The Adoption of Social Innovation in Rural Tourism in Morocco: Towards Sustainable and Equitable Tourism

Abstract

The development of sustainable tourism in rural areas brings new challenges in terms of innovation and social inclusion. In this context, this study examines the adoption of social innovation by managers of rural guesthouses in Morocco. The objective is to identify the factors that influence their intention to adopt such practices, which may strengthen local cohesion, enhance cultural resources, and meet visitors’ expectations regarding sustainability. The analysis draws on the conceptual framework of the diffusion of innovation, which allows for the exploration of perceptions related to the relative advantage, compatibility, ease of use, visibility, and trialability of innovative practices. The research is based on a questionnaire survey conducted with a sample of 174 managers. The data collected underwent confirmatory factor analysis to validate the theoretical dimensions of the model, and were then analyzed using an ordered Logit model to account for the ordinal nature of the dependent variable measuring the intention to adopt. The empirical results indicate that several perceived factors—namely the superiority of the innovation, its economic or symbolic benefits, its cultural compatibility, its simplicity of understanding and use, and the visibility of its effects—have a significant influence. Other dimensions, such as technical compatibility or risk perception, do not show a notable effect. The study also highlights the role of education level and gender in the propensity to adopt social innovation.

1. Introduction

Rural tourism plays an increasingly important role in territorial development strategies in Morocco, where it contributes to the enhancement of heritage, the preservation of local know-how, and the creation of income-generating activities. Rural guesthouses, as local stakeholders, play a central role in this dynamic by offering a form of hospitality rooted in cultural authenticity. Rural tourism in Morocco is undergoing a strong revival, becoming a key lever for sustainable and territorial development.

This momentum is driven by the 2023–2026 tourism roadmap, targeting 17.5 million tourists and 120 billion dirhams in revenue. A 188-million-dirham program aims to transform 16 rural villages into tourism hubs based on local culture and community experiences.

However, in order to address contemporary challenges related to sustainability, social equity, and diversification of the tourism offer, these establishments are called upon to integrate forms of innovation that go beyond purely economic considerations. Social innovation—understood as the introduction of new practices aimed at improving collective well-being—emerges as a relevant lever. It strengthens community cohesion, supports the empowerment of local actors, and aligns tourism activity with the actual needs of rural populations. Yet, the spread of such innovation remains uneven, and its adoption varies significantly according to context and the perceptions of guesthouse managers.

In this context, it becomes essential to explore the factors that facilitate or hinder the adoption of social innovation in rural guesthouses. Beyond material or institutional conditions, it is often the representations held by actors—such as perceived compatibility with their values, ease of use, or the visibility of its outcomes—that influence their willingness to engage with such practices. The intention to adopt may also depend on individual characteristics like educational attainment, professional experience, or gender. Understanding these dynamics helps to better identify the drivers of innovation in rural environments and to tailor support initiatives to the diversity of individual paths and profiles. By shedding light on these dimensions, this study aims to contribute to a broader reflection on the conditions for inclusive rural tourism that combines local development, social sustainability, and the valorization of endogenous resources.

2. Literature Review

2.1. Diffusion of Innovation Theory: Application to Social Innovation

The gradual adoption of social innovation in rural guesthouses across Morocco can be understood through the framework of the diffusion of innovation theory (Rogers, 1962), which identifies several key factors influencing the speed and scale of adoption. The first crucial factor is relative advantage: when local actors perceive that the innovation—such as sustainable tourism practices, improved resource management, or offers that enhance cultural heritage—brings greater benefits than existing solutions, whether economic (such as increased revenue or attractiveness) or symbolic (such as social prestige or alignment with sustainability values), they are more inclined to adopt it. Compatibility is equally decisive: social innovations that respect Morocco’s local cultural values, rural traditions, and community ways of life are more easily embraced.

Adoption is further facilitated when these practices require no major break from existing skills or infrastructure, making integration smoother and less disruptive for rural stakeholders. On the other hand, if an innovation is seen as too complex—either in its conceptual understanding or its daily use—it can act as a barrier to adoption. Therefore, perceived simplicity and learning support are essential. Testability, meaning the possibility of experimenting on a small scale (for instance, in a single guesthouse), allows actors to reduce perceived risks and gain confidence through direct experience. Observability also plays a reinforcing role: when the positive impacts of the innovation—such as increased tourist flows, improved local living conditions, or visible economic gains—are clearly seen and easily shared within the community, the innovation gains social appeal and spreads more widely.

2.2. Perceived Added Value of Innovation

(a)

Relative advantage

Neumeier (2012) explains that in rural areas, a social innovation is adopted when it meets local needs more effectively than existing solutions, by mobilizing diverse networks of actors and fostering collective learning. He emphasizes the necessity of compatibility with local values to overcome resistance and support diffusion. This perspective is echoed by Pue et al. (2015), who highlight the importance of both agency-driven and structural drivers in innovation processes, while acknowledging that the perceived superiority of new solutions—even when difficult to quantify—remains essential for addressing social needs. In a similar vein, Mihalic (2016) introduces the concept of “responsurable” tourism, which combines responsibility, sustainability, and concrete action. He shows that the synergy between collective awareness, local initiatives, and public policies strengthens support for innovative practices. Antonelli et al. (2017) argue that social innovation, by drawing on identity resources such as landscapes and biodiversity, supports rural territorial development and thereby enhances the acceptability of locally rooted tourism initiatives.

Thorburn (2005) states that in rural tourism, identifying distinctive features perceived as innovative is crucial to stimulating the intention to adopt. This perceived superiority over existing offers, linked to the expectations of visitors or local stakeholders, creates a competitive advantage. He also highlights the importance of customer feedback and horizontal cooperation in consolidating adoption. Tajeddini et al. (2017) point to the role of female entrepreneurs in rural areas, showing how their creativity introduces novel forms of organization that respond to community needs. These initiatives, focused on sustainability and the preservation of cultural resources, contribute to tourism practices adapted to local specificities. Vercher et al. (2023) remind us that social innovations can foster social change, particularly by enhancing community cohesion and identity revitalization in rural zones. Howaldt & Schwarz (2010) add that the adoption of a social innovation depends on its integration into collective practices, noting that both economic (jobs, diversification) and symbolic (identity valorization, recognition) returns foster its acceptance in rural contexts.

Partanen and Sarkki (2021) show that social innovations—such as food waste recovery or professional inclusion—generate social and environmental value that can support rural tourism by creating synergies with local dynamics. These initiatives position tourism not only as an economic driver but also as a tool for territorial cohesion. Blackstock (2005) warns that tourism approaches driven solely by instrumental logic can create power imbalances and marginalize local populations. This negative perception of social and economic impacts may hinder innovation uptake when it is seen as imposed. Neumeier (2017) reiterates that social innovation relies on solutions perceived as more effective than current practices, and that it can spread beyond initiator groups if it genuinely meets local expectations. Aquino et al. (2018) demonstrate that social entrepreneurship in tourism, by leveraging external resources and ensuring equitable stakeholder participation, supports adoption. When the benefits are visible—such as increased income or cultural revitalization—community involvement in rural areas is strengthened. Based on these elements, the following hypothesis can be formulated:

H1. 

