Residents’ Perceptions toward Tourism Development: A Case Study from Grand Canyon National Park, USA

Abstract

Although the impacts and challenges of tourism in towns and cities near protected areas have been studied extensively, there is a lack of both data and understanding that limits progress towards generalizable solutions, planning strategies, and guidance for addressing the increasing pressures affecting these communities. This article compares the factors influencing residents’ perceptions and local support of tourism in five gateway communities to Grand Canyon National Park. Importance–performance analysis (IPA) and structural equation modeling (SEM) were used to assess the proposed measurements of perceptions and hypotheses concerning local support and to compare the relationships among selected variables, such as community participation (CP), living environment (LE), trust in tourism institutions (TT), tourism benefits (TB), community satisfaction (CS), and perceived tourism cost (TC). Four groups of factors influenced residents’ perceptions; these were classified into four stages based on their management priority. A gap between the desires of community residents for the development of national parks and community tourism and the current state of development was identified, suggesting that these communities would benefit from management measures to mitigate the impacts of tourism. Through SEM, five factors were verified as drivers of local support for national park tourism development, including community participation, living environment, trust in tourism institutions, tourism benefits, and community satisfaction. Perceived tourism cost was not found to be a significant driver.

1. Introduction

Gateway communities are small towns and cities that provide goods and services to the visitors at or near the entrances and boundaries of national parks and other types of protected areas [1,2]. Throughout western North America, these communities are becoming increasingly popular places to live and visit. As a result, many gateway communities are experiencing a range of pressures and challenges.

Gateway communities may have relatively rich tourism resources and facilities to accommodate visitors, and are key promotors of national park tourism and resource protection. Tourism is an essential tool for engaging local communities around national parks and garnering support for ecological protection but, at the same time, poorly managed tourism can fail to produce the desired social and economic benefits to the surrounding communities and can threaten conservation [3]. Recent research that estimated local economic effects from an iterative set of surveys tracing the flow of tourism money found that tourism accounts for ~40% of local household income and at least half of business growth in the gateway community near South Luangwa National Park, Zambia [1].

Researchers have studied many aspects of gateway communities around national parks. Howe et al. [4] were amongst the first to analyze the characteristics of gateway communities and the opportunities and challenges that gateway communities may face. Beunen et al. [5] explored visitor management in two gateway communities in the Netherlands. They analyzed differences in the application of the gateway concept to national park management between North America and Europe. The main functions of most national parks and reserves in North America are nature protection and recreation. The number of gateway communities is limited. However, there are many entrances and roads to protected areas in Europe, which are used not only by tourists but also by people who live or work locally. Gateways may concentrate tourists and traffic flows and become a means of managing natural areas. Frauman and Banks [6] studied the perceptions of different residents in Watauga—a gateway community in North Carolina—regarding the impacts of tourism development. They found that the area was unique in terms of what people considered to be important and of concern, and they attributed this to the environmental characteristics of the area. Stoker et al. [7] identified a need to better understand planning and development challenges in western gateway communities in the USA.

In a previous study [8], the impacts and spatial characteristics of tourism on gateway communities were identified and discussed. These impacts can be divided into economic, sociocultural, and environmental. The development of tourism can increase the employment of residents and improve living standards, but at the same time change the industrial structure and increase the cost of living. Many studies have investigated tourism’s impacts on local communities near protected areas and, although the impacts of tourism expansion have been documented to some extent through case studies of particular communities, there is a lack of understanding of the exact nature of the processes involved. This limits progress towards the development of generalizable solutions, strategies, and guidance for addressing the increasingly acute pressures affecting these natural-amenity-rich communities. In this study, we investigated the perspectives of residents living within a gateway area in an effort to contribute to the development of general solutions, strategies, and guidance for planning and management. This research builds on the previously identified impacts and issues of common concern [8]. It applies importance–performance analysis—in which residents’ perspectives of various economic, sociocultural, and environmental impacts are examined—in the main gateway communities of Grand Canyon National Park in the USA

To further explore residents’ perceptions, we also used a hypothetical framework to apply structural equation modeling to evaluate the extent of local support for national parks. The model outlines the relationships between tourism support and community participation, living environment, residents’ trust in tourism institutions, tourism profits, tourism costs, and community satisfaction. Residents’ attitudes toward tourism are the result of a comprehensive evaluation of costs and benefits during the tourism development process [9], and community support reflects the positive attitudes of residents toward tourism development. When residents feel that the benefits of tourism will exceed the costs, they will likely have a positive attitude toward tourism; otherwise, they may oppose and resist tourism. In this study, hypotheses were put forward regarding the relationships between community participation (CP), the living environment (LE), trust in tourism institutions (TT), tourism benefits (TB), perceived tourism costs (TC), community satisfaction (CS), and support for tourism (ST).

