(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



Adoption of climate-smart agricultural practices by smallholder farmers in rural Ghana: An application of the theory of planned behavior [1]

['Jonathan Atta-Aidoo', 'Department Of Environmental Science', 'Kwame Nkrumah University Of Science', 'Technology', 'Kumasi', 'Philip Antwi-Agyei', 'Andrew John Dougill', 'Department Of Environment', 'Geography', 'University Of York']

Date: 2022-11 

Climate-Smart Agricultural (CSA) practices are crucial in managing climatic shocks faced by smallholder farmers in sub-Saharan Africa. However, evidence on the socio-psychological drivers of farmers’ adoption of CSA practices remains limited. This study employed the Theory of Planned Behavior framework to analyze smallholder farmers’ intention and adoption behavior toward CSA practices in rural Ghana. The study sampled 350 smallholder farmers from the Upper East and North-East Regions of Ghana and employed the Structural Equation Model to understand smallholder farmers’ intention and adoption behavior toward CSA practices. Results showed that farmers’ attitudes (notably their beneficial evaluation of CSA practices) had a significant impact (0.25) on their intention to adopt CSA practices. Social pressure exerted on farmers to use CSA practices (Subjective norm) also had a significant impact (0.52) on farmers’ adoption behavior. Perceived behavior control which measures the controllability and use of CSA practices also had a significant impact on both the intention (0.43) and adoption behavior (0.20) of smallholder farmers. Findings highlight the role socio-psychological factors play in explaining the adoption of CSA practices in rural Ghana. We recommend the need to create awareness of CSA practices by sharing relevant information more widely on CSA practices through community leaders, chief farmers, assembly members, and clan heads in order to exert influence on farmer’s adoption of CSA practices.

Funding: The work was supported by the Royal Society, London, through study received funding from the FLAIR Collaboration Grant by the Royal Society, London [FCG\R1\211025 to AJD and PA-A]. This work was also supported by the Future Leaders-African Independent Research (FLAIR) Fellowships funded by the Royal Society, London [FLR\R1\201640 to PA-A]. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Copyright: © 2022 Atta-Aidoo et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

We contribute to the literature on the adoption of CSA practices by identifying the relative significance of the Theory of Planned Behavior constructs on farmers’ behavioral intention and adoption behavior towards CSA practices in dryland farming systems. Insights can inform policymakers the areas of possible interventions that can be impactful at the household level to positively alter farmers’ behavior and enhance their adoption of CSA practices.

Several studies have examined the determinants of the adoption and impact of CSA practices in SSA countries [ 23 – 25 ] and in Ghana specifically [ 20 , 22 , 26 ]. The determinants identified by these studies were mostly socio-demographic factors. Other determinants identified were access to extension services, awareness of climate change/variability, agricultural insurance, membership of farmer-based organization, and location of the farmer. Some adoption-related studies identified economic incentives as the major determining factor of the adoption of climate smart agricultural practices [ 27 , 28 ]. However, the factors affecting adoption of agricultural practices goes beyond just socio-demographic factors and economic incentives and are largely influenced by individual and intrinsic motivations [ 29 , 30 ] and other perceptions which can best be explained by psychological theories [ 31 , 32 ]. As yet, there is a dearth of empirical studies on the influence of individual and intrinsic motivation on the adoption of CSA practices in SSA and Ghana in particular. This study addresses this gap by using the Theory of Planned Behavior developed by [ 33 ] to examine the behavioral intention and actual adoption behavior of smallholder farmers toward CSA practices in Ghana. The Theory of Planned Behavior was chosen for this work because it provides socio-psychological basis for understanding human behavior [ 34 , 35 ] in diverse fields to encourage behavior change [ 36 ]. The main aim of the study was to identify the socio-psychological factors that influence farmers’ behavioral intention and actual behavior towards the adoption of CSA practices in rural Ghana using the Theory of Planned Behavior. The specific objectives of this study were to:

Ghana, like many SSA countries, has sought to promote CSA through its sustainable agricultural development policy [ 15 ]. A National Climate-Smart Agriculture and Food Security Action Plan was developed with the aim of facilitating and operationalizing the National Climate Change Policy for effective incorporation of climate change into food and agriculture sector development policies and programs [ 16 ]. The action plan sought to provide a multi-sectoral institutional mechanism for climate-smart agriculture [ 16 ]. Over the years, numerous efforts have been made by the Government of Ghana and international organizations to promote the adoption of CSA practices to help mitigate the impact of climate change [ 17 , 18 ]. Despite these efforts CSA adoption remains low among smallholder farmers in many parts of Ghana [ 19 ]. However, the few farmers that have adopted practices attest to their effectiveness in increasing farm productivity and incomes, enhancing food security, and conserving the natural resources in Ghana [ 20 – 22 ].

