The Landscape Assessment Scale: A New Tool to Evaluate Environmental Qualities

Abstract

This study contributes to the growing interest in evaluating environmental qualities and characteristics for the enhancement of social and individual well-being by introducing and validating the Landscape Assessment Scale (LAS), a standardized tool designed to assess key environmental qualities across both natural and urban landscapes within metropolitan settings. The scale comprises 30 items related to 10 key environmental components: coherence, complexity, ephemera, imageability, naturalness, safety, visual scale, stewardship, disturbance, and historicity of places. In study 1, the LAS was first tested on 327 participants, who evaluated either a natural (N = 176) or urban (N = 151) environment. Exploratory Factor Analysis (EFA) revealed three correlated factors: Landscape Disharmony, Landscape Organized Complexity, and Landscape Naturalistic Impact. In study 2, participants (N = 185) were asked to select and to assess two environments (natural and urban) using the shortened LAS and the Perceived Restorativeness Scale (PRS). A Confirmatory Factor Analysis (CFA) was used to investigate the invariance of the LAS factor structure in both natural and urban environments, and the correlational analysis was used to investigate LAS convergent validity. The CFA supported the three-factor structure and showed significant correlations between LAS and PRS components, supporting convergent validity. By capturing key perceptual dimensions that are relevant across landscape types, the LAS offers a practical and scientifically robust tool for informing evidence-based urban planning and landscape design.

1. Introduction

With over half of the global population residing in urban areas [1], rapid urbanization presents significant environmental challenges that affect urban sustainability and human well-being [2,3]. Urbanization often results in the loss of natural habitats [4,5], increased pollution (e.g., air pollution [6], noise [7], artificial light [8]), elevated temperatures [9], and diminished access to green spaces [10,11,12]. Much of the existing literature has primarily emphasized the negative effects of urban spaces, which are often associated with increased stress, anxiety, and adverse emotions [13,14,15,16,17]. These factors collectively have a detrimental effect on the quality of life of urban residents, contributing to a range of environmental and health challenges.

However, certain urban areas may also offer opportunities for individual restoration. In particular, spaces of cultural or historical significance have been proposed as environments that could foster unique emotional benefits, depending on their design, aesthetic appeal, and cultural relevance [6,18]. These settings support a sense of belonging, identity, or connection, potentially benefiting emotional states. Moreover, incorporating natural elements into urban landscapes can significantly improve urban living conditions by providing opportunities for leisure, sports, aesthetic enjoyment, etc. Over time, the focus of urban development has shifted from merely improving living standards to enhancing environmental quality and quality of life. For many, particularly in economically developed countries, the modern city is not only viable but also a preferred habitat, associated with comfort, opportunity, and a high standard of living [6].

In this context, urban planning that prioritizes environmental sustainability and human well-being has become a crucial factor in advancing the United Nations Sustainable Development Goals of “good health and well-being” and “sustainable cities and communities” (SDG11 [1]). Achieving these goals requires a bottom-up, participatory approach [19] that actively involves citizens in the development of tools and strategies that highlight and enhance the environmental features most crucial to their well-being.

1.1. Building Biophilic Cities: The Need to Incorporate Nature into Urban Spaces

Parks and green spaces, as integral components of urban environments, play a critical role in enhancing urban emotional well-being [20]. The presence of parks, indeed, can mitigate the adverse effects of the urban environment by lowering air pollution, alleviating urban heat islands, and reducing noise levels [21,22]. Additionally, urban parks promote psychological restoration. According to the Stress Reduction Theory (SRT), natural environments are more effective in reducing stress and fostering positive emotions compared to urban landscapes [23]. Also, the Attention Restoration Theory (ART) suggests that exposure to stimulating and enriching natural environments enhances cognitive performance and helps restore attention [24]. Both theories are grounded in the biophilia hypothesis, which posits that humans have an innate connection to nature due to evolutionary factors [25]. To this aim, by the early 21st century, biophilia became a key concept in architecture, driven by the growing recognition of humans’ emotional need for nature in urban spaces. This led to the rise of biophilic design, which aims to integrate natural elements into architectural settings [26,27,28] to improve user satisfaction and well-being. Biophilic frameworks were developed to this aim: one of the most adopted is proposed by Kellert [29], whose dimensions emphasize both the incorporation of natural features within built environments (e.g., natural shapes and forms, natural patterns and processes) and the consideration of how individuals subjectively experience these features (e.g., evolved human–nature relationships, place-based relationships). Prior researches suggest that buildings designed with biophilic principles significantly enhance individual well-being across a variety of settings, from residential to healthcare environments [30,31,32].

As such, biophilic architecture is seen as contributing to sustainability by encouraging greater interaction with nature and promoting efficient management of natural resources [33,34]. However, not all natural environments and features offer equal benefits; variations in landscape types [35], elements [36], and spatial characteristics [37,38] influence perceptions and the health impacts of these spaces [39,40,41,42]. Specific features (such as tree canopy density [43], biodiversity [44], and landscape configuration [45]) can facilitate different types of interactions, supporting diverse activities from physical exercise to relaxation [46].

1.2. Understanding Environmental Preferences for Restorative Urban Design: The Main Theoretical Models

In the last decades, understanding how landscape characteristics impact well-being and sustainability has become a priority in urban planning. Many studies have analyzed environmental preferences, identifying essential visual components that shape people’s appreciation for landscapes [47,48,49]. Importantly, preferred environments are often those that facilitate psychological recovery from stress and mental fatigue—particularly among individuals who are already stressed or fatigued, because these are most sensitive to the benefits of restoration [47,48,49]. This link between preference and restoration is also supported by the environmental self-regulation theory, which suggests that individuals use their physical surroundings to manage emotional states, with natural settings often favored due to their capacity to enhance positive affect [50,51,52]. Indeed, natural environments offer supportive contexts for a range of emotion regulation strategies and processes [53,54], as individuals frequently seek out nature to cope with stress and manage emotions [55,56,57]. Therefore, understanding the environmental components underlying environmental preference is not only relevant for aesthetic and theoretical considerations, but also has concrete implications for designing urban spaces that promote collective restoration and well-being.

As stated, the biophilia hypothesis suggests that humans are biologically evolved to respond positively to natural settings [25,58]. Numerous studies and reviews support this preference for natural over built environments, with aesthetic and emotional experiences emerging as the primary benefits [24,47,59,60]. Moreover, several evolutionary theoretical models have been proposed to explain the consistent human preference for certain types of landscapes, particularly natural ones. For example, Orians’ savannah hypothesis [61] suggests that humans evolved a preference for landscapes resembling the East African savannah, characterized by open spaces, scattered trees, and nearby water sources, because such environments offered optimal conditions for survival. Similarly, Appleton’s prospect–refuge theory [62] posits that people are drawn to environments that provide a balance between visibility (prospect) and concealment (refuge), reflecting an evolutionary need to detect potential threats while remaining protected.

Complementing these evolutionary approaches, Kaplan and Kaplan’s [24] information-processing theory emphasizes the role of cognition in environmental preference. According to their model, humans favor environments that are easy to interpret (coherence), easy to navigate (legibility), rich enough to sustain interest (complexity), and promise further exploration (mystery). These theoretical models have supported the development of indicator frameworks as well as objective and subjective tools that are widely implemented in landscape monitoring, evaluation, and decision-making processes [63,64].

Despite substantial theoretical support, research in this area still demands an integrated approach. Holistic tools for understanding and evaluating environmental qualities and individual preferences are essential for designing urban and natural spaces that effectively promote psychological restoration, enhance well-being, and support sustainable development.