Relative advantage has a positive impact on the adoption of social innovation in rural tourism.

H1a. 

A high perception of the superiority of the innovation over existing solutions has a positive impact on the intention to adopt social innovation in rural tourism.

H1b. 

A strong perception of the economic or symbolic benefits of the innovation has a positive impact on the intention to adopt social innovation in rural tourism.

(b)

Compatibility

Roman et al. (2020) emphasize the importance of grounding innovation in local traditions to ensure its acceptance among rural populations. This approach is echoed by Ćurčić et al. (2021), who show that integrating traditional knowledge into modern tourism offerings contributes to both the economic and cultural sustainability of these initiatives. By linking ancestral practices to the tourist experience, they highlight the value of constructing authentic offerings that align with local values. Nogués-Pedregal et al. (2017), for their part, stress the importance of storytelling based on local history to prevent destination identities from being distorted by marketing imperatives. By promoting an endogenous perspective of tourism development, they argue that perceived authenticity serves as a legitimizing factor for innovation. Along the same lines, Vujko et al. (2017) demonstrate that traditional gastronomy can be a lever for innovation by strengthening community participation and visitor engagement.

Altinay et al. (2016) highlight the importance of adjusting innovations to local social dynamics, advocating for progressive acculturation through stakeholder dialogue. They show that this adaptation fosters ownership of the projects by local actors and secures their legitimacy. This idea is extended by Matteucci et al. (2022) who underline the role of cultural and territorial specificities in transforming tourism into a form of endogenous development. They insist that innovation cannot be separated from collective aspirations, especially in rural settings where social ties influence commitment. Tresiana and Duadji (2022) emphasize the connection between traditional knowledge and tourism governance systems, showing how this linkage strengthens community resilience and the adoption of sustainable practices. Vrana (2023) underscores the need to align social innovations with local socio-economic structures to ensure feasibility.

Gajdošík et al. (2017) stress the need to integrate innovations, particularly digital ones, in a way that is coherent with existing rural infrastructures. According to them, such compatibility preserves heritage while enhancing territorial competitiveness. Astawa et al. (2018) extend this logic by emphasizing the importance of harmonizing innovations with local cultural norms to ensure project sustainability. Their study shows that adoption is easier when initiatives are rooted in shared practices and beliefs, thus ensuring socio-economic continuity. Raimo et al. (2022) reveal that technological solutions introduced in tourism—especially in response to the pandemic—were accepted because they addressed real needs and were adapted to practical constraints, particularly in rural areas. Bell and Jayne (2010) add to this perspective by showing that the arrival of external actors in rural areas can create synergies between new practices and existing community systems, generating a hybrid environment that is conducive to innovation. These studies suggest that integration into existing uses and infrastructures strongly conditions the local appropriation of social innovation. This leads to the following hypothesis:

H2. 

Compatibility with local specificities has a positive impact on the adoption of social innovation in rural tourism.

H2a. 

A strong coherence between the innovation and local cultural values has a positive impact on the intention to adopt social innovation in rural tourism.

H2b. 

A high level of compatibility between the innovation and existing uses or infrastructures has a positive impact on the intention to adopt social innovation in rural tourism.

2.3. Ease of Use and Reliability, and Scope for Experimentation

(a)

Complexity

Van der Have and Rubalcaba (2016) show that social innovations, by reshaping technological and organizational dynamics, also alter social practices in tourism territories. When the focus is placed on collective value rather than individual gain, the commitment of rural communities increases—provided the initiatives are understandable and aligned with local expectations. Najda-Janoszka and Kopera (2014) emphasize that clearly identifying obstacles and establishing explicit methodological frameworks fosters adoption within rural tourism organizations. Petrou and Daskalopoulou (2013) add that transparency in innovation processes plays a central role in acceptance by local actors, by activating the social capital needed for change. Alkier et al. (2017) support this view, highlighting that collaborative models—incorporating cross-disciplinary knowledge and local resources—make social innovation more intelligible and engaging in rural settings, enhancing the capacity of tourism territories to appropriate such approaches.

Dargan and Shucksmith (2008) stress the structural role of social networks and collaborative practices in generating innovation projects in rural areas. A culture of collective learning, which enables clearer dissemination of ideas, supports better integration of innovations into local tourism practices. Secco et al. (2018) point out that the success of innovative tourism projects depends on a social reconfiguration that local actors must be able to understand and support. Parag and Janda (2014) highlight the role of intermediary actors in clarifying the benefits of innovations for users, especially in technical fields such as energy efficiency. Their mediation facilitates the practical use of social innovations in rural hospitality infrastructures. Wang et al. (2023) propose a visual and interactive approach to encourage stakeholder participation in tourism villages, making planning tools more accessible and thus easier to adopt in rural environments.

Bas and Guillo (2015) stress the need to design social innovations based on users’ real-life experiences. Their model, which emphasizes accessibility, intuitiveness, and alignment between expectations and lived reality, helps identify barriers linked to excessive complexity. Applied to rural areas, this approach would align innovative tourism projects with the cognitive and operational capacities of hosts, easing their day-to-day use. Carayannis et al. (2019) underline that economic models with a social purpose—driven by collaboration between universities, governments, businesses, and citizens—can simplify the understanding of innovations. In rural areas, such collective governance helps better communicate the purpose of tourism projects. Damanpour and Schneider (2006) remind us that the attitude of local managers—particularly their openness and ability to view innovations as easy to implement—directly influences their willingness to adopt them in fragile socio-economic contexts. Winand et al. (2016) confirm that when an innovation is perceived as feasible and compatible with available resources, its implementation is strengthened, especially when local actors face pressure to modernize their tourism offerings. Based on this, the following hypothesis can be proposed:

H3. 

Complexity has a negative impact on the adoption of social innovation in rural tourism.

H3a. 

Ease of understanding of the innovation has a positive impact on the intention to adopt social innovation in rural tourism.

H3b. 

Ease of use of the innovation has a positive impact on the intention to adopt social innovation in rural tourism.

(b)

Testability

Richter et al. (2019) describe how social enterprises, by supporting pilot projects and gradual funding, make it possible to experiment with solutions on a small scale—an approach particularly effective in rural areas. This method fosters community support by integrating their feedback from the earliest stages. This logic of experimentation is also advocated by Peng et al. (2012), who emphasize the value of collaborative design methods in involving stakeholders from the testing phase. They demonstrate that iterative adjustment of solutions to expressed needs leads to better adoption grounded in local realities. Malek and Costa (2015) adds that community participation in tourism planning through inclusive mechanisms facilitates ownership of social innovations. This dynamic proves more effective when proposals are tested in their real-life settings. Mosedale and Voll (2017) expands on this by stressing that bottom-up approaches, rooted in local practices, help develop innovation models perceived as legitimate. By adopting small-scale experimental formats, rural actors gain confidence and adapt innovations to their specific circumstances.

