Micro-Irrigation Technology Adoption in the Bekaa Valley of Lebanon: A Behavioural Model

Abstract

Potato crops are one of the main sources of income for farmers living in the Bekaa Valley of Lebanon. Given the high sensitivity of potatoes to water stress, water shortages can cause considerable losses in terms of potato yield and quality. To overcome this challenge, the use of water-saving technologies such as micro-irrigation systems are very important. However, the adoption of this technique remains quite low among potato farmers in the Bekaa region, who still use ordinary sprinkler systems. In this study, the unified theory of acceptance and use of technology (UTAUT) serves as the conceptual framework for investigating these farmers’ behaviour in adopting a new micro-irrigation system. To achieve this objective, we extended the UTAUT model by considering farmers’ risk perception of the use of a new micro-irrigation technology. The moderators tested were age, previous experience, voluntariness of use, gross unit margin and educational level. Examining the standard regression coefficients, i.e., the β weights, the results indicate that performance expectancy raised behavioural intention for investment in micro-irrigation (β = 0.29) while for effort expectancy the β weight value was 0.24. Overall, an increase of 1 standard deviation of the behavioural intention strongly impacted investment in micro-irrigation systems, β = 0.8 standard deviation of the effective adoption of the technology. Risk perception (β = −0.08) negatively affected farmers’ performance expectancy, i.e., the higher the perceived risk, the lower the perceived performance of the investment, which in turn affected their intention to use micro-irrigation systems. Age (β = 0.11) exerted a significant effect on effort expectancy. Finally in this paper, the policy implications of the results are discussed.

1. Introduction

Climate change is seriously threatening the agricultural sector and poses major risks for developed as well as developing countries [1,2]. High temperatures, increased evaporation and fluctuations in precipitation are altering water availability and reducing crop yields [2,3]. These factors affect the management of farms, especially in arid and semi-arid regions [4].

In Lebanon, current levels of water consumption are not sustainable in view of population growth, industrial development, the extension of irrigated agricultural land and the rising unregulated use of groundwater [5]. According to a recent report from the United Nations Development Programme [6], Lebanon is expected to face an increase in mean annual temperatures of between 1.2 and 1.7 °C by mid-century, as well as a 4–11% decrease in precipitation by 2100. The increased frequency of heatwaves and incidence of drought conditions are also projected to affect the country [6,7]. Thus, various conditions threatening water balance make adaptation to climate change particularly difficult in Lebanon.

The Bekaa Valley region of Lebanon is rapidly facing the effects of drought and decreased water availability [8]. This region lies to the east of the Lebanon Mountains and is a fertile valley approximately 16 km wide and 129 km long that gently slopes from north to south from an altitude of 900 to 1100 m and represents 42% of Lebanon’s area. It is divided into three main zones: North Bekaa, Central Bekaa and West Bekaa. The Bekaa Valley is the most important production area in Lebanon, accounting for the highest percentage of seasonal cultivations (60%), which include cereals, potatoes, vegetables and grapevines. The production of potatoes typically ranks first among the top 10 commodities produced in Lebanon each year, with a total production of 390,000 tonnes in 2017 [9]. Two-thirds of Lebanon’s potato production comes from the Bekaa Plain, which is entirely irrigated [10]. Potato is one of the most sensitive crops to soil moisture stress and requires a systematic irrigation schedule [11].

Unfortunately, long- and short-term environmental challenges in the Bekaa Valley are related to water shortages, water quality problems, groundwater table depletion and the impacts of climate change [10]. All of these factors severely threaten the sustainability of irrigated crops in the Bekaa Valley. Recurrent droughts in 2008 and 2014 characterised by low precipitation augmented the effects of water stress [12]. In this context, the improvement of efficient water use in irrigation and the conservation of water resources are becoming strategic priorities. Currently, the majority of farmers in the Bekaa Valley use surface water and ordinary sprinkler irrigation to irrigate potatoes affected by water availability, especially in spring and summer [13]. Micro-irrigation, particularly mini-sprinklers, could be a solution to the above-mentioned climate-change related problems [14]. Mini-sprinklers are small sized static sprinklers with a flow varying between 150 and 300 L per hour and a pressure of 1.5 bars [15] inducing a water cooling canopy [16]. Notably, the adoption of micro-irrigation in potato cultivation can result in irrigation water savings of up to 40% [17,18], energy savings, and increased crop quality and yields [19,20].

This study investigated the importance of farmers’ perceptions, motivations and socio-economic factors in affecting their investment in and adoption of a new micro-irrigation system. To this end, we adopted the unified theory of acceptance and use of technology (UTAUT) model [21], which facilitates the analysis of individual acceptance and the use of new technology by disentangling and evaluating the influencing factors. Furthermore, many researchers have identified individual risk behaviour as a key guide of technology acceptance [22,23]. Accordingly, we modified the UTAUT model to include the impact of risk factors on the acceptance of a new micro-irrigation system.

The partial least squares structural equation model (PLS-SEM) [24,25] was employed to quantify the relevance of these factors for a sample of farmers producing potato crops in the Bekaa Valley. From these results, we provide certain policy interventions and management recommendations to enhance the use of a new micro-irrigation system. To the best of our knowledge, this is the first study to use a modified UTAUT model and the PLS-SEM methodology to shed light on the impact and importance of behavioural factors in influencing the adoption and use of a micro-irrigation system.

The paper is organised as follows. In Section 2, we briefly analyse the UTAUT model, extend it to consider risk variables, and illustrate our research hypotheses. Section 3 describes the methodological approach, while Section 4 presents the results obtained using a PLS-SEM estimation method. In Section 5, we discuss the results and Section 6 presents the main conclusions. Finally, in Section 7 we examine some research limitations.

2. Behavioural Models and Research Hypotheses

A number of behavioural theories have been proposed to explain individuals’ adoption (i.e., use behaviour) of new technologies. The present study applied UTAUT [21], which integrates previous technology acceptance models [26,27,28,29,30,31,32,33]. The model suggests four major determinants that act on a person’s ‘use behaviour’ when adopting a technology: the ‘performance expectancy’, the ‘effort expectancy’, the ‘social influence’ and the ‘facilitating conditions’. The first three constructs influence use behaviour through a behavioural intention latent variable, while the fourth construct directly impacts the use behaviour. Furthermore, the model supposes that factors such as age, gender, experience and voluntariness of use considerably moderate the effect of the crucial parameters [21]. Figure 1 presents the standard UTAUT model.