The factors that affect local support have been studied extensively by tourism researchers. Residents’ attitudes toward the impacts of tourism [10,11], community attributes [12], and perceived benefits and costs of tourism [13,14] may all affect local support for tourism development. In this research, we sought to use residents’ perceptions as a foundation and means to determine the support of residents living within the gateway communities. Lee [15] investigated variables such as community dependence, community participation, interest perception, and perceived cost to evaluate Taiwanese community support for sustainable tourism development. Community dependence and participation were decisive factors affecting the support for sustainable tourism development.

Jurowski et al. [15] proposed that community dependence, economic benefits, tourism resources, and attitudes toward the environment all affect residents’ perceptions of tourism which, in turn, affect support for tourism. Gursoy et al. [16] studied the effects of perceived benefits and costs on community support. Later, this type of model was expanded, and the impacts of tourism were divided into five types, including social benefits, economic benefits, social costs, cultural costs, and cultural benefits. Lee [17] posited that community dependence, community participation, perceived costs, and perceived benefits affect residents’ support for tourism. In addition, social exchange relies on the trust between the two parties of the exchange [18]. Therefore, some scholars have incorporated the trust between community residents and tourism agencies into theoretical models of communities’ support for tourism [19]. Community residents’ satisfaction with community life is also related to their perceptions of tourism, and the community living environment affects their satisfaction with the community; thus, community satisfaction and the community living environment are also factors that require consideration [12].

The impacts of tourism development are borne by the community residents of the tourism destination [20] and, among different stakeholder groups, the residents are often marginalized [21]. Since the 1990s, with increasing interest in “sustainable tourism” and “community tourism”, the goals of sustainable tourism have become inseparable from community participation [22], and community participation is one of the decisive factors affecting residents’ support for tourism [17]. Community participation refers to the degree of community participation during the crucial stages of tourism development, such as planning, development, management, and decision-making. Studying community participation may help local governments to understand the impacts of tourism and develop plans that reduce conflict between tourists and residents [23], reduce the negative impacts of tourism on culture and the environment, and help build more harmonious community groups. Residents of highly visited areas and a mature tourism industry often have high participation and relatively positive perceptions, and most studies indicate that the greater the participation, the higher the levels of satisfaction [24,25].

Residents’ perceptions of the benefits and costs of tourism are important factors affecting support for tourism [26]. Tourism development introduces employment opportunities, raises income, increases the number of local entertainment facilities and opportunities, enriches the cultural value of the community, promotes cultural exchange, and improves the cultural identity and quality of life of residents [27,28]. In general, residents’ perceptions of the positive impacts of tourism and their support for tourism are positively correlated, and perceptions of the tourism economy are mostly positive [12,26,29,30]. Conversely, perceptions of the sociocultural impacts of tourism—such as increased living costs, rising land and housing prices, reduced community security, and environmental pollution—are often negative [12]. Overall, perceptions of the costs associated with tourism may reduce residents’ support for the tourism industry.

Community satisfaction is an effective measure of residents’ perceptions of tourism [31] and is an important component of community development and planning [32]. Community satisfaction affects residents’ attitudes; residents who are satisfied with the status quo of community development may have more positive attitudes toward tourism development, while residents who are dissatisfied may believe that tourism development has negative consequences [12]. There is a direct negative relationship between community satisfaction and tourism development costs [15]. However, there have been few studies of community satisfaction and community support, and no conclusion has been reached [33]. This may be due to the consideration of community satisfaction as a single variable. Community satisfaction consists of several aspects, including trust in tourism institutions [34] and satisfaction with the living environment [32]. Some believe that the higher the community satisfaction, the stronger the support for tourism [12].

Tourism development has both positive and negative effects on the living environment [29], including local transportation [35], scenery [36], environmental quality, and security measures to prevent crime [37]. Despite the risks and realities, tourism development may not only improve the community economy, but also improve the long-term and sustainable construction of the community. Satisfaction with the living environment is a measure of satisfaction with the economic and social functions of the community [34]. A poor living environment may have serious consequences for the overall quality of life of the community, and satisfaction with the living environment is positively correlated with community satisfaction.

Based on the literature above, the hypotheses outlined in

Table 1. The hypotheses included in the structural equation model.

Hypothesis	Content
H1	CP positively impacts TB
H2	CP negatively impacts TC
H3	CP positively impacts CS
H4	LE positively impacts TB
H5	LE negatively impacts TC
H6	LE positively impacts CS
H7	TT positively impacts TB
H8	TT negatively impacts TC
H9	TT positively impacts CS
H10	TB positively impacts ST
H11	TC negatively impacts ST
H12	CS positively impacts TB
H13	CS negatively impacts TC
H14	CS positively impacts ST

Note: CP—community participation, LE—living environment, TT—trust in tourism institutions, TB—tourism benefits, TC—perceived tourism costs, CS—community satisfaction, ST—support for tourism.

were proposed.