Banerjee and Shaban (2019) emphasize the importance of micro-initiatives as drivers of transformation in rural territories, noting that their effectiveness, demonstrated through experimentation, encourages communities to embrace innovative models. This view is supported by (Christmann, 2020) who argues that progressive experimentation helps identify and correct the limitations of prototypes while reinforcing the legitimacy of the solutions among users. Such an iterative process facilitates adoption, as it is based on practical, co-constructed adjustments with beneficiaries. Bock (2016) cautions that innovation-related risks must be actively managed, or they may generate roadblocks. In rural tourism, risk governance becomes a key lever to avoid the abandonment of promising solutions. Tha (2017) reinforces this by explaining that fear of failure often undermines entrepreneurial efforts, particularly when the social or symbolic stakes of risk are heightened. A well-structured testing phase can therefore help remove these barriers by making the process feel more secure.

Christmann (2020) also notes that while social innovation introduces new solutions, it can provoke strong resistance in rural areas due to deeply rooted habits. The prospect of change is often associated with perceived loss, which hinders collective buy-in. In this context, the presence of actors who can provide reassurance by explaining the incremental benefits of experimentation plays a key role. Brogaard (2019) highlights additional obstacles, such as skill asymmetries and institutional misalignments in partnerships, which may exacerbate perceptions of risk. For rural tourism actors, these tensions weaken the co-construction of innovative projects. Bock (2016) complements this view by pointing out that some innovations create uneven effects across different groups, generating feelings of disadvantage that discourage engagement. Testability, by anticipating and addressing such gaps, becomes a tool to reassure and mobilize stakeholders. Brown and Osborne (2013) finally emphasizes that risks which are poorly assessed or inadequately discussed promote inertia, especially in sectors entrenched in professional routines. The opportunity to test innovation in real-world conditions, under a structured framework, thus emerges as a key enabler of transition toward new practices in rural ecosystems. This supports the following hypothesis:

H4. 

Testability has a positive impact on the adoption of social innovation in rural tourism.

H4a. 

The possibility to experiment with innovation on a small scale has a positive impact on the intention to adopt social innovation in rural tourism.

H4b. 

A low perception of risk during the testing phase has a positive impact on the intention to adopt social innovation in rural tourism.

(c)

Observability

Edwards-Schachter and Wallace (2017) emphasize that social transformations generated by innovation, when they produce tangible outcomes in terms of equity or sustainability, are likely to be replicated in other contexts. This demonstrative dimension is essential in encouraging other rural communities to adopt new practices. Neumeier (2017) shows that social innovations in rural areas gain credibility when they address visible local needs and are supported by strong collaborative networks. Concrete results, such as the inclusion of marginalized groups or the improvement of living conditions, provide reference points for similar territories. Urban and Kujinga (2017) highlight that observable success in specific environments enhances the attractiveness of hybrid models. Local grounding thus becomes a lever for replication in other rural regions. Similarly, (Blanco et al., 2009) demonstrate that environmental initiatives in tourism, when they prove their economic and organizational effectiveness, motivate other actors to engage in comparable practices. In this sense, the visibility of positive outcomes plays a unifying role in structuring the adoption of social innovation within rural tourism.

Vercher et al. (2023) remind us that successful experiences in citizen participation or heritage valorization in rural areas can generate social transformations that inspire further local initiatives. These dynamics become benchmarks for other project leaders seeking replicable models, particularly in fields like agritourism. Aquino et al. (2022) show that socially innovative tourism can lead to lasting changes at personal, structural, and ecological levels. When these effects are visible, they foster a dynamic of engagement and stronger collective ownership. For Schor and Fitzmaurice (2015), it is essential to maintain active communication on social benefits to prevent economic logic from becoming dominant. This ongoing narration of collective gains encourages long-term adoption. Wittmayer et al. (2019) stress the importance of participatory narratives, which, by integrating diverse local actor perspectives, increase the legitimacy and social reach of innovations. In rural tourism, such shared storytelling ensures broader and more locally grounded diffusion.

Pikkemaat et al. (2018) explain that structured interactions among stakeholders, supported by collaborative networks, extend the reach of social innovations by facilitating their diffusion into similar environments. Rural tourism, often organized around interconnected local actors, can benefit greatly from these reinforced cooperation dynamics. Howaldt and Schwarz (2010) argue that for a social innovation to become sustainably embedded, it must be widely disseminated and accepted as a collective practice. This requires promoters to have adaptive communication strategies capable of engaging varied audiences. Grabs et al. (2016) emphasize the exemplary role of local pioneers, whose successful practices generate trust and foster imitation. In rural areas, these inspiring figures are often the most effective agents in broadening the reach of innovation. Healey (2006) reinforces this by highlighting the importance of inclusive, multi-format communication that resonates with diverse audiences. In rural tourism, such communicative diversity helps convey messages clearly and secure community engagement with social innovation. This leads to the formulation of the following hypothesis:

H5. 

Observability has a positive impact on the adoption of social innovation in rural tourism.

H5a. 

A high visibility of the positive results of innovation has a positive impact on the intention to adopt social innovation in rural tourism.

H5b. 

A strong capacity for social diffusion and communication around the innovation has a positive impact on the intention to adopt social innovation in rural tourism.

3. Methodology

3.1. The Model

To explain the determinants of the adoption of social innovation in rural guesthouses in Morocco, an econometric model was developed. This model is based on a series of key variables derived from the theory of innovation diffusion and specifically adapted to the rural context under study. It thus makes it possible to identify the main factors influencing the intention to adopt these innovative practices among guesthouse managers. The model is presented as follows:

ADOPT_i = β₀ + β₁RELVA_i + β₂RELSY_i + β₃COMPV_i + β₄COMPU_i + β₅SIMPL_i + β₆USEEZ_i + β₇TRIAL_i + β₈RISKT_i + β₉VISIB_i + β₁₀SOCOM_i + β₁₁AGEHH_i + β₁₂EDUHH_i + β₁₃GENDR_i + ε_i

The dependent variable ADOPT captures the manager’s intention to adopt social innovation in the management of rural guesthouses. It is measured on a five-point Likert scale ranging from 1 (“very low intention”) to 5 (“very high intention”). The explanatory variables are derived from the key attributes of the Diffusion of Innovation theory (Rogers, 1962) and are operationalized through perception-based items measured on five-point Likert scales. A confirmatory factor analysis (CFA) was conducted to retain only the items with a convergent validity above 0.50, and the average of the retained items was used to construct a composite score for each dimension. RELVA (relative advantage) measures the perceived superiority of the innovation compared with existing practices through four items: (RELVA1) the innovation improves the quality of services offered by the guesthouse; (RELVA2) the innovation responds better to visitors’ expectations; (RELVA4) the innovation increases the attractiveness of the guesthouse; and (RELVA6) the innovation provides clear advantages compared with traditional practices.

