (CSATs) were first introduced in three villages in Vietnam, namely, Ma Village in Yen Bai Province, My Loi Village in Ha Tinh Province and Tra Hat Village in Bac Lieu Province, with the support of the CGIAR Program on Climate Change, Agriculture and Food Security in Southeast Asia (CCAFS SEA). These villages are representative of different agro-ecological regions of Red River Delta, North-Central and Mekong River Delta, respectively (Simelton et al. 2017). My Loi Village in Ha Tinh Province in the uplands of North-Central Vietnam was chosen as one of three villages for CSATs piloting in Vietnam because of its high level of exposure to multiple extreme weather events (temperature and water stress, storm and typhoon) and the potential for climate-smart solutions (Aggarwal et al. 2018; Campbell et al. 2016; Duong et al. 2016; CCAFS-SEA 2016b). Seventeen CSATs were introduced in My Loi Village to make smallholder farmers more productive and resilient to climate change impacts. One important intermediate outcome identified in My Loi Village was the adoption of certain CSATs by farmers due to their enhanced knowledge and favorable attitude (Ferrer and Bernardo 2020), and their attitudes toward climate risks (Jin et al. 2020; Nguyen and Ho 2021). Although CSA practices has been documented and well-recognized method for adjusting agricultural output to the new conditions brought on by climate change, it has been emphasized that smallholder farmers find it challenging to adopt and apply CSATs. As such, a clear understanding of the factors that influence farmers’ CSATs adoption is important to support farmers to increase their resilience to climate change and to promote effective climate-smart actions in the agricultural sector.

Previous studies conducted in different developing countries provide empirical evidence on the factors influencing farmers’ adoption of CSATs, such as age, gender, ethnicity, education, number of children, family agricultural labor, off-farm work, diversified crops, farm size, experience, land ownership, community meetings, credit, yield, revenue, extension, training, institutions, household resources and household dependents (Usha et al. 2022; Atta-Aidoo et al. 2022; Abegunde et al. 2020; Solomon 2020; Jin et al. 2020; Liu et al. 2018; Maguza-Tembo et al. 2017; Di Falco and Veronesi 2013; Teklewold et al. 2013; Grothmann and Patt 2005). Studies found that adoption decisions are location-specific and are influenced by different key drivers. Due to differences in culture, awareness, resource endowments, objectives, preferences and socio-economic backgrounds, and psychological factors, farmers in different countries varied in their willingness to adopt new technologies. Farmers also modified or combined the original CSATs to address specific conditions and strategies they face.

Studies conducted in Vietnam focused on the effectiveness of potential and priority CSATs (e.g., Nguyen et al. 2020; CCAFS-SEA 2016b; Nghia et al. 2015) as well as the factors that influence farmers to adopt one or a combination of CSATs (e.g., Tran et al. 2020; Luu 2020; Nguyen and Nguyen 2019; Le et al. 2014a, b). Key factors affecting farmers’ adoption of CSATs were found to include socio-demographic features, risk perception, social capital, or geographical location. However, no studies have addressed the issues of the continuation (or intensity) of farmers’ adoption of CSATs in Vietnam, specifically in areas, where CSATs were formally introduced and promoted for use by farmers, such as in My Loi Village. This study attempts to fill this gap and expands the definition of successful adoption by incorporating this element. The present paper contributes to the literature on household’s farming CSATs adoption behavior in developing countries by analyzing data from My Loi Village using a binary logit regression model for the initial uptake and an ordinary least squares (OLS) model for the continuation of adoption of CSATs. The OLS particularly looks at the factors affecting farmers’ decision to continue or suspend the use of CSATs and the number of CSATs still in use after 5 years of introduction. This research strategy was proposed to examine farmers’ decisions in the long run. Adaptation to climate change is a long process that requires not only initial adoption but also maintaining the practices without discarding them in the short-to-medium run. To the best of our knowledge, this study is the first attempt to identify key determinants of continuous adoption of CSATs. Moreover, understanding the factors driving farmers’ CSATs adoption is crucial for proposing timely and effective policies and interventions. The findings of this study can guide policymakers in developing plans and programs for disseminating and promoting CSATs adoption and mitigating the detrimental impacts of climate change on the agricultural sector.

The paper proceeds as follows. The next section presents an overview of the study site. Section 3 describes the methodology, including the data and regression models. Section 4 presents the analysis and discusses the results. Section 5 provides a brief summary and areas for future research. Section 6 presents the policy implications of this study.

