Tourists’ Visual Attention and Stress Intensity in Nature-Based Tourism Destinations: An Eye-Tracking Study During the COVID-19 Pandemic

Abstract

Tourists’ visual attention has a central function in constructing their visual experiences and affects their perceptual and cognitive processes. Visual attention might be affected by environmental factors; however, the effects of environmental factors on visual attention are still vague in the literature. Moreover, visual attention might influence tourists’ stress intensity. This study explores how tourists’ visual attention patterns vary under environmental factors and quantifies the effects of visual attention on stress intensity by a mixed- methodology involving observations, eye-tracking experiments, and post-experiment surveys. Findings suggest that crowding is an important environmental factor affecting tourists’ visual attention patterns. Moreover, natural sounds enhance tourists’ visual attention to natural landscapes and mitigate tourists’ stress simultaneously. Mask-wearing can reduce tourists’ visual attention to human crowds but cannot reduce stress intensity directly. Our findings extend the attention restoration theory by a multi-sensory perspective and the transactional theory of stress through eye-tracking analytics.

Introduction

Tourists’ visual attention is critical for destination managers because the intangible travel experiences, the preferences for tourism activities and services, and the tourist-environment relationship can be manifested in tourists’ visual attention and visual processing information (Scott et al., 2019; Wang & Sparks, 2016). Visual attention is a selectively cognitive process that elicits concentrating visual scenes and information (Jonides, 1983). Visual attention to physical environments and urbanization lifestyles may link to attention fatigue (Ohly et al., 2016). Nature-based tourism destinations, however, serve as ideal environment settings for attention restoration (Qiu, Jin, & Scott, 2021; Qiu, Sha, & Scott, 2021) because natural environments have the capacity to draw attention without fully occupying mindsets (Kaplan, 1995; Ohly et al., 2016). Although visual attention has a central function in constructing tourists’ visual experiences that have effects on their perceptual and cognitive processes (i.e., attention restorative capacity, cognitive appraisal of stress), limited studies have examined tourists’ visual attention patterns in natural environments and how visual attention might influence tourists’ stress.

Traveling and recreation activities provoke and mediate stress (Jordan & Vogt, 2017; Jordan et al., 2019). Positive travel experiences (e.g., visiting natural or cultural scenery and participating in desired activities with satisfactory experiences) can reduce tourists’ stress (Jordan et al., 2019). On the contrary, negative or overwhelming travel experiences may provoke tourists’ stress (Zhu et al., 2020). Research has found that environmental factors are antecedent factors in stress-related mechanisms (Berto, 2014; Schuster et al., 2006). For example, crowding has been identified as one of the most commonly experienced stressors during traveling (Schuster et al., 2006; Xiao et al., 2021). Tourists’ cognitive appraisals of environmental conditions involve complicated processes (Jordan et al., 2019; Schuster et al., 2006), and visual attention to environmental conditions may influence the intensity of stress and available coping strategies for the stressful conditions (Qiu, Jin, & Scott, 2021). However, research on visual attention to tourism-related environments is in a notably nascent stage (Wang & Sparks, 2016). During the COVID-19 pandemic, increasing demands for nature-based tourism have intensified the virus transmission risks (Hamidi & Zandiatashbar, 2021; Shoari et al., 2020; Venter et al., 2020). Increasing levels of human crowds, inappropriate social distance, or inadequate implementation of preventive strategies may elicit more visual attention from tourists, which changes the cognitive appraisal processes of tourists in nature-based tourism destinations (Xiao et al., 2021). Understanding how tourists’ visual attention and stress intensity respond to different levels of crowding stimuli can provide guidance for nature-based tourism destinations to fascinate the attention restorative capacity and enhance tourists’ quality of life as a long-term outcome.

Although a great deal of research has documented that natural environments can provide stimuli for attention restoration based on attention restoration theory (Basu et al., 2019; Kaplan, 1995; Ohly et al., 2016), the relationship between natural environments’ attentional restorative capacity and stress mediation is still a vague picture. Natural environments provide multi-dimensional sensory stimuli for tourists, including visualscape, soundscape, smellscape, etc., and these sensescapes provide multisensory connections between tourists and natural environments and might collectively influence tourists’ attention restoration (Buzova et al., 2021; Qiu, Jin, & Scott, 2021). During the COVID-19 pandemic, social-distancing preventive measures (e.g., stay-at-home order and strict domestic and international travel policies) limited individuals’ mobility for long-distance travel, making proximate natural tourism destinations hotspots for attention restoration (Hamidi & Zandiatashbar, 2021; Shoari et al., 2020; Venter et al., 2020). However, studies on the factors impacting tourists’ visual attention and stress intensity in nature-based tourism destinations from the multisensory perspective are very limited in the current literature (Qiu, Jin, & Scott, 2021).

To address these research gaps, this study aims to investigate three overarching research questions:

• Research Question 1: What are the patterns of tourists’ visual attention under varying environmental factors (e.g., crowding stimuli, preventive measures, natural sounds) in nature-based tourism destinations during the COVID-19 pandemic?

• Research Question 2: How will environmental factors (e.g., crowding stimuli, COVID-19 preventive measures, natural sounds) influence tourists’ visual attention and stress intensity in nature-based tourism destinations during the COVID-19 pandemic?

Findings from this study have important theoretical and practical implications for tourism researchers, practitioners, and stakeholders. Theoretically, this study seeks to extend the attention restoration theory and the transactional theory of stress by examining tourists’ visual attention to different types of environmental factors in nature-based tourism destinations. Understanding how tourists’ visual attention changes among different types of environmental factors and how the stress intensities are related will expand the knowledge about visual processing and emotional responses toward various tourism environments, in particular under the challenges of the COVID-19 pandemic. For practitioners and destination managers, understanding how tourists’ stress is related to visual attention and multiple types of sensory stimuli can inform strategies to manage social density conditions, optimize landscape design, and mitigate negative psychological outcomes for nature-based tourism destinations under the impacts of the COVID-19 pandemic.

Literature Review

Visual Attention Research in Tourism

Extensive research efforts on visual attention have been focused on advertisements (Pfiffelmann et al., 2020; Simola et al., 2020), packaging (Pieters & Wedel, 2004), and websites (Pan et al., 2013). The eye-tracking approach is the primary methodology used to measure visual attention to advertisements and web-based information. Eye-tracking devices can capture the real-time data of eye movement (Scott et al., 2019). Common indicators of the eye-tracking approach include fixation duration (the length of time a person gazes on a specific area of interest), fixation count (total number of fixations on a specific area of interest), and dwell time (the length of time a person spends looking at one image) (Li et al., 2020; Scott et al., 2019). Although previous eye-tracking studies apply artificial visual stimuli designed (e.g., advertisement, semiotic, and text) to experimental images, the real-world materials (e.g., webpages, advertisement without alternations) are increasingly used in recent eye-tracking studies to better present the visual information in the appropriate context (Li et al., 2016).

