Mediating and moderating effects in golf tourism: Evidence from Hainan Island

Abstract

This article examines the relationship between golf activities and tourist perceptions and intentions using data from 592 golf tourists on Hainan Island, China. The results obtained from structural equation modeling show that the attractiveness of the destination can improve the identity of the location, and both factors significantly increase the travel intentions of golf tourists (i.e. revisit intentions and word-of-mouth recommendations). The results also show that place identity has significant mediating effects between destination attractiveness and travel intentions. In addition, the results reveal that the disposable income of golf tourists has significant moderating effects on the attractiveness of the destination, the identity of the destination, and the intention to revisit. This study has some important economic implications for golf destination management. Golf course managers can improve the attractiveness of the destination and strengthen the identity of the location to enhance the behavioral intentions of golf tourists.

Keywords

behavioral intention disposable income golf tourism mediating effect moderating effect place identity

Introduction

As a niche and profitable sector, the golf tourism industry has been growing rapidly, contributing to an increase of tax revenues, regional employment, and economic opportunities (Bárcena-Martín et al., 2017). Since 2000, the number of golf tourists worldwide has increased by 50%, and there are more than 60 million golf tourists across the world. Golf tourists spend a great amount of money on transport, hotels, entertainment, and other golf-related expenses, on average approximately US$280 per visit and 20% more than mass tourists (Song et al., 2017). It is likely that golf tourists will take an overnight trip to visit a golf course or club if it is more than an hour’s driving distance away from their place of residence.

With regard to the significant development of the global golf industry, golf tourism has become a growing research field (Jorge and Monteiro, 2011). Based on the data from ProQuest, ScienceDirect, EBSCO, Wiley, and Taylor & Francis, there were approximately 150 research articles on golf tourism over the period 2009–2018. Scholars have focused on golf tourists’ behavioral intentions (Hutchinson et al., 2009; Pereira et al., 2015) and golf destination management (Humphreys and Weed, 2014; Kim and Ritchie, 2012).

Given the significant economic impacts of golf tourism on destinations, it is becoming increasingly necessary for government authorities, industry practitioners, and scholars to better understand golf destination attributes and golf tourists’ behavioral intentions (Correia et al., 2007; Petrick and Backman, 2002). In the emerging golf tourism markets, some golf courses have been suffering a decline in their revenue due to the competition between destinations (Song et al., 2017). This study attempts to examine the determinants of golf tourists’ behavioral intentions, and how these intentions are affected by golf place identity (PI) as a mediator variable and by disposable income as a moderator.

Literature review

Destination attractiveness

Destination attractiveness (DA) is defined as the perceived ability of the destination to fulfill tourists’ goals or satisfy their needs (Hu and Ritchie, 1993). The more a destination is able to meet tourists’ needs, the more it is perceived to be attractive, and the more that particular destination is likely to be chosen (Henkel et al., 2006). Moreover, over the years, the importance of DA for tourists’ behavioral responses has received increasing attention (Moore and Graefe, 1994). Cheng et al. (2013) found that the extent of the attractiveness of a destination perceived by tourists is positively associated with PI and place dependence. It has been further suggested that DA is one of the critical determinants of tourists’ behavioral intentions (Sangpikul, 2018).

From a demand-side perspective, DA reflects tourists’ perceived value of the destination, which is likely to be an essential factor for golf tourists’ PI (Kim and Ritchie, 2012). The golf tourism destination, as an attractive tourism site, has been given increasing attention because of its influence on golf tourists’ behavioral intentions (Hutchinson et al., 2009). More specifically, golf tourism-related information often comes from friends’ recommendations (Song et al., 2017). When a golf course is perceived to be attractive, satisfied golf players are most likely to make positive word-of-mouth (WOM) recommendations for the destination they have visited (Moital et al., 2013). On the basis of these considerations, three hypotheses were proposed:

H1a: Golf tourism destination attractiveness positively influences golf tourists’ place identity.

H1b: Golf tourism destination attractiveness positively influences golf tourists’ revisit intentions.

H1c: Golf tourism destination attractiveness positively influences golf tourists’ WOM recommendations.

Sense of place and PI

Sense of place can be considered as a general attitude toward a spatial setting or physical environment, and a complex psychosocial structure that contains emotions, self-referent beliefs, and behavioral commitments (Jorgensen and Stedman, 2001; Pretty et al., 2003). The tripartite model of sense of place comprises PI, place attachment, and place dependence. Place attachment can be considered as an individual’s emotional connection to a spatial setting, and place dependence can be defined as the perceived behavioral advantage of a spatial setting relative to other settings (Jorgensen and Stedman, 2006).

