Main Factors of Professional Experience on People’s Visual Behavior and Re-Viewing Intention in Different In-Forest Landscapes

Abstract

Background: With the acceleration of urbanization, the demand for people to return to forests and their desire for nature is also increasing annually. However, whether the visual behaviors and aesthetic preferences of users vary with background attributes when viewing forest landscapes remains to be further explored. This information will help forest landscape planners and managers to create a forest landscape space suitable for different groups of people. Purpose: This study mainly discusses whether people’s professional background will affect their visual behavior, satisfaction preference and re-viewing intention of a landscape and discusses the relationship among them. Main results: (1) Under the background of an in-forest landscape, the visual behavior of users in different layouts presents great similarity. (2) Satisfaction preference for landscapes is not significantly influenced by the user’s background, but there is a significant linear relationship between satisfaction preference and re-viewing intention (Adj. R² = 0.412 − 0.697, sig. = 0.000). (3) The spatial cognition that affects users’ visual behavior and satisfaction preference in a scene will change with the landscape layout and users’ professional background. Therefore, we suggest that landscape planners and managers should consider not only the spatial layout of the landscape itself but also the user’s own attributes (whether they have professional knowledge of the landscape) when optimizing the visual behavior and satisfaction preference of in-forest landscape space. This will ensure that users have a high sense of identity and attachment to the scene in a targeted way, thus arousing a greater “desire” to re-view the scene. In addition, our results can identify a more satisfactory course for tour routes according to the user’s professional background, thus enhancing the rate of tourists who decide to “visit again”.

1. Introduction

1.1. Utilization Trend in Forest Landscape Resources

At present, in addition to wood production and conservation, the ecological service functions of forests, such as popular science education, health care and leisure tourism, have been affirmed by various international conventions [1]. Reports pointed out that in the past five years in China, 122 “National Forest Tourism Demonstration Cities and Counties” have been named and 100 national key construction bases for forest experience and forest health have been identified [2]. In addition, according to the statistics of the China Forestry and Grass Bureau in 2020, the number of tourists in the forest tourism industry reached 1.8 billion in 2019, accounting for approximately 30% of domestic tourists [3]. That is, with the increase in forest area, people’s demand for forest tourism increased at the same time, which also poses new challenges to landscape planners and forestry workers, especially in cities and surrounding areas [4].

On the one hand, with the normalization of the COVID-19 epidemic situation and the control measures taken by governments in various countries to fight against the spread of the epidemic, people tended to cancel long-distance travel plans for safety and health and choose short-distance travel for sightseeing and recreation [5,6]. On the other hand, outdoor recreation activities are widely regarded as one of the more environmentally friendly and socially suitable ways to use natural areas, which may promote the long-term sustainability of these important resources [7].

However, the development and planning policies of landscape resources are usually driven from top to bottom. Planning strategies are often formulated by experts, and public opinions are not fully considered [8,9,10].

As feedback on this phenomenon, many researchers have indicated that users’ preferences should be considered in the decision-making process for landscape resource planning and design. After all, users are the experience and benefit groups of scenic resources in the final stage [11,12].

1.2. Importance of Meeting the Needs of Diversified Tourists

Visitors to forest parks are usually diverse, and their purposes for visiting are different. These tourists inevitably gather in forest parks. However, research by Francesca pointed out that the mismatch between the design of landscape resources and the feeling (cognition) of the user will hinder the embodiment of the recreational value of forest landscape [13].

Therefore, when planning a forest, landscape planners should balance people’s landscape demands and the preference of different groups to combine landscape resources with the needs of tourists [11]. Meanwhile, tourists’ preferences and overall evaluation of the landscape should be investigated. Using these surveys, planners can determine the preferences and needs of different tourists in forest areas [14].

Recently, related research on people’s perception and evaluation of outdoor recreation spaces and their use found that tourists who stay in forests for a long time have higher satisfaction than those who stay in forests for a short time [11]. Research by De Meo [15] conducted in natural recreation areas in Italian mountainous areas showed that analyzing the perception and preferences of local community residents can improve the sustainability of forest landscape resource utilization and reduce conflicts among forest users.

Petucco et al. [4] incorporated tourist characteristics and recreation preferences into the sustainable use of forests. Meanwhile, Ghimire and other scholars gradually began to pay attention to research on the landscape preferences and behavior characteristics of tourists with different attributes and pointed out that there are significant differences in users’ preferences for outdoor recreation space in terms of gender, marital status, educational background and working experience [16,17,18,19,20].

Karahalil [1] showed that the outdoor entertainment experience of a natural recreation place is influenced by demographic characteristics and tourists’ cognition. This research also pointed out that understanding the preference and perception of different tourists will better provide information for developing an integrated management system that considers both resource protection and the tourist experience. In the planning and design of natural recreation areas, tourist attribute information and entertainment experience should be considered [1].

Incorporating the preferences and expectations of different users into the decision-making process for forest landscape resource planning and design is an effective method and a means for sustainable natural resource management. Thus, it can increase the social acceptance of decision-making and reduce conflicts among users [21,22].

