论文信息

Chaima Zouaoui / Hassen Abdelhafidh / Houda Mazhoud 等 4 人

Frontiers in Sustainable Food Systems

论文解读

AI 辅助整理
方法
对53户农户问卷调查,用有序probit模型和PLS-SEM分析。
发现
主观规范影响最强,其次为感知行为控制和态度,社会经济因素也显著。
机会点
可扩大样本并结合数字推广渠道,研究社会规范如何通过信息化手段促进气候适应技术采纳。

论文摘要

在气候变化对农业系统影响日益加剧的背景下,本研究探讨了突尼斯中部凯鲁万省农户采用良好农业规范(Good Agricultural Practices, GAP)意愿的影响因素。本研究旨在为更有效的农业推广方法和支持政策提供证据,以促进气候韧性型和可持续耕作方式的采用。研究在舍比凯乡对53名农户开展了实地调查,将农场与农户的描述性特征分析同两种基于计划行为理论(Theory of Planned Behavior, TPB)的互补方法相结合:有序概率模型(ordered probit model)和偏最小二乘结构方程模型(Partial Least Squares Structural Equation Modeling, PLS-SEM)。有序概率模型结果表明,TPB的三个核心维度与农户采用GAP的意愿显著相关,同时部分社会经济与制度特征也与之相关。这些发现凸显了行为因素与农户社会经济背景在解释GAP采用意愿方面的相关性。PLS-SEM分析证实了TPB三个构念对采用意愿具有直接且显著的影响,并显示主观规范的影响最强,其次是感知行为控制和态度。探索性发现表明,态度、主观规范和感知行为控制均与农户采用相关实践的意愿相关联。这些关系均为正向,即各因素水平越高,意愿往往越强。但样本量较小,因此在得出结论时应保持谨慎。这些发现强调了社会环境和感知能力在塑造GAP采用决策中的核心作用,并指出需要采取参与式、扎根于社会关系的推广策略,同时开展更大规模的调查以确认本研究所识别的社会经济决定因素。

全文

Frontiers in Sustainable Food Systems

Frontiers in Sustainable Food Systems 01 frontiersin.org
TYPE Original Research
PUBLISHED 16 November 2026
DOI 10.3389/fsufs.2026.2008966
OPEN ACCESS
EDITED BY
Siphe Zantsi,
Agricultural Research Council of South
Africa (ARC-SA), South Africa
REVIEWED BY
Giulio Cascone,
University of Catania, Italy
Utkal Sapkota,
Wageningen University and Research,
Netherlands
CORRESPONDENCE
Houda Mazhoud
houda.mazhoud@inrat.ucar.tn
RECEIVED 24 September 2026
REVISED 23 October 2026
ACCEPTED 27 October 2026
PUBLISHED 16 November 2026
CITATION
Zouaoui C, Abdelhafidh H,
Mazhoud H and Chemak F (2026)
Determinants of farmers’ intention to
adopt good agricultural practices under
climate change: evidence from the
Chebika community, governorate of
Kairouan, Tunisia.
Front. Sustain. Food Syst. 10:2008966.
doi: 10.3389/fsufs.2026.2008966
COPYRIGHT
© 2026 Zouaoui, Abdelhafidh, Mazhoud
and Chemak. This is an open-access
article distributed under the terms of the
Creative Commons Attribution License
(CC BY). The use, distribution or
reproduction in other forums is
permitted, provided the original
author(s) and the copyright owner(s) are
credited and that the original publication
in this journal is cited, in accordance
with accepted academic practice. No
use, distribution or reproduction is
permitted which does not comply with
these terms.
Determinants of farmers’
intention to adopt good
agricultural practices under
climate change: evidence from
the Chebika community,
governorate of Kairouan, Tunisia
Chaima Zouaoui 1, Hassen Abdelhafidh 1, Houda Mazhoud 2,3
and
Fraj Chemak 2,3
1Agricultural Production Systems and Sustainable Development Laboratory (LR03AGR02), Higher
School of Agriculture of Mograne, University of Carthage, Mograne, Tunisia, 2Rural Economy
Laboratory (LR16INRAT07), National Institute of Agricultural Research of Tunisia (INRAT), Tunis, Tunisia,
3University of Carthage (UCAR), Tunis, Tunisia
Against the backdrop of the growing impacts of climate change on agricultural
systems, this study investigates the factors influencing farmers’ intention to adopt
Good Agricultural Practices (GAP) in the Kairouan Governorate of central Tunisia.
The study aims to generate evidence to inform more effective agricultural exten-
sion approaches and support policies that can foster the adoption of climate-
resilient and sustainable farming practices. A field survey was carried out among
53 farmers in the delegation of Chebika, combining a descriptive characteriza -
tion of farms and farmers with two complementary approaches grounded in the
Theory of Planned Behavior (TPB): an ordered probit model and Partial Least
Squares Structural Equation Modeling (PLS-SEM). The ordered probit results indi-
cate that the three core TPB dimensions are significantly associated with farmers’
intention to adopt GAP, alongside selected socioeconomic and institutional char-
acteristics. These findings highlight the relevance of both behavioral factors and
farmers’ socioeconomic context in explaining GAP adoption intentions. The PLS-
SEM analysis confirms the direct and significant effect of the three TPB constructs
on adoption intention and shows that subjective norms exert the strongest influ-
ence, followed by perceived behavioral control and attitude. The exploratory
findings show that attitudes, subjective norms and perceived behavioral control
are all linked to farmers’ intention to adopt practices. These relationships are
positive meaning levels of each factor tend to go with stronger intentions. Still the
sample size is small so we should be careful when drawing conclusions. These
findings underline the central role of the social environment and of perceived
capacity in shaping GAP adoption decisions, and point to the need for participa-
tory, socially-anchored extension strategies, together with larger-scale surveys
to confirm the socio-economic determinants identified here.
KEYWORDS
adoption, climate change, determinants, good agricultural practices, intention,
Kairouan, Theory of Planned Behavior