PI refers to the link established between tourists and destinations (Wang and Chen, 2015), reflecting the degree of personal values associated with a particular environment and a person’s emotional connection to a certain destination place (Kyle et al., 2003). Tourism is essentially a place-based phenomenon involving the production of destination identity at different scales (Hall, 1998; Urry, 1990). For example, the scale of a destination can be a national, regional, or local place. Thus, destination PI is related to the stakeholders’ desire to position its uniqueness competitively in the policy-making, marketing, and promotional activities at different levels (Dredge and Jenkins, 2003).

In the golf tourism industry, Song et al. (2017) examined the impacts of PI on golf tourists’ revisit intentions (RI), suggesting that the golf tourism destination should upgrade its infrastructure to enhance golf tourists’ destination image, to strengthen their RI. The more favorable the perception by golf tourists of a golf destination, the more likely they are to intend to revisit and/or recommend that specific destination to other golf tourists (Hutchinson et al., 2009). Thus, it is essential for golf tourists to identify with a certain golf tourism destination, rather than others, to strengthen their behavioral intentions (Petrick and Backman, 2002). Accordingly, the following hypotheses were proposed:

H2a: Golf tourism place identity positively influences golf tourists’ revisit intentions.

H2b: Golf tourism place identity positively influences golf tourists’ WOM recommendations.

Behavioral intentions

Behavioral intentions, as a multidimensional outcome variable, consist of intentions to revisit a certain destination in the future and to spread positive WOM recommendations to potential tourists (Lee et al., 2013). It has been suggested that RI have a positive effect on WOM recommendations, and tourists with RI are more likely to share their travel experience and tourism information with other potential tourists (Su et al., 2014). Although RI have attracted substantial attention from tourism scholars, WOM recommendations have rarely been empirically tested in a tourism context.

In the golf tourism field, players’ intentions to visit the same golf course are considered as RI (Petrick and Backman, 2002). Golf tourists often play with friends or family, so, if they are satisfied with a certain golf course, they usually recommend that course to those people they know (Song et al., 2017). It is likely that golf tourists would recommend a particular golf destination they intend to revisit to people who are planning a golf vacation (Hutchinson et al., 2009). Therefore, it is likely that RI contribute to WOM recommendations. On the basis of the empirical studies above, a hypothesis concerning the relationship between RI and WOM recommendations was proposed:

H3: Golf tourists’ revisit intentions positively influence their WOM recommendations.

Mediating effect of PI

PI, as an individual emotion, involves bonding to the social or physical environment and has been found to have a significant mediating effect between destination and travel behavior (Cheng et al., 2013). Kyle et al. (2003) examined the relationship between PI, place dependence, and visitors’ attitudes concerning their spending preferences and found that PI was a significant mediator between destination and tourists. In line with this, it has been suggested that PI mediates the relationship between DA and tourists’ behavioral intentions (Cheng et al., 2013).

In golf tourism, Song et al. (2017) found that place attachment has a significant mediating effect between destination image and golf tourists’ RI. Given the close relationship between RI and WOM recommendations, it has been suggested that it is necessary to examine both of these factors in the context of golf tourists’ behavioral intentions (Hutchinson et al., 2009). Those two studies suggest that PI may mediate in the relationship between DA and golf tourists’ behavioral intentions. Hence, the following hypotheses were proposed:

H4a: Golf tourism place identity has a positive mediating effect between destination attractiveness and golf tourists’ revisit intentions.

H4b: Golf tourism place identity has a positive mediating effect between destination attractiveness and golf tourists’ WOM recommendations.

Moderating effect of disposable income

The term, “disposable income,” refers to the amount of money available for a household to spend and save after paying income taxes (Hultkrantz, 1995). In tourism economics, as a way of examining tourists’ potential consumption, disposable income, a significant factor in affecting tourists’ behavior, is a relative concept compared with absolute personal income (Wang et al., 2018). Furthermore, Leonidou et al. (2015) found that income has a moderating effect on the relationship between tourists’ intentions and eco-friendly attitudes. Akinci and Aksoy (2019) further examined the moderating effects of household income on behavioral intentions via a moderated mediation model.