RELSY (economic and symbolic benefits) captures the perceived financial and reputational benefits associated with the innovation through three items: (RELSY2) the innovation can increase the profitability of the guesthouse; (RELSY5) the innovation enhances the reputation and prestige of the establishment; and (RELSY6) the innovation strengthens the positioning of the guesthouse within sustainable tourism. COMPV (compatibility with local values) measures the alignment of the innovation with the cultural and social context through three items: (COMPV1) the innovation is consistent with local traditions; (COMPV4) the innovation respects the cultural identity of the territory; and (COMPV6) the innovation corresponds to the expectations of local communities. COMPU (technical compatibility) evaluates the degree to which the innovation can be integrated into existing infrastructures and working practices through three items: (COMPU1) the innovation is compatible with the current equipment of the guesthouse; (COMPU4) the innovation can be integrated into existing routines; and (COMPU6) the innovation can be implemented without major structural changes.

SIMPL (ease of understanding) reflects the cognitive accessibility of the innovation through three items: (SIMPL3) the innovation is easy to understand; (SIMPL4) the functioning of the innovation is clear; and (SIMPL6) the innovation does not require complex technical knowledge. USEEZ (ease of use) captures the practical simplicity of implementation through three items: (USEEZ2) the innovation is easy to implement in daily operations; (USEEZ3) the innovation can be used without significant effort; and (USEEZ5) the innovation does not require specialized external support. TRIAL (trialability) measures the possibility of experimenting with the innovation on a small scale through three items: (TRIAL4) the innovation can be tested before full adoption; (TRIAL5) the innovation can be implemented gradually; and (TRIAL6) the innovation can be experimented with in limited conditions. RISKT (perceived risk) evaluates the perceived level of uncertainty associated with experimentation through four items: (RISKT2) the innovation involves financial risk; (RISKT4) the innovation may create operational difficulties; (RISKT5) the innovation may generate uncertainty about outcomes; and (RISKT6) the innovation may involve reputational risk. VISIB (observability) captures the visibility of the innovation’s outcomes through three items: (VISIB2) the positive results of the innovation are clearly observable; (VISIB4) the benefits of the innovation appear rapidly; and (VISIB6) the innovation produces concrete and visible improvements. Finally, SOCOM (social communication) measures the capacity of the innovation to circulate socially within the tourism ecosystem through three items: (SOCOM1) the innovation can be easily communicated to other guesthouse managers; (SOCOM3) the innovation can be shared within local tourism networks; and (SOCOM5) the innovation can inspire imitation by other actors in the community. Three control variables are also included in the model: AGEHH (age of the manager), EDUHH (education level expressed in years of schooling), and GENDR (gender, coded 0 for female and 1 for male).

3.2. Variable Construction

CFA was conducted to validate the dimensions of the model. Only items with a convergent validity above 0.50 were retained. The results show that all dimensions meet the required methodological thresholds: Cronbach’s alpha ≥ 0.70, composite reliability ≥ 0.70, and AVE ≥ 0.50. As presented in

Table 1. CFA: Scale Validation and Reliability.

Items	Convergent Validity	Alpha de Cronbach	Composite Reliability	AVE
RELVA 1	0.624232	0.774649	0.723569	0.570269
RELVA 2	0.735756
RELVA 4	0.540458
RELVA 6	0.754985
RELSY 2	0.890609	0.966044	0.889444	0.628919
RELSY 5	0.836023
RELSY 6	0.575522
COMPV 1	0.872688	0.815664	0.803635	0.603448
COMPV 4	0.772554
COMPV 6	0.534078
COMPU 1	0.611894	0.955536	0.849973	0.627745
COMPU 4	0.835721
COMPU 6	0.837384
SIMPL3	0.618680	0.985115	0.930655	0.659765
SIMPL4	0.882876
SIMPL6	0.816548
USEEZ 2	0.671367	0.764319	0.174383	0.646772
USEEZ 3	0.751045
USEEZ 5	0.888389
TRIAL 4	0.538517	0.791017	0.742858	0.573550
TRIAL 5	0.660638
TRIAL 6	0.810252
RISKT 2	0.961374	0.816290	0.830071	0.617917
RISKT 4	0.738281
RISKT 5	0.642308
RISKT 6	0.590788
VISIB 2	0.658115	0.773572	0.713738	0.520703
VISIB 4	0.695821
VISIB 6	0.614630
SOCOM 1	0.632455	0.791284	0.763219	0.598009
SOCOM 3	0.596955
SOCOM 5	0.792080

Source: authors’ elaboration.

the RELVA dimension includes four valid items, with a Cronbach’s alpha of 0.774, a composite reliability of 0.724, and an AVE of 0.570. RELSY shows three loadings ranging from 0.575 to 0.891, with a Cronbach’s alpha of 0.966, a composite reliability of 0.889, and an AVE of 0.629. COMPV is constructed from three items (ranging from 0.534 to 0.873), with a Cronbach’s alpha of 0.816, a composite reliability of 0.804, and an AVE of 0.603. COMPU reaches loadings between 0.612 and 0.837, a Cronbach’s alpha of 0.956, a composite reliability of 0.850, and an AVE of 0.628.

SIMPL shows loadings ranging from 0.619 to 0.883, with a Cronbach’s alpha of 0.985, a composite reliability of 0.931, and an AVE of 0.660. USEEZ includes three valid items, with a Cronbach’s alpha of 0.764 and an AVE of 0.647. TRIAL is based on three loadings (ranging from 0.539 to 0.810), with a Cronbach’s alpha of 0.791, a composite reliability of 0.743, and an AVE of 0.574. RISKT presents four loadings (from 0.591 to 0.961), a Cronbach’s alpha of 0.816, a composite reliability of 0.830, and an AVE of 0.618. VISIB consists of three items (with loadings between 0.615 and 0.696), a Cronbach’s alpha of 0.774, a composite reliability of 0.714, and an AVE of 0.521. Finally, SOCOM is validated with three items (ranging from 0.597 to 0.792), a Cronbach’s alpha of 0.791, a composite reliability of 0.763, and an AVE of 0.598.

After confirming the convergent validity and reliability of the scales, a discriminant validity analysis was conducted to ensure that each dimension measures a distinct construct. This step is essential to confirm that the dimensions do not overlap and that they capture specific conceptual realities. The Fornell and Larcker method was applied: the square root of the average variance extracted (AVE) for each dimension must be greater than the correlations between that dimension and all others. The correlation matrix analysis confirms that this condition is met for all dimensions (see

Table 2. CFA: Discriminant Validity.