1.3. Tools to Assess Environmental Qualities

Existing frameworks and tools in landscape monitoring and evaluation incorporate both objective and subjective methods [63,64]. Objective measures often involve quantifiable metrics and indicators. For instance, Frank et al. [65] used a set of landscape metrics to examine land-use patterns, including aspects like naturalness and landscape diversity. Similarly, Tveit [66] tested visual preference through objective photo-based indicators, such as the percentage of open land in the view and the size of landscape rooms. Ozkan [67] calculated average visual quality scores for landscapes using both pixel- and object-based texture measures. On the other hand, studies similar to Sowińska-Świerkosz and Chmielewski [68] have emphasized subjective assessments, introducing the Landscape Quality Objectives (LQOs; European Landscape Convention) and bridging social and expert perspectives in landscape evaluation. Sowińska-Świerkosz [69] further developed the Index of Landscape Disharmony (ILDH) as a quantitative measure of landscape harmony, blending scenic beauty analysis and the ecological aspect. Zhang et al. [70] combined objective indicators with subjective preferences through an aesthetic assessment approach for urban landscapes.

However, recognizing the importance of adopting a holistic approach to assess environmental quality and its effects on well-being and sustainability, Tveit et al. [71] were among the first to propose such a more comprehensive framework. Rather than focusing on isolated attributes like naturalness or visual scale, the authors’ approach integrates multiple landscape visual dimensions across both natural and urban contexts. To this end, the VisuLands framework, initially developed by Tveit et al. [71] for rural landscapes, provides a structured and multidimensional basis for evaluating environmental quality in a holistic manner.

This framework, indeed, offers a systematic review of relevant aspects aimed at assessing visual environmental quality, identifying nine foundational components: stewardship, naturalness, complexity, imageability, visual scale, historicity, coherence, disturbance, and ephemera. These categories connect landscape attributes to theoretical foundations, facilitating a comprehensive approach to landscape assessment [71,72]. Later adaptations by Tveit and Ode [73] expanded this framework to capture the unique elements of urban landscapes, integrating the additional component of safety, and making the framework highly relevant for metropolitan settings.

Several studies illustrate the effective application of the VisuLands framework in evaluating landscape aesthetics. For example, Keleay and Atik [74] employed it to evaluate the aesthetic qualities of historical and cultural landscapes in Edirne, Turkey. They used a structured method with a panel of experts—including landscape architects, architects, and urban planners—who were provided definitions for each structural and textural parameter and asked to rate photographs of historically significant sites on a scale from 1 (lowest) to 5 (highest). Similarly, Özhancı and Yilmaz [75] applied this framework to rural landscapes, focusing on visual identity and aesthetic appeal to guide planning decisions that preserve rural aesthetics. Their study compared expert evaluations of rural sites against public perceptions, offering insights into the alignment between professional assessments and community views on landscape quality.

Hence, the VisuLands framework has been applied in these studies, largely depending on subjective, expert-driven evaluations of landscape attributes. Nevertheless, this approach may limit broader applicability and generalizability, since it did not involve a validated measurement scale; rather, assessments were based on operational definitions, which may have been subject to varied interpretations by the experts, potentially leading to inconsistencies or bias in the results. To address this limitation, our study aims to develop a more accessible, quantifiable, and standardized tool—the Landscape Assessment Scale (LAS)—designed for broader public application. Building on the revised VisuLands framework [73], the present study introduces the LAS as a means to assess both urban and natural environments within metropolitan contexts, both exploring potential relations with restorative experiences, aiming to provide a comprehensive understanding of the LAS components’ impact on human well-being. Theoretically, the LAS advances landscape and environmental psychology by providing a unified framework that bridges urban and natural settings, thus allowing for the exploration of how perceived landscape qualities may influence restorative experiences and well-being across diverse contexts. Practically, the LAS can offer a valuable instrument for researchers, urban planners, and policymakers to systematically evaluate and compare landscape characteristics in metropolitan environments, supporting evidence-based decisions that promote mental health and environmental quality.

1.4. Aims and Methods of the Present Work

The present contribution aims to develop a measure that captures environmental qualities common to both natural and urban landscapes (the LAS). Two independent studies were conducted to achieve this aim. In this work, we adopt the term landscape with a precise and intentional scope, following established definitions in landscape theory. In contrast to the broader and more abstract concept of environment, which encompasses everything surrounding us, landscape refers specifically to the portion of the environment that is directly perceived and experienced in a given moment [76].

In study 1, we explored LAS alternative factor structures and selected the best items reflecting the variability of landscape latent factors common to two different contexts: a nature and an urban landscape. Two independent samples were involved in this study that differed in terms of the task requested of participants: in the first sample, participants were asked to rate a natural landscape; in the second sample, participants were asked to rate an urban landscape.

In study 2, a third independent sample was collected to confirm the factor structure developed in study 1. In this study, to maximize the chances that selected items of LAS can be employed to evaluate different contexts, we used a repeated factor research design by asking participants to rate two different contexts using the developed scale: a natural and an urban landscape. Finally, in study 2, we also pursued testing the convergent validity properties of the LAS. Figure 1 illustrates the overall research framework of the present studies. In addition, since a large number of abbreviations are used in this paper, a table with acronyms and their respective acronyms has been inserted in Appendix C for ease of reading.

2. Study 1

The main aim of study 1 was to identify the latent factors that underlie the perception of landscape characteristics shared by both natural and urban environments. This involved determining the key dimensions that consistently capture how individuals experience and evaluate environmental features, providing a theoretical and empirical foundation for the development of a standardized tool for landscape assessment.

2.1. Sample Size Planning

To determine the optimal sample size for the Exploratory Factor Analysis (EFA) of the LAS, we ran an a-priori power analysis. Considering the results from MacCallum, Widaman, Zhang, and Hong [77], the sample size in EFA is mainly affected by the communality (proportion of item’s variance explained by common factors) and by the over-determination of factors (i.e., number of items having high factor loadings on the same factor). In the study by MacCallum et al. [77] and Hogarthy et al. [78], a communality of 0.20, corresponding to a factor loading of about 0.45, was considered low, while highly over-determined factors are factors having high loadings on three to four items. Combining these constraints to develop a measure that is able to assess environmental qualities common to nature and urban landscapes, we considered collecting two independent samples of participants: one assessing a natural landscape and the other assessing an urban landscape.

We considered three alternative factor structures for the initial pool of 30 LAS items: a one-factor structure, a two-factor structure (with 15 items on each factor), and a three-factor structure (with 10 items on each factor).

For the one-factor structure, assuming a sample size of N = 150 participants, with 30 items loading on a single factor with an average factor loading of 0.45, and a residual variance of 0.80, the statistical power was always over 0.90. By lowering the average factor loading to 0.30 (i.e., assuming a communality of about 0.09), and residual variance of 0.91, the statistical power ranged from a minimum of 0.87 to a maximum of 0.92.

For the two-factor structure, assuming a sample size of N = 150 participants, 15 items loading on each factor, an average factor loading of about 0.45 with a residual variance of 0.80, and a correlation among factors of about 0.30, the statistical power was always over 0.90 for factor loadings. By lowering the average factor loadings to 0.30 (with a residual variance of 0.91), the statistical power ranged from a minimum of 0.80 to a maximum of 0.85.

For the three-factor structure, considering an average factor loading of 0.45 on just one factor, with a residual variance of about 0.80, and an average correlation among factors of about 0.30, the average estimated power was higher than 0.90 with a sample size of N = 150. By decreasing the average factor loadings to 0.30 per item (and a corresponding communality of about 0.09), the statistical power was always over 0.90 for each loading.

2.2. Participants, Procedure and Measures

For the first study, we collected a convenient sample of N = 176 students who rated a nature landscape (Women n = 101) with an average age of M = 21 years (SD = 21). For the second study, we collected a sample of student participants (N = 151) who rated an urban landscape (Women n = 80) with an average age of M = 22 years (SD = 22). The data were collected between April and May 2024.