Research on visual attention in tourism context is still at a nascent stage (Scott et al., 2019). Three primary bodies of research have developed over the last several decades, including (1) tourists’ visual attention to tourism destination images; (2) perceived effectiveness of advertisements and destination images on travel or consuming intentions on tourism products; (3) tourists’ visual attention to different types of environment settings. Tourists’ visual attention was predominantly measured by eye-tracking methodologies, where eye-tracking experiments were primarily conducted in indoor/lab settings (Scott et al., 2019). Considerable visual attention research efforts have been focused on tourism destination images. Research findings suggest that spatial location and modality can influence tourists’ visual attention to destination images (Zhao et al., 2022). Visual attention to tourism destination images may vary among cross-cultural contexts; that is, the focal objectives, focal speed, fixation duration, and fixation counts of tourism destination images are different among participants with various ethnical/cultural backgrounds (Chua et al., 2005; Wang & Sparks, 2016).

The second pipeline of research focuses on assessing the relationship between tourists’ visual attention and the perceived effectiveness of advertisements and branding interests. The dual-process theory is one of the most widely used theories in this pipeline of research. The dual-process theory suggests that human responses to stimuli are dependent on habituations and sensitizations (Groves & Thompson, 1970). Moreover, stimuli with intense and distinct features are more likely to generate higher attention (Li et al., 2020). Multiple types of stimuli have been integrated into tourism destination images in this line of research (Li et al., 2016; Pan et al., 2013; Scott et al., 2016; Zhao et al., 2022), including texts embedded in the scenes (Li et al., 2016, 2017), country brand logo and slogan (Lourenção et al., 2020), high-arousal activities (Wang & Sparks, 2016), and cartoon effects (Li et al., 2020). These empirical studies suggest that text embedded in natural scenes, images with high arousal activities, and semiotic signs have higher perceived effectiveness on tourism advertisements (Li et al., 2016; Lourenção et al., 2020; Wang & Sparks, 2016).

The third pipeline of research primarily examines tourists’ visual attention patterns across different environment settings. The cognitive evaluation of the environment settings is often measured as a complementary process to visual attention data to examine how visual attention may influence tourists’ perception of landscape, emotions, and preferences for the environment settings. For example, Wang and Sparks (2016) compared tourists’ visual attention to tourism destination images of natural scenes and built scenes, and found that tourists have higher attention to natural scene images and prefer natural environments more than built environments. The perceived complexity of natural scenes can draw more visual attention than built scenes, yielding more positive perceptions of images with natural scenes (van Den Berg et al., 2016). However, this line of research in tourism is still in a notably nascent stage. Tourism environment has multiple dimensions of elements (e.g., natural scenes vs. cultural scenes, crowded vs. un-crowded, quiet vs. noisy). Yet, previous studies have not examined how visual attention varies among these elements and whether visual attention might affect tourists’ perceptions and stress.

Attention Restoration Theory

Attention restoration theory (ART) examines the role of the natural environment in restoring diminished or impaired attentional capacity (Kaplan, 1995). ART proposes that individuals may have opportunities to reduce attention fatigue in natural environments, including (1) “being away” from daily stress, (2) exposure to adequate space and expansive context, (3) engaging in activities compatible with intrinsic motivation; and (4) experiencing stimuli that are “softly fascinating” (Kaplan, 1995). Although various settings and activities can facilitate attention restoration, opportunities are not equally restorative among different settings and activities (Kaplan, 1995). For example, the restorativeness of typical “escape” relaxing modes (television watching, social media interactions, etc.) is different from the “soft fascination” by experiencing natural environments since the escape relaxing mode forcefully grasps one’s attention and leaves little capacity for mental activity and reflection (Basu et al., 2019). Natural environments, by contrast, capture attention loosely and leave room for contemplation (Basu et al., 2019). ART is primarily used to explore human-nature interactions (Ohly et al., 2016). The natural environment facilitates higher restorative attention than the built environment (Stack & Shultis, 2013). A few studies have used ART to examine the impacts of natural environments on stress and found that visiting natural environments, urban green spaces, and refuges can reduce an individual’s stress intensity (Grahn & Stigsdotter, 2010; Tyrväinen et al., 2014; van Den Berg et al., 2007).

In the tourism research field, ART is primarily applied to identify scales and factors associated with destination fascination (Gill et al., 2019; Liu et al., 2017) or restorative qualities (Lehto, 2013). However, limited efforts have been made on the effects of environmental factors on tourists’ attention, in particular changes in tourists’ visual attention patterns under varying environmental factors. Crowding, as one of the most important environmental factors in tourism destinations, may lead to substantial changes in tourists’ visual attention patterns. Gallup et al. (2012)’s study indicates that pedestrians’ visual attention direction was significantly associated with the size of human crowds, and increasing density of pedestrians can lead to longer visual attention to human crowds. Similarly, individuals’ visual attention directions might be determined by the rule of majority, therefore, the effects of human crowds on visual attention directions are more salient when the sizes of human crowds are larger (Sun et al., 2017).

Building on the ART and the impacts of human crowds on visual attention patterns, the following hypotheses are proposed:

• H₁. Crowded tourism images elicit more attention on people and less attention on natural landscapes.

• H_1a. Crowded tourism images elicit longer fixation durations on people and shorter fixation durations on natural landscapes than un-crowded tourism images.

• H_1b. Crowded tourism images elicit higher fixation counts on people and lower fixation counts on natural landscapes than un-crowded tourism images.

Although ART highlights the attentional restorativeness potential of the natural environment, the elements and attributes that natural environments can provide as “soft fascination” stimuli for attention restoration are vague in the literature (Basu et al., 2019). Rather than just providing single-dimensional stimuli from visual landscapes, natural environments are often multisensory and can provide multi-dimensional stimuli, including soundscapes, smellscapes, tastescapes, and hapticscapes (Buzova et al., 2021; Dann & Jacobsen, 2003; Qiu, Jin, & Scott, 2021). The multisensory features of natural environments can affect the perceived restorative quality collectively (Lu et al., 2022). For instance, a recent study found that natural sounds (e.g., sounds of birds, rain, ocean, etc.) can directly affect the perception of visualscape characteristics and play as a mediating factor between visualscape and perceived restorative environments (Qiu, Jin, & Scott, 2021), which collectively affect the visitors’ quality of life.