Zouaoui et al. 10.3389/fsufs.2026.2008966
Frontiers in Sustainable Food Systems 02 frontiersin.org
1 Introduction
Tunisia is among the countries with the lowest water resource
endowment in the Mediterranean basin (Chouchane et al., 2015). This
water constraint is characterized by strong spatial variability, with the
northern part relatively well supplied in surface water, while the cen-
tral and southern regions face significant scarcity (Albouchi, 2003;
Arfa et al., 2025 ). This disparity is further aggravated by climate
change, which represents a serious threat to agriculture and food secu-
rity. Rapid and unpredictable impacts, rising temperatures, irregular
rainfall, prolonged droughts, and extreme weather events continue to
intensify the vulnerability of agricultural systems (Gharbi et al., 2025;
IPCC, 2022). In Tunisia, this vulnerability is particularly pronounced
in arid and semi-arid areas and is expected to increase in the absence
of adaptation strategies, especially in agriculture (Arfa et al., 2025).
Within this framework, irrigation has gradually become a key
lever for securing and intensifying agricultural and food production
(Ben Mechlia, 2004; Abdelhafidh et al., 2022). In 2024, the total irri-
gated area in Tunisia reached approximately 442,500 hectares, includ-
ing 250,500 hectares of public irrigation schemes and 192,000 hectares
of private schemes, requiring nearly 2.8 billion m3 of water per year
which is about 75% of total water resources mobilized for use
(MARHP , 2024). Irrigated agriculture thus contributes to food secu-
rity, income stability, and the development of high-value agricultural
production (Samaali and Mjejra, 2020). This dynamic is particularly
evident in the governorate of Kairouan, where irrigated areas cover
approximately 61,670 hectares, or about 14% of national irrigated land
(CRDA, 2024). The expansion of irrigated areas is mainly based on
increased exploitation of groundwater resources, resulting in signifi-
cant economic and social transformations at the farm level (Llamas
and Martínez-Santos, 2005). This intensification nonetheless faces
several constraints. From a socio-economic perspective, farms
encounter difficulties related to market access and financial resources,
which weaken the resilience and sustainability of agricultural systems
(Gharbi et al., 2025). From an institutional and technical perspective,
collective management structures particularly water use associations
(WUA) face administrative and financial management difficulties, as
well as problems in maintaining often degraded irrigation networks.
The accumulation of debt and the non-payment of water fees have, in
some cases, led to the cessation of these structures’ activities and to a
loss of collective access to water (Jerbi, 2018).
In parallel, farmers adopted individual coping strategies, such as
the development of private wells and deep boreholes, sometimes with-
out official authorization (Jerbi, 2018). This intensification of ground-
water exploitation has led to a gradual decline in groundwater levels
since the 1980s, now exceeding natural recharge capacities. In 2024,
the groundwater exploitation rate in the governorate of Kairouan
reached 115% of available resources (CRDA, 2024), highlighting the
critical pressure exerted on these aquifers. The most depleted deep
aquifers in the region include the Kairouan plain, Sisseb-El Alem,
Chougafia, Bouhafna, and Aïn Bou Morra ( MARHP , 2020)
(Tables 1–3).
The combination of water scarcity, rising agricultural irrigation
demand, climate change impacts, and the limitations of existing
irrigation systems makes it necessary to reconsider water manage -
ment in agriculture ( El Amri et al., 2013 ; Chebil et al., 2026 ). To
meet the growing challenges that climate change brings using GAP
has become an approach to make agricultural systems more resil -
ient and sustainable (Ben Nasr and Oueslati, 2026 ). Climate change
TABLE 1 Measurement of the TPB constructs.
TPB construct Code Item description
Attitude Att_ Rdt If I adopt GAP , it will
increase my farm’s yield
Att_Resis If I adopt GAP , this will
strengthen my production’s
resistance to the dry
season.
Att _choix Adopting GAP on my farm
is a good choice
Att_utili I am convinced of the
usefulness of GAP
Subjective norms NS _Ins The staff of support
institutions encourage me
to adopt GAP
NS _Ami My friends expect me to
adopt GAP
NS _Voi My neighbors expect me to
adopt GAP
NS _fam My family thinks I should
adopt GAP on my farm
Perceived behavioral
control
CCP _compe I have the necessary skills
to adopt GAP on my farm
CCP_RF I have the necessary
resources to adopt GAP .
Intention Int_satis My regular adoption of
these good agricultural
practices to date has been
satisfactory.
Int_cont I intend to continue
adopting these good
agricultural practices in the
coming years.
Int_enc I will confidently
recommend or encourage
other farmers to adopt
these good agricultural
practices.
TABLE 2 Socioeconomic variables included in the ordered probit model.
Variable Description
Age Farmer’s age (years)
Farming experience Number of years of agricultural activity
Education level Illiterate
Primary
Secondary
Higher education
Access to training Ye s
No
Access to subsidies Ye s
No

Zouaoui et al. 10.3389/fsufs.2026.2008966
Frontiers in Sustainable Food Systems 03 frontiersin.org
is making farming by raising temperatures shifting rain patterns
causing more droughts and making water more scarce. These
changes threaten crop growth and the livelihoods of farmers. In this
situation many GAP help farmers adapt to climate change. They do
this by using resources efficiently and by making farming systems
stronger so they can handle weather changes. For example, water-
efficient irrigation reduces water loss. Helps farmers cope with less
water. Water collection and storage let farmers capture and keep
water for use during times. Modern farming tools, such as soil-
moisture sensors and digital decision-making tools let farmers
manage water better when the weather changes. GAP also include
scheduling irrigation according to crop water requirements, soil
moisture conditions, and weather forecasts. This helps reduce water
stress during periods of drought. Mulching the soil keeps it cooler
reduces evaporation and holds moisture. Adding soil amendments
improves soil structure, fertility and the soil’s ability to hold water,
which helps crops survive drought. Diversifying crops spreads risk
so that one bad weather event does not ruin all produce. Choosing
the crop species and varieties that fit local climate, water availability
and higher temperatures also helps. Together all these GAP
strengthen the ability of systems to adapt lower their vulnerability,
to climate-related risks and support more sustainable farming when
the climate changes ( FAO, 2013; Ben Nasr and Oueslati, 2026 ).
Within this framework, the adoption of GAP represents a relevant
strategy ( Requier-Desjardins, 2015 ). They rely on existing knowl -
edge to make agricultural systems environmentally, economically,
and socially sustainable, while ensuring the quality and safety of
both food and non-food products ( Taamallah, 2010).
The theoretical framework used to analyze farmers’ adoption deci-
sions is the Theory of Planned Behavior (TPB), first introduced by
Ajzen (1985) and formalized in Ajzen (1991), which is widely used to
predict and explain human behavior in specific contexts (Ben Nasr
and Oueslati, 2026 ; Sapkota et al., 2025 ; Bhandari et al., 2025 ;
Feisthauer et al., 2024). Its core constructs: attitude, subjective norms,
and perceived behavioral control directly shape an individual’s inten-
tion to perform a given behavior, intention being itself the most
immediate antecedent of actual behavior.
This study aims to assess the intention of farmers in the governor-
ate of Kairouan to adopt GAP in the context of climate change, with a
particular focus on their social and behavioral determinants. In par-
ticular, the behavioral determinants of farmers’ intentions to adopt
GAP remain insufficiently explored in this context. This study
addresses this gap by examining farmers’ adoption intentions through
the TPB framework, complemented by an analysis of socioeconomic
correlates. Its contribution is exploratory, providing context-specific
evidence to inform future research and policies promoting sustainable
agricultural practices.
2 Methodology
2.1 Theoretical framework: the Theory of
Planned Behavior
The TPB, introduced by Ajzen (1985) and formalized in Ajzen
(1991), is widely used in the social sciences to explain and predict
human behavior in specific contexts; its relevance has been confirmed
across several domains, including research on agricultural practices
(Senger et al., 2017). The framework assumes that behavior is primar-
ily driven by an individual’s intention: a stronger and more favorable
intention increases the likelihood that the behavior will be performed,
and the factors shaping this intention depend on the type of behavior
considered and on the context in which it occurs (Ajzen, 1991). The
model is structured around three constructs. Attitude refers to an indi-
vidual’s overall evaluation of a behavior favorable or unfavorable based
on beliefs about its attributes and expected outcomes. Subjective norm
reflects perceived social pressure, particularly the expectations of sig-
nificant others, which may shape an individual’s intention. Perceived
behavioral control relates to an individual’s perception of their own
ability to perform the behavior, taking into account available resources
and anticipated constraints (Ajzen, 1988).
TABLE 3 Demographic characterization of respondents (n = 53).
Variable Average/
percentage (%)
Age of respondents average (years) 47
Gender Male 100%
Female 0%
Secondary activities(Job, business, remittance,
farm labor)
26.40%
Education level Illiterate 5.70%
Primary level 60.30%
Secondary level 28.30%
High education 5.70%
Farming experience
(years)
Up to 10 20.8%
11–20 22.60%
21–30 28.30%
31 and above 28.30%
Credit Access Ye s 17.00%
No 83,00%
Access to subsidies Ye s 19%
No 81%
Member of a
professional
oraganisation
Ye s 36.00%
No 64.00%
Agricultural Training Ye s 28.3%
No 71.7%
Farm size (ha) From 1 to 15 83.30%
From 16 to 30 11 10%
31 and above 5 70%
Production system Rain-fed 0 00%
Irrigated 86 8%
Mixed 13.2%
Irrigation source Well 1.9%
Borehole 66.0%
Water course 24.5%
Public network 7.5%
Breeding Ye s 68%
No 32%