2 The study area

My Loi Village is located in the uplands of Ky Son Commune, Ky Anh District, Ha Tinh Province on the North-Central coast of Vietnam (Fig. 1). There were 230 households in the village in 2021, each with approximately four members. The main products of the village are cassava, peanuts, and acacia. It has a total land area of 195 hectares, of which 55 hectares is farmland for such annual crops as peanuts (30 hectares), paddy rice (8.5 to 9.5 hectares), maize, green beans, and sweet potatoes. About 90% of households raise a small number of animals for household consumption. My Loi has faced a range of extreme weather events from cold spells to hot spells, droughts and floods, dry Foehn winds, tornadoes, and tropical storms and typhoons. During floods, polluted water sweeps over fields or ends up in wells.

Fig.1
figure 1

Location of My Loi Village

Full size image

Seventeen CSATs to increase productivity and income, enhance the resilience of livelihoods and ecosystems, and reduce or remove greenhouse gas emissions into the atmosphere were designed and introduced to My Loi Village (Bonilla-Findji and Yen 2019; Simelton et al. 2017). They specifically included alley cropping (non-N-fixing trees), the production and use of compost, crop type changes, diet management, improved pigsties and animal cages, manure treatment, intercropping (non-legume), mulching, improved cooking stoves, crop rotation (alteration of legume and non-legume crops), drip irrigation, silvo pasture, multi-strata agroforestry, parklands, complex crop rotation, biogas, and biochar.

3 Methodology

3.1 The theoretical framework

The discipline of economics assumes that economic agents, when making decisions, are rational and maximize their self-interest. However, Simon (1972) challenged this classical thinking by presenting the theory of bounded rationality which states that decision-making was about ‘satisficing’ rather than ‘optimizing’. It was argued that people were limited by their “cognition” and made decisions using limited information to produce a satisfactory result instead of using all available information needed to make rational decisions.

According to this theory, a farmer considering whether or not to adopt recommended CSATs is influenced by the information received either formally or informally as well as the cognitive level and attitude of that farmer regarding CSA. The theory of diffusion of innovation by Rogers (1962) suggests five stages through which a farmer makes a decision about CSATs: (i) the farmer becomes aware of CSATs; (ii) the farmer forms an attitude about CSATs; (iii) the farmer decides to accept or reject CSATs; (iv) the farmer initiates the use of CSATs for testing; and, (v) the farmer continues the use of CSATs.

3.2 Empirical models

To identify the drivers of the adoption and continuation of the practices, a two-way empirical strategy was adopted. First, the drivers of initial adoption are studied using a binary logit choice model. Second, the key drivers for the continuation of the adoption are examined using OLS regression model. The two empirical models are as follows.

3.2.1 Model 1

To explore the factors affecting the adoption of CSATs, a logistic regression model was applied. The model is similar to the one developed by Di Falco and Veronesi (2013) and Bryan et al. (2013), which has become one of the common models in technology adoption research. The logistic procedure is used to capture nominal categorical responses on a set of discrete and/or continuous predictor variables allowing us to estimate the CSATs adoption probability (Wooldridge 2012). The model can also identify the main barriers and potential constraints to adoption.

There are only two choices in farmers’ decisions regarding the introduced technology. The farmers were asked whether or not they had used any of the CSATs. The dependent variable in this study is a binary variable that takes the value of either 1 (adopted) or 0 (not adopted). Independent variables included personal characteristics of the farmer, family level variables, farm variables, and institutional and social variables (Table 1).

Table 1 Variables in the research models
Full size table

In this model, the dependent variable becomes the natural logarithm of the odds when a positive choice is made:

$$\ln \, \left( {\frac{{p_{i} }}{{1 - p_{i} }}} \right) = \beta_{0} + \beta_{1} x_{1i} + \beta_{2} x_{2i} \, + \, ... \, + \, \beta_{k} x_{ki} + e_{i}$$
(1)

where pi is the probability of ith observation in the sample of adoption; (1-pi) is the probability of ith observation in the sample of non-adoption; i is ith observation in the sample; β0 and ei stand for the intercept parameter and the error term, respectively; β1, β2 … βk are regression coefficients of the independent variables; k is the number of independent variables; x1i, x2i,… xki are the independent variables or characteristics of ith farm household/ith observation in the sample.