As an important stimulus in natural environments, natural sound may affect individuals’ visual attention patterns. For instance, individuals’ visual attention to the images of traditional villages with natural sounds embedded was focused more on the natural landscapes compared to the images without sound effects (Ren et al., 2017). Also, the soundscape may influence the visual attention areas of forest landscapes, that the images with natural sounds embedded can expand the visual attention areas on woody plants, while the images with human and traffic sounds expand the visual attention areas on artificial buildings (Liu et al., 2020). Yet, no empirical studies have examined the effects of natural sounds on tourists’ visual attention patterns based on our knowledge.

Guided by the perspective of multi-sensory stimuli of natural environments, the following hypotheses are proposed:

• H₂. Natural sounds embedded images elicit more attention on natural landscapes and less attention on people than normal images.

• H_2a. Natural sounds embedded images elicit longer fixation durations on natural landscapes and shorter fixation durations on people than normal images.

• H_2b. Natural sounds embedded images elicit higher fixation counts on natural landscapes and lower fixation counts on people than normal images.

A few studies suggest that natural sounds in tourism destinations can facilitate tourists’ flow experience, where time distortion (time passes in an enjoyable way rather than the normal feeling) and concentration in landscape viewing could be found (Lu et al., 2022; Zhuang et al., 2020). Based on previous studies on the effects of soundscape on the tourists’ perceptions and experiences, the following hypothesis is proposed:

• H₃. Natural sounds embedded images elicit a longer dwell time than normal images.

The Transactional Theory of Stress

Stress refers to the perception of a threatening or overwhelming person-environment relationship (Lazarus & Folkman, 1984). Tourists’ perception of stress is not a one-step evaluation process; rather, a transactional process of cognitive appraisal of the person-environment relationship to define stressors and identify coping responses (Jordan et al., 2015; Schneider & Wilhelm Stanis, 2007; Schuster et al., 2006). Specifically, when the person-environment relationship is perceived as taxing or overwhelming for an individual, a transactional process of stress response occurs to achieve equilibrium (Schneider & Wilhelm Stanis, 2007). The transactional theory emphasizes the impacts of personal and environmental factors on the appraisals of conditions. For instance, encountering too many tourists during the COVID-19 pandemic might be appraised as a “crowded” and “vulnerable” health environment, which impairs tourists’ experience and enjoyment in tourism destinations (Xiao et al., 2021). As a result, stress is experienced since tourists’ goals are interfered with by others’ behaviors (Schneider & Hammitt, 1995).

Considerable research efforts of the transactional theory have been focused on the cognitive appraisal of stress (Jordan & Vogt, 2017). The cognitive appraisal process identifies whether a person-environment relationship is stressful or not. This person-environment relationship may be determined by environmental factors, such as crowding stimuli and the quality of landscape design. As indicated in the Equilibrium Model of Human Response to Crowding (Stokols, 1972), tourists exposed to crowding scenes or undesirable environments may lead to cognitive inconsistency, where stress may occur as the result of cognitive appraisal. On the contrary, exposing to natural environments may influence the cognitive appraisal process and mitigate tourists’ stress intensity. Based on the transactional theory of stress, we propose the following hypotheses about the impacts of environmental factors on tourists’ stress:

• H₄. Visual attention to natural landscapes can significantly reduce tourists’ stress intensity.

• H₅. Natural sounds embedded in landscapes can significantly reduce tourists’ stress intensity.

In the context of the COVID-19 pandemic, the factors affecting cognitive appraisal of stress might be more complicated, including increased uncertainties of health risks related to other tourists’ behaviors and implementation of preventive strategies (Zheng et al., 2021). Preventive strategies might mitigate tourists’ stress or negative emotions (Xiao et al., 2021). Mask-wearing, as an additional object of human crowds, might enhance tourists’ visual attention on masks and reduce visual attention to human crowds based on the bottleneck theory of attention, which suggests that individuals pay attention to one object of interest and filter out other objects (Friedenberg et al., 2021). Therefore, we propose the following hypotheses:

• H₆. Images with the preventive measure (mask-wearing) elicit less attention on human crowds than normal images.

• H_6a. Images with the preventive measure (mask-wearing) elicit shorter fixation durations on human crowds than normal images.

• H_6b. Images with the preventive measure (mask-wearing) elicit lower fixation counts on human crowds than normal images.

• H_6c. Images with the preventive measure (mask-wearing) elicit longer dwell time than normal images.

• H₇. COVID-19 preventive measures can significantly reduce tourists’ stress intensity.

Conceptual Framework

Based on the transactional theory of stress and preceding theories related to antecedents of stress (the dual-process theory and ART) from the perspective of visual attention, we propose the conceptual framework of this research (Figure 1). The overall hypothesis of this study is that tourists’ visual attentions vary among images displaying different crowding scenes, and environmental factors can influence tourists’ visual attention, which collectively activates the cognitive appraisal of stress and influences the stress intensity. As outlined in earlier studies, the person-environment relationships may be associated with multiple environmental factors (e.g., density of use, landscape, natural sounds, etc.). Tourists’ visual attention to environmental factors is influenced by the social norms (characteristics of habituation) and the intensity of stimuli (sensitization), which may affect attention restorativeness capacity and ultimately influence the stress intensity. Specifically, attention to human crowds may yield cognitive inconsistency, which may intensify the stress. On the contrary, attention to natural environments (including natural landscape and natural sounds) can restore impaired attentional capacity and protect tourists against environmental stressors (Basu et al., 2019; Kaplan & Berman, 2010).

Figure 1.

Conceptual framework of environmental stimuli, visual attention, attention restoration, and stress intensity.

Methods

Study Site Selection and Description

To investigate tourists’ visual attention and stress intensity during the COVID-19 pandemic, our study selects the Leiqiong UNESCO Global Geological Park as the study site. The park is one of the most highly visited nature-based tourism destinations in Haikou, China, and the annual visit was nearly 400,000 before the outbreak of the COVID-19 pandemic. The park was reopened in April 2020 after the COVID-19 pandemic was well-controlled in Haikou. The volcano cave and landscape are the primary attractions of the park. The bottom of the volcano cave often faces the challenges of concentrated tourist flow, which might affect tourists’ view of the natural landscape and serve as an ideal location to investigate the relationship between visual attention and cognitive appraisal of stress during the COVID-19 pandemic. Therefore, our study selects the volcano cave of Leiqiong UNESCO Global Geological Park as the study location.