Zouaoui et al. 10.3389/fsufs.2026.2008966
Frontiers in Sustainable Food Systems 04 frontiersin.org
Within the scope of this study, these constructs are used to exam-
ine GAP adoption by farmers. Attitude reflects how farmers evaluate
GAP; subjective norm captures the perceived social influence sur -
rounding their adoption; and perceived behavioral control refers to
farmers’ perceived ability to implement GAP , given the resources
available to them and the constraints they may encounter. On this
basis, the following hypotheses were formulated and tested
empirically:
H1: Subjective norms have a significant, positive effect on farm-
ers’ intention to adopt GAP .
H2: Farmers’ attitudes toward GAP have a significant, positive
effect on their intention to adopt these practices.
H3: Perceived behavioral control has a significant, positive effect
on farmers’ intention to adopt GAP .
H4: Subjective norms have a positive relationship with attitude
towards GAP .
H5: Farmers’ attitudes toward GAP have a positive relationship
with their perceived behavioral control regarding the adoption
of these practices.
H6: Subjective norms a positive relationship with Perceived
behavioral control over GAP .
2.2 Study area and sampling
This study was conducted in the delegation of Chebika, governor-
ate of Kairouan, central Tunisia. The delegation has an estimated total
cultivable area of 35,381 ha, managed by 1,579 farmers, reflecting the
strong dependence of rural households on agricultural activity in the
area (CRDA, 2024). Chebika belongs to a semi-arid zone character-
ized by hot, dry summers and relatively mild winters. These climatic
conditions impose significant constraints on agricultural production,
particularly regarding water availability and soil degradation. Average
annual rainfall is estimated at approximately 285 mm, below the gov-
ernorate average of 303 mm, reflecting a structural water deficit. The
crop structure is clearly oriented toward arboriculture, dominated by
olive cultivation (2,382 ha), followed by cereals (670 ha), legumes
(262 ha), and forage crops (30 ha). The agricultural intensification
rate, estimated at 121%, reflects considerable pressure on land and
water resources (CRDA, 2024).
A sample of 53 farmers was surveyed across three localities of the
delegation of Chebika (Hamed, El Karma, and Sidi Ali Ben Salem).
The questionnaire used in this survey was tested and validated in con-
sultation with national and regional stakeholders and experts. Farmers
were selected randomly, primarily with the assistance of the Regional
Commissariat for Agricultural Development (CRDA) of Kairouan by
giving the farmers list and showing their location. The survey was
conducted in face to face mode during April 2026. The questionnaire
aimed to explore farmers’ perceptions of the importance of GAP in
coping with climate change, as well as their perceptions of climate
change itself. The TPB constructs were measured using 10 items
assessed on a five-point Likert scale (1 to 5): four items for attitude,
four for subjective norms, and two for perceived behavioral control.
2.3 Data analysis
This study is structured around two complementary levels of
analysis. The first descriptive stage characterizes the socio-eco -
nomic profile of the surveyed farmers, their level of GAP adoption,
and the distribution of the TPB constructs, with a particular focus
on farmers’ intention to adopt GAP as a response to climate
change. An average intention score was computed for each farmer
and classified into three levels: low, moderate, and high using an
equal-interval approach thresholds. Mean scores for attitude, sub -
jective norms, and perceived behavioral control were then com -
pared across intention levels, both to illustrate how these
constructs vary with intention and to support the subsequent
modeling.
Building on this descriptive foundation, two complementary
modeling techniques were used: an ordered probit model and struc-
tural equation modeling (SEM).
2.3.1 Ordered probit model
The ordered probit model was used to estimate the effect of socio-
economic characteristics and TPB constructs on farmers’ adoption
intention, accounting for the ordinal nature of the dependent variable
(low, moderate, high intention). The dependent variable INT ᵢ* is
explained by a vector of TPB constructs (G) and a vector of socio-
economic and institutional variables (Z):
1 12INT iii GZββε∗= ++
Where β1 and β2 are vectors of parameters to be estimated and εᵢ
is the error term.
The observed ordinal categories of INT are defined by two
unknown threshold parameters, N1 and N2, such that INTᵢ = low if
INTᵢ
≤ N1, moderate if N1 < INTᵢ
≤ N2, and high if INTᵢ* > N2.
2.3.2 Structural equation modeling
Partial least squares structural equation modeling (PLS-SEM)
was employed as a complementary approach to the ordered Probit
model to examine the relationships among the latent constructs
underlying farmers’ intention to adopt GAP . Unlike the ordered
Probit model, which focuses on the determinants of categorized
intention, PLS-SEM enables the simultaneous assessment of the mea-
surement model and the structural relationships among the theoreti-
cal constructs.
SEM is a confirmatory statistical methodology designed to test
hypothesized causal relationships among latent constructs ( Byrne,
2010). It is widely used for its capacity to simultaneously assess
assumed causal relationships and latent structures ( Everitt and
Dunn, 2011 ), to clarify relationships between observed variables
(Schumacker and Lomax, 2004), and to examine complex direct and
indirect relationships ( Tabachnick and Fidell, 2013 ) through a
structured sequence of parameter identification, estimation, and
model-fit assessment ( Mueller and Hancock, 2018 ). The analysis
was conducted using SmartPLS 4.2. All constructs were specified
reflectively, meaning that the observed indicators were treated as
manifestations of their respective underlying latent variables. The
measurement model was evaluated in terms of indicator reliability,

Zouaoui et al. 10.3389/fsufs.2026.2008966
Frontiers in Sustainable Food Systems 05 frontiersin.org
internal consistency reliability, convergent validity, and discrimi -
nant validity ( Fornell and Larcker, 1981 ). Specifically, indicator
loadings, Cronbach’s alpha, composite reliability, average variance
extracted (AVE), and the heterotrait–monotrait ratio (HTMT) were
examined, as appropriate. The evaluation of the model included
examination of the estimated path coefficients and their statistical
significance. Additionally, the evaluation considered the coefficient
of determination written as R squared and relevant measures of
explanatory and predictive performance. Statistical significance
assessment used a nonparametric bootstrapping procedure with five
thousand resamples, in Smart PLS. therefore, no missing-data impu -
tation or deletion procedure was required. The structural model was
evaluated based on the estimated path coefficients, their statistical
significance, and the explanatory power of the model. He signifi -
cance of the estimated relationships was assessed using bootstrap -
ping with 5,000 resamples.
3 Main results
3.1 Farmers and farm characteristics
Descriptive analysis highlights farmer characteristics closely
related to their intention to adopt GAP . The average age of surveyed
farmers is 47 years, corresponding to a mature age group; age is
generally taken as an indicator of maturity, from which one can infer
greater capacity for informed decision-making, although younger
and middle-aged farmers tend to be more receptive to new tech -
nologies and practices ( Waaswa et al., 2021 ). The relatively high
average age observed in the sample does not, therefore, exclude a
genuine willingness to adopt GAP , particularly among younger
farmers.
Regarding gender, all of the household heads were male. This does
not mean that women do not play a principal role in the decision
making process. Commonly, in Tunisia, men have traditionally been
associated with the role of family provider and representative, while
women have often assumed responsibility for household chores, child-
care, and managing the family’s daily needs. These traditional roles
can influence the definition of the “head of the household, ” particu-
larly in rural areas and within families where patriarchal structures
remain influential.
Farming experience averages 24 years, reflecting a compara -
tively strong professional background that may foster a better
understanding of farming practices and a greater ability to assess
and adopt innovations. Educational attainment, by contrast,
remains generally low: the majority of farmers have a primary edu -
cation level (60.3%), followed by secondary education (28.3%),
while only 5.7% have reached higher education and 5.7% are illiter -
ate. Education is widely recognized as a key driver of innovation
and of the adoption of improved practices in agricultural and rural
development ( Strauss et al., 1991 ); in this context, the low educa -
tional attainment observed may constrain farmers’ understanding
of GAP related issues and limit their adoption intention, highlight -
ing the need for awareness-raising and extension activities tailored
to this profile. Access to agricultural support services also remains
limited. Only 28.3% of farmers report having benefited from agri -
cultural training, even though access to extension services is known
to directly influence farmers’ awareness and confidence in applying
new techniques ( Dhehibi et al., 2022 ). This limited technical sup -
port restricts the dissemination of knowledge on GAP and reduces
farmers’ capacity to implement them effectively. Access to subsidies
is even more limited, at 6.3% of farmers, representing a major
financial constraint on investment in improved agricultural
practices.
The distribution of farms according to utilized agricultural area
reveals a predominance of small farms, those under 15 hectares
accounting for 83.3% of the farmers surveyed. Medium-sized
farms, ranging from 15 to 30 hectares, make up 11.1% of the
sample, while large farms exceeding 30 hectares represent only
5.6% of the total.
Regarding the production system, no farm relies exclusively on
rain-fed agriculture: 86.8% of farms are fully irrigated and 13.2%
combine irrigated and rain-fed production. This strong depen -
dence on irrigation makes water management a central issue for
GAP adoption. Livestock breeding is practiced by 68% of the
respondents, which shows that most farms are diversified and that
crop and livestock activities are closely linked, with implications
for manure management, input use and farm hygiene. As for the
irrigation source, groundwater is dominant: boreholes supply
66.0% of farms and wells 1.9%, while watercourses account for
24.5% and the public network for only 7.5%. This reliance on
groundwater raises concerns about the sustainability of water
resources and the water quality, which are directly relevant to GAP
(Figures 1, 2).
3.2 Farmers’ climate change perception
Results show that 70% of the investigated farmers are aware of
climate change.
FIGURE 1
Conceptual model of the Theory of Planned Behavior (TPB), adapted from Ajzen (1991) and Chepng’etich et al. (2025).