3.2.2 Model 2

The continuation (or intensity) of the adoption is measured by the number of CSATs currently used. Because the dependent variable (DCi) is the continued adoption of the CSATs, which is a count variable, the ordinary least squares (OLS) method was used (Wooldridge 2012). The use of OLS in this study is similar to previous studies using continuous dependent variable in conservation practice adoption or conservation program enrollment (Liu et al. 2018), agricultural technologies adoption (e.g., Chuang et al. 2020). The list of explanatory variables is shown in Table 1. The regression equation thus is

$$\begin{gathered} \hfill \\ DC_{i} = \beta_{0} + \beta_{1} x_{1i} + \beta_{2} x_{2i} \, + \, ... \, + \, \beta_{k} x_{ki} + e_{i} \hfill \\ \end{gathered}$$
(2)

where DCi is the number of CSATs currently adopted by a ith farmer; i is ith observation in the sample; β0 and ei stand for the intercept parameter and the error term, respectively; β1, β2 … βk are regression coefficients of the independent variables; k is the number of independent variables; x1i, x2i,… xki are the independent variables or characteristics of ith farm household/ith observation in the sample.

3.2.3 Selection of independent variables

Five groups of independent variables employed in this study are personal characteristics of the farmer, family level variables, farm variables, and institutional and social variables. The selection and categorization of variables is based on the CSATs being investigated and literature review. Personal characteristics of the farmer or demographic variables (e.g., age, gender, education, farming experience, training, and information) have been discussed substantially in previous empirical studies regarding the factors influencing farmers’ adaptive behaviors in response to climate change (Nguyen et al. 2022; Hoa et al. 2022; Maguza-Tembo et al. 2017; Tran et al. 2020; Trinh et al. 2018; Le et al. 2014a, b; Ozor et al. 2012; Weinstein 1989). The age of the farm household head, farming experience, education level were reported to affect the CSATs adoption decisions of farming households. The differences between the adaptive responses of farmers may be accounted for by disparities in their land ownership, off-farm job opportunities, health and welfare, education, socio-cultural context, access to credit, information and other resources. The factor-related socioeconomic and farming characteristics such as farm income, family labor force, farm size, and land tenure were found to have strong impacts on farmers’ decision to adapt to climate change (Hoa et al. 2022; Nhemachena and Hassan 2007; Tran et al 2020; SeinnSeinn et al. 2015). Some institutional and social variables have been incorporated in the models, including the farm households’ access to formal credit and status of participation in membership in social/agricultural groups. The two variables are all dummies and have been employed in previous studies of Maguza-Tembo et al. (2017), Teklewold et al. (2013) and Mignouna et al. (2011). Table 1 describes variables used in Eqs. 1, and 2.

3.3 Study participants

The study participants were 215 farming households in My Loi Village. During the data collection in September 2021, My Loi Village had 230 households. The 15 households under quarantine due to COVID-19 were not included in the survey. Women accounted for 73% of the study participants. The data collection period coincided with the rice farming off-season, when many men left the village to find temporary jobs elsewhere and for their wives to take care of the farm. The average female-to-male ratio in agricultural labor in Ha Tinh Province is about 65:35, but it may reach 75:25 in some places.

3.4 Data collection method and instrument

Mixed data collection methods were employed. These were key informant interviews with CSA experts of local government agencies and face-to-face interviews with farming households in My Loi Village. The selection of key informants was based on involvement in agriculture, rural development, and CSA. The key informant interviews were conducted prior to the survey to generate information necessary in designing the interview schedule used in the survey.

We collected data from the household survey as follows:

First, the interview schedule was pilot-tested with 5 households in My Loi Village to check the clarity and appropriateness of the questions, possible alternatives to the questions, the difficulty of the questions, the probability that a large number of questions would go unanswered, and the length of the interview.

Second, the interview schedule was revised and finalized to address the concerns raised by the pilot testing participants. It also became an opportunity for the data collection team to gain experience in working with the farm households and to find the best interview strategy.

Third, 2-day training was provided to the enumerators focusing on familiarization with the interview schedule, how to approach the farmers, and how to conduct the interviews to ensure reliable answers.

Finally, a team of trained data collectors conducted the survey in September 2021.