Climate-Smart Agriculture aims to achieve three pillars: (1) sustainably increase agricultural productivity and incomes; (2) enhance farmers’ adaptive capacity and build resilience; and (3) reduce the emission of greenhouse gases (GHGs) [ 12 , 13 ]. It has become imperative for farmers in developing countries to adopt and use CSA practices since they include numerous inexpensive farm-based sustainable agricultural land management techniques such as water management, zero/minimum tillage, residue management, and agroforestry among others. Additionally, CSA practices mostly include traditional practices and indigenous knowledge that are widely known to, and used by, farmers in addressing climatic risks [ 3 , 14 ].

Agriculture in Ghana, like most SSA countries is largely rainfall-dependent and employs about 75% of the rural population [ 7 ], but extreme weather events arising from climate change pose a serious threat to the agricultural sector and agri-based livelihoods. Projections from climate models point to a worsening situation in Ghana. For example [ 8 ], reported that the annual mean temperature is projected to increase by 2.0°C and 3.9°C while rainfall is also projected to decrease by 10.9% and 18.6% by the years 2050 and 2080, respectively. Historical data indicate a worrying trend of shifting climatic conditions that encompass erratic and declining rainfall patterns and a warming trend across all the agro-ecological zones of Ghana [ 9 ]. These climatic changes are estimated to reduce cassava and rice yields by 13.5% and 8% by the year 2050 [ 10 ]. As such, crop yields will continue to decline unless farmers adopt and utilize Climate-Smart Agricultural (CSA) practices [ 11 ].

Increased rainfall variability and drought associated with climate change poses the greatest challenge to the food systems and sustainable agricultural development of sub-Saharan Africa (SSA) [ 1 , 2 ] and to the region’s food and nutrition security [ 3 ]. SSA is regarded as the most vulnerable region to the negative impacts of climate change, because of structural and institutional weakness, high poverty levels and the low adoption of modern techniques that limits farmers’ capacity to adapt [ 4 – 6 ].

The TPB was extended with additional two hypotheses (H 5 and H 6 ) which showed a direct relationship between attitude and behavior, and subjective norm and behavior. Sapp et al. [ 66 ] argue that, behavior intentions may be ill-informed at certain times leading to inconsistency between intention and actual behavior. It is therefore critical to examine the attitude–behavior and subjective norm–behavior relation to provide a better understanding of their impact on actual behavior because such relation has been largely ignored in the literature. Studies such as [ 67 , 68 ] have asserted that psychological factors such as attitudes and subjective norms are not always mediated by intention but can have a direct influence on actual behavior.

Perceived behavior control relates to the perceived ease or difficulty in performing a particular behavior. Perceived behavior control concerns itself with the existence of control factors that may hamper or enable the performance of a particular behavior [ 43 ]. These control factors may be in the form of money, skills, time as well as cooperation with others [ 62 ] and these may determine the farmers’ ability to carry out a particular behavior. A farmer’s engagement in a given behavior is subject to the farmer’s belief in the likelihood of having access to the required resources and opportunities [ 44 ]. Perceived behavior control is an essential predictor of farmers’ intention to adopt farm practices [ 60 , 63 , 64 ]. By extension, Perceived Behavior Control has a direct influence on intention and behavior [ 54 , 65 ].

Subjective norm includes perceived social influence from internal and/or external sources to carry out or not to carry out a particular behavior. Such pressure may arise from internal sources such as family members and relatives or external sources such as friends and personnel from a government agency or an NGO [ 56 ]. The perceived approval of behavior by important people within a community also serves as a source of pressure that induces individuals’ intention of performing that particular behavior [ 57 ]. Subjective norm, therefore, measures the influence of the society on the decision-making process of a farmer [ 58 ]. Subjective norm has been estimated to be the most important determining factor of farmers’ intention to adopt new practices [ 59 – 61 ].

Attitude refers to the favorable or unfavorable assessment of behavior. The overall assessment of behavior and belief in its desired results determine the attitude towards a behavior [ 50 ]. By implication, a more positive attitude towards a behavior leads to a better intention of carrying out that behavior [ 51 ]. Several studies [ 47 , 52 , 53 ] have indicated the role of attitude in predicting farmers’ intention to adopt farm practices. Attitude can be regarded as a significant determinant of an individual’s intention and behavior [ 54 , 55 ].