Data were collected from a sample of 620 residents in the gateway communities around Grand Canyon National Park, USA. The model was empirically tested via structural equation modeling (SEM). While we only examined the experiences of gateway communities around Grand Canyon National Park, our study likely has relevance for gateway communities elsewhere in western North America, and this merits further study.

2. Materials and Methods

2.1. Study Area

Bordering several famous attractions, the gateway communities of Grand Canyon National Park refer to the main residential areas close to the north and south gates of the national park, which received around 6 million visitors in 2019 [38]. At one time economically depressed, the Grand Canyon region has become a tourist hotspot. Non-permanent residents (i.e., second and third homeowners) account for just under 50% of the property owners, and the population has increased by >30% in the last decade [39]. With distinctive seasons, ample outdoor recreational opportunities, a wide variety of artistic and cultural offerings, and the Colorado River flowing through the national park, the gateway communities are facing tremendous pressure and growth. Many have not been planned for the numbers of visitors now coming. This study aimed to evaluate the perceptions of people living in several gateway communities on both the North and South Rims of Grand Canyon National Park. We chose five of the significant gateway communities listed by the Arizona Trail Association that mainly offer food services and accommodation to visitors of the park [40], including Page, Kanab, Flagstaff, Tusayan, and Williams. These five communities are all within the range of the Grand Circle, which encompasses 10 national parks that are close to one another and are noted for their natural beauty [41]. Grand Canyon National Park also lies within the Grand Circle. Page and Kanab are communities on the North Rim’s gateway, and both are major tourist destinations (Figure 1).

Page was established in 1957 to house workers building the Glen Canyon Dam on the Colorado River, and has a population of 7545, with 2840 households [42]. The main attraction in Page is Lake Powell, but the area hosts a number of hiking, biking, and equestrian trails, as well as recreational experiences in the nearby Grand Canyon National Park. Kanab, just over the border in Utah, has all of the facilities needed for visitors, and was therefore included as a study site. This community is popular with outdoor enthusiasts visiting nearby attractions such as Grand Canyon National Park. The population of Kanab is 4970, with 2093 households [43]. Flagstaff, Tusayan, and Williams are gateway communities for the South Rim. Tusayan is just south of the park entrance and has only 580 residents and 185 households [44]. Flagstaff has become one of the greatest trail towns in the American Southwest, and is also the largest community in this study, with a population of 75,038 [45]. In Flagstaff, we concentrated on the city blocks lying on the route to the national park—specifically Downtown Flagstaff and Flagstaff Townsite. The total population of these districts is 1759, with 839 households [46]. Flagstaff has had a thriving tourism industry since the early 1900s, primarily stemming from its proximity to Grand Canyon National Park and other scenic areas [47]. The development of Tusayan as a Grand Canyon gateway town has captured some of Flagstaff’s overnight tourists. Flagstaff itself also competes with nearby towns for access to the Grand Canyon, several of which (e.g., Tusayan) are focusing on their proximity to the Grand Canyon [48]. Williams is located on the gateway to the South Rim of the Grand Canyon. It is the starting point for the Grand Canyon Railroad and has a population of 3248 people and 1519 households [39]. These communities were considered to be representative of the many gateway communities around Grand Canyon National Park.

2.2. Data Collection

2.2.1. Questionnaire Design

A questionnaire was developed to evaluate the determinants of residents’ attitudes toward tourism development and to discern their sociodemographic characteristics. The options in the questionnaire were structured based on an in-depth review of current issues in the national park communities, as well as a literature review on measuring perceptions of tourism’s impacts [49]. The questionnaire consisted of three parts: The first part was related to the economic, social, cultural, and environmental attributes of local tourism development; it included 25 elements, and collected respondents‘ perceptions of the importance of and their satisfaction with each element. The second part contained questions about the seven dimensions in the community tourism support structure model, including community participation (CP), living environment (LE), trust of tourism institutions (TT), tourism benefits (TB), community satisfaction (CS), perceived tourism cost (TC), and support for tourism (ST). The first two parts used a five-point Likert scale. The third part gathered data on respondents’ demographic characteristics, including gender, age, residential status, and education level.

2.2.2. Survey Procedure

A pre-test was conducted with a small sample of tourism-related staff and students to identify problems such as ambiguous wording and to determine how long the questionnaire would take to administer. The questions were simple, general, and non-sensitive, reducing the probability that participants’ responses would be affected by questionnaire bias. The responses were anonymous, reducing participants’ potential concerns about privacy when answering questions. Data were collected in December 2019, and the questionnaires were administered in a direct face-to-face manner. For studies characterized by limited time and resources, the direct face-to-face approach allows distributors to locate and persuade intercepted respondents to participate, thereby improving contact and cooperation. Furthermore, this approach enables respondents to complete the questionnaire on-site. Consequently, face-to-face surveys tend to have high response rates [50]. Some of the people who were approached declined to complete the questionnaire, and with the questionnaires that were accepted some questions were not answered. This opened the possibility of non-response bias. While steps were taken to minimize the potential for non-response bias, we acknowledge that it may be present.