However, at present, most related studies rely too much on questionnaires [23]. This leads to limitations in the research methods. To compensate for this problem, eye-movement technology has been gradually introduced into the research field of landscape architecture because of its better accuracy and ability to record data in real time [24,25,26].

The introduction of eye-movement technology allows landscape designers to understand people’s perception and observation of landscapes more rationally, to obtain public feedback on the landscape and to better plan and manage the forest landscape.

1.3. Application of Visual Behavior Analysis in Forest Landscape Research

Previous research showed that natural forest scenes are more likely to attract people’s visual attention than built artificial scenes [27]. Meanwhile, people’s visual behavior when viewing natural landscapes is obviously different from that of urban scenes. When viewing natural scenes, people have fewer fixation counts, and the viewing involves less labor-saving to viewing [28].

In addition, Gao [29] and Zhou [30] analyzed the relationship among types of forest landscape space, the characteristics of constituent elements, and people’s visual behavior and preferences from the perspective of the composition characteristics of space and users’ attribute background. The results showed that the characteristics of landscape elements significantly affect users’ visual behavior and cognitive preferences. Users spend more time focused on elements with high complexity and proportion. In addition, research by Amati [31] on the vertical design of green space and the characteristics of users’ visual behavior showed that there are individual differences among users and their preferences for the sky, shrubs and stairs.

As De Lucio pointed out, women are more inclined to adopt the overall viewing strategy than men when they view a natural landscape and are more interested in different parts of the scene [32]. Compared with non-professional people, people with professional backgrounds mostly appreciate the landscape from the whole. While people without a professional background spend more time and attention on single objects, especially architectural elements or water elements of the landscape [33,34]. Additionally, users with a landscape background can efficiently pay attention to and analyze the physical characteristics of stimuli, such as texture, shape, color, etc. However, users without a landscape background usually decompose images into simplified and familiar shapes, which excludes visual effects that are not easy to understand or distract attention [35].

Similarly, we found that the existence of professional knowledge affected participants’ visual attention to forest leisure landscape space from the psychological cognitive level. Among them, people with a landscape specialty have a strong understanding of colors in landscape space [36].

From the above, we can see that (1) forest landscape resources have been increasingly demanded by modern people; (2) users of forest landscape resources have diversified demographic attributes and needs; and (3) the application of eye movement technology the in landscape field has become increasingly mature.

In these related research contents, we can also see that (1) previous studies mostly discussed the influence of landscape contents (elements, types, etc.) on users’ visual behavior or (2) from the perspective of different attributes (gender, professional knowledge background) when exploring the differences of users’ visual behavior when viewing the landscape. However, feedback on the visual behavior of people with different attributes, especially for those who enjoy forest landscape spaces, needs further discussion. Therefore, it is significant and very important to discuss the following issues: (1) How do people’s cognitive evaluations and viewing behaviors vary with attribute backgrounds across types of landscape space? (2) Does the relationship among landscape space (environment), preference evaluation and visual behavior change with people’s attributes?

Above all, there are significant group differences in the process of people’s cognitive processing of landscape. Educational experience is closely related to individuals’ cognitive and aesthetic processing [37,38,39]. Then, does professional experience affect individual’s intention to view the landscape again (aesthetic dimension) and their visual behavior (behavioral dimension)? If professional experience has an influence, what are the eye movement rules and characteristics of the influence? How is this influence reflected in landscape environments with different spatial layout? Additionally, other issues need to be further discussed. After all, an in-depth analysis of how users with different backgrounds observe landscapes (or pictures) will help us (scholars or researchers) better understand how differences in landscape vision come about [34].

So, this study mainly discusses whether users’ professional experience affects their visual behavior, satisfaction preference and re-viewing intention when they appreciate in-forest landscape space with different spatial layouts, and analyzes spatial cognitive factors that jointly act on visual behavior and satisfaction preference to arouse users’ re-viewing intention. After all, landscape designers need to understand participants’ preferences and needs for in-forest landscape to design landscape resources more pertinently and reasonably from the perspective of participants.

2. Materials and Methods

2.1. Study Area and Research Materials

Previous studies showed that arbor forests account for a large proportion of forest landscape space, accounting for an absolute advantage [40,41].

Second, the application of eye-tracking technology in the landscape field has matured, and research by Khachatryan proved that landscape photos are an effective substitute for real scenes [42,43,44,45].

Therefore, after on-site investigation and screening of existing forest parks in Liaoning Province, from the representativeness and universality of in-forest landscape space, we finally determined three forest parks as the research objects (Figure 1). An Olympus EM5 camera was used to collect 50 photos in the selected forest park at the same time (the maple season in 2020, sunny day and morning). In addition, we collected the indexes of tree species, DBH, coordinates (X, Y), crown width and tree height (20 × 20 m) of various plots. According to the concept of angular scale (W_i) proposed by Hui and Gadow [46,47], the selected in-forest landscape space was divided into uniform distribution space (UDS, W < 0.475), random distribution space (RDS, 0.475 ≤ W ≤ 0.517) and cluster distribution space (CDS, W > 0.517).