The behavior intention of a farmer can be defined as that farmer’s motivation regarding their plan or conscious decision to apply effort to carry out a particular behavior [ 44 , 45 ]. Behavior intention represents the immediate antecedent and best predictor of performing an actual behavior [ 33 ]. By implication, stronger behavior intention towards a behavior indicates a stronger likelihood of performance of that behavior [ 46 ]. Such behavioral intention can accurately be estimated from the farmer’s attitude towards that particular behavior, subjective norm, and perceived behavior control [ 33 , 43 ]. However, limited studies [ 47 , 48 ] have examined the relationship between behavior intention and actual behavior due to the difficulty in measuring actual behavior. In addressing this, our study used past adoption behavior as a proxy for future adoption behavior particularly because farmers’ adoption of CSA practices shows a high degree of temporal stability [ 47 , 49 ].

This study aimed to explain the adoption of CSA practices using the Theory of Planned Behavior (TPB) as developed by [ 33 ] ( Fig 1 ). Although a plethora of studies indicate the importance of economic incentives in driving the adoption behavior of farmers [ 31 , 37 – 39 ], the TPB has proven valuable in explaining the decision-making process of farmers [ 40 , 41 ]. This is because farmers are not only profit-maximizing entities [ 42 ], but can be influenced by other individual and intrinsic motivations especially when the decision may have both social and environmental consequences [ 29 , 30 ]. The TPB predicts people’s intention to follow a particular behavior based on the assumption that human behavior is regulated by behavioral intentions which are determined by the attitude, subjective norm, and perceived behavior control of individuals [33, 43; Fig 1 ].

Disagreement from respondents on what constituted climate smart practices is a limitation of the current study. The researchers resolved this limitation by providing further explanations as to what CSA practices were and the goals they seek to achieve. Another limitation of the study was focusing solely on the original factors of the Theory of Planned Behavior in explaining adoption of CSA practices. However, the authors saw this as necessary due to the extensive literature available on other factors affecting adoption decision of farmers. In spite of the limitations, the current study has strengths in terms of measuring CSA practices by not limiting it to a simple yes/no response but by measuring the frequency of use of these practices. The use of a Likert scale in measuring the adoption helps to ensure that a farmer who uses any CSA practice on yearly basis has a greater adoption score than a farmer who rarely uses or had used the given practice only once. Future study can build on this study by recategorizing the CSA practices under similar themes so as to measure the impact of the psychological factors on these sub-themes.

The structural modeling involved the estimation of a set of multiple regressions with particular emphasis on the nature and magnitude of the relation between the latent constructs [ 78 , 81 ] in this case attitude (ATT), subjective norms (SN), perceived behavior control (PBC), behavioral intention (BI) and actual behavior (CSA adoption). The predictive power and the ability of the SEM to estimate multiple regressions simultaneously made it the appropriate tool to examine the causal relations that exist among the TPB constructs and to test the underlying hypotheses. The SEM was estimated using the maximum likelihood procedure because maximum likelihood estimation procedure has proven to produce reliable and robust results under different circumstances compared to other estimation procedures [ 82 ].

The study used Structural Equation Modeling (SEM) with latent constructs to analyze the collected data following [ 59 , 78 , 79 ]. The first step involved Confirmatory Factor Analysis (CFA) to acquire a suitable measurement model. Step two covered the development and testing of the structural model. CFA was carried out to assess the validity of constructs as well as to evaluate the fitness of the model. [ 80 , 81 ] indicate the need for conducting CFA because construct validity reveals the extent to which the measured items reflect the hypothetical construct they are intended to measure. The validity of the measurement model was assessed using the overall goodness-of-fit statistics. Overall goodness-of-fit was assessed by checking the chi-squared value, the root mean square error of approximation (RMSEA), the comparative fit index (CFI), and the standardized root mean square residual (SRMR) [ 59 , 78 ]. Cronbach alpha and factor loadings were used to establish the reliability of the constructs and various items.

Perceived behavior control was measured with six items. These items covered the control a farmer had over actions needed to adopt CSA practices. The fifth latent construct (CSA adoption) consisted of eight items covering CSA practices such as; the use of drought-tolerant varieties, cover cropping, zero tillage, no burning of crop residues, mixed cropping, planting early maturing varieties, water management/irrigation, and intercropping with legumes. A four-point Likert scale was used for these items.