The performance expectancy (PE) determinant refers to the user’s level of belief in how advantageous a system usage will be and how it will help to achieve certain benefits [21]. Studies on technology adoption have consistently revealed that when participants believe a technology to be useful, they are more likely to accept it [21,34]. Performance expectancy aggregates aspects of job performance, including usefulness [31,33], job fit [32], relative advantage [30], extrinsic motivation [35] and outcome expectations that are related to the consequences of the behaviour [27,36,37]. Notably, performance expectancy is expected to positively influence behavioural intention (BI) and technology acceptance [21].

The second determinant, effort expectancy (EE), suggests that the ease of use associated with a system is an important element in the adoption of a new technology [21]. Effort expectancy was previously proposed by [21,32]. Earlier works found that the higher the perceived ease of use of a new technology, the higher the intention to adopt it. Thus, a positive relationship is expected between effort expectancy and the behavioural intention to adopt a new technology [21].

The third determinant suggested to influence a person’s behavioural intention to adopt a technology is social influence (SI). Social influence refers to the extent to which individuals perceive they should adopt the new technology based on inputs from people who occupy valuable positions in their life; in other words, ‘the degree to which peers influence the use of a new technology’ [21,38,39,40]. Social influence is usually identified by the use of three constructs: (i) subjective norms (i.e., the person’s perception that individuals who are important to her or him think that they should or should not perform the specific behaviour) [26,28,29,31]; (ii) social factors related to the person’s internalisation of the reference group’s subjective culture and specific interpersonal arrangements that the individual has made with others in particular circumstances [32]; (iii) image, which is the degree to which the use of a new technology is perceived to improve one’s image or status in one’s social system [30]. Social influence has been generally found to positively influence behavioural intention to use a new technology [21,38].

In the UTAUT model, facilitating conditions (FCs) directly influence use behaviour. These conditions represent the organisational and technical conditions and infrastructure that an individual believes would support the use of the system and make it easier for them to apply it [21]. This definition embraces three different constructs: (i) the perceived behavioural control (adapted from [26,31]), which reflects possible internal and external constraints on behaviour, related to resources and technology-enabling conditions; (ii) facilitating conditions, which refer to objective factors in the environment that individuals agree make an act easy to accomplish [32]; (iii) compatibility [30], which implies the degree to which a new technology is perceived as being coherent with the existing needs and experiences of potential adopters. Notably, facilitating conditions have been supposed to positively influence use behaviour [21,41,42,43].

Following [21], the four constructs can be influenced by four moderators: age, gender, experience with similar technology, and voluntariness of use [21,44,45]. In the final model, we also included a gross margin variable moderator that was hypothesised to influence the relationship between facilitating conditions and use behaviour (UB). All other conditions being equal, higher gross margins facilitate the introduction and use of a new technology.

We modified the original UTAUT model to consider factors related to farmers’ risk perception (FRP) associated with the adoption of micro-irrigation systems. We use a simplified version of the model proposed by [22], in which the perceived risk is theorised to affect the adoption of a new technology (the original model of [22] was related to the consumer acceptance of e-services). Various authors [22,23,46] have addressed the importance of risk as a predictor of technology acceptance. According to [22], risk perception can be linked to the UTAUT model in two ways: (i) on the one hand, the perceived ease of use of the new technology may significantly diminish the perceived risk associated with its adoption; (ii) on the other hand, the perception of new technology as risky may reduce its perceived performance expectancy and the likelihood of adoption. In this work, farmers’ risk perception was evaluated by two sub-facets: (i) overall and financial risk, which are intended to consider the level of farmers’ risk aversion; (ii) micro-irrigation implementation risk, which accounts for the perceived risk of adopting a micro-irrigation technology. In summary, through this model, we introduce the hypothesis that the higher the farmers’ risk aversion (i.e., the values of the three risk items), the lower will be their intention to invest in new micro-irrigation systems. Moreover, we hypothesised that farmers’ risk perception can be moderated by education levels, as these may reduce a farmer’s risk aversion [47]. In Figure 2, we present the final research model.

Overall, the present study analysed several hypotheses relating to the adoption of a new micro-irrigation system in the Bekaa Valley. Specifically, we tested the following hypotheses:

H1. 

Performance expectancy has a positive and significant impact on the behavioural intention to adopt a new micro-irrigation system.

H1a. 

Age moderates the relationship between performance expectancy and behavioural intention.

H2. 

Effort expectancy has a positive significant influence on the behavioural intention to introduce a new micro-irrigation system.

H2a. 

Age moderates the relationship between effort expectancy and behavioural intention.

H2b. 

Experience mediates the relationship between effort expectancy and behavioural intention.

H3. 

Social influence has a positive and significant relationship with the behavioural intention to adopt a new micro-irrigation system.

H3a. 

Age moderates the relationship between social influence and behavioural intention.

H3b. 

Experience mediates the relationship between social influence and behavioural intention.

H3c. 

Voluntariness of use mediates the relationship between social influence and behavioural intention.

H4. 

Facilitating conditions have a positive significant impact on the use behaviour of a new micro-irrigation system.

H4a. 

Age moderates the relationship between facilitating conditions and use behaviour.

H4b. 

Experience mediates the relationship between facilitating conditions and use behaviour.

H4c. 

Voluntariness of use mediates the relationship between facilitating conditions and use behaviour.

H4d. 

Gross unit margin mediates the relationship between facilitating conditions and use behaviour.

H5. 

Behavioural intention has a positive and significant impact on the use behaviour of a new micro-irrigation system.

H6. 

Farmers’ risk perception can negatively and significantly affect the behavioural intention to adopt a new micro-irrigation system.

H7. 

Farmers’ risk perception negatively and significantly affects their performance expectancy of micro-irrigation systems.

H7a. 

Educational level mediates the relationship between farmers’ risk perception and performance expectancy.