Seven university students and visiting scholars majoring in forestry and conservation were trained to administer the questionnaire, encourage participation, and collect the survey data. The questionnaires were distributed door-to-door and collected on-site to ensure the objectivity and validity of the data. Based on a review of sampling methods described in the related literature, a systematic sampling method with a sampling interval of 12 was used to select the resident households by their door numbers, where the first household was selected by a draw. These methods are consistent with those of Chiu et al., who applied a systematic sampling method to a study of visitors’ environmentally responsible behavior in ecotourism [51]. The same methods were also adopted by Tosun when collecting data for a perception analysis and SEM modelling [52]. During the surveys, individual respondents from the same household were avoided. Next, the residents were informed about the aims of the study, and their willingness to participate was ascertained.

The minimum sample size was calculated to be 366 with 5% error and a 95% confidence interval. The questionnaires were distributed to a total of 620 residents. In total, 559 questionnaires were validated, including 131 in Page, 180 in Kanab, 21 in Tusayan, 95 in Flagstaff (Townsite and Downtown areas), and 132 in Williams. Incomplete questionnaires were excluded from the statistical analysis. The sample size met the recommendation for a minimum subject-to-item ratio of 5 items per subject in an exploratory factor analysis, but not less than 100 respondents [53].

2.2.3. Data Analysis

The descriptive statistics and importance–performance analysis were performed using Microsoft Excel, PSS version 25.0 and SEM was analyzed using AMOS version 22.0, sourced from IBM New York, NY, USA.

The questionnaire grouped impact factors into environmental, sociocultural, and economic factors, and this formed the basis for evaluating their importance and performance. Importance–performance analysis was conducted by plotting conditions based on the importance of the factors versus how they are perceived to currently exist (i.e., performance). Reliability refers to the consistency and stability of the test results. To measure whether the intrinsic structure of a questionnaire is reasonable, it is usually necessary to analyze the validity of the questionnaire. In this study, Cronbach’s alpha (Cronbach’s α) and the Guttman split-half coefficient were used to test reliability. Cronbach’s α values ≥ 0.7 indicate high reliability. Values between 0.35 and 0.7 indicate general reliability, while values < 0.35 indicate low reliability. A variable is typically considered to be reliable if its split-half coefficient is >0.5. All of the factors included in this study passed the reliability and validity test, with Cronbach’s alpha scores > 0.70 and Guttman split-half coefficients > 0.50 [54].

Based on the sample size (N = 559), the survey results had a 4.14% sampling error with a 95% confidence level. This sample size was adequate for performing the SEM analysis based on studies by Marsh et al. [55] and Westland [56]. The Cronbach’s alpha scores for CS, CP, TC, LE, TT, TB, and ST were 0.910, 0.802, 0.779, 0.849, 0.924, 0.828, and 0.917, respectively. All of the scores for these latent variables exceeded the benchmark of 0.70 [54]. Thus, these scores indicated that the instrument had an acceptable level of internal consistency for items measuring the same construct. SEM analysis was used to estimate parameters using the maximum likelihood method. The direction and significance of the relationships were determined by simultaneously testing all of the hypotheses.

Perceptions are difficult to measure directly, and it is challenging to avoid subjective measurement errors. Structural equations provide an analytic tool for the observation and processing of latent variables that are difficult to directly observe and that could otherwise incorporate unavoidable errors into the model. To this end, SEM was applied to investigate the impacts of different variables on gateway community residents’ perceptions of tourism. SEM specifies how latent variables or hypothetical constructs are assessed in terms of observed variables and represent the validity and reliability of the responses for the latent variables [57]. The details describing the observed variables can be found in Appendix A. The SEM is represented by the following matrix equation:

$$\eta =\beta \eta +\varphi \xi +\lambda$$

(1)

Equation (1) is a structural model in which $\eta$ represents an endogenous latent variable, $\xi$ represents an exogenous latent variable, and $\varphi$ and $\lambda$ connect an endogenous latent variable with an exogenous latent variable via a sum matrix and an error vector, respectively.

$$X={\wedge }_x\xi +\delta$$

(2)

$$Y={\wedge }_Y\eta +\epsilon$$

(3)

Equations (2) and (3) are measurement models, in which Χ represents the observed variables associated with exogenous latent variables $\xi$, Y represents the observed variables associated with endogenous latent variables $\eta$, ${\wedge }_x$ represents a correlation coefficient matrix of exogenous latent variables and their observed variables, and ${\wedge }_Y$ represents a correlation coefficient matrix of endogenous latent variables and their observed variables. Latent variables can be represented by observed variables via the measurement models. According to previous studies [11,30,31,56], the SEM model used in this article is proposed as shown in Figure 2.