Development of Landscape Assessment Items

Building on the revised VisuLands framework [73], we developed the initial pool of items of the LAS using the 10 components (and their definitions) of the VisuLands framework as generative criteria. Each LAS item was constructed to capture people’s perceptions of environmental qualities, not exclusively visual, that apply to both natural and urban landscapes.

The item development process followed these steps:

• Independent development of three items per component by each of the three authors;
• Selection and refinement of the most appropriate items through consensus among the three authors;
• A piloting phase involving seven experts (N = 7), using the 58 items derived from the preliminary selection. In this phase, participants were asked to rank the items for each component based on perceived relevance.

Based on the results of this piloting phase, the three most relevant items for each component were selected for inclusion in the final version of the scale (

Table 1. VisuLands framework component and definitions [73] and the corresponding LAS item.

N	LAS Initial Components	Definition (Tveit et al., 2006) [71]	Developed Items
1	Stewardship	refers to the sense of order and human care within the landscape, which contributes to the perception of the place as well-maintained and harmonious with an “ideal” state.	LAS_STE_1 A sense of order is noticeable in the space. LAS_STE_2 A careful maintenance of the environment is perceived. LAS_STE_3 The maintenance of the space seems very poor.
2	Naturalness	captures the presence of vegetation and natural elements that evoke an unaltered or wilderness quality, enhancing restorative experiences.	LAS_NAT_1 The presence and sound of water are characteristic of the experience in this environment. LAS_NAT_2 The presence of plants is noticeable in the space. LAS_NAT_3 Sounds that bring you back to nature can be heard (e.g., the rustling of leaves or the wind).
3	Complexity	represents the diversity and richness of landscape elements, their spatial distribution, and the level of detail present in the place.	LAS_COM_1 There is a variety of colors in the space. LAS_COM_2 The place is articulated and rich in detail. LAS_COM_3 There is not enough diversity of elements in the environment.
4	Imageability	refers to the visual strength and memorability of a landscape, determined by distinctive features that make a place easily identifiable and memorable.	LAS_IMA_1 I noticed some details that made the experience of the space unique. LAS_IMA_2 I identified landmarks in the environment. LAS_IMA_3 There are no elements that leave a lasting visual impression of the place (e.g., artworks, design objects, architectural features, fountains, large trees, etc.).
5	Visual scale	describes the perceptible units of the landscape that influence how expansive or enclosed the environment feels.	LAS_VIS_1 The view is not obstructed by obstacles (e.g., trees, plants, or buildings). LAS_VIS_2 It is possible to see the horizon. LAS_VIS_3 The location presents open and wide spaces.
6	Historicity	reflects the presence of temporal layers in the landscape and the diversity and condition of cultural elements that may convey a historical narrative.	LAS_HIS_1The historical value of the space is noticeable. LAS_HIS_2 The influence of different cultures is noticeable in the space (e.g., a Japanese garden in a public villa, artworks from other cultures, ethnic shops). LAS_HIS_3 There are no historical-cultural elements in the environment.
7	Coherence	is the unity of the landscape, where features such as color schemes or textures harmonize and align with the surrounding context.	LAS_COH_1 The elements of the space seem to be in harmony with each other. LAS_COH_2 The environment displays a harmony of colors. LAS_COH_3 I did not perceive a sense of uniformity among the elements of the environment.
8	Disturbance	highlights the lack of adaptation and context in the landscape, where elements (often man-made) disrupt the landscape’s flow and coherence.	LAS_DIS_1 The artificial elements introduced into the environment seem poorly integrated with the landscape. LAS_DIS_2 There are elements in the place that seem out of place compared to the rest of the environment (e.g., a tall building near the beach, furniture elements that do not fit well in their location). LAS_DIS_3 I noticed elements that contrast with the surrounding environment, creating a feeling of disturbance.
9	Ephemera	relates to the presence of transient elements, such as seasonal changes or weather conditions, that alter the landscape over time.	LAS_EPH_1 By looking at the environment, typical elements of the season can be recognized. LAS_EPH_2 The presence of flowers is noticeable. LAS_EPH_3 I noticed features in the environment that change depending on the weather and atmospheric conditions.
10	Safety	pertains to the sense of security in a place, influenced by the relationship between vegetation density and visibility.	LAS_SAF_1 The presence of dense elements (e.g., vegetation, objects, buildings, walls) seems excessive to me. LAS_SAF_2 There is a balance between the density of elements (e.g., vegetation, objects, buildings, walls) and visual depth. LAS_SAF_3 The environment seems safe to me.

Note: LAS_STE: Stewardship; LAS_NAT: Naturalness; LAS_COM: Complexity; LAS_IMA: Imageability; LAS_VIS: Visual scale; LAS_HIS: Historicity; LAS_COH: Coherence; LAS_DIS: Disturbance; LAS_EPH: Ephemera; LAS_SAF: Safety.

). The full version of the scale is available in Appendix A.

Participants of study 1 were asked to rate a nature or an urban landscape using this initial pool of LAS items (three per each scale dimension) on a 6-point Likert scale (0 = not at all, 5 = very much). They were instructed to complete the questionnaire while physically present in the selected environment and, subsequently, to upload a photo of the evaluated landscape. Examples of the uploaded pictures are reported in Appendix D. The uploaded images of nature included parks, forests and lake environments, whereas the considered urban landscapes involved urban streets, urban houses and urban parking zones.

2.3. Analysis Strategy

The Exploratory Factor Analysis (EFA) with Ordinary Least Squares (OLS) rotated (oblimin) solution was used to evaluate all alternative LAS factor structures.

To determine the number of LAS factors to retain and to ensure minimal ambiguity between factors, we considered the following criteria for acceptable factors: all selected factors would have an eigenvalue greater than one and should be included in the maximum number of factors indicated by the parallel analysis, and (2) each selected factor would have a minimum of four items loading [79]. Based on these criteria, we considered several alternative factor structures compatible with these criteria.

To select items, we ran the EFA multiple times, and items with estimated factor loadings below 0.30 or non-congruent with other items loading on the same factor were eliminated from the factor solution. Moreover, as half of the participants rated a natural landscape and the other half rated an urban landscape, we selected only those clusters of items common to both samples (nature and urban) and loaded on the same factor in both samples.

To rate the LAS factors’ congruence obtained from the nature and urban factor solutions, we computed Tucker’s factor congruence coefficients [80]. Values higher than 0.90, assuming fair similarity among corresponding factors, and values higher than 0.95 imply strong equality between corresponding factors.

The reliability of the LAS factors was estimated using McDonald’s omega index. Factors with reliability estimates above 0.70 were considered satisfactory.

We used the psych library ([81], version 2.4.12) for EFA parameter estimates in the R environment [82].

2.4. Results

Results from the parallel analysis (with 5000 random samples) ran on the two samples indicating that the nature sample should select up to six factors. In contrast, the urban sample results indicated selecting a maximum of four factors.

Also, results of EFA on the LAS 30 items showed that eight (for the nature landscape sample) and nine (for the urban landscape sample) factors reported an eigenvalue greater than 1.

Inspecting scatter plots from nature and urban EFA (Figure 2), a two or three-factor solution emerged as a viable solution. Based on this pattern of results, we focused on the solutions for the two and the three correlated factors.

EFA Factor Solutions

The two rotated factor solutions that emerged after excluding nine items for the nature and the urban landscape assessment samples were challenging to interpret.

Regarding the nature landscape sample, the first factor explained 25.82%, while the first factor explained 23.78% in the urban landscape sample. This factor loaded 14 items: three items from naturalness, three items from ephemera, two from the visual scale, two from complexity, one from stewardship, one from imageability, and two from coherence.

The second extracted factor explained 12.09% in the nature landscape sample and 12.09% in the urban landscape sample. On this factor, six items were loaded: three items from disturbance, one item from stewardship, one item from coherence, and one from safety.