Zouaoui et al. 10.3389/fsufs.2026.2008966
Frontiers in Sustainable Food Systems 06 frontiersin.org
Farmers’ Perception of the trend of climatic parameters are
presented in Figure 3 . Based on a survey of farmers in the study
area, the respondents were questioned about their perceptions of
climate change. The results indicated that rising average tempera -
tures and hotter days is the most marked event with 45% of the
responses. The change in seasonal patterns is the second reported
event where 42% of farmers report shifts in the timing and dura -
tion of seasons, affecting agricultural activities. The third most
frequently reported event was the emergency of diseases, account -
ing for 37% of responses. This was followed by a decrease in pre -
cipitation and an increase in the frequency of droughts, each
reported by 34% of respondents. The increase of extreme events
and the increase of water scarcity did not exceed 18% of responses.
The results indicate that farmers primarily perceive climate
change through rising temperatures and changes in seasonal
cycles, followed by the emergence of diseases and reduced
precipitation.
3.3 Categories of GAP and their adoption
intentions
Table 4 presents the mean intention scores regarding various
GAP . These scores, measured on a five-point Likert scale (ranging
from 1 “strongly disagree” to 5 “strongly agree”), reflect the farm -
ers’ willingness to adopt specific GAP . The mean intention scores
are high, indicating that farmers are, overall, prepared to imple -
ment these good practices. The degree of willingness to adopt
them varies moderately from one practice to another. Water-
efficient irrigation received a score of 4.7 (M = 4.7; SD = 0.5). Next
came the use of soil amendments (M = 4.5; SD = 0.7) and irriga -
tion planning (M = 4.4; SD = 0.8). In contrast, water collection
and storage recorded the lowest mean (M = 3.1; SD = 0.4).
Mulching (M = 3.7; SD = 0.9) and crop diversification (M = 3.8;
SD = 0.6) also received scores. Overall, farmers appear to attach
importance to water-efficient irrigation and soil management,
FIGURE 2
Geographic location of the delegation of Chebika in the governorate of Kairouan.
FIGURE 3
Farm householders’ perception about climatic parameters change.

Zouaoui et al. 10.3389/fsufs.2026.2008966
Frontiers in Sustainable Food Systems 07 frontiersin.org
although their choices vary depending on the specific methods
considered.
3.4 Farmers’ intention to adopt GAP
To facilitate the interpretation of farmers’ intention to adopt GAP ,
mean intention score was classified into three categories using an
equal-interval approach. Given the observed score range from 3 to 5,
the interval width was calculated by dividing the range (5–3 = 2) into
three equal parts, yielding an interval width of approximately 0.66.
Accordingly, scores were classified into low, moderate, and high inten-
tion categories using thresholds of approximately 3.66 and 4.32. This
classification provides a straightforward descriptive framework for
distinguishing levels of intention within the observed score range.
However, the thresholds are mathematically derived rather than
empirically validated behavioral cutoffs; consequently, the categories
should not be interpreted as absolute measures of farmers’ willingness
to adopt GAP . Furthermore, categorizing the continuous score entails
a loss of information, which should be considered when interpreting
the results of the subsequent regression analysis.
The distribution of farmers’ intention scores indicates that 54.7%
of respondents fall within the moderate-intention category, compared
with 28.3% in the low-intention category and 17.0% in the high-
intention category. The predominance of the middle category sug-
gests that respondents’ scores are concentrated within the
intermediate interval defined by the study. However, this distribution
alone does not establish the strength of farmers’ willingness to adopt
GAP in absolute terms, nor does it identify the factors underlying
their intention. The ordered probit analysis complements these
descriptive findings by examining the characteristics and perceptions
associated with the probability of belonging to a higher or lower
intention category.
3.5 Determinants of adoption intention:
ordered probit results
The ordered probit model shows a good overall fit: the likelihood-
ratio test is highly significant (LR χ2(8) = 58; p < 0.001), indicating that
the set of explanatory variables jointly explains farmers’ intention to
adopt GAP .
3.5.1 Effects of the TPB constructs
The results confirm the relevance of the TPB in explaining farm-
ers’ adoption intention, consistent with previous applications of the
model to agricultural contexts (Mahdavi, 2021). A positive attitude
toward GAP significantly increases the probability of medium and
high intention levels while reducing the probability of low intention
(β = 1.1082; p = 0.007); farmers with more favorable perceptions of
GAP are, accordingly, more likely to consider adopting them. The
high-intention marginal effect is statistically significant at the 1% level.
Among the three TPB constructs, attitude has the strongest positive
estimated coefficient, although coefficient magnitudes should not be
compared mechanically across variables measured on different scales.
Subjective norms also have a positive and significant effect (β = 0.576;
p = 0.067): social influence from family, neighbors, and institutional
actors plays an important role in adoption decisions, and the percep-
tion of support or approval from one’s social environment strengthens
the likelihood of adoption, consistent with Nguyen and Drakou (2021)
among coffee producers in Vietnam. The marginal effects indicate a
decrease of 7.62 percentage points in the probability of low intention
and an increase of 6.03 percentage points in the probability of high
intention. These effects are statistically significant at the 10% level
according to your table. This suggests that perceived social expecta-
tions and support may be associated with stronger intention, although
the statistical evidence is relatively weak (Tables 5, 6).
Finally, perceived behavioral control has a positive and significant
effect at 10% level (β = 0.616; p = 0.088) on intention, indicating that
the more farmers believe they possess the skills, resources, and ability
required to implement GAP , the stronger their intention to adopt
them. Its marginal effects indicate an 8.15-percentage-point decrease
in the probability of low intention and a 6.63-percentage-point
increase in the probability of high intention. Both reported marginal
effects are statistically significant at the 5% and 10% levels, respec-
tively. This pattern is consistent with the idea that farmers who feel
more capable of implementing GAP tend to report higher intention,
although the evidence for the overall coefficient is marginal at the 10%
significance level.
3.5.2 Effects of socio-economic factors
Among socio-economic variables, age has no significant effect on
farmers’ adoption intention, a result consistent with Salhi et al. (2012),
who reported no significant impact of age on the adoption of drip
irrigation in the Mitidja plain, Algeria. In contrast, access to subsidies
has a positive and significant effect: farmers who benefit from financial
support show a higher propensity to adopt new practices, consistent
with El Intidami and Benamar (2022), who highlight the key role of
subsidies in the adoption of agricultural innovations, notably localized
irrigation technology. The marginal effects suggest that access is asso-
ciated with a 16.52-percentage-point decrease in low intention and a
TABLE 5 Level-based classification of farmers’ adoption-intention scores.
Intention level Score range Percentage of
observations
Low <3.66 28.3%
Moderate 3.66–4.32 54.7%
High >4.32 17%
TABLE 4 Farmers intentions to adopt GAP.
GAP Mean Std. Dev.
Water-efficient irrigation 4,7 0,5
Water collection and
storage
3,1 0,4
Modern farming tools 4,3 0,6
Scheduling irrigation 4,4 0,8
Mulching the soil 3,7 0,9
Adding soil amendments 4,5 0,7
Diversifying crops 3,8 0,6
Choosing adaptive
varieties
4,0 0,7