Finally, nine representative pictures selected by 10 experts and 20 non-experts were selected as the final experimental materials (Figure 1A–I) in three type of spatial layout.

2.2. Experimental Process

First, we analyze whether users’ attributes (landscape specialty and non-landscape specialty) affect their visual behavior when viewing the landscape. Therefore, in the data analysis, we mainly selected five eye-movement indicators that have often been used in previous studies: average fixation count (AFC), average lateral visual span (ALV), average portrait visual span (APV), average first fixation duration (FFD), and average pupil diameter (APD) [29,36,48,49,50,51,52]. With the help of Origin 2018 software, the spatial distribution of users’ fixation points in the scene was visualized.

Second, to avoid causing greater psychological load to users in the experiment, we referred to the relevant research on landscape cognition at home and abroad (following the principles of comprehensiveness, representativeness and designability), and combined with spatial characteristics of the scene selected in this study, we established a questionnaire [28,53]. The questionnaire mainly includes three aspects (

Table 1. Relationship among specialty, eye-movement index, satisfaction preference and re-viewing intention.

	AFC	ALV	APV	FFD	APD	SAP	RVI
Major	−0.094 *	−0.064 ns	−0.022 ns	0.045 ns	0.105 *	−0.044 ns	−0.133 **
N	477	477	477	477	477	477	477

Note: ** At the level of 0.01 (double tail), the correlation is significant. * At the level of 0.05 (double tail), the correlation is significant. ns. The correlation is not significant. AFC: average fixation count; ALV: average lateral visual span; APV: average portrait visual span; FFD: average first fixation duration; APD: average pupil diameter; SAP: satisfaction preference; RVI: re-viewing intention.

):

Part 1: Basic attributes, including user’s gender, age and specialty;

Part 2: Spatial cognitive index, using the 7-point Likert scale to quantify 7 indexes including richness of landscape content (RCO), near–far stereoscopic impression (NSI), richness of color (RCO), brightness of color (BCO), scene permeability (SPE), spatial uniformity (SUN) and sense of hierarchy (SHI);

Part 3: Overall evaluation of the scene, including satisfaction preference (SAP) and re-viewing intention (RVI).

Finally, we recruited 60 students without eye diseases to carry out the experiment. Studies have confirmed that it is feasible and representative to choose students to carry out eye-movement experiments (students have certain aesthetic cognitive ability) [27,54,55]. The experiment mainly includes four parts (Figure 2):

Part 1: Fill in the informed consent form (in line with the ethical statement of Shenyang Agricultural University), wear eye-movement instrument (German head-mounted SMI Glass2 glasses eye-movement instrument), debug and calibrate it (5-point calibration); meanwhile, read the instructions to user.

Part 2: Eye-movement experiment (Environmental Psychology Laboratory A, College of Forestry, Shenyang Agricultural University).

In a silent and dimly lit laboratory, subjects were asked to sit 3 m in front of the projection wearing an eye-tracker and viewing 1 filler picture, 9 experimental pictures and 9 blank pictures (in the form of PowerPoint). Finally, the single eye-movement data of each subject were obtained for approximately 180 s.

Part 3: Spatial cognitive experiment (Environmental Psychology Laboratory B, College of Forestry, Shenyang Agricultural University).

After the eye-movement experiment, we took the subject to laboratory B, played the corresponding 9 experimental materials again, and asked the subject to complete the corresponding cognitive questionnaire for each scene.

Part 4: After completing the spatial cognition questionnaire, we checked and collected the questionnaire to ensure that there were no missing answers, gave the subject a small gift and guided him or her to leave the laboratory.

3. Result

In this study, we recruited 60 students aged 18–24 years, and finally, 42 valid data points were obtained from each in-forest landscape space (excluding the incomplete data in the questionnaire) and 159 were obtained from the in-forest landscape of each spatial layout, a total of 477 valid data points. The effective rate was 88.3%. The participants included 261 landscape professionals and 216 non-landscape professionals (27 of forestry, 36 of horticulture, 45 of economic management, 108 of land and environment), and the ratio of landscape professionals to non-landscape professionals was 1:0.83.

Meanwhile, from Table 1, we find that there is a significant correlation among users’ professional knowledge of landscape, AFC, APD and RVI (p < 0.05).

That is, when users view in-forest landscape spaces, their own attributes (whether they have professional knowledge of landscape) will affect their visual behavior and re-viewing intention. Therefore, it is necessary to explore the influence mechanism among users’ professional knowledge, visual behavior, satisfaction preference and re-viewing intention of in-forest landscape space.

3.1. Visual Behavior of Users with Different Professional Specialties in In-Forest Landscape

First, we take the pixel size (4068 pixels × 3456 pixels) of the collected photos as the X axis and the Y axis and the fixation duration of a user in the scene as the Z axis to form a fixation point spatial distribution map (Figure 3), which revealed the following:

(1)

Overall, the visual range of a user with landscape professional knowledge is wider than that of a user without landscape professional knowledge. Additionally, the vertical visual range of users with landscape knowledge is relatively stable (X: 1500–3500 pixels; Y: 1000–3000 pixels).