Golf has been defined as a sports activity that requires the players to have a solid financial position (Bárcena-Martín et al., 2017), since it is associated with a large cost (e.g. purchasing golf equipment and golf membership) and the availability of disposable income (Yoh et al., 2006). It has been suggested that golf tourists’ disposable income has a significant influence on their destination selection and behavioral intentions (Jorge and Monteiro, 2011). In line with those findings, Humphreys and Weed (2014) have also claimed that golf tourists’ decision-making is influenced by their perceived value and disposable income. It is believed that golf tourists’ disposable income affects the relationship between their travel behavior and golf destination. Accordingly, the following hypotheses were proposed:

H5a: Golf tourists’ disposable income has a moderating effect between golf tourism destination attractiveness and revisit intentions.

H5b: Golf tourists’ disposable income has a moderating effect between golf tourism place identity and revisit intentions.

In summary, DA may positively influence golf tourists’ PI and behavioral intentions, contributing to their perceived values of the golf tourism destination. Furthermore, golf tourism PI may mediate the relationship between DA and golf tourists’ behavioral intentions. Meanwhile, golf tourists’ disposable income may have a moderating effect between golf tourism DA/PI and RI. Based on the analyses described above, this research proposes five groups of hypotheses (10 subhypotheses) in a conceptual model of golf tourism (see Figure 1).

Figure 1.

Hypothesized conceptual model.

Data description

Sampling

Hainan Island is well known as “the capital of golf tourism in China,” with 57 golf courses, which account for 12% of the golf courses in China. More than a million golf tourists (domestic and international) have played golf on the Hainan Island golf courses, resulting in expenditures of US$25 billion (Hainan Provincial Bureau of Statistics, 2018). Golf is considered as one element of the national strategic plans for Hainan Island. A set of golf-related events, such as the Blue Bay Masters 2018, the China Open 2017, and the Women’s Open 2017, were held in Hainan, to highlight its status as a golf tourism destination. In 2019, Hainan won the favorite golf tourism destination of Asian golf tourists. Due to its unique geographical location, weather conditions, and special natural resources, golf tourism is popular on Hainan Island, especially in the winter. For golfers, Hainan Island is a place with beautiful natural environment and attractive Golf Course. It is likely that Hainan Island could be regarded as a typical golf destination for island tourism.

The items in the questionnaire were originally developed in English and then translated into Chinese. A bilingual expert (English and Chinese) reviewed the translation to confirm that the translated version accurately reproduced the meaning of the original items. To confirm the validity of the questionnaire, five professors in the golf, tourism, and hospitality industry examined the content of the questionnaire. Then, the questionnaire was tested via a pilot test of 100 respondents to avoid any misunderstandings and bias (Podsakoff et al., 2003). Twelve students majoring in tourism management at the International Tourism College, Hainan University, were trained for 2 days, and 8 of them were eventually chosen to assist in the formal investigation. The data collection was conducted through an online platform (www.sojump.com) at the local airports and golf clubs on Hainan Island from June 1 to December 30, 2018, to cover both the off-season and the peak season.

In an attempt to guarantee the quality of the results, the IP address of the respondents was limited to Hainan Island. A total of 690 golf tourists met the requirement and were chosen to complete the survey. Furthermore, any incomplete survey was considered invalid, and those completed in less than 2 min were dropped. After filtering, 592 samples were coded and used for preliminary data analyses.

As presented in Table 1, the respondents included 196 women (33%) and 396 men (67%). The majority of golf tourists (72%) were aged 18–39 years and were well-educated (with 76% having higher education) and had a monthly disposable income of over US$2000. Nearly all the golf tourists have had golf tourism experience on two or more occasions in the past 2 years.

Table 1.

Characteristics of respondents.

Observed variables		N	%
Gender	Male	396	66.9
Gender	Female	196	33.1
Age (years)	18–29	185	31.3
	30–39	238	40.2
	40–49	98	16.5
	≥50	71	12.0
Occupation	Management	103	17.4
	Researcher	108	18.2
	Profession	97	16.4
	Sales and service	65	11.0
	Administration	44	7.4
	Education	43	7.3
	National servant	30	5.1
	Self-employed	48	8.1
	Production and technology	27	4.6
	Agriculture-related fields	16	2.7
	Other	11	1.8
Education	High school	62	10.5
	College student	63	10.6
	Bachelor	317	53.5
	Graduate and above	133	22.5
	Other	17	2.9
Disposable income/month	≤US$1000	43	7.3
	US$1001–2000	245	41.4
	US$2001–3000	196	33.1
	≥US$3001	108	18.2
Frequency	Less thanthree times in 6 months	43	7.3
	Once a month	139	23.5
	Twice a month	177	29.9
	Three times a month	109	18.4
	Four times a month	64	10.8
	Greater than four times a month	60	10.1
Golf experience in the past 2 years	Less than or equal to one time	17	2.9
	Two times	222	37.5
	Three times	172	29.0
	Four times	94	15.9
	Greater than four times	87	14.7
Sum		592	100