Zouaoui et al. 10.3389/fsufs.2026.2008966
Frontiers in Sustainable Food Systems 08 frontiersin.org
13.45-percentage-point increase in high intention. The high-intention
marginal effect is reported as significant at the 5% level. Nevertheless,
because few respondents received subsidies, these estimates should be
treated as exploratory. They do not establish that subsidies increase
intention.
Likewise, education level shows a positive association with farm-
ers’ adoption intention, although the effect is statistically significant
only at the 10% level and should therefore be interpreted with caution.
This finding suggests that more educated farmers may have better
access to and understanding of information related to climate change
and GAP , potentially enhancing their capacity to make informed
adoption decisions. Farming experience also has a positive and sig-
nificant effect, consistent with Ewulo et al. (2025): greater experience
strengthens farmers’ ability to adjust their practices and to appreciate
the benefits of new techniques, resulting in a higher propensity to
adopt them. The marginal effects indicate that an additional year of
farming experience is associated with a 1.03-percentage-point
decrease in the probability of low intention and a 0.84-percentage-
point increase in the probability of high intention. The high-intention
marginal effect is significant at the 1% level.
The training variable has a negative effect, which is different
from previous studies showing that training usually encourages
technology adoption. However, this result should be interpreted
carefully because only 28.3% of the 53 farmers received training
(about 15 farmers). The small sample makes the result sensitive to
unusual cases and reduces statistical reliability. Therefore, the nega-
tive effect does not necessarily mean that training discourages GAP
adoption. A larger and more balanced sample is needed to confirm
this finding.
3.5.3 Robustness and sensitivity analysis
To examine whether the main findings are sensitive to the cat -
egorization of farmers’ intention to adopt GAP , the ordered probit
results were compared with those obtained from an OLS regression
using the original intention scores for the 53 respondents ( Table 7).
This alternative specification retains the continuous information
contained in the individual scores and provides a supplementary
assessment of the consistency of the estimated relationships.
Overall, the comparison indicates partial robustness. Farming expe -
rience, attitude, subjective norms, and perceived behavioral control
exhibit positive and statistically significant coefficients in both
models, suggesting that their associations with adoption intention
are relatively stable across the two outcome specifications.
Education also displays a positive association in both models,
although the statistical evidence is comparatively weak. By contrast,
access to subsidies and access to training are statistically significant
only in the ordered Probit model, indicating that the evidence con -
cerning these variables is sensitive to the outcome specification. Age
remains statistically insignificant in both models. Taken together,
these findings provide encouraging evidence that several key asso -
ciations are not solely attributable to the categorization of inten -
tion scores.
3.6 Structural equation model results
3.6.1 Measurement model evaluation
Indicator reliability was first assessed through factor loadings
(Table 8 ); most loadings exceed the commonly accepted 0.60
threshold ( Hulland, 1999 ), supporting the adequacy of the mea -
surement items. Multicollinearity was examined using variance
inflation factors, all of which remain below the critical threshold of
5, confirming the absence of collinearity concerns ( Henseler et
al., 2014).
Table 9 shows the reliability and convergent validity results for
the constructs. The constructs Attitude, Subjective Norms and
Intention show internal consistency. Their composite reliability,
rho_a and Cronbach’s alpha all go above the threshold of 0.70.
Their AVE values are also above 0.50 which confirms validity
TABLE 6 Summary of the ordered probit results (LR χ2(8) = 58, p < 0.001).
Variables Coefficient Std. Err Marginal effects
Low intention Medium
intention
High intention
Age −0.0214 0.0252 0.0028 −0.0005 −0.0023
Education level 0.5697* 0.3426 −0.0754* 0.0140 0.0613
Farming experience 0.0779*** 0.0280 −0.0103 *** 0.0019 0.0084***
Access to subsidies 1.2491* 0.6623 −0.1652 ** 0.0307 0.1345**
Access to training −1.4782 ** 0.6143 0.1956** −0.0364 −0.1592 ***
Subjective norms 0.5760* 0.3148 −0.0762* 0.0142 0.0603*
Perceived behavioral
control
0.6162* 0.3619 −0.0815 ** 0.0152 0.0663*
Attitude 1.1082*** 0.4081 −0.1466 *** 0.0273 0.1193***
/Cut1 9.787 2.394
/Cut2 13.321 2.841
Log likelihood −23.402
Pseudo R2 0.55
***p < 0.01; **p < 0.05; *p < 0.1. Statistically significant marginal effects estimated from an ordered probit regression.

Zouaoui et al. 10.3389/fsufs.2026.2008966
Frontiers in Sustainable Food Systems 09 frontiersin.org
(Fornell and Larcker, 1981 ). Perceived Behavioral Control shows
reliability compared to the others. Its composite reliability is 0.731
and Rho_a is 0.710. Cronbach’s alpha is only 0.572, which is below
the acceptable level. Still the AVE for Perceived Behavioral
Control is 0.577, which’s, above the 0.50 threshold so it does meet
the requirement for convergent validity. Because the Cronbach’s
alpha is low and there are two indicators used for Perceived
Behavioral Control the internal consistency should be interpreted
carefully. The results support reliability and convergent validity
for most constructs. However special attention should be given to
the measurement quality of Perceived Behavioral Control.
3.6.2 Structural model evaluation
The structural model explains 49.6% of the variance in adoption
intention (R2 = 0.496), reflecting a moderate level of predictive rele-
vance. Explanatory power is moderate for perceived behavioral con-
trol (R2 = 0.220) and comparatively weaker for attitude (R2 = 0.084),
highlighting differences in the contribution of the endogenous con-
structs. The standardised root mean square residual (SRMR) was
0.075 below 0.08 falls within the range considered acceptable in the
literature below from 0.08 ( Mahdavi, 2021); the normed fit index
(NFI) reaches 0.756, below the conventional 0.90 threshold (Bentler
TABLE 7 Robustness check using the original intention score.
Independent
variables
Ordered probit
coef.
p-value probit OLS using
original scores
coef,
p-value OLS Assessement
Age −0.021 0.396 −0.006 0.464 Consistent
Education level 0.569 0.096 0.206 0.07 Consistent but weak evidence
Farming experience 0.078 0.005 0.020 0.022 Strong consistency
Access to subsidies 1.249 0.059 0.214 0.341 Direction consistency
Access to training −1.478 0.016 −0.328 0,093 Consistent but weak evidence
Subjective norm 0.576 0.067 0.238 0.028 Consistent
Perceived behavioural control 0.616 0.072 0.266 0,09 Consistent
Attitude 1.108 0.007 0.266 0.034 Consistent
TABLE 8 Measurement model results: outer loadings and construct reliability.
Construct Indicators Original
sample
Sample
mean (M)
Standard
deviation
(STDEV)
T statistics p-values
Attitude Att_ Rdt 0.878 0.867 0.063 13.886 0.000
Att_Resis 0.907 0.896 0.047 19.461 0.000
Att _choix 0.834 0.827 0.070 11.953 0.000
Att_utili 0.735 0.730 0.089 8.232 0.000
NS NS _Ins 0.916 0.911 0.034 26.598 0.000
NS _Ami 0.903 0.899 0.037 24.185 0.000
NS _Voi 0.589 0.588 0.106 5.552 0.000
NS _fam 0.917 0.915 0.029 31.118 0.000
CCP CCP _comp 0.806 0.791 0.130 6.196 0.000
CCP_RF 0.710 0.683 0.189 3.754 0.000
Intention Int_satis 0.911 0.908 0.029 31.653 0.000
Int_cont 0.888 0.884 0.035 25.288 0.000
Int_enc 0.912 0.908 0.030 30.812 0.000
TABLE 9 Reliability and validity of constructs.
Construct Composite reliability
(Rho_c)
Composite reliability
(Rho_a)
Cronbach’s alpha AVE
Attitude 0.906 0.860 0.86 0.707
NS 0.905 0.874 0.852 0.71
CCP 0.731 0.710 0.572 0.577
Intention 0.931 0.888 0.877 0.816