In Figure 2, the seven ellipses containing the terms TB, LE, TT, CP, TC, CS, and ST represent the hidden variables, the 32 squares represent the observed variables, and the 35 small circles represent the residual variables.

AMOS 22.0 software was used to fit the model. According to the standard SEM application, the model should be tested to assess whether it produces violating estimates prior to assessing the overall fit of the model based on each modified model [58]. This means that the requirement for a normal data distribution must be met. If the absolute value of a variable’s skewness coefficient is >3 and its kurtosis coefficient is >10 (the stricter standard is 8), the data may not be normally distributed. SPSS 25.0 was used to validate the data.

The absolute values of the skewness coefficients of all variables in this study were <1, and the absolute values of the kurtosis coefficients were <2; therefore, the data conformed to the assumption of normality. Consequently, the default maximum likelihood estimation method in AMOS software was chosen for parameter estimation.

3. Results

Table 2. The demographic characteristics of the sample residents.

Variable	Variable Meaning	Sample Size	%	Variable	Variable Meaning	Sample Size	%
Gender	Male	224	40	Education	Elementary	10	1.8
	Female	335	60		Secondary	52	9.3
Age	<18	30	5.4		Post-secondary	194	34.7
	19–30	130	23.3		Graduate	304	54.2
	31–45	116	20.8	Relationship to the community	Permanent resident	461	82.5
	46–60	154	27.5		Part-time resident	46	8.2
	>60	129	23.7		Temporary resident	52	9.3
Duration of residency	<5 years	132	23.6	Annual income	<USD 1000	38	6.8
	5–9 years	56	10		USD 10,000–24,999	111	20
	10–14 years	58	10.4		USD 25,000–49,999	159	28.4
	15–19 years	70	12.5		USD 50,000–74,999	115	20.5
	20–29 years	93	16.6		USD 75,000–99,999	62	11.1
	>30 years	165	29.5		>USD 100,000	74	13.2

summarizes the demographic characteristics of the sample residents surveyed. Among the sample residents, 40% were male and 60% were female. The highest proportion of respondents (27.5%) was within the 40–60-year-old age range. The education level of the respondents was generally relatively high, and 54.2% of respondents had a graduate school education. Approximately 82.5% of the respondents reported being permanent residents, and 29.5% of the respondents reported living in the community for at least 30 years. The respondents were largely middle class, and 28.4% reported having an annual income of between USD 25,000 and USD 49,999. The detailed demographic information is available in Appendix A.

3.1. Importance–Performance Analysis

3.1.1. The Common Characteristics of Residents’ Perceptions

The mean perceived importance and performance of tourism development were calculated based on the survey responses (

Table 3. IPA of residents’ overall perceptions in the gateway communities (I represents the average importance score, p represents the average performance score, and positive/negative scores are calculated according to a five-point Likert scale from −2 to 2).

		Total
		p	I	p-I
Environmental	1. Decreased environmental quality	−0.30	1.04	−1.35
	2. Increased garbage and litter	−0.64	1.15	−1.79
	3. Increased water and air pollution	−0.58	1.04	−1.62
	4. Occupied farmland landscape	−0.08	0.48	−0.56
	5. Destroyed undeveloped area/wildness	−0.35	0.84	−1.20
	6. Damaged natural resources	−0.38	0.87	−1.25
Sociocultural	7. Increased traffic problems	−0.75	1.01	−1.76
	8. Impeded infrastructure development	−0.31	0.63	−0.94
	9. Promoted cultural identity	0.55	0.86	−0.31
	10. Promoted cultural exchange	0.70	0.93	−0.23
	11. Expanded the influence of local traditions	0.59	0.87	−0.28
	12. Changed your lifestyle positively	0.41	0.60	−0.19
	13. Increased popularity of the community	0.75	0.86	−0.11
	14. Improved neighborhood interpersonal relationships	0.28	0.58	−0.31
	15. Protected local dialect/language	0.37	0.79	−0.42
	16. Improved women’s social status	0.22	0.65	−0.43
	17. Improved residents’ morality	0.23	0.61	−0.38
Economic	18. Increased employment opportunities	1.07	1.25	−0.18
	19. Increased business/investment opportunities	0.87	1.15	−0.28
	20. Improved living conditions	0.39	0.98	−0.59
	21. Increased local family incomes	0.67	1.10	−0.44
	22. Increased household consumption level	0.44	0.71	−0.27
	23.Reduced number of young people working outside community	0.35	0.70	−0.34
	24. Enriched recreational entertainment	0.69	1.00	−0.31
	25. Benefited most residents	0.69	1.03	−0.34

). The highest overall importance and performance scores were those for item 18, “Increased employment opportunities”. The lowest overall importance and performance scores were those for items 4, “Occupied farmland landscape”, and 7, “Increased traffic problems”. The item with the largest overall variation in perception in the two portal communities was item 2, “Increased garbage and litter”. In contrast, the item with the smallest variation was item 13, “Increased popularity of the community”.