	1	2	3	4	5	6	7	8	9	10
RELVA (1)	0.75516
RELSY (2)	0.08614	0.79304
COMPV (3)	0.08055	0.09235	0.77682
COMPU (4)	0.09347	0.15909	0.14127	0.79230
SIMPL (5)	0.10286	0.00478	0.20658	0.03887	0.81226
USEEZ (6)	0.07261	0.16490	0.02975	0.17198	0.24119	0.80422
TRIAL (7)	0.13455	0.19572	0.02661	0.06231	0.04182	0.05743	0.75733
RISKT (8)	0.08034	0.10790	0.03579	0.04102	0.06319	0.07258	0.15558	0.78608
VISIB (9)	0.06772	0.19433	0.02779	0.03558	0.16537	0.08682	0.24975	0.12297	0.72160
SOCOM (10)	0.06935	0.05384	0.04057	0.23238	0.08074	0.21511	0.05498	0.01381	0.05050	0.77331

Source: authors’ elaboration.

).

For RELVA, the square root of the AVE is 0.755, exceeding all inter-dimensional correlations, the highest being 0.135 with TRIAL. RELSY reaches a value of 0.793, with a maximum correlation of 0.195. COMPV shows a value of 0.777, well above its correlations. COMPU has a square root of AVE of 0.792, compared to a maximum correlation of 0.232 with SOCOM. For SIMPL, the value is 0.812, higher than its strongest correlation of 0.207. USEEZ reaches 0.804, versus 0.242 with SIMPL. TRIAL (0.757), RISKT (0.786), VISIB (0.722), and SOCOM (0.773) also meet this requirement, with their square roots of AVE remaining higher than all inter-dimensional correlations. These results confirm the discriminant validity of all the dimensions analyzed.

3.3. Sample Description

The empirical sample used in this study consists of 174 managers of rural guesthouses operating in Morocco, who represent the relevant decision-makers regarding the potential adoption of social innovation practices within rural tourism activities. These establishments constitute a central component of rural tourism development, as they combine accommodation services with the valorization of local cultural heritage, community interaction, and environmentally oriented tourism practices. Because an exhaustive and updated national registry of rural guesthouses is not available and because a considerable share of these establishments operate partially outside formal institutional frameworks, the identification of respondents relied on a snowball sampling procedure. Initial participants were contacted through professional networks within the tourism sector, local associations, and informal contacts with rural tourism operators. These first respondents were subsequently invited to identify other managers operating in similar contexts. This progressive referral process made it possible to reach operators located in geographically dispersed rural territories, including areas that are relatively remote or weakly integrated into formal tourism networks.

Data collection was conducted through a structured questionnaire administered using a mixed-mode approach, combining face-to-face interviews and online distribution. The face-to-face mode was particularly useful for reaching operators located in remote rural areas where digital communication remains limited, while the online mode allowed broader territorial coverage and facilitated participation for managers who could respond remotely. In total, 311 questionnaires were distributed during the data collection process. Among these, 192 questionnaires were returned, corresponding to a response rate of 61.74%. After a detailed verification of the responses to ensure completeness and internal consistency, 174 questionnaires were retained as valid observations, whereas 18 questionnaires were excluded because they contained missing responses or inconsistencies in key sections of the survey. Consequently, the valid response rate relative to the number of questionnaires distributed amounts to 55.94%, which is considered satisfactory given the methodological difficulties typically associated with surveying small-scale tourism operators in rural areas, particularly when a portion of the population operates outside fully formalized institutional channels as shown in

Table 3. Response rate to questionnaire.

Distributed	Unreturned	Returned	Valid	Invalid	Response Rate Valid/Distributed
311	119	192	174	18	55.94%

Source: authors’ elaboration.

.

The resulting sample presents a diversity of managerial profiles and territorial situations, which strengthens the analytical relevance of the dataset. The respondents differ in terms of age, educational attainment, and professional trajectories, reflecting heterogeneous levels of managerial experience and exposure to innovation-related practices. Some managers entered the sector relatively recently, often through small-scale entrepreneurial initiatives linked to local development or family activities, whereas others have accumulated several years of experience in operating rural guesthouses and interacting with tourism markets. The education level of respondents varies, ranging from secondary education to higher education degrees, allowing the empirical analysis to capture possible differences in openness to innovation, managerial capacities, and the ability to interpret new tourism practices related to sustainability and social innovation.

The sample also includes both male and female managers, reflecting the growing participation of women in rural tourism entrepreneurship in Morocco. This gender diversity provides an opportunity to explore whether managerial characteristics influence the perception and adoption of social innovation practices within tourism establishments. In addition to individual characteristics, the establishments represented in the sample are located in different rural tourism environments, characterized by varying levels of tourism development, accessibility, and integration into regional tourism circuits. Some guesthouses operate in territories where rural tourism is already relatively structured and connected to national tourism flows, while others are located in areas where tourism activities remain emerging and strongly embedded in local community dynamics.

This heterogeneity in socio-demographic characteristics, professional experience, and territorial contexts is particularly important for the objectives of the study. It allows the empirical model to capture potential variations in how managers perceive the advantages, compatibility, complexity, and visibility of social innovation practices. By incorporating a variety of managerial and contextual situations, the sample provides an appropriate empirical basis for examining how both individual attributes and contextual factors influence the intention to adopt social innovation within rural tourism establishments in Morocco.

3.4. Choice of Methodology

The use of the ordered Logit model in this study is based on the ordinal nature of the dependent variable, which measures the intention to adopt social innovation through a Likert scale ranging from 1 to 5. This scale expresses a progression in intention, but the intervals between categories are not assumed to be equal. In other words, the difference between levels 1 and 2 does not necessarily reflect the same intensity of change as that between levels 4 and 5. This characteristic makes linear regression models unsuitable, as they assume a continuous variable, and also renders binary models inadequate, as they only account for two categories.

The ordered Logit model is specifically designed to handle ordinal variables, respecting the natural order of responses without imposing assumptions about the distance between categories. Its use ensures a better fit between the nature of the variable and the chosen method, while allowing for an accurate interpretation of the effects of different dimensions on the expressed levels of intention. This choice is particularly relevant in the context of rural guesthouses, where the adoption of social innovation does not occur abruptly but often follows a gradual process. The ordered Logit model thus makes it possible to understand the determinants of adoption by accounting for intermediate stages in the process, while maintaining the consistency of the effects of explanatory variables across all response levels.

4. Results

4.1. Robustness Analysis

Table 4. Specification test: Ramsey RESET.