Since both extracted factors represent complex blends (hard to interpret) of environmental qualities of the landscape, we considered alternative solutions.

Turning to the three correlated factors solution (

Table 2. Three correlated factor solution for the nature and urban landscape samples.

	Nature					Urban
			F1	F2	F3			F1	F2	F3
	M	SD	λ	λ	λ	M	SD	λ	λ	λ
LAS_VIS_3 The location presents open and wide spaces.	4.20	0.98		0.65		3.49	1.17			0.48
LAS_COH_1 The elements of the space seem to be in harmony with each other.	3.76	1.09		0.45	−0.34	3.07	1.07	0.37	−0.34	0.31
LAS_NAT_2 The presence of plants is noticeable in the space.	4.40	0.99		0.49		3.23	1.29			0.67
LAS_IMA_2 I identified landmarks in the environment.	3.61	1.05	0.38			3.44	1.28	0.35
LAS_COM_2 The place is articulated and rich in detail.	3.22	1.13	0.62			3.03	1.26	0.57
LAS_COH_3 I did not perceive a sense of uniformity among the elements of the environment.	1.96	1.39			0.32	2.05	1.24		0.48
LAS_HIS_2 The influence of different cultures is noticeable in the space (e.g., a Japanese garden in a public villa, artworks from other cultures, ethnic shops).	0.99	1.34	0.36		0.43	1.27	1.54	0.63
LAS_IMA_1 I noticed some details that made the experience of the space unique.	2.93	1.44		0.47		2.05	1.55	0.68
LAS_EPH_2 The presence of flowers is noticeable.	2.93	1.42		0.31		1.89	1.43			0.55
LAS_NAT_1 The presence and sound of water are characteristic of the experience in this environment.	1.71	1.83		0.36		0.75	1.33	0.37
LAS_SAF_3 The environment seems safe to me.	3.24	1.58	0.47			3.69	s1.16			0.30
LAS_DIS_1 The artificial elements introduced into the environment seem poorly integrated with the landscape.	1.78	1.42			0.46	2.13	1.39		0.56
LAS_STE_3 The maintenance of the space seems very poor.	1.97	1.42	−0.68			2.03	1.33	−0.32	0.52
LAS_EPH_3 I noticed features in the environment that change depending on the weather and atmospheric conditions.	2.74	1.39		0.43		2.20	1.47			0.54
LAS_HIS_1 The historical value of the space is noticeable.	2.40	1.65	0.29			1.91	1.79	0.68
LAS_COH_2 The environment displays a harmony of colors.	3.37	1.20		0.56		2.79	1.33	0.49
LAS_SAF_2 There is a balance between the density of elements (e.g., vegetation, objects, buildings, walls) and visual depth.	1.57	1.19	−0.31		0.40	1.61	1.34		0.54
LAS_STE_1 A sense of order is noticeable in the space.	2.99	1.26	0.65			2.79	1.23		−0.47
LAS_VIS_1 The view is not obstructed by obstacles (e.g., trees, plants, or buildings).	2.48	1.39			−0.24	2.13	1.45		−0.17
LAS_NAT_3 Sounds that bring you back to nature can be heard (e.g., the rustling of leaves or the wind).	3.81	1.35		0.58		2.11	1.47			0.52
LAS_VIS_2 It is possible to see the horizon.	2.48	1.65		0.32		1.54	1.52			0.35
LAS_SAF_1 The presence of dense elements (e.g., vegetation, objects, buildings, walls) seems excessive to me.	2.13	1.34			0.38	2.79	1.43		0.29
LAS_HIS_3 There are no historical-cultural elements in the environment.	1.89	1.74	−0.11			2.13	1.90	−0.21
LAS_COM_1 There is a variety of colors in the space.	3.05	1.24	0.39	0.33		2.64	1.16	0.43
LAS_DIS_2 There are elements in the place that seem out of place compared to the rest of the environment (e.g., a tall building near the beach, furniture elements that do not fit well in their location).	1.49	1.35			0.77	1.55	1.31		0.64
LAS_DIS_3 I noticed elements that contrast with the surrounding environment, creating a feeling of disturbance.	1.38	1.23			0.79	1.46	1.17		0.76
LAS_COM_3 There is not enough diversity of elements in the environment.	2.05	1.39	−0.52			2.14	1.24	−0.62
LAS_IMA_3 There are no elements that leave a lasting visual impression of the place (e.g., artworks, design objects, architectural features, fountains, large trees, etc.).	1.99	1.60	−0.17			2.20	1.65		0.24
LAS_STE_2 A careful maintenance of the environment is perceived.	2.86	1.32	0.74			2.36	1.18	0.41	−0.32
LAS_EPH_1 By looking at the environment, typical elements of the season can be recognized.	3.19	1.27		0.48		2.48	1.49			0.71

Note: λ = unstandardized factor loading; LAS_STE: Stewardship; LAS_NAT: Naturalness; LAS_COM: Complexity; LAS_IMA: Imageability; LAS_VIS: Visual scale; LAS_HIS: Historicity; LAS_COH: Coherence; LAS_DIS: Disturbance; LAS_EPH: Ephemera; LAS_SAF: Safety.

), a total of 17 items were excluded: three items with a factor loading < 0.30 (LAS_VIS_1, LAS_IMA_3, LAS_HIS_3); thirteen items were excluded for not having the main factor loading on the same factor in both samples (LAS_COH_1, LAS_IMA_1, LAS_EPH_2, LAS_SAF_3, LAS_STE_3, LAS_HIS_1, LAS_COH_2, LAS_STE_1, LAS_EPH_1, LAS_VIS_2, LAS_IMA_2, LAS_SAF_1, LAS_NAT_1); and finally one item with high factor loading (>0.30) on multiple factors (LAS_HIS_2).

Based on the final rotated factor solution (

Table 3. Selected items from the three correlated factor solution for the nature and urban landscape samples.

	Nature			Urban
	LD	LOC	LNI	LD	LOC	LNI
LAS_DIS_3 I noticed elements that contrast with the surrounding environment, creating a feeling of disturbance.	0.82	0.01	−0.05	0.86	0.06	0.04
LAS_DIS_2 There are elements in the place that seem out of place compared to the rest of the environment (e.g., a tall building near the beach, furniture elements that do not fit well in their location).	0.77	0.05	−0.01	0.61	−0.04	0.19
LAS_DIS_1 The artificial elements introduced into the environment seem poorly integrated with the landscape.	0.53	−0.07	−0.03	0.60	−0.17	−0.16
LAS_SAF_2 There is a balance between the density of elements (e.g., vegetation, objects, buildings, walls) and visual depth.	0.43	−0.14	0.08	0.57	0.07	−0.05
LAS_COH_3 I did not perceive a sense of uniformity among the elements of the environment.	0.40	−0.15	−0.05	0.40	−0.04	−0.04
LAS_COM_2 The place is articulated and rich in detail.	0.10	0.71	0.02	−0.01	0.63	−0.08
LAS_STE_2 A careful maintenance of the environment is perceived.	−0.18	0.66	−0.06	0.04	0.55	−0.01
LAS_COM_1 There is a variety of colors in the space.	0.12	0.59	0.19	−0.03	0.52	0.10
LAS_COM_3 There is not enough diversity of elements in the environment.	0.21	−0.49	0.10	0.13	0.39	0.13
LAS_VIS_3 The location presents open and wide spaces.	−0.04	−0.05	0.78	0.12	0.08	0.64
LAS_NAT_2 The presence of plants is noticeable in the space.	−0.02	0.07	0.63	0.05	0.01	0.56
LAS_NAT_3 Sounds that bring you back to nature can be heard (e.g., the rustling of leaves or the wind).	−0.11	0.18	0.36	0.11	0.12	−0.55
LAS_EPH_3 I noticed features in the environment that change depending on the weather and atmospheric conditions.	0.13	0.12	0.31	−0.28	0.22	0.31

Note: factor loading < 0.30 highlighted in grey; LAS_STE: Stewardship; LAS_NAT: Naturalness; LAS_COM: Complexity; LAS_IMA: Imageability; LAS_VIS: Visual scale; LAS_HIS: Historicity; LAS_COH: Coherence; LAS_DIS: Disturbance; LAS_EPH: Ephemera; LAS_SAF: Safety.