4 Results and discussion

4.1 Characteristics of the study participants

Among the 215 farmers who participated in the study, 74% (159 farmers) adopted at any one time at least one CSAT introduced since 2014 (Table 3). While most study participants were women (73%), the women-non-adopters were higher in proportion than the women-adopters (80% vs. 70%). Almost all of the participants were married (88%). On average, they were in their late 40 s but the non-adopters were younger (41 years) than the adopters (50 years). This suggests that the adoption of new technology is more attractive to older farmers. The study participants, on average, had 8 years of formal education with the non-adopters staying a little longer in school than the adopters (9.41 years vs. 8.04 years). The majority of them (68%) studied or finished junior high school education at the minimum (Table 2).

Table 2 Profile of the farmers in My Loi Village who participated in the study
Full size table

The number of adopted CSATs varied through the years. The most commonly adopted CSATs were ally cropping (75%), compost (52%), and crop type change (49%) (Table 3). At least one-third of the adopters–farmers attempted improved pig sty or animal cage and diet management. About one-fourth of the adopter–farmers practiced manure treatment, intercropping, or mulching. The least popular CSATs were multi-strata agroforestry, parklands, rotation, biogas, and biochar. At the time of the survey, the number of adopted CSATs ranged between 1 and 12 among 159 adopter–farmers with an average number of four practices. The proportion of farmers who had both heard about and actually attended at least one of the training sessions and the proportion of farmers who heard about the training sessions of almost all of the CSATs were both higher among the adopters than among the non-adopters.

Table 3 Adoption of CSATs by My Loi farmers
Full size table

4.2 Factors influencing the adoption of CSATs

The factors that significantly influence a farmer’s decision to adopt at least one of seventeen CSATs were identified using a binary logit regression for initial adoption (“ever adopted”) and OLS regression for the continuation (intensity) of adoption. The summary statistics for the two dependent variables and independent variables are presented in Table 4.

Table 4 Descriptive statistics
Full size table

4.3 Factors influencing adoption at any time of any CSATs

Table 5 presents the results of the binary logit regression model that identified the factors influencing the adoption of any CSATs at any time (ever adopted) since the introduction of CSATs in 2014. The model fits the data reasonably well (Prob > chi2 = 0.000) with correct prediction at 89.77%. The significant factors positively influencing the decision of a farmer to adopt or not adopt a CSAT were attendance to training on CSATs (p < 1%); having a fellow farmer as a source of information (p < 5%); rice cultivation, farmer’s experience as a source of information, and the number of crops grown (each at p < 10%). The two factors that significantly and negatively influenced adoption decision were having more working men in the family (p < 5%) and membership in a farming organization (p < 10%).

Table 5 Logit regression results: Adoption of CSATs (Ever adopted)
Full size table

Attendance to training on any CSAT was highly significant (1%) and had a strong positive impact on adoption behavior. The odds ratio indicated that the farmer’s attendance of training greatly increased the likelihood of adopting any CSAT by a factor of 40.70 compared to those who did not attend. This highlights the significance of training when introducing new farming technologies and practices. This may also reflect the quality of the training provided to the farmers in this case. These results were consistent with the findings of Nguyen et al. (2022), Nguyen et al. (2020), Nguyen and Nguyen (2019), and Trinh et al. (2018) that training not only helps farmers increase their awareness of the severity of the risks and benefit of each CSAT but also provides detailed guidance on how to practice the proposed methods, leading to a positive response to the risks. Grothmann and Patt (2005) showed that farmers will not respond to warning information if they perceive the risk to be low, with little intention of taking suggested actions.

Having a fellow farmer (or farmers) as a source of information was also significant (5%) and positively associated with adoption behavior. The odds ratio showed that the likelihood of adopting any CSAT is boosted by a factor of 7.13 than otherwise. This result was similar to those of Tran et al. (2020) and Teklewold et al. (2013), which confirmed that consultation with other farmers who are well-informed about the new ways of farming is an important strategy to scale up CSATs. Thus, the CSA roving workshops, where farmers meet other farmers to share practices as well as any other gatherings, where practices and experiences are similarly exchanged among farmers, are highly recommended. The farmer’s own “experience as a source of information” was also strongly associated with adoption behavior, with an odds ratio of 4.098. This variable can be construed as a proxy for the confidence of the farmer as a farmer. This is consistent with the results of Nguyen et al. (2022), Trinh et al. (2018), Le et al. (2014a, b), Ozor et al. (2012), and Weinstein (1989), which concluded that past risk experience positively influences the self-protection behavior of farmers.