3.1.2. The Importance–Performance Analysis of the Perceptions of Specific Residents

The total average perceived importance of the park to the gateway communities was 0.87, while the total average perceived performance was 0.24. The perceived importance of the factors and the average scores of performances were mapped as coordinates to an IPA chart in order to obtain a four-quadrant dot matrix chart containing all factors (Figure 3). In the chart, performance is shown on the X-axis, importance is shown on the Y-axis, and the total average value of performance and importance is zero.

Quadrant I reflects both high importance and high performance, and the factors located in this quadrant are mainly economic and sociocultural factors. Factors 10, 11, 18, 19, 20, 21, 24, and 25 are in quadrant I, indicating that tourism development, economic development, and sociocultural development complement one another.

The factors in quadrant II are ones that the residents perceive as requiring improvement, as their performance values are lower than their importance values. Factors 1, 2, 3, 6, and 7—which are all environmental impact factors—are located within this quadrant, indicating that the development of tourism has imposed serious problems on the local environment.

Quadrant III is the low-priority area, and the factors located in this quadrant (factors 4, 5, 8, 16, and 17) are characterized by intermediate performance. The residents’ perceptions of these factors are not strong, but these factors do not require prioritized improvement.

The factors in quadrant IV are those that are perceived as having been overemphasized, and include factors 9, 12, 13, 14, 15, 22, and 23. These indicate that tourism promotes cultural development, improves people’s lifestyles, increases the popularity of the community, protects the local dialect, increases consumption, and gives young people more opportunities.

3.2. SEM

Results of Model Fitting Evaluation and Correction

The initial hypothetical model was tested and required further revision. Based on the initial model test results, only TB and CS had a statistically significant impact on ST. Therefore, model correction was carried out by using TB, TC, and ST for model construction, and the resultant model fit well. LE was then added to the model, but the path coefficient of the model remained significant. Therefore, TC was also added, and was found to have no significant effect on TB and CS. TC was then temporarily removed and replaced with TT, yielding an acceptable model fit and accurately estimated path coefficients. After adding CP, some path coefficients were not significant. This insignificant relationship was therefore removed, and the final model was obtained (Figure 4).

The parameters were estimated using AMOS software, and the fit indicators of the model are shown in

Table 4. Model global fit metric.

Fit Index	Judging Criteria	Actual Fitting Value	Suitability Evaluation
Absolute Fitting Index
GFI	GFI > 0.90	0.733	Is
RMR	RMR < 0.05	0.142	Is
RMSEA	RMSEA < 0.1	0.085	Is
Value-Added Fitting Index
TLI	tli > 0.90	0.839	Is
NFI	Closer 1, better model suitability	0.789	Close
CFI	Closer 1, better model suitability	0.853	Close
IFI	Closer 1, better model suitability	0.854	Close
Simple Fit Index
PCFI	PCFI > 0.50	0.782	Is
PNFI	PNFI > 0.50	0.724	Is
CN	CN > 200	455	Is
Chi-squared degree-of-freedom ratio	<5	2.795	Is

.

The absolute fit, relative fit, and simple fit indices were used to evaluate the overall suitability of the model. The specific indices, standards, index values, and compliance are reported in Table 4. All of the fit indices met the corresponding standards, indicating that the overall fit of the model was very good.

After the final model was established, the path coefficients in the structural model were estimated using the maximum likelihood estimation method. The hypotheses of the theoretical models were verified (

Table 5. The results of hypothesis verification.

Hypothesis	Content	Parameter Estimate	Critical Ratio Value	Significance	Results
H1	CP positively impacts TB	0.299	5.335	***	Validated
H2	CP negatively impacts TC	N/A	N/A	N/A	No impact
H3	CP positively impacts CS	0.057	0.974	0.33	Failed
H4	LE positively impacts TB	0.167	2.698	0.007	Failed
H5	LE negatively impacts TC	N/A	N/A	N/A	No impact
H6	LE positively impacts CS	0.596	5.638	***	Validated
H7	TT positively impacts TB	0.339	5.475	***	Validated
H8	TT negatively impacts TC	N/A	N/A	N/A	No impact
H9	TT positively impacts CS	0.246	3.835	***	Validated
H10	TB positively impacts ST	0.748	6.868	***	Validated
H11	TC negatively impacts ST	0.072	0.971	0.331	Failed
H12	CS positively impacts TB	N/A	N/A	N/A	No impact
H13	CS negatively impacts TC	N/A	N/A	N/A	No impact
H14	CS positively impacts ST	0.404	5.53	***	Validated

Note: CP—community participation, LE—living environment, TT—trust in tourism institutions, TB—tourism benefits, TC—perceived tourism costs, CS—community satisfaction, and ST—support for tourism. *** Significant at p < 0.001.

).