(2)

Professional knowledge of landscape will affect their visual attention of in-forest landscape space. When viewing the in-forest landscape of UDS, users with landscape specialty have a larger horizontal and vertical search range than users without landscape knowledge. Moreover, users with landscape specialty (LAS) have a wider vertical search range, while those without landscape specialty (NLS) have the same horizontal and vertical search range (X_LAS: 1000–4000 pixels; Y_LAS: 1000–3000 pixels; X_NLS: 1500–3000 pixels; Y_NLS: 1000–2500 pixels).

In the in-forest landscape of RDS, users with landscape specialty have a wider vertical search range than users without landscape specialty. In addition, the horizontal and vertical visual ranges of those with landscape specialty are relatively consistent, and users without landscape specialty tend to search for “information” in a wide horizontal range (X_LAS: 500–2500 pixels; Y_LAS: 1000–3000 pixels; X_NLS: 500–3500 pixels; Y_NLS: 1000–2500 pixels).

In the in-forest landscape of CDS, users with landscape specialty have a wider horizontal and vertical search range than users without landscape knowledge. Meanwhile, users with landscape specialty have the same range of horizontal and vertical search information, while users without landscape specialty have a larger vertical search range (X_LAS: 1500–3500 pixels; Y_LAS: 1000–3000 pixels; X_NLS: 1500–2000 pixels; Y_NLS: 1000–2000 pixels).

That is, when users view the in-forest landscape space, their visual behavior is not only changed by the layout of plants in the space, but also the users’ own landscape specialty.

Second, to explore the differences of users’ visual behavior, we analyze the eye-movement data of users with different professional experiences with one-way ANOVA (Figure 4). Overall, regardless of whether users have professional knowledge of landscape, most of the eye movements produced by them when viewing in-forest landscape space have no significant difference (p > 0.05). However, when viewing the in-forest landscape of UDS, the average first fixation duration of user with a landscape specialty background is significantly longer than that of a user with non-landscape specialty (p < 0.05; AFD_LAS: 305668; AFD_NLS: 227227).

That is, in the in-forest landscape of UDS, a user with landscape specialty is less sensitive to the scene than a user without landscape specialty, possibly due to the “freshness” and “curiosity” of a user without landscape specialty to explore the scene faster.

3.2. Satisfaction Preference and Re-Viewing Intention of In-Forest Landscape for Users with Different Specialties

We analyzed the satisfaction preference of users with different specialties when they viewed the in-forest landscape space using one-way ANOVA. Additionally, to explore the relationship between users’ satisfaction preference and re-viewing intention, we performed a linear regression analysis with satisfaction preference as the independent variable and re-viewing intention as the dependent variable (Figure 5).

First, from Figure 5, we find that users with a landscape specialty have a slightly higher overall satisfaction preference for the scene than users without a landscape specialty. Meanwhile, it can be seen that users, whether they have a landscape specialty or not, show higher and more consistent satisfaction preference for forest landscape of USD (SAP_LAS: 4.149, SAP_NLS: 4.014) and CDS (SAP_LAS: 4.345, SAP_NLS: 4.208), and lower satisfaction for RDS (SAP_LAS: 3.701, SAP_NLS: 3.667).

Second, when users viewed in-forest landscape spaces with different spatial layouts, their satisfaction preference for the scene was not significantly different due to professional specialty.

In addition, we found that the more satisfied user is with the scene, the higher their re-viewing intention, regardless of whether they have a landscape specialty (Adj. R² = 0.412 − 0.697, Sig. = 0.000).

That is, both landscape specialty and non-landscape specialty present a consistent aesthetic evaluation (satisfaction preference) of the in-forest landscape space. Additionally, the better their first impression of the scene, the higher their attachment to the scene, thus stimulating their “desire” to re-view the scene.

3.3. Satisfaction Preference and Re-Viewing Intention of In-Forest Landscapes with Different Spatial Layout

First, we used Spearman correlation analysis to clarify the relationships among user’s specialty, spatial layout type, eye movement index, satisfaction, re-viewing intention and spatial cognitive factors (

Table 2. Relationships among spatial layout types, eye-movement indicators, satisfaction, re-viewing intention and spatial cognitive characteristics of in-forest landscape.