Moreover, being a rice farmer also increased the likelihood of adoption with an odds ratio of 3.466. Although a few farmers in My Loi Village have diversified their crops from rice to cash crops, fruit trees, and forest trees, still 80% of the farmers grew rice. Rice as the most important agricultural product may allow farmers to receive more relevant information and support than the growers of other crops, which may make the former relatively more open and willing to improve their practices. In addition, the odds ratio (1.394) of the variable “number of crops grown” indicated that one additional crop grown also increased the likelihood of adopting any CSAT by 39%.

On the other hand, having male family members in the labor force (odds ratio = 0.455) reduced the likelihood of adoption. This can be understood in the context of My Loi Village, where the men usually leave the village temporarily and seasonally, especially between farming seasons or more permanently to find work in other regions. Farm labor is dominated by women. In this study, 73% of the survey participants were female farmers. The mobility of men who spend much of their time outside the village may make them pay less attention to farming and more conservative about the adoption of new methods. This result corroborated the findings of Nhemachena and Hassan (2007) and Tran et al (2020) that female-headed households more willingly adopt climate change adaptation methods in farming than male-headed ones.

Adoption of CSATs was negatively affected by involvement in community organizations (odds ratio = 0.360). In My Loi Village, men were usually members of the farmer’s organization, which can explain the negative influence of gender as explained above. This result, however, was contrary to what Vo et al. (2021) found, where membership in local organizations encourages farmers’ adaptation to climate change. Nonetheless, the results of the current study were generally consistent with previous findings on factors affecting the adoption of good agricultural practices (e.g., Di Falco et al. 2011; Jin et al. 2020; Tran et al. 2020; Trinh et al. 2018).

4.4 Factors influencing the continued adoption of CSATs

OLS regression was used in the second stage to identify the significant factors affecting the continuation or intensity of CSATs adoption. The dependent variable is the number of CSATs the farmer was adopting at the time of the interview. It should be noted that this measure of adoption covers the period from 2014 to 2021. It is acknowledged that counting the number of adopted practices is not a perfect representation of the frequency or duration of adoption of each CSAT. However, the measure should still be useful in studying the extent to which farmers continue to use CSATs. The model has an adjusted R-squared of 41%, showing that the independent variables can collectively explain 41%, a relatively large portion, of the variation of the dependent variable.

The factors that significantly and positively influenced the continued adoption behavior were attendance to CSATs training (p < 1%); the agriculture extension officer as a source of information and TV as a source of information (at p < 5% each); and, upward mobility, farm ownership, and the number of crops grown (at p < 10% each). The factors that significantly (at p < 10%) and negatively influenced this decision were having more male family members in the labor force and the ease of finding farm labor.

Attending to a CSA training course was likely to increase the number of CSATs adopted by a factor of three, which was similar to the results of the ‘ever adopted’ regression. It indicated a strong association between training and adoption behavior. The positive role of the agriculture extension officer on the adoption behavior of the farmers was also brought to the fore. Survey data confirmed that the agricultural extension officer was a common source of information for matters related to production inputs, such as crop variety, fertilizer, and pesticide, as well as soil and livestock management. This suggests the importance of a highly skilled agricultural extension officer who can share information and skills with the farmers. Similarly, sourcing farming information from TV increases the number of CSATs adopted by one. For farmers, TV was the main source of information on the daily weather forecast and one of the major sources of the seasonal forecast. This points to the importance of TV as a medium to disseminate better farming technologies and practices in Vietnam. The presence of popular information media perceived to be reliable by farmers, such as TV, can improve their awareness of the threats posed by the changing climate conditions (Di Falco et al. 2011). This can positively influence farmer’s responses against such risks (Li et al. 2022; Le et al. 2014a, b; Di Falco et al. 2011).

In addition, the upward mobility of farmers, i.e., their positive attitude toward better ways of farming, increased CSA adoption rate. The same was true for farmers who owned their land. The positive relationship between land ownership status and CSATs intensity of adoption revealed in this study implies that farmers are more likely to manage self-owned land in a more favorable manner than rented land. Our research finding is in line with Dung et al. (2018), who found that household land tenure status positively influences the CSATs adoption. Since 1988, Vietnam recognized farm households as autonomous economic units, freed up markets for inputs and outputs, recognized their right on land, and made allowance for long‐term allocations of land to farming households. Farmers can use agricultural land for agricultural production purposes. They can also transfer, exchange, lease, inheritance and mortgage the allocated land. This policy has resulted in a boost in agricultural output and significant improvement in living standards in rural areas (World Bank 2016).