4. Discussion

4.1. Prioritizing Tourism Management Based on the Results of the Importance–Performance Analysis

If the average performance score is significantly smaller than the average importance score in IPA, corresponding management measures should be implemented [59]. For all of the data analyzed in this study, the differences between the average performance and average importance were negative. This indicates a gap between community residents’ desire for the development of national parks and community tourism and the current state of development. This disparity should attract the attention of local government, community managers, and national park officials and motivate them to implement corresponding management measures to improve the performance of these impact factors and achieve sustainable management and development of the gateway communities. The specific perceptions of these impact factors are discussed below.

Tourism development in the park has resulted in economic development, increased employment opportunities, increased local incomes, and enriched recreational activities, and has promoted the development and inheritance of local traditional culture. Thus, the aspects plotted in quadrant I in Figure 3 should be maintained. Residents believed that the problems of local garbage disposal and environmental pollution were serious, and they were aware of the deterioration of natural resources. Residents were also aware of the importance of the environment to tourism development, and believed that local environmental problems had yet to be resolved.

The findings also indicated that the surveyed communities did not have strong perceptions of tourism development occupying farmland and undeveloped land, improving the morality of residents, or improving the status of women in the community. The lack of attention to some environmental and sociocultural factors indicated that the residents cared less about these aspects, and that these problems in the communities were not serious enough to seek to resolve.

Economic factors generally fell in quadrant I in Figure 3, indicating that the residents assigned importance to the economic benefits of tourism and were satisfied with the economic status of tourism development. While their attention to sociocultural factors was slightly lower, sociocultural factors performed well in these communities. However, residents paid more attention to environmental issues and believed that environmental factors were an important aspect of tourism development. The environmental aspects in this study did not perform well, indicating that the environmental issues require management measures.

4.2. Verified Key Factors That Influence Local Support

Structural models reflect the relationships among latent variables. The results of this research indicated that community participation had a significant positive impact on tourism benefits, and H1 was therefore supported (Table 5). Among the community participation variables observed, “you or your family work in the national park or a tourism-related department” (for detailed information, refer to Appendix A.) had the greatest impact on living environment, with a factor load of 0.986, indicating that this was the most influential variable in improving overall community participation. Living environment had a significant positive impact on community satisfaction, and H6 was therefore supported (Table 5). This conclusion is consistent with results of other studies [11]. Satisfaction with the living environment is an important factor that determines the overall satisfaction of community residents, and improving the living environment can increase overall satisfaction. Therefore, the positive impact of tourism on the living environment will also indirectly and positively affect satisfaction. Trust in tourism institutions had a direct and positive impact on the perception of tourism benefits, and H7 was therefore supported (Table 5). Tourism agencies and companies are key players in the development of national park tourism in gateway communities. The formulation and implementation of tourism planning and related policies can largely guide the direction of local tourism development and determine the quality of tourism. Tourism agencies win the trust of residents, which can enable residents to benefit from the development of—and then to actively support—local tourism. Trust in tourism institutions had a significant positive impact on community satisfaction, and H9 was therefore supported. Residents’ trust in tourism agencies and their policies is one of the determinants of community satisfaction. The benefits introduced to community residents by the policies and behaviors of tourism agencies help win the trust of residents, resulting in high community satisfaction. The verification of H9 indicates that institutional trust has an important influence on residents’ attitudes, and it reflects the practice of social exchange theory in the development of local community tourism. Tourism’s benefits had a significant positive impact on support for tourism, and H10 was therefore supported. Social exchange theory holds that as long as residents believe that the profits of tourism are greater than its costs, they will tend to support the development of the tourism industry, indicating that residents’ attitudes and behaviors are consistent. Residents who perceive more benefits are more likely to support local tourism’s development and welcome tourists, and are more willing to accept tourism investment from operators outside of the community. Community satisfaction had a positive effect on support for tourism, and H14 was therefore supported. Community residents who are satisfied with tourism’s development are more likely to perceive its positive effects and support it. This shows that community satisfaction is an important variable for understanding residents’ behavior.

Eight of the hypotheses were not supported, most of which were related to perceptions of the costs of tourism. Although community participation had a positive impact on community satisfaction and living environment had a positive impact on tourism’s benefits, these impacts were not significant. The perception of perceived tourism costs had little relationship with other variables. The history, type, and mechanisms of tourism development have led to differences in the results of previous research [11,29,60]. Unlike destinations such as Mauritius, where tourism has matured, the tourism of the gateway communities of Grand Canyon National Park is still in its infancy [60]. Although Grand Canyon National Park has a long history and is visited by millions of tourists every year, the negative impacts of tourism in the gateway communities are not obvious. Community residents’ perceptions of tourism’s benefits were stronger than their perceptions of costs, which is consistent with the findings of Vargas [33], who concluded that positive perceptions of the tourism industry tend to have stronger impacts on residents’ attitudes than negative perceptions. H2 was also not supported, indicating that perceived tourism costs had little relationship with community participation. Based on the results of the validity and reliability analyses, community participation had no significant impact on community satisfaction, and H3 was therefore also not supported. Different gateway communities are at different stages of tourism development, and the degrees of residents’ participation were correspondingly different, as were residents’ perceptions of community satisfaction and tourism costs. Therefore, in terms of the overall analysis, there were insufficient conditions to demonstrate the impacts of community participation on community satisfaction and perceived tourism costs. Although living environment had a positive impact on tourism’s benefits, H4 was not supported because the impact was not significant. Communities in different geographical locations are characterized by differences in transportation and tourism resources. In terms of the overall analysis, the impacts of different living environments on the perceptions of tourism’s benefits could not be accurately identified.