Following the TPB guidelines, constructs for behavioral intention, attitude, subjective norm, and perceived behavioral control followed the principle of compatibility to avoid the occurrence of weaker and less-robust correlations among the constructs [ 77 ]. These constructs were defined in terms of the same element (i.e., CSA practices) to ensure construct compatibility and we also ensured that measurement scales were compatible across study sites to achieve scale compatibility [ 34 , 43 ]. Behavioral intention to adopt CSA practices was measured by four items, which enquired about farmers’ willingness to utilize CSA practices (with or without support) and their willingness to overcome barriers in terms of finance and information. Attitude towards CSA was measured using six items, three of which were concerned with the importance, convenience, and practicability of CSA practices. The other three considered the possible contributions of CSA practices in terms of increases in yield, on-farm income, and reputation. Subjective norm toward CSA was measured using six items, three of these items were about the motivation to use CSA practices while the other three covered the perception of others concerning adopting CSA practices.

Smallholder farmers were randomly selected using the Census and Survey Processing System (CSPro) software in the seven farming communities. The survey was conducted between August 2021 and September 2021 using locally trained enumerators. Interviews were conducted at the convenience of the farmers at their homes and lasted between 45 to 60 minutes. The survey instrument consisted of a questionnaire that solicited information on the socio-demographic characteristics of the respondents, and questions framed base on the theory of planned behavior about CSA practices ( S1 File ). Four of the latent constructs (i.e. behavioral intention, attitude, subjective norm, and perceived behavioral control) were measured using twenty-two items adopted and modified from [ 75 , 76 ]. A five-point Likert scale was used for all the items (Part 1 in S1 File ).

Ethical approval for this study was provided by the Humanities and Social Sciences Research Committee (HuSSRECC) of the Kwame Nkrumah University of Science and Technology, Ghana. HuSSRECC subjected the protocol to a thorough review and, among other things, observed that the necessary precautions have been taken to ensure that the participants in study will be well protected from risks and other distasteful occurrences they may face in the administration of questionnaire in particular. Formal consent for participation was obtained verbally from each study participant after the study objectives have been interpreted to them in their local dialect. Study participants were assured of anonymity and confidentiality.

Three hundred and fifty (350) household surveys were conducted in the seven study communities. A total of 87 households (38 in Sagadugu and 49 in Mimima) were interviewed in the West Mamprusi Municipality. Eighty-eight (88) households (46 in Ayelbia, 20 in Sinabisi and 22 in Feo-Asabere) were interviewed in the Bongo District while 175 households (87 in Yikene and 88 in Zaare) were interviewed in the Bolgatanga Municipality.

Three districts, namely West Mamprusi Municipality in the North East Region and Bongo District and Bolgatanga Municipality in the Upper East Region of Ghana, which have significant rural populations with agriculture as the main source of livelihoods were purposively selected. These districts were selected because they host several CSA demonstration fields of the Ghana Agricultural Sector Investment Program (GASIP). Subsequently, with the assistance of district agricultural officers, Sagadugu and Minima in the West Mamprusi Municipality, Yikene and Zaare in the Bolgatanga Municipality, and Ayelbia, Sinabisi, and Feo-Asabere in the Bongo District were selected.

These districts were selected because they are among the most vulnerable to drought in Ghana and the majority of the populace are dependent on rain-fed agriculture for their livelihood [ 73 , 74 ]. Consequently, several projects and interventions such as the knowledge systems and advisory services supporting CSA aimed at enhancing farmers’ adoption of CSA practices have been instituted in these areas.

The Bolgatanga Municipality is located between longitudes 0°30’W and 1°00’W and latitudes 10°30N and 10°50’N. The Bolgatanga Municipality has a total population of 139,864, comprising 66,607 males and 73,257 females. Despite being a relatively urbanized municipality, livestock farming and crop production continue to be the main economic activity employing over 60% of the labor force within the municipality [ 72 ].

The Bongo District is located between longitudes 0°W and 1°30’W and latitudes 10°30’N and 11°N. The Bongo district has a total population of 120,254, comprising 56,920 males and 63,334 females [ 69 ]. Subsistence agriculture involving the production of sorghum, millet, rice, groundnuts, and maize is the main economic activity in the district [ 71 ].

The West Mamprusi Municipality lies between longitudes 0°35’ W and 1°45’ W and latitudes 9°55’ N and 10°35’ N. The municipality has a total population of 175,755, comprising 85,712 males and 90,043 females [ 69 ]. West Mamprusi municipality is rural with agriculture being the mainstay of the local economy [ 70 ]. The main agricultural activities in the municipality include the rearing of livestock and the production of maize, millet, sorghum, and groundnuts.