Measurement scales

All items in the scales were measured using a five-point Likert-type scale ranging from 1 (completely disagree) to 5 (completely agree). DA was measured by 10 items relating to the destination’s accessibility, amenities, infrastructure, scenery, local community, and natural environment (Mikulić et al., 2016). Following Song et al. (2017), a five-item scale was used to test golf PI. A four-item scale was utilized to measure golf tourists’ RI (Song et al., 2017) and a four-item scale was utilized to test WOM recommendations to potential tourists (Gholipour Soleimani and Einolahzadeh, 2018). This study also applies a four-item scale to both golf tourists’ RI and WOM recommendations. The respondents were also asked to provide information about their disposable income, to empirically test its moderating effect on the relationship between DA, PI, and RI.

Empirical results

Confirmatory factor analysis

Using SPSS 22.0 to analyze 592 valid questionnaires, the Kaiser–Meyer–Olin (KMO) value and Bartlett’s test confirmed the suitability and reliability of the samples. The values of KMO (.926) and Bartlett’s test of sphericity (p < 0.001) demonstrate the suitability of the factor analysis (Field, 2009). The total variance explained for all constructs was 71.5%, which further indicates a satisfactory convergent validity (Hair et al., 2013). The number of missing values for every variable was below 5%, showing that the missing values occurred on a random basis (Parent, 2013).

Confirmatory factor analysis (CFA) was used to examine the validity of the measurement model (maximum likelihood estimation method) by testing the discrete nature and consistency of the latent variables (Anderson and Gerbing, 1988). All analyses were conducted with IBM SPSS Statistics 24.0 using the variance–covariance matrices. Measurement models are required to have acceptable goodness-of-fit indices (Cliff, 1983), including the Chi-squared/degrees of freedom (χ²/df) < 3.000, the goodness-of-fit index (GFI) > 0.900, the root mean square error of approximation (RMSEA) < 0.080, the comparative fit index (CFI) > 0.900, the Tacker–Lewis index (TLI) > 0.900, and the incremental fit index (IFI) > 0.900. The indices of the measurement model in this study are acceptable, with χ²/df = 2.952, p < 0.050, GFI = 0.913, RMSEA = 0.058, CFI = 0.964, TLI = 0.958, and IFI = 0.961.

As presented in Table 2, Cronbach’s α values which exceed 0.800 indicate high internal consistency (Fornell and Larcker, 1981). The composite reliability estimates of all constructs exceed the suggested threshold of 0.700, and all values which exceed 0.800 indicate that the measures are reliable (Bacon et al., 1995). The convergent validity of the measurement model was measured using standardized factor loading. All standardized factor loadings were above the threshold value of 0.500 (significant at p < 0.001).

Table 2.

Descriptive statistics and CFA results for the measurement model.

Dimension and items	Mean	SD	St FL	SE	p	Cronbach’s α	CR
DA						0.903	0.926
Sports and recreational options	4.031	0.045	0.859	0.047	***
Well-developed environment	4.044	0.043	0.759	0.023	***
Friendly host and locals	4.075	0.042	0.830	0.027	***
Good accommodation and restaurants	4.100	0.042	0.805	0.030	***
Shopping options	4.046	0.042	0.831	045	***
Excellent landscape of the golf courses	4.041	0.044	0.729	0.069	***
Diversity of cultural activities	4.196	0.040	0.687	0.700	***
Quality of tourist information	4.013	0.044	0.710	0.055	***
Unique local specialties and souvenirs	4.128	0.040	0.593	0.070	***
Convenient transportation	4.065	0.043	0.613
PI						0.918	0.919
Hainan Island is a part of me	3.957	0.048	0.822
Hainan Island is very special to me	3.965	0.045	0.853	0.042	***
I identify strongly with Hainan Island	4.087	0.040	0.824	0.039	***
I am very attached to Hainan Island	4.105	0.042	0.832	0.040	***
Hainan Island means a lot to me	3.995	0.045	0.831	0.042	***
RI						0.879	0.883
I may revisit Hainan Island	4.102	0.041	0.822
I plan to revisit Hainan Island	4.089	0.042	0.857	0.037	***
I want to revisit Hainan Island	4.084	0.043	0.868	0.038	***
I intend to revisit Hainan Island	4.099	0.042	0.845	0.035	***
WOM recommendations						0.892	0.908
Recommend Hainan golf courses to others	4.138	0.040	0.857
Encourage others to travel to Hainan Island	4.096	0.042	0.846	0.036	***
Introduce Hainan golf courses to others	4.008	0.045	0.831	0.035	***
Spread positive words about Hainan Island	4.081	0.041	0.838	0.036	***

Note: CFA: confirmatory factor analysis; DA: destination attractiveness; SD: standard deviation; St FL: standard factor loading; SE: standard error; CR: composite reliability; WOM: word-of-mouth; PI: place identity; RI: revisit intentions.