Ramsey RESET Test Equation: UNTITLED Specification: ADOPT C RELVA RELSY COMPV COMPU SIMPL USEEZ TRIAL RISKT VISIB SOCOM AGEHH EDUHH GENDR
Omitted Variables: Squares of fitted values
Test	Value	df	Probability
t-statistic	0.081826	159	0.9349
F-statistic	0.006695	(1, 159)	0.9349
Likelihood ratio	0.007327	1	0.9318
F-test summary:
Test	Sum of Sq.	df	Mean Squares
Test SSR	0.000614	1	0.000614
Restricted SSR	14.57023	160	0.091064
Unrestricted SSR	14.56961	159	0.091633

Source: authors’ elaboration.

presents the results of the Ramsey RESET specification test, used to assess the functional validity of the ordered Logit model. This test helps determine whether omitted variables or nonlinear forms should have been included in the model. The results show a t-statistic of 0.0818, an F-statistic of 0.0067, and a likelihood ratio statistic of 0.0073, all associated with very high p-values (p > 0.93). These values clearly indicate that the addition of quadratic terms does not significantly improve the explanation of the dependent variable. In other words, the model’s structure is correctly specified and does not suffer from major omissions.

Table 5. Multicolinearity detection: Variance Inflation Factors (VIF).

Variable	Coefficient Variance	Uncentered VIF	Centered VIF
C	0.019232	36.74691	NA
RELVA	0.007300	4.545345	1.097977
RELSY	0.006891	4.022734	1.019682
COMPV	0.005995	3.525917	1.041516
COMPU	0.006932	4.296341	1.093751
SIMPL	0.007043	4.521645	1.039039
USEEZ	0.007060	4.478608	1.077500
TRIAL	0.007055	4.232182	1.034272
RISKT	0.008019	4.538503	1.099686
VISIB	0.006452	4.295287	1.098259
SOCOM	0.006586	4.409044	1.071482
AGEHH	0.005966	3.714479	1.075408
EDUHH	0.006535	3.837372	1.050532
GENDR	0.006872	4.540093	1.064137

Source: authors’ elaboration.

presents the results of the multicollinearity analysis using Variance Inflation Factors (VIF). Centered VIF values help detect excessive correlations between explanatory variables. In this model, all VIF values are well below the critical threshold of 5, indicating no problematic multicollinearity. The lowest VIFs are observed for RELSY (1.02), COMPV (1.04), and TRIAL (1.03), while the highest—such as those for RELVA (1.10) or VISIB (1.10)—remain well below critical levels. This confirms that the variables are sufficiently independent from one another.

Figure 1 presents the comparison between observed values, fitted values, and residuals from the model. A strong alignment is observed between the actual and predicted values, indicating a satisfactory fit. The model accurately reproduces the data’s dynamics without major discrepancies. The residuals, which represent the differences between actual and predicted values, are generally centered around zero and show no specific trend. Their dispersion is relatively consistent, with most falling within the range of −2 to +2, which is a sign of stability. No recurring pattern appears in the residual series, suggesting a correct model specification. Thus, Figure 1 confirms the model’s reliability and the consistency between modeled values and actual observations.

Figure 2 illustrates the distribution of the model’s residuals along with the results of the Jarque–Bera normality test. The test shows a distribution that is generally symmetrical and centered around zero, with no marked skewness or excessive kurtosis. The statistical indicators support this: the mean is close to zero (0.0198), the median is also very low (−0.0072), skewness (0.14) and kurtosis (2.33) remain within acceptable ranges. The Jarque–Bera statistic is 3.87 with an associated probability of 0.144, which is well above the 0.05 threshold. This means that the hypothesis of normality for the residuals cannot be rejected, confirming that the errors are normally distributed.

Table 6. Heteroscedasticity test: Harvey.

Test Statistic	Value	Degrees of Freedom	p-Value
F-statistic	0.723187	(13, 160)	0.7385
Obs*R-squared	9.656648	Chi-Square (13)	0.7217
Scaled Explained SS	6.862073	Chi-Square (13)	0.9091

Source: authors’ elaboration.

presents the results of the Harvey heteroskedasticity test, used to determine whether the variance of the residuals remains constant across observations. This test assesses the presence of heteroskedasticity, that is, non-constant error variance, which could affect the reliability of the estimates. The results show an F-statistic of 0.7232 (p = 0.7385), an Obs*R² statistic of 9.6566 (p = 0.7217), and a Scaled Explained SS statistic of 6.8621 (p = 0.9091). The associated p-values, all well above the 0.05 threshold, indicate that the null hypothesis of homoskedasticity cannot be rejected.

Figure 3 presents the evolution of the recursive coefficients estimated from progressively increasing sub-samples. This method is used to assess the stability of the model by examining whether the coefficients follow a consistent path as new observations are added. The results show that all coefficients stabilize quickly after a brief initial phase of fluctuation. The trajectories then become relatively constant, with no visible breaks, and the confidence intervals gradually narrow. This pattern indicates that the estimates are robust and do not vary significantly with changes in sample size.

Figure 4 displays the values of the Hat Matrix, used to identify potentially influential observations in the model estimation. These values measure the individual influence of each observation on the overall fit by quantifying their weight in the construction of the fitted values. The results show that the observations fall within a range between 0.04 and 0.12, well below the commonly used critical threshold of 2k/n, which in this case is approximately 0.16 (with k variables and n observations). No value reaches or exceeds this threshold, indicating that no single observation exerts excessive influence on the model estimation. Thus, the data are well-balanced, and no unit has a disproportionate impact. This reinforces the reliability of the results, the stability of the model, and the overall quality of the fit.

The validity of the explanatory dimensions was confirmed through a confirmatory factor analysis (CFA), retaining only items with a convergent validity above 0.5, and demonstrating satisfactory reliability and discriminant validity according to Cronbach’s alpha, composite reliability, and AVE criteria. Robustness tests indicate that the model is well specified (RESET test), free from multicollinearity (VIF < 5), and that the residuals are normally distributed (Jarque–Bera: p = 0.144), homoskedastic (Harvey test: p > 0.7), and the estimated coefficients are stable. No observation was identified as influential through the Hat Matrix, confirming the overall reliability of the model.

4.2. Ordered Logit Results

Table 7. Ordered Logit model: Estimation results.