The study was carried out in the West Mamprusi Municipality in the North East Region, and the Bongo District and Bolgatanga Municipality in the Upper East Region of Ghana ( Fig 2 ). These districts lie within the Sudan savannah agro-ecological zone and have a single rainfall pattern that lasts from May/June to September/October.

4. Results

4.2. Item measurement in the TPB model Table 2 presents descriptive statistics, factor loadings, and Cronbach alpha for the various constructs of the TPB framework. All the 350 respondents reported that they used at least one of the eight CSA practices most prevalent in their localities. A comparison of the eight items that make up the CSA adoption construct shows that mixed cropping (94.9%) was the most used practice followed by intercropping with legumes (82.9%), planting early maturing varieties (73.1%), no burning of crop residues (67.4%), cover cropping (62.6%), use of drought-tolerant varieties (60.9%), zero tillage (57.4%) and water management/irrigation (17.7%). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Table 2. Descriptive statistics for CSA practices, behavioral intention, attitude, subjective norm and perceived behavior control of respondents towards the adoption of CSA practices. https://doi.org/10.1371/journal.pclm.0000082.t002 In terms of behavioral intention to adopt CSA practices, a cumulative 5% of the sample expressed disagreeable intention to adopt CSA practices while 3% showed neither agreeable nor disagreeable intention to adopt CSA practices. Item b1n6 (“I am willing to learn about CSA practices”) shows the highest mean score while item b1n1 (“I am willing to adopt CSA practices by myself; with or without financial support”) shows the least mean score. Farmers expressed a positive attitude (mean of 4.44) towards the adoption of CSA practices. The majority of farmers interviewed expressed an agreeable attitude towards the adoption of CSA practices. About 2% of the sample expressed a disagreeable attitude towards the adoption of CSA practices, while 6% seem indifferent about the adoption of CSA practices. About 25% of the sample expressed disagreeable subjective norms towards the adoption of CSA practices while about 10% of the sample expressed neither disagreeableness nor agreeableness towards the adoption of CSA practices. Sn6 (“CSA practices are something I speak about with important referents”) showed the highest mean score compared to sn8 (“I feel under pressure from extension agents to integrate CSA practices in my farming”) which received the lowest mean score. In terms of perceived behavioral control, 15% of the sample indicated disagreeableness while about 9% indicated that they were neither agreed nor disagreed with the items under this construct. Pbc4 (“I have the resources to implement the CSA practices”) showed the least mean score while pbc5 (“I can easily command to use CSA practices on my farm”) showed the highest mean score. Factor loadings from the confirmatory factor analysis (Table 2) show that the observed variables were significant at the p < 0.01 level and can be considered adequate, ranging from 0.18 to 0.89. Although, six items recorded factor loadings less than 0.30 as recommended for a sample size of at least 350 [82], they were maintained because they were greater than 0.10 and proved to establish a simple structure [42, 82] and suggested at least good contributions of these items to their respective constructs [83]. The factor loadings (Table 2) indicate that all the five latent variables satisfied the convergent validity test. The Cronbach alpha which was used to test for the reliability of the constructs indicated that all five constructs–attitude, subjective norm, perceived behavior control, behavioral intention and CSA adoption–recorded Cronbach alpha of above 0.60, implying that measurement scales for all the variables were internally consistent and reliable [82, 84].

4.3. Goodness-of-fit statistics Based on the “cut-off” points developed by [82] and presented in Table 3, we chose four measures namely: chi-squared/degrees-of-freedom (χ2 / df), comparative fit index (CFI), root mean square error of approximation (RMSEA), and standardized root mean squared residual (SRMR) to determine the overall model fit. Although a significant χ2 indicates an unfit model, this was expected due to the large sample size and a high number of observed variables hence the χ2 is not sufficient to measure the overall fit of the model [82, 85, 86]. The CFI which is less sensitive to model complexity shows that the model is fit given the “cut off” point of 0.92 for large sample sizes [82]. The observed value for RMSEA which attempts to rectify the tendency of using χ2 to reject models with large sample sizes [82, 87] indicates a good fit given an observed value of 0.069. The observed value of 0.08 for SRMR suggests no problem with the model fit indicating that the estimated model is significant and inferences made can be reliable [82, 88]. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Table 3. Goodness-of-fit indices for (n > 250 and observed variables ≥ 30). https://doi.org/10.1371/journal.pclm.0000082.t003

[END]
---
[1] Url: https://journals.plos.org/climate/article?id=10.1371/journal.pclm.0000082

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/