Zouaoui et al. 10.3389/fsufs.2026.2008966
Frontiers in Sustainable Food Systems 10 frontiersin.org
and Bonett, 1980 ). This result suggests that the model’s overall fit
requires cautious interpretation. This may be a sample-size artifact
rather than evidence of structural misspecification. The impact of
independent variables related to farmers on dependent variables is
presented in Table 10 and Figure 4.
The structural model validates all six hypotheses. Attitude toward
GAP has a positive and significant, though comparatively modest,
effect on adoption intention (β = 0.102), consistent with Coulibaly et
al. (2021) among cocoa producers in Côte d’Ivoire. Subjective norms
constitute the most influential direct determinant of intention
(β = 0.504), underscoring the decisive role of social pressure exerted
by family members and agricultural institutions, in line with Atta-
Aidoo et al. (2022) in Ghana; within the Tunisian context of
Kairouan, this predominance suggests that farmers place consider-
able weight on the expectations and recommendations of their social
environment before changing their practices. Perceived behavioral
control also exerts a positive and significant effect (β = 0.278), reflect-
ing the central role of farmers’ perceived technical, material, and
organizational capacity, consistent with Tama et al. (2021) on the
adoption of conservation agriculture. Beyond these direct effects, the
strong effect of the NS on the intention may be explained by the indi-
rect influence of NS on the Attitude (β = 0.291) and on the perceived
behavioral control (β = 0.277) which indicates that farmers are more
likely to convert a sense of capability into intention when their social
environment supports and encourages GAP adoption this perspec-
tive is based on the idea that farmers can develop greater awareness
and personal decision-making methods. These enhanced reflection
and decision-making skills help them better understand the impor-
tance of effectively implementing GAPs. The positive relationship
between attitude and perceived behavioral control (β = 0.307) high-
lights the importance for farmers of perceiving GAPs favorably in
order to feel more capable of implementing them. This result cor -
roborates the principles of the TPB (Ajzen, 1991) as well as the find-
ings of previous studies indicating that farmers’ attitudes and their
sense of control over their actions are key determinants of their will-
ingness to adopt sustainable farming methods (Zeweld et al., 2017;
Tama et al., 2021). In short, these results indicate that when farmers
view GAPs positively and believe in their ability to implement them,
they are more inclined to adopt them. Overall, these results demon-
strate that psychological factors, social influence, and perceived capa-
bility act jointly rather than in isolation to shape farmers’ intentions
to adopt GAPs; these combined effects remain relatively under-doc-
umented in the empirical literature, which has largely focused on the
direct effects of constructs from the TPB.
TABLE 10 Path coefficients of the structural model.
Path Original
sample (O)
Sample
mean (M)
Standard
deviation
(STDEV)
T statistics p-values 2.5% 97.5%
Att→CCP 0.307 0.321 0.130 2.36 0.018 0.052 0.559
Att→Int 0.102 0.109 0.052 1.96 0.050 0.020 0.311
CCP→Int 0.278 0.273 0.125 2.23 0.026 0.010 0.506
NS→Att 0.291 0.302 0.115 2.52 0.012 0.055 0.499
NS→CCP 0.277 0.283 0.123 2.25 0.024 0.031 0.510
NS→Int 0.504 0.506 0.098 5.17 0.000 0.302 0.683
Bootstrapping with 5,000 subsamples.
FIGURE 4
Structural equation model of the determinants of farmers’ intention to adopt GAP (standardized path coefficients; SmartPLS 4).

Zouaoui et al. 10.3389/fsufs.2026.2008966
Frontiers in Sustainable Food Systems 11 frontiersin.org
4 Discussion
Faced with growing water scarcity and the intensifying impacts of
climate change, understanding what drives farmers’ intention to adopt
GAP is a prerequisite for designing effective agricultural support poli-
cies in the governorate of Kairouan. Combining an ordered probit
model with PLS-SEM within an extended reading of the TPB, this
study confirms that all three core constructs of the theory: attitude,
subjective norms, and perceived behavioral control are positive and
significant determinants of farmers’ adoption intention, a result con-
sistent with prior applications of the TPB to agricultural adoption
decisions (Mahdavi, 2021; Nguyen and Drakou, 2021). The results also
reveal positive relationships among the three TPB constructs, suggest-
ing that farmers’ social perceptions, attitudes, and perceived capacity
to implement GAP are interconnected within the estimated model.
However, given the cross-sectional design of the study, these relation-
ships should be interpreted as associations rather than causal effects.
This finding, echoed by Atta-Aidoo et al. (2022) in the Ghanaian con-
text, suggests that in the agricultural community of Kairouan, farmers
attach considerable importance to the judgment and experience of
their social circle family members, neighbors, and agricultural institu-
tions before committing to new practices. This social validation mech-
anism constitutes a powerful, and currently under-exploited, lever for
extension policy: strategies that mobilize peer farmers, farmer orga-
nizations, and locally respected figures are likely to be more effective
than purely technical or individually targeted extension messages.
The significant relationships further show that attitude, subjective
norms, and perceived behavioral control do not operate indepen -
dently: a positive attitude toward GAP is most likely to translate into
strong adoption intention when it is reinforced by both a supportive
social environment and a genuine sense of capability. This finding, still
comparatively under-documented in the empirical literature on GAP
adoption, suggests that extension and support programs are likely to
be more effective when they simultaneously target farmers’ evaluation
of GAP , their social environment, and their perceived technical and
material capacity, rather than addressing these dimensions separately.
Regarding socio-economic determinants, education level and
farming experience emerge as positive and significant drivers of adop-
tion intention, confirming the role of accumulated knowledge and
skills in facilitating the understanding and appropriation of new prac-
tices (Ewulo et al., 2025; Strauss et al., 1991). Access to subsidies also
has a clear positive effect, consistent with the broader literature on the
role of financial incentives in reducing the investment barriers associ-
ated with agricultural innovation (El Intidami and Benamar, 2022).
The negative coefficient obtained for participation in training pro-
grams is the only result that runs counter to theoretical expectations
and to most of the empirical literature, which typically finds a positive
association between training and adoption ( Wu et al., 2023 ). This
result should not be read as evidence that the training received by
farmers in Chebika was inadequate or poorly targeted; no data were
collected in this survey on the content, duration, or quality of the
training programs attended, so any claim about their adequacy would
be unsupported. The more defensible interpretation is a purely statisti-
cal one: with only 53 farmers surveyed and barely one in three having
received training, the corresponding subgroup is too small, and poten-
tially too heterogeneous in other respects, to yield a stable estimate of
the training effect. This limitation should be explicitly considered by
future research through larger and more balanced samples of trained
and untrained farmers, ideally complemented by information on the
type and quality of the training received, before drawing any policy
conclusion regarding agricultural training in the region.
Beyond the domain of training, the modest overall sample size
(N = 53) is a limitation that applies to the study as a whole. While it is
sufficient for the exploratory PLS-SEM and ordered probit analyses
conducted here both of which are recognized as suitable for small,
non-normally distributed samples (Henseler et al., 2014) it nonethe-
less constrains the statistical power and generalizability of individual
coefficients, particularly those relating to socio-economic subgroups.
This reinforces the case for treating the present findings as a first,
theory-grounded characterization of adoption intention in Chebika,
to be confirmed through a larger-scale survey across the governorate
of Kairouan.
5 Conclusion
Faced with the growing challenges posed by climate change and
increasing water scarcity, the adoption of Good Agricultural Practices
has become a strategic necessity for farmers in the governorate of
Kairouan. This study analyzed the determinants of farmers’ intention
to adopt these practices by combining an ordered probit model and
structural equation modeling within an extended reading of the
TPB. The results show that, despite an overall favorable perception of
GAP , adoption intention remains predominantly moderate within the
studied sample, reflecting the existence of real barriers that limit the
transition from intention to action and that deserve full consideration
in agricultural support policies.
The study highlights the central role of the social environment in
shaping adoption intention: farmers in Kairouan attach considerable
importance to the judgment and experience of their social circle
before committing to new practices, a social validation mechanism
that remains insufficiently mobilized by current support programs. It
also confirms the importance of perceived behavioral control, educa-
tion, farming experience, and access to subsidies, and shows that these
psychological and socio-economic dimensions interact rather than
operate in isolation.
The counter-intuitive negative effect associated with training par-
ticipation should be treated as a methodological caveat linked to the
limited size of the trained subsample, rather than as evidence against
agricultural training itself; it points to the need for a larger, dedicated
survey capable of properly isolating the effect of training programs
and, ideally, of their content and quality on adoption intention.
Ultimately, promoting GAP adoption in the governorate of Kairouan
cannot rely on the individual efforts of farmers alone. It requires a
collective dynamic involving public institutions, professional agricul-
tural organizations, research bodies, and local communities, united
around a shared vision and coordinated actions. The findings offer
several practical implications for promoting farmers’ adoption of
GAP . First, where favorable attitudes are associated with adoption
intention, agricultural extension services should organize on-farm
demonstrations and field days to illustrate the potential economic,
agronomic, and environmental benefits of GAP . Second, to strengthen
subjective norms, farmer-to-farmer learning, peer demonstration
plots, and local farmer groups could be encouraged to facilitate
knowledge exchange and reinforce social support for adoption. Third,
where perceived behavioral control is an important determinant of
adoption intention, policymakers should prioritize accessible