), we found that the first factor in the nature landscape sample explained about 27.33% of the total variance, and in the urban landscape sample, 20.96%. Five items loaded on this factor: three items from the disturbance group, one from safety, and one from coherence. This first factor was named Landscape Disharmony (LD) based on this pattern. Reliability for this factor was good (nature landscape sample, McDonald’s omega index = 0.75; urban landscape sample, McDonald’s omega index = 0.75).

In the nature landscape sample, the second factor explained about 13.82% of the total variance, while in the urban landscape sample, it explained 18.93%. This second factor comprised four items: three from complexity and one from stewardship. This second factor is characterized by landscape diversity—such as the variety of vegetation and the richness of plant life in terms of colors, forms, and textures—and by the care and maintenance of the environment, conveying a sense of organization. For this reason, it was named Landscape Organized Complexity (LOC). Reliability for this factor was good (nature landscape sample McDonald’s omega index = 0.72; urban landscape sample McDonald’s omega index = 0.61).

Finally, the third factor explained about 11.07% of the total variance in the nature landscape sample and about 10.11% in the urban landscape sample. This factor saturated four items: two from naturalness, one from ephemera, and one from visual scale. As this third factor is primarily characterized by the presence of components related to natural features, such as vegetation, natural patterns like visual scale, and natural processes, in terms of the ephemeral cycles of life and environmental variability, we named it Landscape Naturalistic Impact (LNI). Reliability for this factor was good (nature landscape sample, McDonald’s omega index = 0.64; urban landscape sample, McDonald’s omega index = 0.62).

The three correlated factor solution was selected to represent the LAS factor structure and offers greater interpretability if seen within the theoretical perspective of environmental preferences.

Considering the Tucker’s factor congruence indices of the three-factor solution of the two samples, we found that all factors showed high similarity indices, respectively: Landscape Disharmony phi = 0.977; Landscape Organized Complexity phi = 0.91; Landscape Naturalistic Impact phi = 0.908.

Table 4. Correlations between LAS factors for each of the two landscapes rated: natural (below the main diagonal) vs. urban (above the main diagonal).

			LD	LOC	LNI
	M	SD	r	r	r
LD	1.64	0.92		0.043	0.044
LOC	2.79	0.67	−0.129		0.355 *
LNI	3.79	0.80	−0.243 *	0.300 **
M			1.76	2.54	2.76
SD			0.90	0.61	0.92

Note: * p < 0.01; ** p < 0.001; LOC = Landscape Organized Complexity; LNI = Landscape Naturalistic Impact; LD = Landscape Disharmony.

shows the means, the SD, and the correlation between LAS factors for both contexts: natural and urban landscapes. For the natural landscape (below the main diagonal), LNI is negatively correlated with LD, and positively correlated with LOC. LD and LOC do not show a significant relationship. This result is also confirmed for the urban landscape, where we found only a positive and significant correlation between Landscape Naturalistic Impact and Landscape Organized Complexity.

2.5. Discussion

The primary objective of study 1 was to identify the latent factors underlying the assessment of landscape characteristics common to both natural and urban environments. To achieve this, we evaluated two alternative factor structures of the LAS—a two-factor and a three-factor solution—and selected the most representative items that captured shared dimensions of landscape perception across these contexts. The final reduced version of the scale is available in Appendix B.

Ultimately, the three-factor correlated solution was chosen due to a first analysis of the scree plot, clearer interpretability, and more substantial conceptual alignment with the theoretical perspective of environmental preferences. Following item reduction procedures, which led to the exclusion of 17 items due to low or inconsistent loadings, three distinct factors emerged: LD, LOC, and LNI.

LD captures perceptions of components related to coherence, disturbance, and safety, mainly corresponding to spatial harmony, a naturalistic pattern closely related to biophilia, particularly to the naturalistic dimension of Kellert’s [29] biophilic framework. Additionally, the sense of safety experienced in a place—shaped by the balance between vegetation density and visibility—can be connected to both the prospect–refuge theory [83] and the savannah hypothesis [61], highlighting how visibility and spatial arrangement can influence landscape preference. This factor accounted for the largest portion of the variance in both samples and demonstrated excellent internal reliability.

LOC reflects the richness and care evident in the environment, encompassing a variety of visual details and indications of maintenance. This factor can be linked to the individual inclination to prefer environments that foster curiosity and exploration, while preserving the legibility and comprehensibility of the place, as explained by the information-processing theory proposed by Kaplan and Kaplan [24]. Reliability of this factor was acceptable across both natural and urban samples.

LNI represents components related to natural features, such as vegetation, patterns like visual scale, and natural processes, regarding the ephemeral life cycles and environmental variability. This factor closely aligns with the concept of biophilia, specifically with the naturalistic dimension within Kellert’s [29] biophilic framework. Thus, the LNI factor captures those components in the landscape that, according to the biophilia hypothesis [25], reflect our fundamental inclination toward natural surroundings. Reliability for LNI was satisfactory in both contexts.

Using Tucker’s congruence coefficients, comparisons of these factors across natural and urban landscapes indicated a high degree of structural similarity. This suggests strong consistency in how these latent dimensions are perceived across different types of landscapes.

Correlational analyses further highlighted both the distinctiveness and interrelatedness of these factors. In the natural context, LNI was negatively correlated with LD and positively correlated with LOC, indicating that naturalistic attributes of the landscape tend to be experienced more positively when they are perceived as harmonious, and vice versa. Meanwhile, LD and LOC were not significantly related, reflecting their conceptual distinction: LD reflects the perception of spatial discordance or visual incoherence, whereas LOC refers to the presence of a structured and organized complexity—a quality of the environment that is rich in elements but perceived as well-organized and coherent.

A significant positive correlation was found between LNI and LOC in the urban sample, while other relationships remained non-significant. This pattern mirrors that observed in the natural context, underscoring that naturalistic attributes within the urban environment are also more likely to be positively experienced when accompanied by structural organization, and vice versa.

From a practical perspective, these findings suggest that the LAS can serve as a valuable tool for assessing both urban and natural environments. The three identified factors leverage innate human predispositions that influence environmental preferences. For example, LD relates to the sense of safety derived from experiencing a harmonious and coherent environment, where visibility is facilitated; LOC taps into humans’ natural inclination to explore, heightened by environments that stimulate curiosity through variety without compromising legibility; and LNI reflects the innate tendency to connect with life and life-like processes, favoring natural elements for restoration and well-being. Enhancing these dimensions can translate into actionable strategies for urban planners and architects:

• Reducing LD involves integrating artificial elements with the surrounding context and avoiding visual disruptions. For instance, in a coastal area, limiting tall buildings that obstruct open views or selecting materials and colors consistent with the local environment and traditions can strengthen visual harmony. Also, maintaining a balance between the density of vegetation or objects can improve perceived safety and visibility.
• Enhancing LOC requires creating environments that stimulate curiosity through variety and richness while maintaining clear legibility. This can be achieved by incorporating diverse architectural details, colors, or textures in an urban square, ensuring logical spatial organization and sightlines for ease of navigation. Careful maintenance of spaces, clean surfaces, and repaired fixtures also contribute to aesthetic appeal and a sense of stewardship.
• Strengthening LNI focuses on increasing the presence of natural features and processes. Strategies include planting diverse vegetation, integrating water elements, and designing spaces that evolve with seasonal changes (e.g., deciduous trees offering shade in summer and open views in winter). Incorporating natural sounds, such as rustling leaves or flowing water, and preserving open spaces with broad visual horizons can further promote restorative experiences and enhance human–nature connections.