H8. 

Effort expectancy can negatively influence farmers’ risk perception.

H8a. 

Educational level mediates the relationship between effort expectancy and farmers’ risk perception.

3. Materials and Methods

3.1. Sampling and Data Collection

A quantitative study was conducted in the Bekaa Valley of Lebanon by targeting farmers located in its three main districts (North, Central, and West Bekaa). A random sample was selected for the study, comprising potato farmers using the ordinary sprinkler irrigation system in the three main districts of the Bekaa Valley. The total number of potato growers in the area is approximately 500 (identified while interviewing the president of the syndicate of potato growers in the Bekaa Valley), of which 35, 20 and 45% are located in North Bekaa, Central Bekaa and West Bekaa, respectively. The survey was conducted using a Google Forms questionnaire administered face to face by a team of three agricultural engineers who received three days of training in techniques and ethical features of questionnaires, taught by one of the authors (who was also part of the team). A pilot test that involved 40 farmers was conducted in December 2020 to evaluate the reliability and validity of the questionnaire’s items and scales. The main survey was conducted between January and March 2021.

The study’s purpose was initially explained to all participants via telephone calls and again before the surveys. The significance of confidentiality and the privacy of all participants was reasserted at the beginning of the interviews, which lasted between 12 and 15 min.

3.2. Survey Design

The survey was developed based on insights acquired from six previous focus groups conducted in the same area from March to April 2020. The quantitative questionnaire was divided into two sections. The first section included questions related to the farmers’ socio-economic characteristics and the moderating variables proposed in the conceptual framework (i.e., gender, age, educational level, household assets, farming practices, percentage of share of potato land, gross margin, type of distribution channel, potato production quantity, micro-irrigation experience and voluntariness of use). This survey section featured the use of nominal and ordinal scales.

The second section contained questions about the farmers’ risk perception and anxiety, and the major constructs included in the UTAUT model. Specifically, performance expectancy was measured using seven items concerning different perceptions of farmers regarding the micro-irrigation system, including its potential benefits linked to improved water management, increases in potato yield and quality, the possible reduction of energy costs, and improved plant disease management and control through greater pesticide and fertiliser efficiency. Effort expectancy was evaluated using five items related to micro-irrigation of potatoes: perceived ease of use, the volume of effort required, specialised workforce requirements (or lack thereof), time savings and the likelihood of participants’ skilful utilization. Social influence was measured using six items reflecting how the perceptions of people whose opinions are important to the farmers impacted the adoption of micro-irrigation systems, and the degree to which peers could affect the use of this new system. These items also considered the effect of farmers’ moral obligation norms on adopting a micro-irrigation system to preserve water resources for the sake of protecting the environment while continuing to grow potatoes by rain-fed agriculture. Furthermore, the facilitating conditions construct was assessed with four items related to the ability to gain access to required resources, the necessity to obtain advanced training and the support necessary for using the micro-irrigation system.

Since risk behaviour can affect the adoption of new technologies [48], farmers’ risk latent variable was measured by three types of risks (i.e., overall, financial, micro-irrigation implementation) to reveal the presence of risk-averse or risk-taking farmers.

Attitudes, intentions and preferences cannot be quantified directly [49]. However, they can be indirectly quantified through observed and measurable indicators using scaling approaches [50]. To this end, a five-point Likert-type scale ranging from strongly disagree (−2) to strongly agree (2) was used to measure the participants’ beliefs and opinions about the acceptance of a micro-irrigation system. Risk perception items (i.e., overall risk, financial risk and micro-irrigation risk) ranged from ‘extremely risky’ (−2) to ‘not at all risky’ (2).

3.3. Statistical and Econometric Analysis

Structural equation modelling (SEM) was employed to analyse the relationships among the aforementioned variables. The SEM technique has previously been applied to the UTAUT model by various authors [51,52,53,54]. SEM is a multivariate analysis that integrates factor and path analysis [55] and allows researchers to test and estimate a set of hypothesised relationships between numerous independent and dependent variables, each of which can be assessed by a set of indicators based on a theoretical model [50,56]. There are two types of SEM: (i) covariance-based SEM, which is generally used to test theories, and (ii) PLS-SEM, which is generally used to broaden theories in exploratory research [25]. Since PLS-SEM is geared towards theory development and prediction, this study referred to the latter type of SEM.

The specification of a PLS-SEM requires two steps. In the first step, a measurement model is specified that defines the latent variables in terms of the indicators that outline them. In this way, principal component analysis is usually performed by computing the factor loadings to assess the relative importance of the variance of the latent variable. Factor loadings greater than 0.70 show that an indicator loads significantly on a construct providing acceptable indicator reliability [57,58,59]. However, a factor loading greater than 0.50 is also a widely accepted threshold for significance [60]. Moreover, the variance inflation factor (VIF) is often used to evaluate the presence of collinearity among indicators. VIF values should be lower than or equal to 3 to exclude collinearity issues [59]. The internal consistency and reliability of different items as a group are usually evaluated according to Cronbach’s alpha, with values ranging from 0.70 to 0.95 being deemed acceptable [59,61,62,63].

Next, the convergent validity of each construct is measured using the average variance extracted (AVE). The AVE is a measure of the amount of variance captured by each construct in relation to the amount of variance due to measurement error. Values greater than 0.50 reveal that the construct explains at least 50% of the variance of its items [64,65]. AVE statistics are also useful in testing the discriminant validity of each construct. In this case, the Fornell and Larcker (1981) criterion [66] suggests that the square root of each construct’s AVE should be compared to the squared inter-construct correlation. A further analysis of the discriminant validity of each construct can be assessed using the heterotrait-monotrait ratio (HTMT) [67]. The HTMT reflects the mean of the item correlations throughout all constructs relative to the geometric mean of the average correlations for the items assessing the identical construct. Values lower that the threshold of 0.90 are usually accepted [25,59,67].

The second step of PLS-SEM analysis requires the estimation and testing of the structural model. The major purposes of this step are to analyse the links between the variables integrated in the model and to evaluate the hypothesised theoretical relationships [60]. The general assessment criteria are the coefficient of determination (R²) and the statistical significance and relevance of the path coefficients [68].