The remaining hypotheses involving the perceived costs of tourism were also not supported; thus, perceived tourism costs cannot be used as a factor to measure gateway community residents’ support of tourism. Community participation did not affect the positive perception of tourism by community residents, i.e., the perceived cost of tourism was not significant in this area. This differs from the results of Nunkoo [12,60] but is consistent with the results of Gursoy [61]. No significant relationship was found between perceived tourism costs and support for tourism. The type and nature of local tourism development have caused differences in residents’ perceptions, indicating that the gateway community residents’ support for tourism in the park results more from their perceptions of the benefits of tourism than from their perceptions of tourism’s costs. Although community satisfaction had a negative impact on the perceived costs of tourism, H13 was not supported (Table 5), indicating that residents who were dissatisfied with the community may not necessarily believe that the development of tourism will have a negative impact.

In short, local support for tourism development in the gateway communities at Grand Canyon National Park was influenced by CP (community participation), LE (living environment), TT (trust in tourism agencies), TB (tourism benefits), and CS (community satisfaction), and these should be fully considered in national park development strategies and community management to ensure more community support. This is of great significance for the sustainable development of both national parks and gateway communities.

5. Conclusions

In this study, residents living in the gateway communities around Grand Canyon National Park were found to generally express positive perceptions of tourism development. The differences between average performance and average importance were negative for all variables, indicating a gap between the desire of community residents for the development of national parks and the current community tourism development status. Residents’ satisfaction with the current development of tourism could be higher. More specifically, the importance–performance analysis in this study could help with prioritizing tourism management by grouping the impact factors into four stages. The high importance and low performance of the environment factors made local environmental problems rank first in terms of future management priority. The managing departments should pay more attention to these and maintain economic development. The sociocultural factors do not need further emphasis.

For the purpose of exploring the variables influencing local support for tourism development as well as their interrelationships, a hypothetical model of gateway community residents’ overall tourism support was constructed. Via both exploratory and confirmatory factor analyses, the model was revised to evaluate the appropriateness of the inclusion of various indicators in the revised model. Based on the SEM analysis and verification of the theoretical model, 14 hypotheses were proposed. The verified hypotheses showed that community participation has a positive impact on tourism benefit perception, the living environment has a positive impact on community satisfaction, trust in tourism institutions has positive impacts on tourism benefit perception and community satisfaction, tourism profit perception has a positive impact on tourism support, and community satisfaction has a positive impact on tourism support. The positive impact of community participation on community satisfaction was not significant, nor was the positive impact of the living environment on tourism benefit perception, and the perception of tourism’s costs had little relationship with the other variables. Finally, community participation (CP), the living environment (LE), trust in tourism agencies (TT), tourism benefits (TB), and community satisfaction (CS) were verified to be the drivers of local support for national park tourism development. However, the findings did not verify perceived tourism costs as a significant driver of local support in the focal communities.

Based on these findings, the following suggestions can be articulated for future management: (1) For local residents’ satisfaction, government and management agencies should prioritize and optimize the use of environmental resources in tourism development, maintain essential ecological processes, and conserve natural heritage and biodiversity. (2) For enhancing residents’ support, government should improve local living environments, help tourism agencies to build more trust, ensure that tourism benefits residents, improve community satisfaction with local tourism development, and engage community participation.

The most positive impacts of national park tourism on local communities are economic factors, indicating that national park managers should maintain economic development by ensuring viable, long-term economic operations around the park and providing socioeconomic benefits to local people with more stable employment, income/earning opportunities, and social services. In addition, to improve local satisfaction with national park management, the parks should also provide more environmental education programs to both visitors and local people and cooperate with residents in controlling visitation flows.

6. Research Limitations and Future Work

This study has several limitations that should be addressed in future research. The case study was carried out over a relatively short time period and with a relatively small sample. Although systematic sampling can be easy to operate, there are some disadvantages to this method. The door number interval was chosen based on past literature and may not have been suitable for these small, remote communities. COVID-19 also impacted data collection, and a comparison with other national parks could not be completed due to insufficient data.

Future work should also integrate secondary data, including economic and spatial data, to characterize the focal gateway communities. More surveys are necessary to identify issues in the gateway communities, particularly after the pandemic.