Dependent Variable: ADOPT Method: ML—Ordered Logit (Newton-Raphson/Marquardt steps) Sample: 1 174 Included observations: 174 Number of ordered indicator values: 5 Convergence achieved after 7 iterations Coefficient covariance computed using observed Hessian
Variable	Coefficient	Std. Error	z-Statistic	Prob.
C	14.300039	2.961636	4.828426	*** 0.00000
RELVA	0.797762	0.437957	1.821554	* 0.07024
RELSY	0.641179	0.236575	2.710252	*** 0.00740
COMPV	11.545629	4.584974	2.518145	** 0.01270
COMPU	−0.185274	2.842198	−0.065187	0.94810
SIMPL	13.621195	3.691649	3.689732	*** 0.00030
USEEZ	2.162015	0.791776	2.730589	*** 0.00697
TRIAL	4.854434	2.192490	2.214119	** 0.02812
RISKT	4.525237	3.737309	1.210828	0.22760
VISIB	3.466284	1.616448	2.144383	** 0.03339
SOCOM	−1.702154	1.236603	−1.376476	0.17044
AGEHH	6.489004	4.194758	1.546932	0.12370
EDUHH	3.795570	1.202386	3.156698	*** 0.00188
GENDR	−7.740772	3.693138	−2.095988	** 0.03753

*** Significant at 1%; ** Significant at 5%; * Significant at 10%. Source: authors’ elaboration.

presents the technical characteristics of the ordered Logit model estimated to analyze the intention to adopt. The model was estimated using the maximum likelihood (ML) method, with the Newton–Raphson algorithm and Marquardt adjustments. The sample includes 174 observations, with an ordinal dependent variable measured on five levels, corresponding to the degrees of adoption intention on a Likert scale. The model converged after seven iterations, indicating a stable estimation process with no numerical issues. In addition, the variance–covariance matrix of the coefficients was computed using the observed Hessian, ensuring the reliability of the standard errors used for significance testing.

The perception of the superiority of the innovation (RELVA) exerts a positive influence on the intention to adopt, validating hypothesis H1a at the 10% significance level. This suggests that when the innovation appears superior to existing practices, rural guesthouse managers show an increased intention to adopt it. The perception of economic or symbolic benefits (RELSY) plays a significant role, being highly significant at the 1% level, which strongly supports hypothesis H1b. The more the innovation is associated with economic gains or social prestige, the more local actors tend to implement it. Furthermore, alignment with local cultural values (COMPV), significant at the 5% level, confirms hypothesis H2a, emphasizing the importance of cultural compatibility in the adoption of innovations. In contrast, technical or infrastructural compatibility (COMPU) does not appear to be significant, leading to the rejection of hypothesis H2b. This indicates that technical familiarity is not a determining factor in this context.

The simplicity of understanding the innovation (SIMPL) exerts a highly significant influence at the 1% level, validating hypothesis H3a: the easier the innovation is to understand, the more likely it is to be adopted. Similarly, ease of use (USEEZ) is a decisive factor, significant at the 1% level, confirming H3b and highlighting the importance of practicality. The possibility of testing the innovation on a small scale (TRIAL), significant at the 5% level, validates hypothesis H4a by showing that experimentation reassures potential users. However, the perception of risk (RISKT) does not significantly influence the intention to adopt, leading to the rejection of H4b. The visibility of positive outcomes (VISIB) is significant at the 5% level, validating H5a, which indicates that observable results encourage adoption. In contrast, social diffusion (SOCOM) is not significant, resulting in the rejection of H5b. Finally, among the control variables, the manager’s age (AGEHH) has no notable effect, whereas education level (EDUHH), significant at the 1% level, and being female (GENDR), significant and negative at the 5% level, both positively influence the adoption of social innovation.

5. Discussion

The results provide empirical support for several core mechanisms proposed by the Diffusion of Innovation theory (Rogers, 1962). First, the significant effects associated with relative advantage (RELVA) and economic or symbolic benefits (RELSY) confirm that the perceived value of innovation plays a central role in adoption decisions. In Rogers’ framework, innovations are more likely to diffuse when potential adopters perceive them as offering clear improvements over existing practices. In the context of rural guesthouses in Morocco, the adoption of social innovation appears to be strongly influenced by the expectation that such practices can enhance both economic performance and symbolic positioning within sustainable tourism markets. These results are consistent with Neumeier (2012) and Thorburn (2005), who highlight that innovations in rural contexts gain legitimacy when they demonstrate tangible advantages for local actors. In practical terms, this suggests that policies promoting social innovation in rural tourism should emphasize visible economic gains, differentiation in tourism markets, and reputational value associated with sustainability-oriented practices. Concrete policy instruments may include certification schemes for socially responsible rural tourism, targeted financial incentives for innovative practices, or promotional programs that highlight the competitive advantages of socially innovative guesthouses.

The results also highlight the importance of cultural compatibility (COMPV), which significantly influences the intention to adopt social innovation. This finding reinforces the idea that innovation in rural contexts cannot be understood solely through technological or organizational lenses, but must also be interpreted through the cultural embeddedness of local economic activities. Previous studies such as Roman et al. (2020) and Tresiana and Duadji (2022) emphasize that innovations are more easily adopted when they resonate with local traditions, identities, and collective values. In rural tourism environments, where hospitality practices are often closely linked to local heritage and community relationships, innovations perceived as culturally coherent are more likely to gain acceptance. This result suggests that public policies should prioritize locally adapted innovation strategies, rather than standardized tourism development models. For example, support programs could encourage projects that integrate traditional knowledge, local gastronomy, or community participation into tourism services, thereby reinforcing the cultural legitimacy of innovation.

The strong influence of ease of understanding (SIMPL) and ease of use (USEEZ) also confirms one of the fundamental predictions of the Diffusion of Innovation theory: innovations perceived as simple and accessible tend to diffuse more rapidly. In small-scale rural enterprises such as guesthouses, managerial resources and technical expertise are often limited. As a result, innovations that require complex knowledge or substantial organizational changes may encounter resistance. The results therefore suggest that the adoption of social innovation is facilitated when practices are operationally simple and compatible with existing managerial routines. From a policy perspective, this implies that programs supporting innovation in rural tourism should prioritize capacity-building initiatives, practical training workshops, and simplified implementation tools rather than purely technological interventions. By reducing the cognitive and operational barriers associated with innovation, such measures can significantly increase the willingness of local actors to adopt new practices. Another important finding concerns the significant effect of trialability (TRIAL). The possibility of experimenting with innovation on a limited scale appears to play a reassuring role for potential adopters. This result is consistent with Richter et al. (2019), who argue that experimentation mechanisms reduce uncertainty and facilitate the progressive appropriation of innovative practices. In rural tourism settings, where financial margins are often limited and decision-making tends to be cautious, the ability to test new approaches before full implementation may significantly reduce perceived barriers. This suggests that policy interventions should incorporate pilot programs, experimental projects, and demonstration initiatives that allow tourism operators to test innovative practices in real conditions before committing to large-scale changes.