Data Collection

To answer the study research questions, we utilize an experimental design tracking the eye movements of research participants by showing them pictures of a natural area with differing conditions. Three phases of data collection were conducted to fulfill the study design, including real-time observation (Phase 1), eye-tracking experiments (Phase 2), and self-reported surveys (Phase 3). The following paragraphs contain full details about experimental design and data analysis.

Phase 1: Real-time observation of density of use

First, we collected data in real-time observations of the study site by taking pictures every 5 minutes and recording people per view (PPV) before and during the COVID-19 pandemic. The rationale for conducting the real-time observations is to elicit images representing realistic scenarios of the density of use at the study location. In this approach, the images can visually present the density of use before and during the COVID-19 pandemic. Moreover, research has indicated that real-world images can better represent tourism scenarios than artificial materials and have been increasingly used in eye-tracking studies in tourism (Li et al., 2016). We conducted the real-time observations on 12/29/2019 and 06/21/2020, respectively. Through this approach, the different levels of density of use can be represented by the indicator of PPV. PPV is a widely-used indicator for the density of use in tourism destinations (Manning, 2011). We selected six images representing different levels (6-, 12-, 18-, 24-, 30-, 36-PPV) of the density of use in the eye-tracking study.

Phase 2: Eye-tracking experiment design and procedure

Experiment design: The experiment design involves tracking the eye movements of research participants by showing photographs of the study site with varying conditions. We used the within-subjects design for the eye-tracking experiments because this approach is appropriate to measure changes in visual attention patterns of all participants under varying treatments (i.e., environmental factors) (Charness et al., 2012). The full design was 6 (PPV densities) × 2 (mask or non-mask) × 2 (natural sound or no sound) design. In the first experiment, the six-image with different PPVs without masks (normal version) were displayed to participants in random order. A follow-up group of the same images when tourists are wearing masks (environmental factor 1) were displayed to participants. In the second experiment, the image of 6-PPV (low density) without natural sound (normal version) was first displayed, and the same image with natural sound (environmental factor 2) was displayed to participants. In addition, the images of 6-PPV (low density of use) and 18-PPV (medium density of use) without natural sounds (normal version) and with natural sounds (environmental factor 2) were displayed to participants in random order. Three indicators of visual attention will be captured by the eye-tracking devices during these two groups of experiments: fixation duration, fixation count, and dwell time. These indicators are the most widely used indicators in tourism-related visual attention research. By measuring these indicators, the differences in visual attention toward different levels of crowding and the effects of environmental factors (e.g., mask-wearing and natural sound) can be identified by conducting follow-up Chi-square tests.

Participants: Tourists over 18-year-old with normal (or corrected to normal) vision and who have visited Haikou Leiqiong Global Geological Park before were qualified to participate in this study. We distributed recruitment advertisements in several local communities surrounding a public university in Haikou through flyers and posters and recruited 50 participants in our study. More than half of the participants were females (54.9%), and 45.1% were males. The age of the participants ranged from 19 to 57, and more than one-third of the participants have a bachelor’s degree or higher education background. Although small samples are common in eye-tracking studies (Scott et al., 2019), we recruited the participants across different gender, age, and education levels to enhance the representativeness of the samples.

Apparatus: This study used a Pupil Core eye tracker to collect eye movement data. Pupil Core is a mobile eye-tracking headset that can detect, record, and map real-time eye movements with the assistance of Pupil Capture Software (Kassner et al., 2014). Pupil Capture algorithm can achieve a 90% detection rate with an error threshold of 5 pixels. Pupil Capture is considered accurate and capable of collecting information every 4 ms (Kassner et al., 2014).

Procedure: Before participating in the eye-tracking experiment, participants were informed that they would be viewing images representing different levels of density of use at the study site. No specific task was assigned to participants when they were viewing. Rather than viewing images on a computer monitor, our experiment displayed all images on a large screen (144 cm × 81 cm) to better reflect tourists’ experiences in the study location. A calibration process (5-point calibration positions with a 1.5 seconds interval) was conducted before each eye-tracking experiment, and an offset of 5° or lower could indicate successful calibration (Kassner et al., 2014). Each image was displaced for 10 seconds, and images were randomly presented to participants in both two experiments.

Phase 3: Self-reported survey

After the eye-tracking experiments, the participants were asked to rate the perception of crowding and stress intensity for each viewed picture, respectively. The perceived crowding was measured by a 9-point scale, where “1” represents not crowded at all, and “9” represents extremely crowded (Manning, 2011). To make the measurement scale constant between perceived crowding and stress intensity, the 9-point scale was applied to measure the stress intensity by a separate question (“what is your stress intensity when you are viewing this image?”), where “1” represented “not stressful at all,” and “9” represented “extremely stressful.”

Data Analysis

Eye movement data analysis

The fixation count, fixation duration, and dwell time of the eye-movement data were used for eye-tracking data analysis. A typical participant’s raw eye movement data viewing all images (140 seconds) contains more than 20,000 gazing points. We set the threshold of the fixation as 100 ms, which aligns with the threshold of clinical visual cognitive research (Manor & Gordon, 2003). The fixation durations of human crowds viewing and landscape viewing were calculated by defining the area of interests (AOIs) presenting people and natural landscape, respectively. Rather than traditional eye-tracking studies that directly compare the fixation durations among different images, we calculated the fixation durations of different objects by the ratio of total viewing time on the targeted object dividing the total viewing time on all objects. This approach could exclude fixation durations of saccades (rapid eye movements that abruptly change the point of fixation) and fixation durations out of the AOIs (Li et al., 2016). We calculated the average fixation duration ratios and fixation count ratios of people viewing and landscape viewing and the dwell time for each viewed image (14 images in total) from 42 valid participants’ eye movement data to identify visual attention patterns under varying PPVs and environmental factors. The fixation duration ratios ranged from 0% to 100%. The dwell time was calculated as the sum of the fixation duration on each image, ranging from 0.70 to 8.66 seconds.

The heatmap of each viewed image was generalized based on the fixation duration and the normalization coordinates (X, Y) of the eye-movement data using R programming 3.6.4. Unlike the traditional eye-tracking studies using the eye-tracking software embedded plug-in to generate the heatmap for the single participant’s eye movement heatmap, our study generated the heatmaps by grouping all 42 valid participants’ eye-movement data to represent the aggregated heatmap of the participants’ visual attention in general. In this approach, the heatmaps can be more representative of visual attention patterns for the whole group of participants.