	RLC	NSI	RCO	BCO	SPE	SUN	SHI	RVI	TIL
Landscape Specialty
AFC	−0.097	−0.175 **	−0.054	−0.055	−0.185 **	−0.166 **	−0.188 **	−0.136 *	0.027
ALV	0.167 **	0.078	0.119	0.101	0.229 **	0.090	0.162 **	0.079	0.031
APV	0.140 *	−0.029	0.139*	0.047	0.068	0.077	−0.008	−0.059	0.101
FFD	0.092	0.113	0.015	0.005	0.149 *	0.028	0.072	0.011	−0.079
APD	−0.263 **	−0.099	−0.181 **	−0.159 *	−0.157 *	0.043	−0.095	−0.012	−0.196 **
SAP	0.536 **	0.641 **	0.459 **	0.525 **	0.461 **	0.636 **	0.647 **	0.784 **	0.067
RVI	0.475 **	0.585 **	0.451 **	0.474 **	0.346 **	0.546 **	0.627 **	1.000	0.072
TIL	0.432 **	0.120	0.529 **	0.420 **	−0.193 **	−0.316 **	0.047	0.072	1.000
Non-landscape Specialty
AFC	0.054	0.251 **	0.118	0.080	0.010	0.222 **	0.083	0.222 **	−0.043
ALV	0.047	0.158 *	0.037	−0.026	−0.094	0.065	−0.022	−0.060	0.002
APV	0.231 **	0.048	0.163 *	0.097	−0.060	−0.009	−0.022	0.026	0.123
FFD	−0.021	−0.134 *	−0.038	0.059	0.067	−0.083	−0.059	−0.120	0.153 *
APD	−0.016	0.019	0.001	0.046	0.198 **	0.201 **	0.124	−0.018	−0.202 **
SAP	0.311 **	0.385 **	0.363 **	0.438 **	0.353 **	0.527 **	0.517 **	0.694 **	0.071
RVI	0.407 **	0.450 **	0.467 **	0.358 **	0.322 **	0.437 **	0.497 **	1.000	0.095
TIL	0.471 **	−0.004	0.472 **	0.313 **	−0.030	−0.275 **	0.000	0.095	1.000

Note: * Significant correlation with p < 0.05; ** Significant correlation with p < 0.01. TIL: Type of in-forest landscape layout (1 represents uniform distribution space, 2 represents random distribution space, 3 represents cluster distribution space); AFC: Average fixation count; ALV: Average lateral visual span; APV: Average portrait visual span; FFD: Average first fixation duration; APD: Average pupil diameter; SAP: Satisfaction preference; RLC: Richness of landscape content; NSI: Near-far stereoscopic impression; RCO: Richness of color; BCO: Brightness of color; SPE: Scene permeability; SUN: Spatial uniformity; SHI: Sense of hierarchy.

). From the table, we can find that:

(1)

Users’ professional backgrounds are different, and their spatial cognition to arouse their visual behavior of in-forest landscape space is different.

Compared with users who do not have professional knowledge of the landscape, BCO and SHI can also arouse the visual behavior of users with professional knowledge of landscape.

(2)

In the forest landscape of in-forest space, when the distribution form (spatial layout) of plants changes, the visual behavior of individuals in response to the scene also changes accordingly. Additionally, the APD of users with a landscape specialty changes (negatively), while the FFD of users without a landscape specialty also changes (positively).

That is, whether users have professional knowledge of the landscape or not, when the layout of the in-forest landscape space they are viewing changes, their cognitive load on the scene will change. Additionally, their cognitive load is the largest in the UDS and the smallest in the CDS.

(3)

Users’ fixation count of in-forest landscape space can reflect their re-viewing intention of the scene. For users with a landscape specialty, the more they look at the scene, the lower their re-viewing intention. In contrast, users who do not have professional knowledge of the landscape show high re-viewing intention.

We speculate that the reason for this difference is that users with a landscape specialty explore the effective information of the scene quickly within the specified viewing time (15 s), and after exploring the effective and interesting “information” of the scene, they will have invalid and negative fixation. Users without a landscape specialty may be relatively slow to explore the scene because of the “freshness” of the scene, and the probability of negative fixation is low during the 15 s viewing.

Second, we took eye movement data and SAP of users with different specialties when viewing in-forest landscape spaces with different spatial layouts as independent variables and spatial cognitive indicators as dependent variables for Spearman correlation analysis, to explore the spatial cognitive factors that jointly act on visual behavior and satisfaction preference (Figure 6).

Figure 6 shows that users with different professional experiences have different spatial cognitive indicators that affect their visual behavior and satisfaction preference when viewing the in-forest landscape.

(1)

When users with a landscape specialty appreciate the landscape of UDS, RLC, NSI, SPE and SHI are positively correlated with SAP (p < 0.05) and negatively correlated with AFC and APD (p < 0.05), while SHI and SUN are positively correlated with SAP, AFC and APD for people without a landscape specialty (p < 0.05).

This mean that users with a landscape specialty are more inclined to enjoy the landscape of UDS with a rich, three-dimensional, transparent and strong sense of hierarchy, and such landscapes can reduce their “negative” fixation times and reduce the cognitive load on the scene. However, users without a landscape specialty prefer the landscape of UDS with a strong sense of hierarchy and neatness, to induce more “positive” visual behavior and less cognitive load.

(2)

When users with a landscape specialty appreciate the landscape of RDS, RLC, RCO, BCO, SPE and SHI are positively correlated with SAP, ALV and APV (p < 0.05) and negatively correlated with APD (p < 0.05). Meanwhile, when users without a landscape specialty view such scenes, RLC, NSI, RCO, BCO, SPE and SUN are positively correlated with SAP, AFC, APV and APD (p < 0.05) and negatively correlated with FFD (p < 0.05).