4. Results

4.1. Socio-Demographic and Economic Characteristics of Respondents

The survey was completed by a total of 220 randomly selected farmers. Overall, 20 questionnaires were eliminated due to being incomplete or containing compilation errors. Therefore, the analysis was conducted using a total sample size of 200 respondents. In the absence of official statistics, we interviewed the president of the syndicate of potato growers in the Bekaa Valley to obtain information on the total number of potato growers in this area. He reported that a total of approximately 500 potato growers were working in the area. Thus, the sample size of 200 farmers included in the analysis is statistically appropriate for representing the potato grower population located in the Bekaa Valley, permitting a ±5% margin of error.

In

Table 1. Summary statistics of the socio-economic variables.

Socio-Economic Variables	Response Scale	North Bekaa (N = 69)		Central Bekaa (N = 36)		West Bekaa (N = 95)		Overall Bekaa (N = 200)
		Mean	Std.Dev	Mean	Std.Dev	Mean	Std.Dev	Mean	Std.Dev
Age	0: Less than 45 years 1: 45–60 years 2: More than 60 years	0.99	0.10	1.22	0.14	1.14	0.06	1.14	0.06
Educational level	0: Not attended school 1: Primary 2: Secondary 3: University	1.93	0.11	1.94	0.14	1.82	0.10	1.88	0.06
Number of family members	Number	4.51	0.25	4.72	0.24	4.71	0.21	4.64	0.14
Number of household members on-farm work	Number	1.29	0.07	1.44	0.12	1.23	0.06	1.29	0.04
Number of household members in off-farm sector	Number	0.58	0.11	0.80	0.17	0.77	0.10	0.71	0.07
Farming experience	0: Less than 10 years 1: 10–30 years 2: More than 30 years	1.07	0.10	1.27	0.13	1.18	0.07	1.16	0.05
Other financial income	0: No 1: Yes	0.32	0.06	0.58	0.08	0.44	0.05	0.43	0.04
Total land area	Hectares	125	72.8	248.4	148.7	58.7	14.7	158	37.4
Potato cultivation area (ha)	Hectares	45.4	15.7	222	147.7	41.2	11.1	75.2	27.8
Share of potato land	Percent	80.30	3.48	78.00	5.42	86.61	5.00	82.88	2.83
Land management	0: Rented land 1: Private land 2: Both private and rented	1.03	0.10	1.11	0.14	0.93	0.08	0.99	0.06
Wholesale channel	0: No 1: Yes	0.45	0.06	0.47	0.08	0.45	0.05	0.46	0.04
Intermediaries/agents channel	0: No 1: Yes	0.68	0.05	0.86	0.06	0.72	0.05	0.73	0.03
Export channel	0: No 1: Yes	0.25	0.05	0.11	0.05	0.26	0.05	0.23	0.03
The overall number of channels per farmer	Numbers	1.39	0.07	1.44	0.09	1.42	0.06	1.42	0.04
Gross margin (%)	Percent	12.15	2.14	10.00	2.09	8.74	1.52	10.15	1.10
Social participation	0: No 1: Yes	0.30	0.06	0.58	0.08	0.29	0.05	0.35	0.03
Micro-irrigation experience	0: I don’t have experience 1: I have experience	0.07	0.03	0.19	0.07	0.12	0.03	0.12	0.02

, we present the demographic and socio-economic characteristics of the participants. Out of the 200 respondents, 69, 36 and 95 participants were from North, Central and West Bekaa, respectively. This composition broadly reflects the distribution of potato farmers in the Valley as indicated by the president of the potato growers’ syndicate. All participants were males, since no women were running farms in the area. Farmer age was measured following three categories: less than 45 years, 45–60 years and more than 60 years old. In the Bekaa region overall, the farmers were mostly aged from 45 to 60 years. They had an average family size of five people (s = 0.14), and on average only one member of the household worked on the farm (overall Bekaa: $\overline{x}$ = 1.29, s = 0.04), while some households had a family member involved in off-farm work (overall Bekaa: $\overline{x}$ = 0.71, s = 0.07). Moreover, approximately 44% (s = 0.04) of farmers declared that their income also depended on external activities. The educational level of the participants was evaluated based on the following scale: farmers who did not attend school; those who attended primary school; those who earned a secondary or university diploma. The results presented in Table 1 indicate that in the three studied regions, the majority of participants possessed a secondary education and nearly all farmers had between 10 and 30 years of farming experience.

Previous micro-irrigation experience was assessed by a dichotomic variable where 0 was associated with ‘don’t have any micro-irrigation experience’ and 1 was associated with ‘I have experience with micro-irrigation’. Most farmers declared not having any experience with micro-irrigation (overall $\overline{x}$ = 0.12, s = 0.02). Moreover, the majority of farmers stated that they had no social participation as a member of an agricultural organisation or association (overall $\overline{x}$ = 0.35, s = 0.03). Furthermore, the share of potato land against the total cultivated land showed an average value of 82.9% (s = 2.83) in the three main districts, thereby confirming that potato cultivation is the main crop produced in the area. The average farmer cultivated 75 hectares of potato land ($\overline{x}$ = 75.2, s = 27.8) with private management.

There were three main potato distribution channels: wholesale, agents and export channels. On average, each potato grower in the region used approximately two channels ($\overline{x}$ = 1.42, s = 0.04). The gross margin for potato sales over the last 3 years was similar in the three main districts, with a mean value of 10.15% (s = 1.10) for the Bekaa Valley. Finally, the results did not seem affected by geographical differences between the three regions since mean pairwise test statistics did not reveal a rejection of the null hypothesis of equal means (for brevity, we do not report these values; however, they are available upon request).

4.2. Results for UTAUT Behavioural Variables

4.2.1. Measurement Model

The measurement model was assessed for indicator reliability, internal consistency and convergent validity using confirmatory factor analysis.

Table 2. Summary statistics of the micro-irrigation related items and latent components.