Overall, these findings reinforce the multidimensional nature of landscape experience, emphasizing that people’s perception of natural and urban environments can be meaningfully described through three core dimensions: disharmony, organized complexity, and naturalistic impact. These dimensions reflect fundamental aspects that individuals tend to recognize when making evaluations across environments. The stability of this factor structure across contexts suggests a shared environmental framework underlying landscape appraisal.

3. Study 2

In the present study, we aimed to confirm the LAS factor structure identified in study 1. In that initial study, the LAS structure was developed using two independent samples of participants who each assessed a different type of landscape—natural or urban—to select items capable of capturing features relevant to both settings. Building on this, the current study seeks to confirm the factor structure and test whether the same LAS structure holds consistently across both natural and urban landscapes. To this end, participants in study 2 were asked to rate both a natural and an urban landscape.

3.1. Sample Size Planning

For power analysis, we used the library pwrSEM developed by Wang and Rhemtulla [84]. We considered a repeated LAS structural equation model, including three factors assessing a nature and three factors assessing an urban landscape, with an assumed average factor loading of 0.40, an average correlation among the same latent factors but between types of landscapes of about 0.50 (i.e., LD nature and LD urban), and an average correlation of 0.30 for those correlations among different factors both within and between types of landscapes (i.e., LD nature and LNI nature or urban), and a residual average variance for manifest variables of about 0.84, and a residual average covariation between the same manifest variables in the two tasks of 0.10. Based on this model, with 1000 Monte Carlo simulations, and a sample of N = 150 participants, the power for estimating factor loadings ranged from a minimum of 0.85 to a maximum of 0.92.

3.2. Participants, Procedure and Measures

For study 2, N = 185 student participants were involved in the study (Women n = 110) with an average age of M = 21 years (SD = 3).

All participants completed the following list of questionnaires:

The Landscape Assessment Scale: The LAS that was developed in study 1. Using the LAS, all participants were requested to rate two landscapes, one LAS for nature and one for urban landscape.

Perceived Restorativeness Scale (PRS): An adapted 11-item version of the Perceived Restorativeness Scale (PRS [85]) was used to assess participants’ perceived restorativeness of natural and urban environments, rather than the full 26-item version. Participants rated each item on a 5-point Likert scale (1 = not at all, 5 = very much). The items represented the four components outlined by the Attention Restoration Theory (ART) as essential for a restorative environment: being-away, fascination, coherence, and compatibility [24,86], along with an additional component reflecting perceived regenerative potential. The item adaptation was based on the validated Italian version of the PRS [87]. Coherence items were reverse-coded, and a total average PRS score was calculated for analysis, where higher scores indicated a greater perceived restorative effect in the virtual nature setting.

Finally, all participants completed a socio-demographic sheet including gender, age, level of education, and occupation.

3.3. Analysis Strategy

To confirm the factor structure developed in study 1, we ran a Confirmatory Factor Analysis (CFA) considering alternative models other than the three-factor model developed in study 1: a one-factor model and a two-factor model by aggregating LOC and LNI factors into the same latent factor. As in study 1, we asked participants to rate two landscapes, and a repeated confirmatory factor model was estimated with two identical and correlated factor structures: one for the natural and one for the urban landscape. As for longitudinal models, we correlated the residual variance of the same manifest variable in the two landscape types. Finally, the latent correlations between the same (LSD nature and LSD urban) and different latent factors, both within the same landscape (LSD nature and LOC nature) and across landscapes (LSD nature and LOC urban), were estimated.

Before running the CFA, we checked the normality of the distribution assumption for each LAS item with the Shapiro–Wilk test. As all items reported a significant Shapiro–Wilk test (minimum S–W test = 0.55, p = <0.001; maximum S–W test = 0.93, p = <0.001), we used the robust WLS estimator for parameter estimates. For evaluating the goodness of fit of the model, we used the following list of indices [88]: the Chi-squared test; the Root Mean Square of Approximation (RMSEA) and the Standardized Root Mean Square Residual (SRMR) (both below 0.08); the Comparative Fit Index (CFI) and the Tucker Lewis Index (TLI, both indices above 0.90). To determine whether the three-factor structure of the LAS provided a better fit than the alternative solutions, we compared its fit indices with those of alternative models (i.e., one- or two-factor models). If the Chi-squared difference test between the two models—one with more parameters (i.e., the two-factor model) and one with fewer parameters (i.e., the one-factor model)—is not significant, then the more parsimonious model is preferred. On the contrary, if the Chi-squared difference is significant (p < 0.05), the less parsimonious model provides a better fit.

Furthermore, to investigate whether the LAS factor structure was invariant across repeated landscape tasks (nature vs. urban), we also tested the LAS factors following the recommendations by Wu and Estabrook [89], where invariance is probed considering the following series of increasingly restrictive conditions: configural dimensionality of LAS factor structure; the invariance of item responses’ thresholds; the invariance of LAS factors loadings across landscapes, the invariance of latent response intercepts, and the invariance of latent response scales. To decide whether a more restrictive (as opposed to the less restrictive) condition holds, we considered the following differential tests between the more and the less restrictive models: a non-significant χ² difference test (∆χ² [90,91]), the difference in CFI ≥ −0.01 (∆CFI [92]), and a difference in RMSEA indices ≤ 0.01 (∆RMSEA [93]). It is important to note that while we expected the LAS structure to apply to both types of landscapes, we also anticipated that the landscapes might differ along specific LAS dimensions, given their distinct characteristics. As a result, full measurement invariance between the two contexts might not be achieved. We consider this a critical feature of the LAS, which should be carefully addressed before evaluating its convergent and discriminant validity.

Latent correlations between LAS factors and the PRS dimensions were used to assess convergent aspects.

Further, to examine whether the LAS factors effectively differentiate between natural and urban landscapes, we conducted a series of paired-samples t-tests comparing participants’ scores on each LAS factor across the two landscape types. This approach allowed us to test the scale’s sensitivity in detecting experiential differences between the natural and urban conditions.

For the CFA, we used the lavaan ([94], version 0.6–19) and semTools [95] libraries in the R environment [82].

3.4. Results

Table 5. Comparing the fit indices of the three CFA models.

					RMSEA 90% CI
Model	Robust χ2	df	p	RMSEA	LL	UL	CFI	TLI	Δ χ2	Δ df	Δ p
[M1] One-factor model	1142.97	285	0	0.116	0.108	0.124	0.712	0.672	--	--	-
[M2] Two correlated factors model	770.68	280	0	0.085	0.077	0.094	0.846	0.822	−5590.85	5	1.00
[M3] Three correlated factors model	481.08	271	0	0.076	0.067	0.085	0.882	0.859	390.80	9	<0.01
[M3a] Three correlated factors model	429.631	268	0	0.065	0.055	0.075	0.915	0.897	84.59	3	<0.01

Note: Satorra and Bentler scaled difference Chi-squared test was used to determine the best-fitting model; χ² = Chi-squared; df = degree of freedom; RMSEA = Root Mean Square Error of Approximation; C.I. = confidence interval; SRMR = Standardized Root Mean Square Residual; CFI = Comparative Fit Index.

presents the comparison of the fit indices of competing confirmatory models of LAS: the one-factor model (M1), the two correlated factors model (M2), and the three correlated factors model (M3). In all cases, we compute robust parameter estimates. The one-factor model (M1) reported the worst model fit among all models. The three-factor model (M3) fit was significantly better (Δχ² (9) = 390.80, p < 0.01) than that of the two-factor model (M2). However, all the three tested models reported low values of both CFI and TLI, even if the three factors model fit indices were slightly below the cutoff of 0.90. Inspecting the residual covariance matrix to improve the fit indices of the three LAS factor model, we decided to release three residual covariances between item STE_2 and SAF_2 related to the LAS nature landscape, between DIS_2 and DIS_3 items related to the LAS urban landscape, and between LAS_DIS_2 and LAS_COM_1 for the natural landscape. This last model (M3a) showed an improved RMSEA and a CFI value above 0.90, while TLI is now closer to 0.90.