That is, users with a landscape specialty prefer diverse landscape content and colorful, bright, transparent and hierarchically randomly distributed scenes, thus arousing greater “horizontal + vertical” processing modes and less cognitive load. Users without a landscape specialty prefer space with diverse landscape contents and rich colors and which are bright, three-dimensional, transparent and neat, thus inducing more “positive” fixation, greater cognitive load, improving their sensitivity to the scene and a wide range of “vertical” processing modes.

(3)

When users with a landscape specialty appreciate the landscape of CDS, SPE, SHI and SUN are positively correlated with SAP, ALV and APV (p < 0.05) and negatively correlated with AFC (p < 0.05). However, when users without a landscape specialty view such scenes, NSI is positively correlated with SAP and AFC (p < 0.05).

That is, users with a landscape specialty prefer transparent, hierarchical and neat cluster distribution space, which leads to a wide range of “horizontal + vertical” exploration modes and inhibits the generation of “negative” fixation. Users without a landscape specialty prefer space with a strong sense of hierarchy, thus promoting the generation of “positive” fixation.

4. Discussion

This study discusses whether users’ professional experiences affect their visual behavior, satisfaction preference and re-viewing intention when viewing in-forest landscape spaces with different spatial layouts and analyzes the spatial cognitive factors that act on visual behavior and satisfaction preference. These clarified the problems of “ look at what ?” and “how to look at ?” the contents of the in-forest landscape when users with different educational backgrounds appreciated it, thus revealing the viewing and cognitive patterns of users with different educational backgrounds on the in-forest landscape.

Bagozzi proposed the theory of self-regulating attitude and pointed out that people’s attitude towards a scene can be divided into three stages: the evaluation of the scene, the emotion generated and the behavior in the scene. Evaluation promotes emotional generation and further affects the individual’s behavior or behavioral intention, and the overall performance is the process of evaluation–emotion–behavior. Additionally, Kevin Lynch also pointed out that the environment plays an important role in people’s feelings and practices. Landscape cognition emphasizes the process through which users perceive the environment and then identify the landscape and choose a behavior through interaction with the environment [56].

Our research results shows that different spatial distributions of in-forest landscapes have different degrees of visual cognitive load and preference evaluation on the subjects with different professional education. Among them, the UDS results in a greater cognitive load, and the CDS results in a lower cognitive load. Although people’s cognitive processing of landscape content of in-forest environment is influenced by their educational background, people’s overall preference evaluation of in-forest landscapes and their visual viewing mode have not changed. At the same time, such an overall evaluation has promoted users’ re-viewing intention for in-forest landscapes. Additionally, the better their first impression of the scene, the higher their “attachment” will be, thus stimulating their “desire” to view the scene again (Figure 7).

That is, although the processes of information cognition and processing of landscape are significantly influenced by educational background, it does not affect users’ overall preference for in-forest landscape and “return visit rate”. People’s interaction with the landscape environment is also a process of continuous cognition of the landscape, which leads to a series of visual behaviors and then forms a “self-cognition” of the landscape to show their preference for the landscape and produce “dependence”, and then produces an “attachment” to the landscape, which finally arouses people’s desire to re-view it. That is, it enhances the “stickiness” between people and the landscape (Figure 7).

Although existing research showed that people with different professional background will exhibit different visual behaviors when viewing landscape environment [57,58], in our research, we found users with different professional backgrounds showed similar visual behavior characteristics except for the first fixation duration in the landscape of UDS (Figure 4). This discovery is also consistent with the conclusions of Shi [59] and Paraskevopoulou [60], who find that professionals and non-professionals present the same visual behavior characteristics of people in spaces where plants are collocated.

This phenomenon can be attributed to the following aspects:

(1)

For spatial attributes, the visual structure and spatial attributes of landscape are the basic framework for aesthetic cognition [61]. For woody plant space, its main expressions are the form of space and the texture of plants [62,63]. This study discusses people’s visual behavior characteristics of in-forest space from the perspective of different spatial layouts. Although there are differences in spatial layout, they all show more homogeneity in shape, texture and spatial attributes, which is one reason why people have similar visual behavior characteristic in landscape space with strong homogeneity.

(2)

Regarding participants’ attributes, all participants recruited in this study are college students aged 20–26. Although they are considered to have good aesthetic judgment [27,29], in terms of age, such people all belong to young people. Similar age group attributes will make them have the same behavior trend, which is the second reason for their similar visual behavior characteristics.

(3)

From the background of participants’ current education and living environment, as students in an agricultural and forestry college, they will inevitably be greatly influenced by relevant knowledge of forestry or agriculture. After all, imparting knowledge from courses and teaching will make participants have similar knowledge orientation. This is also the third reason why they have a similar understanding of the forest, which leads to the formation of similar visual behavior.