Micro Irrigation (MI) Items and Latent Components	Loading	North Bekaa (N = 69)		Central Bekaa (N = 36)		West Bekaa (N = 95)		Overall Bekaa (N = 200)
		Mean	Std.Dev	Mean	Std.Dev	Mean	Std.Dev	Mean	Std.Dev
Performance Expectancy		0.52	0.13	0.35	0.22	0.43	0.11	0.45	0.08
I think MI would increase my yield	0.89	0.43	0.16	0.22	0.27	0.33	0.14	0.35	0.10
I think MI enhances the potato quality	0.88	0.33	0.17	0.11	0.26	0.30	0.14	0.28	0.10
I find MI would reduce energy costs	0.76	0.90	0.13	1.11	0.16	0.87	0.09	0.93	0.07
I find MI allows efficiency in fertilizers’ and pesticides’ use	0.85	0.65	0.16	0.39	0.25	0.40	0.13	0.49	0.09
I think MI reduces disease incidence	0.86	0.28	0.17	−0.08	0.27	0.25	0.14	0.20	0.10
Effort Expectancy		0.70	0.15	0.22	0.27	0.43	0.14	0.49	0.10
I find MI does not need a lot of effort	0.94	0.70	0.17	0.11	0.28	0.29	0.15	0.40	0.11
I think MI would save time in respect to my actual irrigation system	0.94	0.71	0.16	0.33	0.28	0.56	0.14	0.57	0.10
Social Influence		0.90	0.12	0.85	0.16	0.86	0.08	0.87	0.06
I feel a moral obligation to modify my current irrigation system in order to save water to face the impact of climate change	0.96	0.91	0.13	0.89	0.16	0.88	0.09	0.90	0.07
I feel a moral obligation to use MI in order not to be forced to move from growing potatoes to a rain-fed agriculture	0.97	0.88	0.12	0.81	0.17	0.83	0.09	0.85	0.06
Facilitating Conditions		1.38	0.08	1.13	0.16	1.19	0.08	1.25	0.05
I need subsidies to be able to implement the MI system	0.75	1.59	0.08	1.11	0.20	1.32	0.10	1.38	0.07
I need training to raise my awareness about the benefits of the MI and to technically know how use it in a proper way	0.89	1.17	0.10	1.14	0.17	1.07	0.09	1.12	0.06
Farmers’ Risk Perception		0.01	0.12	0.56	0.16	0.07	0.10	0.14	0.07
In general, how risky would I say are my behaviour and the decisions I take?	0.92	0.07	0.14	0.83	0.17	0.19	0.11	0.27	0.08
For the implementation of a micro-irrigation system in my farm, how risky would I say are my behaviour and the decisions I take?	0.95	0.16	0.12	0.78	0.16	0.21	0.10	0.30	0.07
With regards to finance, how risky would I say are my behaviour and the decisions I take?	0.91	−0.20	0.13	0.08	0.19	−0.18	0.11	−0.14	0.08
Behavioural Intention
I am very likely to adopt the MI system for potato cultivation in the next 12–24 months		−0.04	0.14	−0.22	0.24	0.24	0.15	0.06	0.10
Use Behaviour		17.28	1.70	14.08	2.33	15.09	1.34	15.67	0.96
Percentage of my land on which I will adopt MI	0.92	34.06	3.26	27.78	4.43	29.74	2.57	30.88	1.84
I really want to use micro-irrigation to improve my potato cultivation	0.94	0.51	0.18	0.39	0.29	0.45	0.15	0.46	0.11

summarises the descriptive statistics of the UTAUT model components and the differences between the constructs. All items were evaluated using a five-point scale ranging from −2 to 2. Differences were examined by performing pairwise t-tests (not reported here, for the sake of space) on the average summation scores for items obtained after having assessed reliability using Cronbach’s α for each construct.

The reliability of each indicator was tested by examining the loadings and the VIF. Factor loading values greater than or equal to the threshold level of 0.70 and VIF values lower than or equal to 3 were treated as significant, as recommended by [59].

The performance expectancy scores indicate that farmers agreed that the micro-irrigation system provides benefits. The results highlight that farmers had a high performance expectancy of micro-irrigation systems, especially concerning the yield increase and better quality of potato production. They also believed that a micro-irrigation system will help them reduce energy costs and potato disease incidence while improving the efficient use of pesticides and fertilisers. No differences were noted among the three zones when applying pairwise t-tests.

Concerning effort expectancy, participants perceived the implementation of a micro-irrigation system as a strategy to reduce efforts related to the actual time spent on irrigation management. Relating to the social influence construct, the findings showed that farmers perceived the new system as a way to avoid moving to rain-fed irrigation and as a moral obligation to preserve water. Concerning the facilitating conditions construct, potato farmers reported that it is important to receive effective training to raise their awareness about the use of micro-irrigation systems. Moreover, they believe that certain types of subsidies could help them to facilitate the introduction of micro-irrigation systems in their land. No significant differences were observed between the three main districts since pairwise tests did not reject the null hypothesis of equal means among the different areas of the Bekaa Valley.

The farmers’ risk perception construct showed that farmers in Bekaa Valley generally show risk-averse attitudes not only towards general decisions but also concerning the implementation of micro-irrigation systems in their fields. Risk-taking decisions only seemed to emerge from the analysis with respect to financial risk.

Regarding the behavioural intention to invest in micro-irrigation systems, participants in Central Bekaa were pessimistic ($\overline{x}$ = −0.22, s = 0.24) about the possible implementation of this technology. This result could be partly related to the economic and social instability of the country, initiated with the economic setback of October 2019 and aggravated by the COVID-19 pandemic and peaking with the devastating Beirut port explosion in August 2020. Concerning the use behaviour construct, farmers showed a neutral position regarding the desire to implement micro-irrigation. In the case of possible adoption, most stated that they would begin to implement it on an average of 30% of their land.

The construct reliability was assessed by evaluating the composite reliability (CR) and Cronbach’s alpha (CA). The results presented in

Table 3. Reliability and validity measures: Composite reliability (CR), Chronbach’s alpha (CA), and Average variance extracted (AVE) of latent variables.