Growing one additional crop also increased the adoption of CSAT by a factor of 0.21. On the other hand, the ease of finding farm labor negatively influenced the adoption of CSATs. This may be attributable to labor mobility in My Loi Village, in particular, and the rural areas of North-Central region of Vietnam, in general. Another possible explanation could be that the available workforce in rural area in North-Central Vietnam is normally less educated and has fewer information-comprehension skills, and is, therefore, is less aware of new farming technology. As a result, local farmers are less likely to adopt CSATs that predominantly require unskilled labor. This also might mean that proactive farmers constantly look for ways to improve their farming activities by continuously seeking agricultural information and training, thus providing them with knowledge and skills. This result corresponds to the study of Kangogo et al (2021), who found that CSA adoption to be negatively associated with unskilled labor. Similarly, having more male family members in the labor force negatively influenced CSATs adoption. This result can again be attributed to the mobility of village men who migrate to other areas temporarily or permanently to find work Table 6.

Table 6 Regression results: Continuation of adoption of CSATs
Full size table

5 Conclusions

The study focused on farming households’ adoption and practice of CSATs in My Loi Village in the North-Central region of Vietnam. Among 215 farming households interviewed, 159 had adopted at least one CSAT since 2014. At the time of the survey (September 2021), the average number of adopted and practiced CSATs over farming households which adopt and practice at least one CSAT is around 4. Factors are identified that positively and negatively influence farmer’s decision to adopt and continue to adopt CSATs.

It is proved that positive impacts are given by training on CSATs and having fellow farmers who can spread new ways of farming to others, well-informed and highly skilled agricultural extension officers, and reliable media of information for farmers including TV. More specifically, the following results for initial adoption and continued adoption are particularly noteworthy.

First, initial adoption is enhanced by attendance to any training course on CSATs, presence of a fellow farmer or farmers as a source of information, rice cultivation, farmer’s own experience in farming, and number of crops grown. On the other hand, the two factors that significantly and negatively influence adoption are having more working men in the family and membership in the village farming organization.

Second, variables that enhance the decision to continuously adopt CSATs are attendance to any CSATs training, support by an agriculture extension officer, TV as a source of farming information, upward mobility of farmers, farm ownership, and the number of crops grown. The factors that negatively influence continuous use are the large number of working men in the family and the ease of finding farm labor.

This research is not without limitation. The data for the study were collected from farmers based on recall. Future research would benefit from the examination of another research design, such as longitudinal experiments and other regression estimation methods. In addition, the study only looked at the factors that influence CSATs adoption and intensity adoption among smallholder farmers in My Loi Village in North-Central Vietnam. Further research should be carried out in other regions of Vietnam.

6 Policy implications

The study highlights the importance of both the quality and quantity of information available to farmers and as well as easy access to it. A key policy message suggested is effective and accessible information provision. Trust-building communication among stakeholders is the area that needs more policy effort and action. While everyone needs to work to raise public awareness of CSA, the role and leadership of the government are particularly important.

More concretely, the government must provide high-quality information through appropriate training programs on CSA. This is a relatively simple and inexpensive yet highly efficient way to produce changes in the farmer’s attitude and behavior toward the adoption of CSATs. Training activities on CSATs and educational opportunities should be made more accessible for female farmers. Utilizing the power of the media—including TV but also radio, posters, websites, others—to strengthen the awareness of CSA and communicate effective response strategies is crucial.

In addition, the local government leaders responsible for the agriculture sector should formulate a plan to increase staffing for agricultural extension and support services as well as staff development programs to further accelerate the adoption of CSATs. Extension activities should be widened and made more accessible to all households particularly to those with larger number of male members. The technical know-how on CSATs by the agricultural extension and support staff is significantly important in promoting new technologies and upgrading the skills of farmers, thereby speeding the adoption of CSATs.

Finally, since adaptation to climate change is a location-specific issue, understanding the features and constraints of farming households is vital and needs more attention at the policy, research, and practice levels for introducing new CSATs and designing proper incentive mechanisms to encourage local farmers to adopt them.