(4)

In addition, it is very important that Cordon Allport (1937) put forward the trait theory that personality traits can be divided into common traits and personal traits [64]. It also points out that as an intermediary variable, people’s behavior is consistent in a certain social and cultural form or in a certain group. Similarly, ecological psychologists, represented by Barker, also pointed out that the characteristics of the environment support certain fixed behavior patterns. Although the user in them is constantly changing, fixed behavior patterns will be repeated in a certain period of time [65]. This also explains why the visual behavior of college students who viewing the landscape space of forests is similar in this study.

4.1. When Viewing Landscape Space with the Same Spatial Layout, the Overall Satisfaction Evaluation of the Landscape Space Is Not Affected by the Professional Background, and the Aesthetic Preference of the Scene Will Stimulate People’s High Desire to Visit Again

The outline, shape, composition, size, color and texture of elements in a landscape space will affect people’s perception and understanding of the landscape [62,63].

Although previous studies showed that users, as the main audience of the landscape, are influenced by culture and psychology [66,67,68], our research still find that people’s overall aesthetic preference is not influenced by professional education when the layout structure of in-forest landscape space is the same. This discovery is also supported by previous studies, in which people’s aesthetic preference for different plant structure spaces showed the same tendency in the analysis of gender [59,60].

In our opinion, on the one hand, the reason is that there is similar consistency in the spatial layout of forests. On the other hand, the volunteer group in this study has a high education level (university education). In other words, the same level of education narrows the differences of people with different professional educational backgrounds in the overall preference evaluation of a landscape. Alternatively, a higher education level allows people with different professional backgrounds to form a more unified overall aesthetic evaluation of a landscape space with the same structure.

Moreover, we also found that, whether people have a landscape specialty or not, their overall aesthetic preference (satisfaction preference) for a scene will promote people’s higher desire to visit it again (Figure 5), which is a very interesting phenomenon. It reflects that landscape appreciation, as a viewing experience, can influence and promote people’s desire for the next viewing of the landscape. This also confirms the research results of Kunst-wilson [69] and Zajonc [70]: repeated familiarity with stimuli will enhance people’s emotional preference for the stimuli. That is, people’s aesthetic processing depends on some internal memory effects, and people’s aesthetic preferences are influenced by the familiarity of stimuli.

This also means that forest landscape space, as an environmental scene with rich natural elements, is willing to explore and understand it more deeply with the improvement of its viewing times or aesthetic evaluation. In our previous research, we also found that the longer tourists stay in a forest park, the higher their overall satisfaction preference [11]. In other words, the more people look at forest landscape space, the more they like it, which also reflects the high charm of the forest landscape.

4.2. Spatial Cognitive Factors That Work Together on Visual Behavior and Satisfaction Preference Change with Differences in Spatial Layout and Professional Background

The feature theory in Bruner’s theoretical study of cognitive structure holds that when stimulated by the outside world, people will analyze these features and compare them with their memories and connections when they grow up to reflect them [71]. Our research found that when the spatial layout of a forest changes, the evaluation factors acting on visual behavior and overall satisfaction preference also change (Figure 6). That is, in forest landscape spaces with different layout characteristics, although people’s overall aesthetic evaluation presents the same trend (Figure 4 and Figure 5), participants have different evaluation dimensions of cognitive evaluation indicators for various types of spaces. In other words, in a landscape environment with the same in-forest background, with the change of spatial structure layout characteristics, the evaluation dimension of viewers will change, but this will not affect people’s overall aesthetic evaluation of this kind of space. This also means that although people have different evaluation mechanisms for different landscape spaces, the overall aesthetic effect of in-forest landscapes in people’s cognition presents the same trend.

In addition, attention recovery theory holds that the complexity and coherence of the natural environment can promote the soft charm of space and then awaken people’s attention [72]. Additionally, previous studies showed that a change in plant structure characteristics will have an important impact on people’s perception of a natural landscape [73], and when the composition structure of vegetation changes in space, it can produce higher aesthetic value [74].

In our research, we found that users with a landscape education background and those without a landscape education have the same trend in visual behavior and overall satisfaction evaluation of various types of spaces under different layouts (Figure 4 and Figure 5). However, the spatial cognitive factors affecting their visual behavior and satisfaction evaluation show different trends with the change of spatial layout characteristics (Figure 6). This is consistent with our previous research results [36]. These can also be supported by the previous conclusion that there are significant group differences in the cognitive processing of landscape by people [37,38,39].

From the results of spatial cognitive factor indicators that can promote visual behavior and overall preference evaluation in this study, we can conclude the following:

(1)

Cognitive indicators (layering and stereoscopic impression) represented by the structure of space can jointly act on the visual behavior and preferences of users with different educational backgrounds. That is, people’s cognitive index of spatial structure is the core element to promote people’s positive visual behavior and overall evaluation.

(2)

When the layout of an in-forest landscape space changes, the contribution (or influence) of content richness and ribbon diversity derived from structural cognition to visual behavior and overall satisfaction evaluation also changes.