	CR	CA	AVE	PE	EE	SI	FC	FRP	BI	UB	AGE	ExpMI	VoUS	UNMARG
PE	0.92	0.90	0.71	0.84
EE	0.94	0.87	0.88	0.00	0.94
SI	0.95	0.92	0.93	0.61	0.61	0.95
FC	0.81	0.54	0.68	0.00	0.47	0.00	0.82
FRP	0.95	0.92	0.86	−0.19	−0.03	0.00	0.08	0.93
BI	1.00	1.00	1.00	0.00	0.00	0.00	0.00	0.00	1.00
UB	0.92	0.84	0.86	0.71	0.74	0.61	0.42	−0.26	0.83	0.93
AGE	NA	NA	NA	0.00	0.00	0.00	0.00	0.00	0.00	0.00	1.00
ExpMI	NA	NA	NA	0.00	−0.06	0.00	0.00	0.00	−0.11	0.00	0.02	1.00
VoUS	NA	NA	NA	0.45	0.60	0.45	0.55	−0.07	0.68	0.72	−0.14	−0.20	1.00
UNMARG	NA	NA	NA	0.04	0.08	0.07	0.05	−0.02	0.04	0.00	−0.05	−0.03	0.00	1.00
Educ	NA	NA	NA	0.00	0.00	0.00	0.00	0.00	0.06	0.00	−0.62	0.00	0.00	0.00

Note: Diagonal elements are the square root of the Average Variance Extracted (AVE). PE: performance expectancy; EE: effort expectancy; SI: social influence; FC: facilitating conditions; FRP: farmers’ risk perception; BI: behavioural intention; UB: use behaviour; AGE: age; ExpMI: previous experience with micro-irrigation; VoUS: voluntariness of use; UNMARG: gross unit margin; Educ: educational level; CR: composite reliability; CA: Chronbach’s alpha; AVE: average variance extracted; NA: not applicable.

suggest that the CR and CA results for each variable are above the accepted threshold levels, which demonstrates the presence of internal consistency. The convergent validity was assessed using the AVE. Table 3 also shows that the AVE results for each construct are higher than the threshold of 0.5, which determines convergent validity.

To assess discriminant validity, the Fornell–Larcker criterion and the HTMT were each applied. Table 3 presents the Fornell–Larcker results. The square roots of the AVE for each construct (performance expectancy (0.84), effort expectancy (0.94), social influence (0.95), facilitating conditions (0.82), farmers’ risk perception (0.93)) were all greater than the correlations of these constructs with other latent variables. This reveals that all constructs are applicable measures of unique concepts.

The HTMT values are presented in

Table 4. The Heterotrait-Monotrait Ratio (HTMT) values.

	PE	EE	SI	FC	FRP	BI	UB	AGE	ExpMI	VoUS	UNMARG
EE	0.00
SI	0.67	0.69
FC	0.00	0.70	0.00
FRP	0.22	0.04	0.00	0.13
BI	0.00	0.00	0.00	0.00	0.00
UB	0.81	0.86	0.69	0.60	0.30	0.90
AGE	0.00	0.00	0.00	0.00	0.00	0.00	0.00
ExpMI	0.00	0.06	0.00	0.00	0.00	0.11	0.00	0.02
VoUS	0.46	0.65	0.47	0.74	0.07	0.68	0.78	0.14	0.20
UNMARG	0.05	0.09	0.07	0.07	0.02	0.04	0.00	0.05	0.03	0.00
Educ	0.00	0.00	0.00	0.00	0.00	0.06	0.00	0.62	0.00	0.00	0.00

Note: PE: performance expectancy; EE: effort expectancy; SI: social influence; FC: facilitating conditions; FRP: farmers’ risk perception; BI: behavioural intention; UB: use behaviour; AGE: age; ExpMI: previous experience with micro-irrigation; VoUS: voluntariness of use; UNMARG: gross unit margin; Educ: educational level.

. Notably, all HTMT values were equal to or lower than the threshold level of 0.90. Therefore, we can state that all the constructs exhibit evidence of discrimination.

4.2.2. Estimation Results—Structural Model

In Figure 3 and

Table 5. Structural model with path coefficients and R-squares for models with UTAUT and UTAUT and Perceived Risk, with direct (D) effects only, and with direct and interaction effects (D + I).

	UTAUT		UTAUT + Farmers’ Risk Perception
	D	D + I	D	D + I
Behavioural intention
R2 Adj.	0.55 **	0.65 **	0.55 **	0.65 **
Performance expectancy (PE)	0.32 **	0.31 **	0.31 **	0.29 **
Effort expectancy (EE)	0.42 **	0.21 *	0.45 **	0.24 **
Social influence (SI)	0.08	0.04	0.05	0.01
Farmers’ risk perception (FRP)			0.06	−0.08 *
Age		0.07		0.05
Experience in micro-irrigation (ExpMI)		−0.01		−0.04
Voluntariness of use (VoUS)		0.45 **		0.44 **
PE × Age		−0.08		−0.09
EE × Age		0.09		0.11 *
EE × ExpMI		0.05		0.05
SI × Age		0.03		0.02
SI × ExpMI		−0.07		−0.07
SI × VoUS		0.10 *		0.10 *
Use Behaviour
R2 Adj.	0.71 **	0.74 **	0.71 **	0.74 **
Facilitating conditions (FC)	0.13 **	0.14 **	0.13 **	0.14 **
Behavioural intention (BI)	0.79 **	0.80 **	0.79 **	0.80 **
Age		−0.06		−0.06
Experience in micro-irrigation (ExpMI)		0.11 *		0.11 **
Gross unit margin (UNMARG)		0.11 *		0.11 **
FC × Age		−0.03		−0.03
FC × ExpMI		0.02		0.02
FC × UNMARG		0.10 *		0.10 *
Risk Perception
Educational level (Educ)				−0.28 **
Effort expectancy (EE)				−0.01
EE × Educ				−0.16 *
Performance expectancy
Educational level (Educ)				0.13 *
Farmers’ risk perception (FRP)				−0.14 *
FRP × Educ				−0.16 **

Note: * p-values < 0.05, ** p-values < 0.01; all other path coefficients are not significant. PE: performance expectancy; EE: effort expectancy; SI: social influence; FC: facilitating conditions; FRP: farmers’ risk perception; BI: behavioural intention; UB: use behaviour; AGE: age; ExpMI: previous experience with micro-irrigation; VoUS: voluntariness of use; UNMARG: gross unit margin; Educ: educational level.