Zouaoui et al. 10.3389/fsufs.2026.2008966
Frontiers in Sustainable Food Systems 12 frontiersin.org
technical training, reliable extension services, and improved access to
the financial and material resources required to implement
GAP . Educational initiatives should also be adapted to farmers’ differ-
ent learning needs, although the education-related finding should be
interpreted cautiously because statistical significance was observed
only at the 10% level. Finally, given the exploratory nature of the study,
the relatively small sample of 53 farmers, these recommendations
should be considered preliminary. Further research using larger and
more diverse samples is needed to assess the robustness and wider
applicability of the findings.
6 Limitations and future research
This study has limitations. First the sample is small. 53 Farmers
and comes from one delegation. That makes it hard to say if the
results apply broadly. Second the study is cross-sectional meaning it
only looked at things at one point in time. Also the data came from
farmers telling us their intentions to adopt GAP . That means we
cannot say for sure that their intentions cause changes in behavior.
There’s also a chance people gave answers they thought were more
acceptable, not their thoughts. We did not measure GAP adoption
directly so we do not know if what people said matches what they
actually do.
Also the measure for perceived control (PBC) used only two ques-
tions. That’s not very detailed. Using an equal-interval approach
threshold to define for grouping the responses might have lost some
information. Important factors like access to credit or support from
agricultural extension services were not included in the study. Because
of all this the findings are best seen as exploratory. A starting point,
not an answer.
For research it would help to use bigger and more varied samples.
A longitudinal design. Following farmers over time. Would show how
adoption changes. Measuring actual adoption behavior directly would
give results. Using detailed measures for PBC and including other
important factors, like credit or extension support would also improve
the study. These changes could help us understand what really drives
farmers to adopt GAP .
Data availability statement
The raw data supporting the conclusions of this article will be
made available by the authors, without undue reservation.
Author contributions
CZ: Conceptualization, Data curation, Formal analysis,
Investigation, Methodology, Writing – original draft, Writing – review
& editing. HA: Formal analysis, Investigation, Methodology, Project
administration, Supervision, Validation, Writing – review & editing.
HM: Conceptualization, Data curation, Methodology, Project admin-
istration, Validation, Writing – original draft, Writing – review & edit-
ing. FC: Data curation, Funding acquisition, Investigation,
Methodology, Project administration, Resources, Supervision,
Validation, Visualization, Writing – review & editing.
Funding
The author(s) declared that financial support was received for this
work and/or its publication. This work was supported by the Ministry
of Higher Education and Scientific Research of Tunisia through the
TWISTT project, funded under the PRIMA Partnership.
Conflict of interest
The author(s) declared that this work was conducted in the
absence of any commercial or financial relationships that could be
construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that Generative AI was not used in the
creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this
article has been generated by Frontiers with the support of artificial
intelligence and reasonable efforts have been made to ensure accuracy,
including review by the authors wherever possible. If you identify any
issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors
and do not necessarily represent those of their affiliated organizations,
or those of the publisher, the editors and the reviewers. Any product
that may be evaluated in this article, or claim that may be made by its
manufacturer, is not guaranteed or endorsed by the publisher.
References
Abdelhafidh, H., Gharbi, I., and Maaroufi, A. (2022). Irrigation water demand management
to address water scarcity. Emir. J. Food Agric. 34, 792–799. doi: 10.9755/ejfa.2022.v34.i9.2944
Ajzen, I. (1985). “From intentions to actions: a theory of planned behavior, ” in Action
Control: From Cognition to Behavior , eds. J. Kuhl and J. Beckmann (Cham: Springer).
Ajzen, I. (1988). Attitudes, Personality, and Behavior. Milton Keynes: Open University Press.
Ajzen, I. (1991). The theory of planned behavior. Organ. Behav. Hum. Decis. Process. 50,
179–211. doi: 10.1016/0749-5978(91)90020-T
Albouchi, L. (2003) Gestion de L ’ eau en Tunisie: D’une Politique de Mobilisation à une
Politique de Réallocation de la Ressource Selon sa Valorisation Économique. Cas du Bassin
Versant du Merguellil, Tunisie Centrale. Université Montpellier I.
Arfa, L., Abdelhafidh, H., and Harzli, I. (2025). Assessment of farm households’ percep-
tion of and adaptation strategies to climate change risks in Tunisian semi-arid areas. SAGE
Open 15:317. doi: 10.1177/21582440251405317
Atta-Aidoo, J., Antwi-Agyei, P ., Dougill, A. J., Ogbanje, C. E., Akoto-Danso, E. K., and
Eze, S. (2022). Adoption of climate-smart agricultural practices by smallholder farmers
in rural Ghana: an application of the theory of planned behavior. PLoS Clim. 1:e0000082.
doi: 10.1371/journal.pclm.0000082
Ben Mechlia, N. (2004). Perspectives de valorisation de l’ eau d’irrigation dans les pays du
Maghreb. Options Méditer Série B 47, 169–185.
Ben Nasr, J., and Oueslati, A. (2026). The adoption of good agricultural practices
(GAP) in the Tunisian cereal sector: an extended theory of planned behavior (TPB)