That is, although there is no significant difference in visual behavior and the overall satisfaction evaluation of users with different educational backgrounds, the cognitive processing modes (or cognitive processing mechanisms) acting on overall preference evaluation and viewing do show different trends.

This can also explain why an increasing number of people like in-forest landscapes, because it can meet public demand (or diversified public demand). This is not only influenced by the different attributes of a user, but also by the landscape characteristics derived from the structural layout in the landscape space. Therefore, as a designer, when planning and designing landscape resources, it is very important to consider not only the overall evaluation effect of space, but also the difference in cognitive mechanisms brought by the change of spatial characteristics to users of different groups.

4.3. Feasibility and Limitations

4.3.1. Feasibility

First, previous studies have fully proved that landscape photos can be used instead of landscape scenes to study the aesthetic preference and visual behavior of landscapes, and it is feasible [44,75].

Second, for research on landscape aesthetic preferences and visual behavior, eye trackers and cognitive questionnaires have been widely used and fully proved to be scientific, valid and feasible [43,44,59,76].

Furthermore, research by Ghimire and others showed that it is necessary to study users’ attributes to improve landscape quality [16,18,19,20,77]. After all, landscape planning and design is not “entertaining” for landscape practitioners, so it is necessary for us to understand and master the characteristics and needs of landscape users.

Based on the above, we conclude that our research is feasible and necessary.

4.3.2. Limitations

First, in terms of experimental materials, we chose the forest landscape of an in-forest landscape. Even if the research showed that arbor forests are absolutely predominant in forest landscapes [40,41], there are not only in-forest landscape spaces, but also waterscape spaces, overlooking spaces and other landscape types in forest landscapes. Therefore, it is necessary for us to enrich the sample size to study different types of landscape spaces in future research.

Second, in terms of the survey object, we studied only college students. Because some research showed that the group aged 18–38 is absolutely predominant in forest tourism, it is representative to choose college students as the survey object [54,55]. However, the aging situation in China is becoming more and more serious, and it is expected to enter the stage of severe aging in 2035 [78]; thus, the phenomenon of the elderly returning to nature will gradually become a common phenomenon [79]. Therefore, in the future research, we should expand the scope of the survey object and enhance the universality of the research results.

Furthermore, in terms of user attributes, we studied only the users’ professional background, but user attributes also include gender, age, income and other indicators. In future research, we should also enrich the attributes of users to establish a more comprehensive mechanism of users’ visual behavior and satisfaction preference in forest landscapes.

Finally, in terms of seasons, we studied only the landscape in autumn, but research by Lin [80] has pointed out that seasonal changes have a significant impact on users’ visual behavior and aesthetic preference [51,81,82]. Therefore, as landscape planners and managers, it is necessary to understand the influence of seasons on users’ visual behavior and satisfaction preference in forest landscape spaces. This is another direction for future research.

5. Conclusions

This study takes nine sub-types of in-forest landscape space with three different spatial layouts in three forest parks in northern China as an example and analyzes the relationship among users’ visual behavior, spatial cognitive characteristics, satisfaction preference and re-viewing intention from the perspective of whether users have professional knowledge of landscape. The aim was to clarify users’ (with landscape specialty and non-landscape specialty) visual behavior, satisfaction preference, re-viewing intention of in-forest landscape spaces with different spatial layouts and which spatial cognitive characteristics influence them. The main results are as follows:

(1)

Under the background of an in-forest landscape, the visual behavior of users in different layouts presents great similarity.

(2)

Although users with different professional attributes present similar visual behaviors and satisfaction preferences for in-forest landscapes, the aesthetic preferences of the scene will stimulate people’s higher desire to visit again. (There is a significant linear relationship between satisfaction preference and re-viewing intention (Adj. R² = 0.412 − 0.697, Sig. = 0.000). The more satisfied people are with the first impression of the scene, the higher their desire to visit it again).

(3)

The spatial cognitive mechanism of arousing the visual behavior and satisfaction preference for in-forest landscapes of users with different professional backgrounds is different. On the whole, the color brightness and layering of the scene work together on the visual behavior and satisfaction evaluation of professional landscape users (p < 0.05), but these two spatial cognition factors have no significant influence on the cognitive mechanism of visual behavior and satisfaction evaluation of non-professional landscape users (p > 0.05).

These research conclusions show that when optimizing and improving the visual behavior and satisfaction preference of in-forest landscape spaces with different layouts, we should not only improve and optimize the spatial cognitive factors of the scene according to the spatial layout of in-forest landscapes but arouse users’ “desire” to visit the scene again. Moreover, it is necessary to comprehensively consider the users’ own attributes (whether they have a landscape specialty) to optimize the spatial cognitive factors in the scene so that they can have an attachment complex to the scene, which enhances users’ “desire” to view the scene again.

In other words, users are expected to change their feelings about the scene from “gone forever” to “reluctant to go”, “lingering” and a strong “desire” to visit it again. In addition, according to our research, we can provide users with a landscape specialty and those without a landscape specialty with a route that makes them more satisfied, to increase the “return visit” rate of tourists.