, we present the PLS-SEM estimation results. As shown in Table 5, several models were tested. The UTAUT and farmers’ risk perception (UTAUT + FRP), that is the research model with interaction effects (D + I) and without them (D), were tested to evaluate whether the moderator’s age, prior micro-irrigation experience, voluntariness of use, gross unit margin and educational level influenced the behavioural intention and use behaviour. The UTAUT without farmers’ risk perception was also tested with interaction effects (D + I) and without them (D).

In all of the tested models, the adjusted R² values for both behavioural intention and use behaviour were higher than 0.25, thereby excluding weak model power [25]. Upon comparing the estimated models, it was shown that the inclusion of moderators increased the adjusted R² for behavioural intention (0.55 vs. 0.65 for UTAUT and UTAUT + FRP, respectively) and for use behaviour (0.71 vs. 0.74 for UTAUT and UTAUT + FRP, respectively). Moreover, upon adding the farmers’ risk perception variable and moderating effects, the adjusted R² for use behaviour increased from 0.71 (UTAUT (D + I)) to 0.74 (UTAUT + FRP (D + I)). Thus, adding the farmers’ risk perception to the UTAUT model with its moderating effects helps to explain variance in the use behaviour construct, better than all other models. Therefore, we focused on the analysis of the main research model (UTAUT + FRP (D + I)). Figure 3 presents the path coefficients of the research model. The bootstrapping technique was used, involving 5000 iterations [24]. The results show that performance expectancy (β = 0.29, p = 0.00) was the most predictive factor of potato farmers’ behavioural intention to invest in micro-irrigation systems, followed by effort expectancy (β = 0.24, p = 0.01). These two latent variables positively and significantly impacted the behavioural intention variable. Interestingly, the results indicated that social influence had no significant impact on their behavioural intention to adopt micro-irrigation systems (β = 0.01, p = 0.45). Furthermore, the results revealed that the facilitating conditions had a significant effect on the potato farmers’ behaviour regarding the use of micro-irrigation systems on their lands (β = 0.14, p = 0.00). The farmers’ risk perception constructs had a significant and negative effect on performance expectancy (β = −0.14, p = 0.03) and behavioural intention among potato farmers (β = −0.08, p = 0.04), revealing the importance of risk aversion in influencing the adoption of micro-irrigation systems. On the contrary, effort expectancy had no significant effect on risk perception (β = −0.01; p = 0.47). As expected, the potato farmers’ behavioural intention had a strong and significant impact on the use of micro-irrigation (β = 0.80, p = 0.00).

Regarding the moderating effects presented in Table 5, only the direct effect of voluntariness of use on behavioural intention was statistically significant (β = 0.44; p = 0.00). The direct effect of prior experience in micro-irrigation (β = 0.11; p = 0.00) and the gross unit margin (β = 0.11; p = 0.00) were statistically significant in terms of use behaviour. Furthermore, educational level was negatively and significantly related to performance expectancy (β = −0.08, p = 0.04) and farmers’ risk perception (β = −0.14, p = 0.03). Regarding the product-indicator results, age moderated the effect of effort expectancy on behavioural intention (β = 0.11, p = 0.05), while voluntariness of use moderated the effect of social influence on behavioural intention (β = 0.10, p = 0.03) and gross unit margin moderated the effect of facilitating conditions on use behaviour (β = 0.11, p = 0.00). Moreover, educational level moderated the effect of effort expectancy on farmers’ risk perception (β = −0.16, p = 0.02) and the effect of farmers’ risk perception on performance expectancy (β = −0.16, p = 0.00). The other interaction effects were not significant (see Table 5).

In summary, the results of the PLS-SEM strongly support the use of the extended UTAUT + FRP model to predict potato farmers’ behavioural intention and use behaviour in micro-irrigation adoption. The research model was able to explain 65% of the variance in their behavioural intention and 74% of the variance in their use behaviour related to adopting micro-irrigation systems.

5. Discussion

Climate change and extreme weather events, such as prolonged droughts combined with repeated heatwaves, are projected to become more frequent in Lebanon and will affect areas of specialised agricultural production such as the Bekaa Valley [6]. Notably, a growing interest has emerged in the need to put water mitigation strategies in place [8,11]. Water demand in the Bekaa Valley is often greater than water supply, which is primarily obtained from groundwater sources that are being diminished [19] due to the semi-arid environment [6]. Water management in agriculture requires technological interventions such as micro-irrigation. Notably, the implementation of micro-irrigation may induce significant benefits in arid and semi-arid regions such as the Bekaa Valley, by saving water during crop production. Benefits have also been noted in terms of crop yield and quality, reduced energy consumption, decreased labour input and more efficient use of fertilisers and pesticides [69]. Thus, improving water use efficiency in potato cultivation through micro-irrigation is a strategic approach to addressing the Bekaa region’s water scarcity.

Table 6. The outcomes of the tested hypotheses.

Hypotheses	Relationship	Result
H1	Performance expectancy has a positive and significant impact on the behavioural intention to adopt a new micro-irrigation system.	Supported
H1a	Age moderates the relationship between performance expectancy and behavioural intention.	Not supported
H2	Effort expectancy has a positive significant influence on the behavioural intention to introduce a new micro-irrigation system.	Supported
H2a	Age moderates the relationship between effort expectancy and behavioural intention.	Supported
H2b	Experience mediates the relationship between effort expectancy and behavioural intention.	Not supported
H3	Social influence has a positive and significant relationship with the behavioural intention to adopt a new micro-irrigation system.	Not supported
H3a	Age moderates the relationship between social influence and behavioural intention.	Not supported
H3b	Experience mediates the relationship between social influence and behavioural intention.	Not supported
H3c	Voluntariness of use mediates the relationship between social influence and behavioural intention.	Supported
H4	Facilitating conditions have a positive significant impact on the use behaviour of a new micro-irrigation system.	Supported
H4a	Age moderates the relationship between facilitating conditions and use behaviour.	Not supported
H4b	Experience mediates the relationship between facilitating conditions and use behaviour.	Not supported
H4c	Voluntariness of use mediates the relationship between facilitating conditions and use behaviour.	Not supported
H4d	Gross unit margin mediates the relationship between facilitating conditions and use behaviour.	Supported
H5	Behavioural intention has a positive and significant impact on the use behaviour of a new micro-irrigation system.	Supported
H6	Farmers’ risk perception can negatively and significantly affect the BI to adopt a new micro-irrigation system.	Supported
H7	Farmers’ risk perception negatively and significantly affects their performance expectancy of micro-irrigation systems.	Supported
H7a	Educational level mediates the relationship between farmers’ risk perception and performance expectancy.	Supported
H8	Effort expectancy can negatively influence farmers’ risk perception.	Supported
H8a	Educational level mediates the relationship between effort expectancy and farmers’ risk perception.	Supported