Visual attention and survey data analysis

Combining the data of fixation duration ratios, fixation counts, dwell time, and the survey data of perceived crowding and stress intensity, the effects of visual attention on perceived crowding and stress intensity can be further analyzed by statistical tests. We used paired-sample t-tests to compare the perceived crowding and stress intensity by environmental factors (mask-wearing vs. non-mask wearing, natural sound vs. no sound). We also employed the mediation analysis to identify the mediation effects between environmental factors and stress intensity through visual attention. Finally, we used the multiple regression model to examine the effects of visual attention indicators, perceived crowding, and social demographic variables on stress intensity.

Results

Visual Attention Toward Tourism Images

Fixation durations

Crowding stimuli

The fixation duration ratios of human crowds (FDRHC) and fixation duration ratios of natural landscapes (FDRNL) by different PPVs are depicted in Figure 2. In general, an increase in density of use yielded a higher FDRHC, whereas a lower FDRNL. Specifically, at the low crowding level (PPV = 6), the average FDRHC and FDRNL were 52.3% and 46.2%, respectively. The average FDRHC at medium crowding level (PPV = 18) was 78.2%, significantly higher than FDRHC at a low crowding level. Similarly, at the high crowding level (PPV = 36), the FDRHC increased to 87.0%, significantly higher than the FDRHC at the low crowding level. Under the high crowding level, only 13% of fixation duration was viewing natural landscape, implying that a high level of crowding could yield a high proportion gazing duration on human crowds and low proportion gazing duration on the natural landscape, supporting hypotheses H_1a.

Figure 2.

Fixation duration ratios under varying levels of crowding stimuli when tourists do not wear masks.

Mask-wearing and non-mask-wearing comparisons

The fixation durations for the same group of images when tourists wore masks were calculated for three objects: people, natural landscape, and mask (Figure 3). Compared to the images when all tourists did not wear masks, the FDRHCs were significantly lower under a few conditions at the 95% confidence level. Specifically, in the medium crowding conditions (PPV = 12 and PPV = 18), the FDRHCs were 52.6% and 54.1% when tourists were wearing masks, which were significantly lower than 62.1% (p < .001) and 74.7% (p < .001) when tourists were not wearing masks, respectively. In the high crowding conditions (PPVs = 30 and 36), the FDRHCs when all tourists wore masks were significantly lower than the FDRHCs when all tourists did not wear masks (p < .001 for both tests). Notably, in extremely crowded condition (PPV=36), the FDRHC was 65.4% compared to 87.0% when tourists did not wear masks, indicating the visual attention on human crowds was significantly reduced when COVID-19 preventive measures were implemented.

Figure 3.

Fixation duration ratios under varying levels of crowding stimuli when tourists wear masks.

The FDRNL for the mask-wearing images did not heavily change compared to non-mask-wearing images. The FDRHCs have significantly increased from the low crowding condition to the medium crowding scenario (p < .001). However, no significant differences in mask viewing fixation duration ratios were found between medium and high crowding conditions. These results imply that mask-wearing, as a preventive measure, could draw tourists’ visual attention under low and medium crowding conditions, but the visual attention to masks did not change heavily in high crowding conditions.

Natural sound effect

The FDRHCs and FDRNLs between normal images and natural sound embedded images were compared in the low and medium crowding conditions (Figure 4). Notably, natural sounds play a significant impact in reducing the FDRHC and increasing the FDRNL. With the embedded natural sounds under the low crowding condition, the FDRHC reduced from 49.2% to 33.9% (p = .020). Similarly, the embedded natural sounds under the medium crowding condition reduced the FDRHC, that 57.2% of fixation duration was targeted on people, compared to 74.7% in the same image without natural sounds (p = .001). These results supported the research hypothesis H_2a that natural sounds can reduce the FDRHC and increase the FDRNL.

Figure 4.

Fixation duration ratios of images with/without natural sound embedded.

Fixation counts

Crowding stimuli

Similar to fixation durations, an increase in density of use would yield a higher FCRHC and a lower FCRNL (Figure 5). Specifically, the average FCRHC (76.0%) at medium crowding level was significantly higher than the FCRHC (49.7%) at low crowding level (p < .001). At the high crowding level (PPV = 36), the FCRHC increased to 86.1%, which was significantly higher than the FCRHC at the medium crowding level (p = .004). These results suggest that a higher level of crowding could yield a high FCRHC and a low FCRNL, supporting hypothesis H_1b.

Figure 5.

Fixation count ratios under varying levels of crowding stimuli when tourists do not wear masks.

Mask-wearing and non-mask-wearing comparisons

Under the mask-wearing conditions, the FCRHCs were generally lower than the FCRHCs under the non-mask-wearing conditions (Figure 6). These differences were manifested in the medium to high crowding conditions. Specifically, in the medium crowding conditions (PPV = 12 and PPV = 18), the FCRHCs were 54.8% and 56.9% when tourists wore masks, which were significantly lower than 64.0% (p = .011) and 75.6% (p < .001) when tourists did not wear masks, respectively. In the high crowding conditions (PPVs = 30 and 36), the FCRHCs when all tourists wore masks were significantly lower than the FCRHCs when all tourists did not wear masks (p < .001 for both tests). Notably, in the extremely crowded condition (PPV=36), the FCRHCs was 68.2% compared to 86.2% when tourists did not wear masks, indicating the fixation count on human crowds was significantly reduced when COVID-19 preventive measures were implemented.

Figure 6.

Fixation count ratios under varying levels of crowding stimuli when tourists wear masks.

Similar to fixation duration, fixation count ratios on masks increase with the number of tourists at low and medium crowding conditions but stay relatively steady from medium to high crowding conditions. These results indicated that in the high crowding conditions, the masking-wearing preventive measure did not draw more attention than in medium crowding conditions.

Natural sound effect

Natural sounds significantly affect the fixation counts on the natural landscapes (Figure 7). With the embedded natural sounds in the low crowding condition, the FCRHC reduced from 48.5% to 31.7% (p < .001). Similarly, the embedded natural sounds in the medium crowding condition reduced the FCRHCs, that 60.0% of fixation count was targeted on people, compared to 76.0% in the same image without natural sounds (p = .001). These results supported the research hypothesis H_2b that natural sounds can reduce the fixation count on people and increase the fixation count on natural landscapes.

Figure 7.

Fixation count ratios of images with/without natural sound embedded.

Dwell time

We compared the dwelling time on images representing different crowding conditions as well as non-mask-wearing and masking-wearing conditions. The results indicated that the dwell time on the images of different PPVs has no significant differences. In general, the dwell time on the images of (6-, 12-, 18-, 24-, 30-, and 36-PPV) ranged from 4.97 to 5.38 seconds. Moreover, when comparing the dwell time on the images of mask-wearing and non-mask-wearing conditions, no significant difference was found for the six pairs of images representing different PPVs. Results reject the research hypothesis H_6c that mask-wearing preventive measures can increase the dwell time on images.