3.4.1. Test for Invariance of LAS Factors Structure Across (Repeated) Landscapes

A further goal of the present study was to investigate the invariance of the LAS factor structure across contexts, i.e., nature and urban landscapes. As shown in

Table 6. Comparison between models’ fit for the invariance hypothesis.

	Robust χ2	df	p	RMSEA	CFI	TLI	Δ χ2	Δ df	Δ p	Δ RMSEA	Δ CFI
M0: Configural model	429.63	268	0	0.07	0.92	0.90
M1: Invariant Thresholds model 1	465.88	302	0	0.06	0.92	0.92	33.35	34	0.50	0.006	−0.007
M2: Invariant Factor loadings model	644.81	312	0	0.06	0.92	0.91	17.40	10	0.07	−0.001	0.006
M3: Invariant scalars model	966.52	325	0	0.14	0.56	0.56	474.97	13	0.00	−0.075	0.360

¹ Two items (VIS_3 and NAT_2) were recoded (respectively to four and three response options) due to zero or close to zero frequencies in response options that caused estimation issues in the invariant thresholds model. Note: χ² = Chi-squared; df = degree of freedom; RMSEA = Root Mean Square Error of Approximation; C.I. = confidence interval; SRMR = Standardized Root Mean Square Residual; CFI = Comparative Fit Index.

, the fit of the invariant thresholds model (M1) was not significantly different from the configural model (M0, i.e., the best fitting model, M3a, shown in Table 5), meaning that thresholds are invariant across repeated landscape tasks. Additionally, the weak metric invariance hypothesis was confirmed, as the fit of the invariant factor loading model (M2) was not significantly different from M1. However, the test for the scalar invariance hypothesis failed; therefore, no further invariance tests were considered.

Table 7. Invariant (across landscapes) and completely standardized factor loadings, S.E., and 95% C.I.

				95% C.I.
Item Label	Latent Factor	Std λ	s.e.	LL	UL
LAS_COM_2 The place is articulated and rich in detail.	LOC	0.48	0.06	0.38	0.59
LAS_COM_1 There is a variety of colors in the space.	LOC	0.66	0.05	0.55	0.76
LAS_COM_3[R] There is not enough diversity of elements in the environment.	LOC	0.45	0.06	0.33	0.57
LAS_STE_2 A careful maintenance of the environment is perceived.	LOC	0.52	0.06	0.40	0.65
LAS_VIS_3 The location presents open and wide spaces.	LNI	0.56	0.07	0.42	0.70
LAS_NAT_2 The presence of plants is noticeable in the space.	LNI	0.65	0.07	0.52	0.79
LAS_EPH_3 I noticed features in the environment that change depending on the weather and atmospheric conditions.	LNI	0.45	0.06	0.33	0.58
LAS_NAT_3 Sounds that bring you back to nature can be heard (e.g., the rustling of leaves or the wind).	LNI	0.80	0.06	0.68	0.91
LAS_COH_3 I did not perceive a sense of uniformity among the elements of the environment.	LD	0.40	0.06	0.29	0.51
LAS_DIS_1 The artificial elements introduced into the environment seem poorly integrated with the landscape.	LD	0.63	0.05	0.54	0.73
LAS_SAF_2 There is a balance between the density of elements (e.g., vegetation, objects, buildings, walls) and visual depth.	LD	0.36	0.06	0.25	0.47
LAS_DIS_2 There are elements in the place that seem out of place compared to the rest of the environment (e.g., a tall building near the beach, furniture elements that do not fit well in their location).	LD	0.70	0.04	0.62	0.79
LAS_DIS_3 I noticed elements that contrast with the surrounding environment, creating a feeling of disturbance.	LD	0.85	0.04	0.77	0.93

Note: λ = unstandardized factor loading; S.E. = standard error for the unstandardized factor loading; LL = Lower Limit for the 95% C.I.; U.L. = Upper Limit for the 95% C.I.; std λ = completely standardized factor loading; LOC = Landscape Organized Complexity; LNI = Landscape Naturalistic Impact; LD = Landscape Disharmony; LAS_STE: Stewardship; LAS_NAT: Naturalness; LAS_COM: Complexity; LAS_IMA: Imageability; LAS_VIS: Visual scale; LAS_HIS: Historicity; LAS_COH: Coherence; LAS_DIS: Disturbance; LAS_EPH: Ephemera; LAS_SAF: Safety.

shows the three-factor model’s invariant (across landscapes) and completely standardized factor loadings.

3.4.2. Correlations Between and Within Landscape Tasks, Reliability, and AVE for LAS Factors

As expected from study 1, the correlations between LOC and LNI factors in the two landscapes were positive and significant (r = 0.627 and r = 0.451, respectively;

Table 8. Cronbach’s alpha (for ordinal variables) reliability estimates and AVE indices for the LAS factors.

		Nature			Urban
		LOC	LNI	LD	LOC	LNI	LD
Nature
	LOC
	LNI	0.627 **
	LD	−0.345 **	−0.201 *
Urban
	LOC	0.269 *	0.068	−0.003
	LNI	0.108	−0.048	0.131	0.451 **
	LD	−0.013	0.123	0.307 **	−0.443 **	−0.017
	Cronbach’s alpha	0.590	0.684	0.712	0.614	0.583	0.724
	AVE	0.285	0.395	0.382	0.289	0.293	0.350

Note: * p < 0.05; ** p < 0.01; LOC = Landscape Organized Complexity; LNI = Landscape Naturalistic Impact; LD = Landscape Disharmony.

). For the nature landscape assessment, LD correlated negatively with LOC and LNI (r = −0.345, and r = −0.201). However, for the urban landscape, LD was negatively related only to LOC (r = −0.443), not to LNI (r = −0.017).

Turning to cross-landscape correlations, interestingly, both LOC and LD reported positive correlations in the two tasks (r = 0.269 and r = 0.307, respectively), meaning that both factors are useful in assessing common elements shared by natural and urban landscapes. However, this is not true for LNI, which showed a non-significant correlation between landscapes (r = −0.048), and emerges as the most discriminating factor between the two landscapes.

The reliability of the LAS factors ranged from a minimum of Cronbach’s alpha = 0.59 (Landscape Organized Complexity of nature landscape) to a maximum of Cronbach’s alpha = 0.72 (urban landscape Disharmony) (Table 8). However, the Average Variance Extracted (AVE) indices were below the expected cutoff of 0.5 [96].

3.4.3. Convergent Validity

Table 9. Correlations among LAS factors and PRS factors.

	Nature						Urban
	LNI		LD		LOC		LNI		LD		LOC
	r	p	r	p	r	p	r	p	r	p	r	p
PRS.FASC	0.510	<0.000	−0.131	0.075	0.602	<0.000	0.168	0.023	0.107	0.147	0.124	0.092
PRS.BEAWAY	0.472	<0.000	−0.121	0.100	0.283	<0.000	0.096	0.195	0.112	0.128	0.073	0.326
PRS.COHER	0.226	0.002	−0.299	<0.000	0.442	<0.000	−0.083	0.262	−0.035	0.641	0.073	0.324
PRS.SCOPE	0.525	<0.000	−0.003	0.972	0.221	0.002	0.060	0.417	0.007	0.925	0.149	0.043

Note: LOC = Landscape Organized Complexity; LNI = Landscape Naturalistic Impact; LD = Landscape Disharmony; PRS.FASC = Fascination; PRS.BEAWAY = Being Away; PRS.COHER = Coherence; PRS.SCOPE = Scope.

shows the correlations between the LAS and PRS. Comparing the correlations between LAS factors and the restorative quality of the environment (i.e., the PRS factors: Fascination, Being Away, Coherence, and Scope) across the two landscapes, it emerges that the assessment of LNI and LOC correlates higher with PRS factors in the natural than in the urban landscape. This pattern of correlations is congruent with our expectations that the environmental quality aspects assessed with LAS factors contribute, at least in part, to the perceived restorativeness of the environment.