Zouaoui et al. 10.3389/fsufs.2026.2008966
Frontiers in Sustainable Food Systems 13 frontiersin.org
using PLS-SEM modeling. Front. Sustain. Food Syst. 10:1815334. doi: 10.3389/
fsufs.2026.1815334
Bentler, P . M., and Bonett, D. G. (1980). Significance tests and goodness of fit in the analy-
sis of covariance structures. Psychol. Bull. 88, 588–606. doi: 10.1037/0033-2909.88.3.588
Bhandari, G., Pandey, A., Sapkota, U., Singh, S. P ., and Muran, H. (2025). Pesticide use
and safety behavior among rice farmers in Nepal: the assessment of theory of planned
behavior and potential health risk. Environ. Dev. Sustain. 6, 1–30. doi: 10.1007/
s10668-025-06385-z
Byrne, B. M. (2010). Structural Equation Modeling with AMOS: Basic Concepts,
Applications, and Programming. Abingdon: Routledge.
Chebil, A., Soula, R., Abdelhafidh, H., and Ltaief, W . (2026). Stakeholders’ preferences for
sustainable groundwater management in the Mahdia region (Tunisia): application of the
analytic hierarchy process. Irrig. Drain. 75, 1214–1224. doi: 10.1002/ird.70097
Chepng’ etich, E., Mbeche, R., Ateka, J. M., and Obebo, F . (2025). Livestock farmers’ inten-
tions to adopt climate-smart agricultural practices in Kenya’s arid and semi-arid lands:
What role do behavioural factors play? Sustainability 17:7688. doi: 10.3390/su17177688
Chouchane, H., Hoekstra, A. Y ., Krol, M. S., and Mekonnen, M. M. (2015). The water
footprint of Tunisia from an economic perspective. Ecol. Indic. 52, 311–319. doi:
10.1016/j.ecolind.2014.12.015
Coulibaly, T. P ., Du, J., Diakité, D., Dabuo, F . T., and Coulibaly, I. (2021). Extending the
theory of planned behavior to assess farmers’ intention to adopt sustainable agricultural
practices: evidence from cocoa farmers in Côte D’ivoire. Agric. Mech. Asia Afr. Lat. Am.
51, 1733–1748.
CRDA (2024) Rapport Annuel sur les Ressources en eau et L ’ agriculture Irriguée du
Gouvernorat de Kairouan. Kairouan, Tunisia: Commissariat Régional au Développement
Agricole de Kairouan.
Dhehibi, B., Dhraief, M. Z., Ruediger, U., Frija, A., Werner, J., Straussberger, L., et al.
(2022). Impact of improved agricultural extension approaches on technology adoption:
evidence from a randomised controlled trial in rural Tunisia. Exp. Agric. 58:e13. doi:
10.1017/S0014479722000084
El Amri, A., Majdoub, R., Ben Ayed, S., and M'Sadak, Y . (2013). Évaluation du fonc-
tionnement d’un périmètre irrigué dans la région de Kairouan après sa mise en valeur.
Larhyss J. 13, 17–29.
El Intidami, M. E., and Benamar, F . (2022). Approche économique de l’analyse des déter-
minants socioéconomiques de l’adoption des innovations en agriculture. Revue AME 2,
237–258.
Everitt, B. S., and Dunn, G. (2011). Applied Multivariate Data Analysis. Hoboken: Wiley.
Ewulo, T. A., Akinseye, F . M., Teme, N., Agele, S. O., Y essoufou, N., and Kumar, S. (2025).
Factors driving climate-smart agriculture adoption: a study of smallholder farmers in
Koumpentum, Senegal. Front. Agron. 7:1552720. doi: 10.3389/fagro.2025.1552720
FAO (2013) Climate-Smart Agriculture Sourcebook. Rome: Food and Agriculture
Organization of the United Nations.
Feisthauer, P ., Hartmann, M., and Börner, J. (2024). Behavioral factors driving farmers’
intentions to adopt spot spraying for sustainable weed control. J. Environ. Manag.
353:120218. doi: 10.1016/j.jenvman.2024.120218
Fornell, C., and Larcker, D. F . (1981). Evaluating structural equation models with unob-
servable variables and measurement error. J. Mark. Res. 18, 39–50. doi:
10.1177/002224378101800104
Gharbi, I., Aribi, F ., Abdelhafidh, H., Ferchichi, N., Lajnef, L., Toukabri, W ., et al. (2025).
Assessment of the agroecological transition of farms in Central Tunisia using the TAPE
framework. Resources 14:81. doi: 10.3390/resources14050081
Hensler, J., Dijkstra, T.K., Sarstedt, M., Ringle, C.M., Diamantopoulos, A., Straub, D.W .,
et al (2014). Common beliefs and reality about PLS: comments on Rönkkö and Evermann.
Organ. Res. Methods 17, 182–209. doi: 10.1177/1094428114526928
Hulland, J. (1999). Use of partial least squares (PLS) in strategic management research: a
review of four recent studies. Strateg. Manag. J. 20, 195–204. doi: 10.1002/(SICI)1097
-0266(199902)20:2<195::AID-SMJ13>3.0.CO;2-7
IPCC (2022). Climate Change 2022: Impacts, Adaptation and Vulnerability. Cambridge:
Cambridge University Press.
Jerbi, H. (2018). Anthropisation des Processus Hydrologiques Autour de L ’ oued Merguellil
(Tunisie Centrale): Caractérisation des Formes D’ évolution et Quantification Des Flux
[Thèse de Doctorat]. Montpellier: Université de Montpellier.
Llamas, M.-R., and Martínez-Santos, P . (2005). Intensive groundwater use: silent revolu-
tion and potential source of social conflicts. J. Water Resour. Plan. Manag. 131, 337–341.
Mahdavi, T. (2021). Application of the theory of planned behavior to understand farmers'
intentions to accept water policy options using structural equation modeling. Water
Supply 21, 2720–2734. doi: 10.2166/ws.2021.138
MARHP . (2020). Rapport National de L ’ eau, Tunis, Tunisia: Ministère de l’ Agriculture, des
Ressources Hydrauliques et de la Pêche.
MARHP . (2024). Rapport National de L ’ eau, Tunis, Tunisia: Ministère de l’ Agriculture, des
Ressources Hydrauliques et de la Pêche.
Mueller, R. O., and Hancock, G. R. (2018). “Structural equation modeling, ” in The
Reviewer’s Guide to Quantitative Methods in the Social Sciences, eds. G. R. Hancock, L. M.
Stapleton and R. O. Mueller (Abingdon: Routledge).
Nguyen, N., and Drakou, E. G. (2021). Farmers’ intention to adopt sustainable agriculture
hinges on climate awareness: the case of Vietnamese coffee. J. Clean. Prod. 303:126828.
doi: 10.1016/j.jclepro.2021.126828
Requier-Desjardins, M. (2015). “Les bonnes pratiques agricoles, ” in CIHEAM-IAMM.
Communication Présentée à l’ atelier Sous-Régional ENPARD Machreq, (Le Caire, Le Caire,
Égypte).
Salhi, S., Imache, A., Tonneau, J.-P ., and Ferfera, M.-Y . (2012). Adoption du Système
Goutte-à-Goutte en Algérie. doi: 10.1684/agr.2012.0598
Samaali, H., and Mjejra, O. (2020) État des Lieux du Périmètre Irrigué de Nebeur. Actes
IA V Hassan II. Rabat, Morocco: Actes IAV Hassan II.
Sapkota, U., Bhandari, G., Sapkota, M., Khanal, S., Poudel, A., Khanal, D., et al. (2025).
Modeling vegetable farmers’ intention to use pesticides in Central Nepal: an extended
version of the planned behavior theory. Environ Challeng 18:101084. doi: 10.1016/j.
envc.2025.101084
Schumacker, R. E., and Lomax, R. G. (2004). A Beginner’s Guide to Structural Equation
Modeling. Mahwah, NJ: Lawrence Erlbaum Associates.
Senger, I., Borges, J. A. R., and Machado, J. A. D. (2017). Using the theory of planned
behavior to understand the intention of small farmers in diversifying their agricultural
production. J. Rural. Stud. 49, 32–40. doi: 10.1016/j.jrurstud.2016.10.006
Strauss, J., Barbosa, M., Teixeira, S., Thomas, D., and Gomes, R. (1991). Role of education
and extension in the adoption of technology: a study of upland rice and soybean farmers
in central-West Brazil. Agric. Econ. 5, 341–359.
Taamallah, H. (2010). Gestion Durable des Terres en Tunisie: Bonnes Pratiques Agricoles,
Tunisia.
Tabachnick, B. G., and Fidell, L. S. (2013). Using Multivariate Statistics. London: Pearson.
Tama, R. A. Z., Ying, L., Yu, M., Hoque, M. M., Adnan, K. M., and Sarker, S. A. (2021).
Assessing farmers’ intention towards conservation agriculture by using the extended
theory of planned behavior. J. Environ. Manag. 280:111654. doi: 10.1016/j.
jenvman.2020.111654
Waaswa, A., Nkurumwa, A. O., Kibe, A. M., and Ng’ eno, J. K. (2021). Communicating
climate change adaptation strategies: climate-smart agriculture information dissemina-
tion pathways among smallholder potato farmers in Gilgil Sub-County, Kenya. Heliyon
7:e07873. doi: 10.1016/j.heliyon.2021.e07873
Wu, F ., Guo, X., and Guo, X. (2023). Cooperative membership and new technology adop-
tion of family farms: evidence from China. Ann. Public Coop. Econ. 94, 719–739. doi:
10.1111/apce.12433
Zeweld, W ., Van Huylenbroeck, G., Tesfay, G., and Speelman, S. (2017). Smallholder farm-
ers’ behavioural intentions towards sustainable agricultural practices. J. Environ. Manag.
187, 71–81. doi: 10.1016/j.jenvman.2016.11.014

论文原文与出版信息

Frontiers in Sustainable Food Systems
查看论文原文

继续探索