In terms of dwell time on images with and without natural sound embedded, the paired t-tests indicate that the dwell times on images with natural sounds were significantly higher than those without natural sound. Specifically, under the low crowding condition (6-PPV), the average dwell time on the image with natural sounds was 6.63 seconds, which was significantly longer than the dwell time (M = 5.07 seconds) on the image without natural sound (p < .001). Similarly, the average dwelling time on the image of medium crowding with natural sounds was 5.64 seconds, significantly longer than the dwell time (M = 4.97 seconds) on the image without natural sounds (p = .040). These results support the research hypothesis H_5c that images with natural sound elicit longer dwell time than normal images.

Heatmaps

We created the heatmaps for different crowding conditions as well as mask-wearing conditions (Figure 8). Overall, natural landscapes draw more attention under low crowding conditions than medium and high crowding conditions. By comparing the heatmaps between mask-wearing and non-mask-wearing conditions, we found that under the mark-wearing conditions, visual attention to human crowds was more scattered than the non-mask-wearing conditions. Moreover, under the mask-wearing conditions, visual attention to natural landscapes is more concentrated than the non-mask-wearing conditions, particularly when the crowding levels are low and medium. These results imply that mask-wearing can potentially disperse tourists’ visual attention to human crowds as well as concentrate visual attention on natural landscapes.

Figure 8.

Heatmaps of fixation duration by crowding stimuli with/without masks.

Effects of Visual Attention on Stress Intensity

Participants were also asked to report the perception of crowding and stress for all images. By comparing the mean values of perceived crowding and stress intensity by varying PPVs, we found that the perceived crowding and stress levels were significantly different under different PPV scenarios (Table 1). Specifically, the perceived crowding and stress levels increased from 2.81 and 2.62 under the 6-PPV scenario to 9.00 and 8.93 under the 36-PPV scenario (p < .001).

Table 1.

Chi-square Tests for Visual Attention Indicators, Crowding, and Stress Under Different Stimuli.

Variables	6-PPV	12-PPV	18-PPV	24-PPV	30-PPV	36-PPV	p-Value
Fixation duration ratio of people viewing	49.73%a	63.98%b	76.44%c,d	77.71%c,d	87.57%d,e	87.57%d,e	<.001
Fixation count ratio of people viewing	48.54%a	62.97%b	75.69%c,d	77.35%c,d	87.35%d,e	86.17%d,e	<.001
Fixation duration ratio of landscape viewing	50.27%a	36.02%b	23.56%c,d	22.29%c,d	12.43%d,e	12.43%d,e	<.001
Fixation count ratio of landscape viewing	51.46%a	37.03%b	24.31%c,d	22.65%c,d	12.65%d,e	13.83%d,e	<.001
Crowding	2.81a	5.40b	6.98c	8.43d	8.93e	9.00e	<.001
Stress	2.62a	5.31b	7.02c	8.51d	8.90e	8.93e	<.001

Note. Superscript ^a,b,c,d,e indicates statistically significant differences among groups at the p < .05 level based on post hoc comparisons using Duncan’s post hoc tests.

We also compared the perceived crowding and stress intensity under non-mask-wearing and mask-wearing conditions. No significant differences were found among all tested scenarios except the scenario of 12-PPV (Table 2). In terms of natural sounds’ effects on perceived crowding and stress levels, the ANOVA tests indicated that natural sounds significantly reduce perceived crowding and stress levels under both 6-PPV and 18-PPV conditions (p < .001 for both two tests).

Table 2.

Paired Sample T-Tests for Crowding and Stress Level by Environmental Stimuli.

Factors*	Crowding		Stress
Comparison groups	Base group	Treatment	Base group	Treatment
6-PPV × mask	2.81	2.52	2.62	2.48
12PPV × mask	5.40	4.76**	5.31	4.74**
18-PPV × mask	6.98	6.62	7.02	6.62
24-PPV × mask	8.43	8.64	8.43	8.60
30-PPV × mask	8.93	8.88	8.62	8.52
36-PPV × mask	9.00	9.00	8.90	8.70
6-PPV × natural sound	2.81	1.76***	2.62	1.64***
18-PPV × natural sound	6.98	4.29***	7.02	3.86**

Note. Asterisks indicate *significant at .05, **significant at .01, and ***significant at .001.

Mediation effects between environmental stimuli and stress intensity

The differences in visual attention patterns and stress intensity under environmental stimuli indicate that mediation effects might exist between (a) mask-wearing and stress intensity and (b) natural sounds and stress intensity. The mediation effects of visual attention (i.e., FDRNL and FCRNL) on stress intensity were analyzed by the Sobel Tests (Table 3). For the mask-wearing stimuli, the FDRNL (p < .001) and FCRNL (p < .001) were significantly affected by mask-wearing, and stress intensity was significantly reduced when FDRNL (p < .001) and FCRNL (p < .001) increases, respectively. However, mask-wearing has no direct impact on stress intensity (p = .062), rejecting H₇. These results suggest that the impact of mask-wearing on stress intensity was primarily driven by the indirect path, and visual attention is a mediator between mask-wearing and stress intensity: when encounters wore masks, tourists were more likely to shift visual attention toward natural landscapes, which significantly reduced tourists’ stress intensity.

Table 3.

Mediation Effects of Visual Attention on Environmental Stimuli and Stress Intensity.

Factors	Unstandardized coefficients	Standardized coefficients	p-value
NS → FDRNL	0.170	0.311	<.001***
FDRNL → SI	−3.356	−0.397	<.001***
NS → SI	−1.499	−0.324	<.001***
NS × FDRNL → SI	−0.255	−0.101	.020*
NS → FCRNL	0.159	0.303	<.001***
FCRNL → SI	−3.618	−0.411	<.001***
NS → SI	−1.494	−0.323	<.001***
NS × FCRNL → SI	−0.255	−0.101	<.001***
MW → FDRNL	0.125	0.249	<.001***
FDRNL → SI	−4.844	−0.479	<.001***
MW → SI	0.472	0.093	.062
MW × FDRNL → SI	−0.606	−0.119	<.001***
MW → FCRNL	0.144	0.234	<.001***
FCRNL → SI	−5.180	−0.497	<.001***
MW → SI	0.457	0.090	.066
MW × FCRNL → SI	−0.746	−0.116	<.001***

Note. NS = natural sounds; FDRNL = fixation duration ratio on natural landscape; FCRNL = fixation count ratio on natural landscape; MW = mask-wearing; SI = stress intensity.