3.4.4. Scale Sensitivity: Differences Between Natural and Urban Landscape

To test whether the LAS factors can capture differences between urban and natural landscapes, we performed a series of paired t-tests comparing natural vs. urban landscape scores (

Table 10. Differences between natural and urban landscape LAS factor scores.

	Nature		Urban
	M	SD	M	SD	t-Test	df	p
LNI	4.07	0.64	2.50	0.97	18.18	184	<0.001
LD	1.48	0.84	1.77	0.88	−3.69	184	<0.001
LOC	3.11	0.81	2.52	0.82	7.74	184	<0.001

Note: LNI = Landscape Naturalistic Impact; LD = Landscape Disharmony; LOC = Landscape Organized Complexity.

). Results showed that the natural landscape was rated higher in naturalistic impact and organized complexity compared to the urban landscape. Conversely, the urban landscape was perceived as more disharmonious and incoherent than the natural landscape. These findings support the sensitivity of the LAS in capturing meaningful differences across landscape types.

3.5. Discussion

The primary objective of study 2 was to confirm the factor structure of the LAS previously developed in study 1, to test its structural invariance across two distinct environmental contexts (i.e., nature and urban landscapes) as well as its sensitivity to landscapes’ differences, and explore the convergent validity of the scale with the PRS. The results provided important evidence supporting both the robustness and contextual sensitivity of the LAS, as well as its convergent validity with restoration measures.

First, our findings confirmed that the three-factor model of the LAS, comprising dimensions of Landscape Disharmony (LD), Landscape Organized Complexity (LOC), and Landscape Naturalistic Impact (LNI), outperformed alternative one- and two-factor models. Although the initial three-factor model (M3) did not fully meet conventional cutoff values for all fit indices, a refined model (M3a), which allowed for theoretically plausible residual covariances, demonstrated an acceptable and improved fit. This supports the idea that the LAS captures multiple distinct but interrelated experiential dimensions of landscape perception, aligning with both the theoretical rationale for multidimensionality and empirical results from study 1.

Second, the LAS showed partial invariance across the nature and urban landscape contexts. Specifically, we found support for configural, threshold, and metric invariance, suggesting that participants interpreted the factor structure and item loadings similarly across the two landscape types. This result is particularly relevant because it confirms that the LAS measures comparable constructs regardless of whether a landscape is natural or urban in character—an important prerequisite for generalizability and cross-context comparison. However, scalar invariance was not supported, indicating that the intercepts of item responses varied between nature and urban landscape ratings. This result suggests that participants understood the constructs similarly across contexts and systematically rated particular landscape features higher or lower depending on the environment type. The lack of scalar invariance is not surprising and should not be viewed as a limitation but rather as an expected outcome based on theoretical grounds. Indeed, we anticipated that participants would perceive and experience nature and urban environments differently along key dimensions of the LAS. The LAS is designed to detect such differences, and this sensitivity is an important feature of the scale.

Additionally, correlational analyses in study 2 support the validity of the LAS. LOC and LNI were positively associated in both nature and urban contexts, suggesting that these dimensions reflect shared elements of perceived landscape richness and complexity. LD, in contrast, showed negative associations with LOC in both landscapes, and with LNI only in the nature condition, indicating that disharmony may interfere more with naturalistic impact in natural settings. Cross-landscape correlations were modest for LOC and LD. At the same time, LNI showed no significant consistency, reinforcing its sensitivity to environmental context and its role as a key factor in distinguishing landscape types.

Convergent validity was partially supported through correlations with the PRS. As hypothesized, LNI and LOC in natural settings were more strongly related to perceived restorativeness (especially fascination and scope). These associations were weaker or inconsistent in urban contexts, suggesting that environmental quality features captured by the LAS contribute more robustly to restorative perceptions in natural environments.

Moreover, findings showed that the internal consistency of the LAS factors was acceptable overall. However, some values were modest, and the AVE indices fell below the recommended threshold of 0.50, suggesting that while the factors capture coherent constructs, further refinement may enhance their convergent validity.

Finally, the paired comparisons confirmed the scale’s sensitivity to environmental context. Nature landscapes were rated significantly higher on LNI and LOC—interpreted as greater in naturalistic impact and organized complexity—whereas urban landscapes were perceived as more discordant (LD). These differences reflect theoretically grounded distinctions in how natural and built environments are experienced and further validate the LAS utility in capturing the experiential signature of different landscape types. Overall, these findings reinforce the multidimensional structure and contextual adaptability of the LAS, as well as its ability to detect both shared and unique experiential qualities across environments. Indeed, with respect to already existing tools, the LAS structure allows for merging different aspects that characterize the experience of the environments in which individuals usually live and with which they interact; it combines objective [65,66,67] and subjective [68,69] approaches used separately in previous studies. Widening the existing literature [70], even if the environmental quality features captured by the LAS contribute more robustly to the restorative perceptions in natural environments, the scale results are a valuable tool for landscape assessment of both the natural and urban environments, following individual environmental preferences.

4. Conclusions

This study contributes to the growing interest in evaluating environmental qualities and characteristics for the enhancement of social and individual well-being by introducing and validating the LAS, a standardized tool designed to assess key environmental qualities across both natural and urban landscapes within metropolitan settings. In contrast to many existing instruments that rely solely on expert-driven subjective evaluations or narrowly focused objective indicators, the LAS adopts a more holistic approach, grounded in the revised VisuLands framework, developing specific items for measuring its theoretical components.

Through two independent studies, the LAS demonstrated a stable three-factor structure—Landscape Disharmony, Landscape Organized Complexity, and Landscape Naturalistic Impact—each interpretable in light of established environmental preference theories.

In the end, theoretically, the LAS advances landscape and environmental psychology by providing a unified framework that bridges urban and natural settings, enabling researchers to explore how perceived landscape qualities influence restorative experiences and well-being across diverse contexts. Practically, the LAS offers a novel instrument for researchers, urban planners, and policymakers to systematically evaluate, compare, and monitor landscape characteristics in metropolitan environments, supporting evidence-based decisions that promote both well-being and environmental quality. Its originality lies in integrating multidimensional perceptual measures within a standardized scale applicable across multiple landscape types, thus filling a critical gap in the assessment of environmental experiences.

Limitations and Future Directions

One limitation of the present studies is that both relied on convenience samples of university students, which may limit the generalizability of the findings. Future research could broaden the sample to include participants from more diverse demographic and geographic backgrounds, as well as expert raters or older adults, to capture a wider range of landscape perceptions and experiential responses. Another limitation concerns the subjectivity of the data: participants evaluated perceived environments at a specific moment; therefore, responses may be influenced by temporary mood, personal experiences, cultural expectations, or familiarity with the environment, which could affect attention to landscape characteristics or perceived restorative potential. Incorporating additional control measures could complement subjective ratings and strengthen the robustness of the findings. Moreover, while convergent validity was examined through correlations with the PRS, future studies should include conceptually distinct measures to assess the divergent validity of the LAS and further complete the validation process. Such work would enhance the scale’s generalizability and its potential to inform evidence-based landscape planning and design across diverse contexts.