The analysis also provides insight into the role of observability (VISIB). The visibility of positive outcomes significantly increases the intention to adopt social innovation, confirming Rogers’ argument that innovations diffuse more rapidly when their benefits are observable to others. In rural tourism environments, where actors are embedded in relatively close social and professional networks, visible success stories can act as powerful signals encouraging imitation. This implies that public authorities and tourism organizations should promote demonstration projects and knowledge-sharing platforms, enabling successful experiences of social innovation to become visible references for other operators. In contrast, some variables expected to influence adoption do not appear to be significant. Technical compatibility (COMPU) does not significantly affect the intention to adopt social innovation. This result suggests that, in the context of rural guesthouses, the adoption of social innovation is not primarily constrained by infrastructural or technological conditions. Rather, adoption decisions appear to depend more on symbolic, cultural, and practical considerations than on technical integration. This finding is theoretically meaningful because it indicates that social innovation in rural tourism often takes the form of organizational or relational practices, which can be implemented without major technological transformations. Consequently, innovation policies should not focus exclusively on technological modernization but also support organizational experimentation and community-based initiatives.

Similarly, the perception of risk (RISKT) does not significantly influence adoption decisions. This result may reflect the incremental nature of social innovation in rural tourism. Unlike large-scale technological investments, many social innovation practices—such as cooperation with local producers, community engagement initiatives, or sustainable resource management—can be implemented gradually and adjusted over time. The availability of experimentation opportunities may therefore reduce the salience of perceived risk. In theoretical terms, this finding suggests that when innovations are incremental, locally adaptable, and socially embedded, risk perceptions become less central in the adoption decision. The results also show that social communication (SOCOM) does not significantly influence adoption. While the diffusion of innovation theory emphasizes the role of communication networks in spreading new practices, the findings suggest that in this context managers rely more on direct observation of outcomes than on information transmitted through social channels. This may reflect the pragmatic decision-making patterns often observed in small tourism enterprises, where actors prefer to adopt practices that have already demonstrated visible results rather than those promoted through discourse alone.

Finally, the control variables provide additional insights into the socio-demographic determinants of innovation adoption. The education level of managers (EDUHH) has a strong positive influence, suggesting that higher educational attainment may enhance the ability to interpret new practices, access information, and engage with innovative approaches. From a policy perspective, this result highlights the importance of training and educational programs targeted at rural tourism entrepreneurs, particularly those with limited formal education. The significant effect associated with gender (GENDR) also suggests that adoption dynamics may vary according to managerial profiles. This finding indicates the potential value of gender-sensitive support programs, including initiatives that strengthen the role of women in rural tourism entrepreneurship and innovation processes. Overall, these results contribute to a better understanding of how social innovation spreads in rural tourism contexts. They suggest that adoption decisions are primarily driven by perceived usefulness, cultural alignment, operational simplicity, experimentation opportunities, and visible outcomes, rather than by purely technological conditions or risk considerations. These findings highlight the importance of designing policy instruments that combine economic incentives, local capacity building, pilot experimentation programs, and the visibility of successful initiatives, in order to foster the development of socially innovative practices in rural tourism.

6. Conclusions

This study examined the determinants of the intention to adopt social innovation in rural guesthouses in Morocco by mobilizing the analytical framework of the Diffusion of Innovation theory. Based on an ordered Logit model estimated from a survey of 174 guesthouse managers, the results highlight several mechanisms that shape the adoption process in rural tourism environments. Consistent with Rogers’ framework, the findings show that adoption is primarily influenced by the perceived usefulness and practical relevance of innovation. In particular, the perception of the superiority of the innovation and the economic or symbolic benefits associated with it significantly increase managers’ intention to adopt such practices. These results suggest that social innovation is more readily adopted when it is perceived as offering tangible advantages compared with existing tourism practices, either through improved service quality, enhanced reputation, or better positioning within sustainability-oriented tourism markets. Cultural compatibility also emerges as an important determinant, indicating that innovations that resonate with local values and traditions are more likely to be accepted by rural tourism actors. Furthermore, the results confirm the importance of simplicity and accessibility, as both the ease of understanding and the ease of use of the innovation strongly support adoption. The possibility of experimenting with innovation on a small scale and the visibility of its positive outcomes also appear to facilitate the diffusion process, as they reduce uncertainty and allow managers to observe concrete benefits before committing to broader implementation.

At the same time, the analysis provides insight into the factors that appear to play a more limited role in the adoption decision. Contrary to some expectations, technical compatibility with existing infrastructures, perceived risk during experimentation, and the capacity of innovations to spread through social communication networks do not significantly influence adoption intention in the context examined. These results suggest that social innovation in rural tourism is not primarily constrained by technological conditions or risk perceptions, but rather by the extent to which innovations are perceived as meaningful, practical, and culturally legitimate within local contexts. In other words, adoption decisions seem to depend more on socially embedded and experiential dimensions than on purely technical considerations. The analysis of the control variables also reveals that education level plays a significant role, with more educated managers displaying a stronger intention to adopt social innovation. This result indicates that educational background may enhance the capacity to understand, evaluate, and implement innovative practices. Gender differences also emerge as significant, suggesting that female managers may display a greater openness to socially oriented innovation strategies in rural tourism.

Beyond these empirical contributions, the findings provide several implications for the design of public policies and support mechanisms aimed at promoting social innovation in rural tourism. First, policy interventions should emphasize the concrete benefits and visible outcomes of socially innovative practices, since perceived usefulness and observability appear to be central drivers of adoption. Demonstration projects, pilot initiatives, and recognition programs for innovative guesthouses could therefore play an important role in strengthening diffusion dynamics. Second, the strong influence of cultural compatibility suggests that innovation policies should be anchored in local cultural and community contexts, encouraging initiatives that integrate traditional knowledge, local heritage, and community participation into tourism development strategies. Third, the importance of simplicity and experimentation highlights the need for practical support instruments, such as training programs, simplified implementation guidelines, and small-scale pilot schemes that allow tourism operators to test innovative practices before adopting them more broadly. Finally, the positive influence of education suggests that capacity-building initiatives and targeted training programs for rural tourism entrepreneurs could significantly enhance the adoption of socially innovative practices.

From a scientific perspective, this study contributes to the literature by extending the application of the Diffusion of Innovation framework to the field of social innovation in rural tourism, which remains relatively underexplored, particularly in the Moroccan context. It provides empirical evidence on the mechanisms through which social innovation spreads within rural tourism systems characterized by small-scale enterprises and strong local embeddedness. By identifying the significant role of perceived usefulness, cultural alignment, operational simplicity, experimentation opportunities, and visible outcomes, the study offers a clearer understanding of the conditions under which social innovation can support sustainable and inclusive rural tourism development. At the same time, the study has some limitations. It is based on cross-sectional data and on a sample limited to 174 rural guesthouse managers in Morocco, which may restrict the generalizability of the findings and does not allow changes in adoption behavior over time to be observed. In addition, the analysis focuses on perceived intention to adopt rather than on the long-term observation of actual adoption practices. Future research could build on these results by using longitudinal designs, larger and more diversified samples, comparative studies across regions or countries, and mixed-methods approaches combining quantitative and qualitative evidence.