^*, and *** represents significant at .05, and .001 level, respectively.

The mediation analysis suggests that natural sounds affect stress intensity through both direct and indirect paths. When viewing images with natural sounds embedded, tourists’ FDRNL (p < .001) and FCRNL (p < .001) significantly increase. Also, natural sounds and visual attention collectively affect stress intensity. These results suggest that visual attention is a mediator between natural sound and stress intensity. Tourists’ stress intensity can be reduced by listening to natural sounds directly (supporting H₅), as well as increasing the duration or frequency of visual attention to natural landscapes indirectly.

Effects of perceived crowding on stress intensity

A multiple regression model was conducted to identify the impacts of visual attention indicators on stress intensity (Table 4). Results indicate that the FDRNL has a significantly negative impact on stress intensity (p = .009). Moreover, the FCRNL has a significantly negative impact on stress intensity (p = .020). These results support H₄ that visual attention to natural landscapes can significantly reduce tourists’ stress intensity. Perceived crowding has a significantly positive impact on stress intensity. The socio-demographic factors, including gender, age, and education level, had no significant impact on stress intensity.

Table 4.

Multiple Regression Model for Stress.

Variables	Unstandardized coefficient	Standardized coefficient	p-Value
Fixation duration ratio of people viewing	1.510	0.159	.009**
Fixation count ratio of people viewing	1.385	0.144	.020*
Crowding	1.011	0.983	<.001***
Gender	0.089	0.018	.074
Age	−0.050	−0.021	.085
Education	−0.062	−0.022	.092
Constant	−0.064		.765

Note. Asterisks indicate *significant at .05, **significant at .01, and ***significant at .001.

Discussion and Conclusions

Discussions

This study examined tourists’ visual attention and perceived stress from tourism environmental stimuli in a nature-based tourism destination during the COVID-19 pandemic. A mixed-methods of data collection that involve real-time observations, eye-tracking technique, and post-experiment questionnaire survey was conducted with 50 tourists who have visited the study site before. Results indicate that crowding stimuli is an important environmental factor affecting visual attention patterns, manifesting in significant differences in FDRHC and FCRHC under varying levels of crowding stimuli. Specifically, the effects of PPVs on visual attention are more salient under low and medium crowding conditions, where the FDRHC and FCRHC of 18-PPV were significantly higher than the FDRHC and FCRHC of 12-PPV, respectively. However, the effects of PPVs on visual attention are minimal under high crowding conditions, where the FDRHC and FCRHC of 30-PPV and 36-PPV have no significant difference. By testing the effects of visual attention, perceived crowding, and socio-demographic variables on stress intensity under varying crowding stimuli, our study suggests that FDRHC, RCRHC, and perceived crowding collectively affect stress intensity in nature-based tourism destinations. These results align with findings from earlier studies that crowding is an important environmental factor for tourists’ stress, and the stress intensity was evaluated by the transactional processes involving cognitive appraisals of environmental factors (Jordan & Vogt, 2017; Schneider & Hammitt, 1995).

Our study also compared visual attention patterns, perceived crowding, stress intensity, and heatmaps under the non-mask-wearing and mask-wearing conditions. Mask-wearing can reduce FDRHC and FCRHC significantly, in particular under high crowding conditions. However, mask-wearing has no significant impact on stress intensity under most PPV conditions. The mediation analysis indicates that the mask-wearing strategy affects stress intensity primarily through the indirect path, where mask-wearing increases visual attention to natural landscapes, yielding lower levels of stress intensity. Results are aligned with findings from earlier studies that COVID-19 preventive strategies can reduce tourists’ stress (Xiao et al., 2021). However, our study quantifies the indirect effects of mask-wearing on stress intensity through visual attention, which highlights the importance of visual attention on stress intensity in the post-pandemic era.

We also tested the effects of natural sounds on visual attention and stress intensity under low and medium crowding conditions. Images with natural sounds embedded elicit significantly higher FDRNL and FCRNL than images without sound effects. In addition, natural sounds elicit longer visual attention (dwelling time) to tourism images, which is aligned with findings from earlier studies that natural sounds facilitate tourists’ flow experiences by time distortion (immersive experiences) (Lu et al., 2022). Moreover, the mediation analysis indicates that natural sound affects stress intensity through both direct and indirect paths. Natural sound, as one dimension of multi-sensory environment, can facilitate the cognitive appraisal process of stress and mitigate stress intensity. More importantly, natural sounds can influence visual attention patterns by increasing the FDRNL and FCRNL, which reduce the stress intensity indirectly. The coherent effects of natural sounds on stress intensity by direct and indirect paths provide empirical support for the ART, that the “soft fascination” dimension for attention restorativeness of natural environments might be facilitated by the compound effects of multi-sensory stimuli (e.g., visualscape, soundscape, etc.) (Qiu, Jin, & Scott, 2021).

Theoretical Implications

Study findings contribute to the transactional theory of stress from the perspective of visual attention. Although previous studies have suggested that crowding can be an environmental factor that leads to cognitive inconsistency and perceived stress (Schneider & Hammitt, 1995; Stokols, 1972), how the crowding stimuli influence the cognitive appraisal of stress has not been verified from the perspective of visual attention. Our study findings show that the FDRNL and FDRNL are significantly associated with stress intensity. These results highlight the fact that tourists’ cognitive inconsistency under crowding stimuli is not an instant perception; rather, it is a visual information evaluation process of environmental stimuli that challenges the cognitive equilibrium. Our study findings suggest that the effects of visual attention to human crowds on stress intensity are more salient under low and medium crowding conditions, whereas the effects are minimal under high crowding conditions. The minimal effects may be explained by the habituation component of the dual-process theory (Groves & Thompson, 1970), that tourists’ responses to high levels of crowding stimuli are more affected by habituation than sensitizations, yielding minimal effects on mitigating the stress intensity. These results highlight the important role of visual attention in the cognitive appraisal of stress; that is, visual attention to different environmental elements (human crowds vs. natural landscape) can affect the perceived stress levels through the transactional process (Lazarus & Folkman, 1984).