presents the outcomes of the tested hypotheses. According to the results of the PLS-SEM model, performance expectancy, effort expectancy and farmers’ risk perception significantly impacted the potato farmers’ behavioural intention, while social influence did not show a significant relationship with their behavioural intentions. Notably, the facilitating conditions variable positively influenced the use behaviour.

Following the path coefficients’ rankings, performance expectancy plays a central role in affecting farmers’ behavioural intention to adopt micro-irrigation. Thus, the farmers were driven to accept the micro-irrigation system based on their confidence in its usefulness. This implies that farmers would accept micro-irrigation based on the expectation that using this system for potato production would help them gain benefits in potato farming, which is consistent with results obtained in previous studies [70,71]. Based on our analysis, it emerged that farmers were convinced that micro-irrigation reduces energy costs. However, they showed hesitancy regarding increased potato yield, the generation of better potato quality, the reduction of disease incidence and the efficient use of fertilisers and pesticides following the adoption of a micro-irrigation system. The prior focus group analysis showed that farmers were hesitant to adopt micro-irrigation due to the lack of information they had about the system’s usage. Having specialists run workshops and seminars to improve farmers’ knowledge of how micro-irrigation works and the advantages it brings is likely to increase levels of acceptance and adoption. Furthermore, agricultural extensions were found to reduce these technical gaps [72], resulting in the more widespread adoption of this new technology. Similarly, [73,74] reported the need to augment extension services to enhance technical knowledge, self-confidence and the ease of using micro-irrigation technology.

Effort expectancy is the second most influential construct affecting behavioural intention. Our results highlighted that potato farmers preferred to adopt systems that require less effort and time than ordinary sprinklers. Similar results were obtained in different areas by [21,75], who found that effort expectancy had a positive and significant relationship with behavioural intention. Thus, to enhance the use of micro-irrigation systems, the challenge facing local non-governmental organisations, agricultural associations and the Ministry of Agriculture is to jointly coordinate the development of initiatives for planning site training. Through such initiatives, pilot area studies and field training can be launched to create awareness and increase farmers’ knowledge on micro-irrigation benefits as well as efficient methods of implementation [76].

Additionally, our study revealed that social influence did not have a significant impact on behavioural intention, which is similar to the results of previous works [77,78]. This suggests that farmers’ moral obligation related to water conservation does not seem to influence their adoption of micro-irrigation.

In contrast, we observed a significant impact of facilitating conditions on farmers’ use behaviour. This result is not new, since [79,80] reported the same finding. This suggests that our respondents were concerned about their surrounding environments and that appropriate training and subsidies could influence their use of micro-irrigation. Factors such as a lack of training, awareness, resources and incentives might be preventing farmers from accepting and adopting micro-irrigation systems in the Bekaa region.

The UTAUT model was extended to include the farmers’ risk perception. The results showed that micro-irrigation, financial and overall risks are salient concerns related to risk perception. Similar to [23], our results reveal that risk perception is an important factor that negatively affects farmers’ performance expectancy and their intention to use micro-irrigation systems. This suggests that the higher the risk aversion, the lower the investment in micro-irrigation would be. As result, policymakers and farmers associations must guarantee that micro-irrigation systems are technically feasible and improve yield and quality in potato crop production.

Furthermore, the age moderator exhibited a significant impact on effort expectancy in terms of behavioural intention, which suggests that older farmers perceived micro-irrigation as easy to use.

Additionally, our study shows a significant and positive impact of previous experience on use behaviour, revealing that farmers who had prior experience with micro-irrigation were more likely than inexperienced farmers to accept and use these systems. Thus, previous experience affected the adoption of new investments. This is unsurprising since the results of [81] indicated that experts are more likely than non-experts to choose non-traditional investments.

Similarly, the education moderator showed that farmers with lower education levels would have less confidence in using the micro-irrigation system [82,83]. Thus, the national government, donors and local authorities’ interventions should target farmers in the Bekaa Valley based on their education level, especially among the younger generations. This strategy is expected to positively impact the introduction of new technologies (e.g., micro-irrigation systems) in the area.

6. Conclusions

The aim of this study was to analyze and investigate potato farmers’ behaviour in adopting a micro-irrigation system. To achieve this objective, we adapted the unified theory of acceptance and use of technology (UTAUT) model by [21] and extended it by using the simplified version of the model proposed by [22], in which perceived risk can impact the adoption of a new technology.

Our outcomes offer visions for the policymaking process, and they bring up insights that can develop this field of research. Firstly, farmers are willing to adopt a micro-irrigation system if they have extended knowledge about the system and technical assistance. Secondly, they are keen to invest if they can engender more gains by reducing time and effort in their farming activities. For a risk-averse farmer population, government or donor policy incentives can encourage the adoption of a micro-irrigation system. Finally, agricultural extensions, field training, pilot area studies and subsidies will be important in increasing farmers’ intention to adopt a micro-irrigation system.

7. Limitations

Although it presents useful data and recommendations, the current study is not without certain limitations. Legal restrictions and safety measures linked to the COVID-19 pandemic sometimes kept us away from certain physical spaces and prevented some face-to-face interviews. Notably, several farmers rejected face-to-face participation in the questionnaires due to the pandemic. Also, the sample only included males since no females operated farms in the study area. Thus, it would be useful to repeat the same analysis and extend the study to other countries and incorporate female participation.