Second, study findings advance the transactional theory of stress by testing the mediation effects of visual attention between environmental factors and stress intensity. Results indicate that stress intensity in nature-based tourism destinations is collectively affected by environmental factors (e.g., COVID preventive strategies and natural sounds) and visual attention patterns. Natural sounds influence stress intensity through both direct path and indirect path of visual attention on natural landscapes. Masking-wearing strategy, however, affects the stress intensity only through the indirect path of visual attention to natural landscapes. Based on ART, higher levels of attention to natural landscapes initiate attention restoration, which can reduce tourists’ stress levels as a positive outcome from natural landscape viewing (Grahn & Stigsdotter, 2010). These findings suggest that attention restoration is an essential antecedent process of cognitive appraisal of stress, and the personal-environment transactions can be partially explained by visual attention patterns under varying environmental factors. Study results also provide theoretical support for the role of natural landscapes in tourists’ stress mitigation and extend the application of the ART to tourists’ transactional stress appraisal process.

This study also reveals the effects of the multi-sensory stimuli of natural environments on tourists’ visual attention and stress levels. Results from this study suggest that the natural sound embedded images draw more attention to the natural landscape and facilitate stress mitigation than normal images. As manifested in higher FDRNLs and FCRNLs, as well as longer dwell time on the natural sounds embedded images under the low and medium crowding conditions, this study highlights natural environments’ attention restorativeness capacity by multisensory stimuli. These findings provide empirical evidence to support the ART. Moreover, mediation analysis suggests that natural sounds embedded in the natural environment can serve as a complementary stimulus of visualscape to initiate “soft fascination” for tourists and reduce stress in nature-based tourism destinations (Basu et al., 2019) collectively. This study’s visual attention and stress appraisal for multi-stimuli tourism images advance the ART and highlight the need for future tourism research to include the multi-scape (sensescape) since tourists’ behaviors are often influenced by multi-types of stimuli in nature-based tourism destinations.

Methodologically, this study employs a mixed-method design and advances analytics to extend the innovations of the eye-tracking research approach. This study is one of the first studies that uses real-time observation images of tourism scenarios in the existing eye-tracking tourism research. The real-time observation image can reflect the realistic tourists’ experiences and can overcome the limitations of earlier tourism eye-tracking studies where the materials, text, and pictures are artificially designed (Li et al., 2016). Second, this study advances the analytic techniques for tourism eye-tracking studies. Rather than generating the heatmaps for the single participant in the earlier studies (Li et al., 2020; Wang & Sparks, 2016), the heatmaps of visual attention in this study present the visual attention fixation of all participants. In this approach, the visual attention heatmap of a specific group can be calculated, which can help tourism destinations optimize environmental and landscape design for the targeted tourist groups. Finally, manipulating the image with COVID-19 preventive measures is an innovative approach to examining the impact of preventive measures on tourists’ attention and stress appraisals. Although tourism destinations have implemented various types of COVID-19 preventive management policies, how tourists view and perceive these strategies is still an incomplete picture in the literature (Xiao et al., 2021). Comparing visual attention patterns between the mask-wearing and non-mask-wearing conditions provide a foundation for subsequent research related to tourists’ stress appraisal and coping strategies under the impacts of public health crisis events.

Management Implications

First, tourists’ visual attention to natural landscapes decreases with the crowding stimuli. Therefore, destination managers should identify the threshold of the density of use and make the appropriate standard of carrying capacity to maintain the attention restorativeness of nature-based tourism destinations and mitigate tourists’ stress. Nature-based parks, green spaces, and refugees have been highly-visited tourism destinations to provide tourists with opportunities for attention restoration and refreshment in short-distance and local trips during the COVID-19 pandemic (Hamidi & Zandiatashbar, 2021; Li et al., 2021). Making appropriate standards of crowding stimuli is essential for tourists to cope with the stress associated with the stay-at-home order by visiting nature-based tourism destinations and minimizing the risks of virus transmission.

Second, our study findings contribute to the body of knowledge on COVID-19 related preventive measures on tourists’ visual attention and stress appraisals. Mask-wearing, as one of the most important COVID preventive measures, can significantly reduce tourists’ visual attention to human crowds under low and medium crowding conditions. Moreover, mask-wearing can reduce tourists’ stress intensity through the mediator of visual attention to natural landscapes. The heatmaps of visual attention indicate that mask-wearing can disperse tourists’ fixation duration on human crowds as well as concentrate fixation duration on the natural landscape. As indicated in the literature, attention to human crowds might lead to attention fatigue (Gallup et al., 2012). Findings from this study confirm the positive effects of COVID-19 preventive measures on tourists’ experience by facilitating concentrated attention on the natural landscape and scattered attention on human crowds in nature-based tourism destinations. The mask-wearing policy can be recommended as an effective way to reduce visual attention to human crowds under the impact of the COVID-19 pandemic.

Findings from this study also highlight the need to strengthen the multi-sensory characters of nature-based tourism destinations. Tourists pay more attention to the natural sound embedded images than the normal images under low and medium crowding conditions, and dwell longer time on the natural sounds embedded images. Tourism destination managers should initiate multi-sensory stimuli, facilities, and programs to collectively attract tourists’ attention and reduce tourists’ stress under crowded conditions. Besides, the natural sound embedded images can provide multiple types of sensescape, which might help attract tourists’ attention to nature-based tourism destinations. Destination managers may promote images with natural sounds embedded to improve the attractiveness of destinations as marketing strategies.

Limitations and Future Research

Despite the contributions of this study to tourists’ visual attention and stress appraisal literature, the study has limitations in the following aspects. First, the laboratory design of eye-tracking experiments may influence tourists’ observation of images. The discrepancies between a laboratory setting and realistic tourism destinations cannot be completely avoided. Second, the scope of this study is focused on nature-based tourism destinations due to their popularity during the COVID-19 pandemic. The visual attention pattern might be different in culture-oriented tourism destinations or indoor tourism destinations. These limitations open several avenues for further research. For instance, future studies can conduct eye-tracking experiments in tourism destinations using mobile eye-tracking devices rather than in the laboratory. It might also be worth comparing the visual attention patterns among different types of tourism destinations and extending this line of mixed-method research to examine the broader connections between visual attention and the stress appraisal process.

Author Biographies

Peizhe Li is a research assistant at the Center for Sustainable Tourism, Arizona State University. His research interests include climate adaptation planning for cultural heritages, tourists’ stress, emotion, and coping. He hopes to continue addressing and exploring environmental and social issues of tourism through research, education, and community engagement.

Dr. Xiao Xiao is an assistant professor in the School of Community Resources and Development at Arizona State University. Her research interests include climate adaptation planning for parks and recreation areas and transportation management in parks and protected areas.

Dr. Evan Jordan is an Assistant Professor in the Department of Health and Wellness Design in the School of Public Health at Indiana University. His research focuses on the impacts of tourism on the physical and mental health of residents of host communities. He is particularly interested in tourism’s impact on stress, emotions, and quality of life and their implications for public health.