*p < 0.05.

**p < 0.01.

***p < 0.001.

To evaluate the discriminant validity, the present study utilized a comparison of the square roots of average variance extracted (AVE) with the interconstruct correlation between variables. All the square roots of AVE values were greater than the interconstruct correlations between variables, indicating that each construct is statistically different from the others, and the constructs have good discriminant validity (Fornell and Larcker, 1981). Table 3 presents the satisfying AVEs of the four constructs: DA, PI, RI, and WOM recommendations.

Table 3.

Discriminant validity matrix.

	DA	PI	RI	WOM
DA	0.747
PI	0.647	0.832
RI	0.529	0.657	0.804
WOM recommendations	0.515	0.638	0.682	0.812

Note: DA: destination attractiveness; WOM: word-of-mouth; PI: place identity; RI: revisit intentions. The off-diagonal elements are interconstruct correlations.

Structural equation modeling

Structural equation modeling (SEM) was further employed, based on the results of the CFA model, to test the proposed relationship among the construct variables. The indicators were used to evaluate how well the structural models fit the data: χ²/df = 2.381, p < 0.050, GFI = 0.933, RMSEA = 0.049, CFI = 0.974, TLI = 0.969, and IFI = 0.974, all suggesting satisfying goodness-of-fit parameters (Hair et al., 2013).

Table 4 presents the SEM results. DA is positively and significantly associated with PI (β = 0.661, t = 12.093, p < 0.001), and thus H1a is not rejected. It is also positively associated with golf tourists’ RI (β = 0.215, t = 2.969, p < 0.010) and their WOM recommendations (β = 0.107, t = 3.202, p < 0.010), so H1b and H1c are not rejected. Furthermore, PI is positively and significantly associated with golf tourists’ RI (β = 0.994, t = 14.797, p < 0.001) and their WOM recommendations (β = 0.271, t = 2.932, p < 0.010), and thus H2a and H2b are not rejected. Last, it was found that golf tourists’ RI had a positive and significant relationship with their WOM recommendations (β = 0.705, t = 8.579, p < 0.001), and thus H3 is not rejected.

Table 4.

Empirical results for the structural equation model.

Hypothesis	Estimate	SE	t-statistic	p	Result
H1a: DA → PI	0.661	0.027	12.093	***	Not rejected
H1b: DA → RI	0.215	0.065	2.969	**	Not rejected
H1c: DA → WOM	0.107	0.055	3.202	**	Not rejected
H2a: PI → RI	0.994	0.048	14.797	***	Not rejected
H2b: PI → WOM	0.271	0.066	2.932	**	Not rejected
H3: RI → WOM	0.705	0.082	8.579	***	Not rejected

Note: DA: destination attractiveness; PI: place identity; RI: revisit intentions; WOM: word-of-mouth recommendations; SE: standard error.

*p < 0.05.

**p < 0.01.

***p < 0.001.

Bootstrapping was further used to measure the mediating effect of PI. To obtain 95% confidence intervals (CIs) (MacKinnon et al., 2007), the Bootstrap estimates of the direct and indirect effects were based on 2000 bootstrap samples (Fritz and MacKinnon, 2008). This study further examined the full mediation role of PI in the relationship between DA, RI, and WOM recommendations. The goodness-of-fit indices of the mediation model were χ²/df = 2.460 (lower than the cutoff value of 3.000); RMSEA = 0.047 (lower than the cutoff of 0.080); and CFI = 0.968, TLI = 0.973, and IFI = 0.953 (all higher than the threshold value of 0.900). Therefore, the complete mediation model was a good fit.

Table 5 presents the 95% CIs (LLCI = 0.082 and ULCI = 0.144) for DA-PI-RI: indirect effect (β = 0.660), which means that, with a mediating effect, the indirect effect does not reach 0 with a 95% CI. It was found that DA indirectly affects golf tourists’ intentions through PI with indirect effects 0.660 (p < 0.001). Thus, H4a is not rejected. The indirect effects of DA and WOM through PI are also significant (β = 0.553, LLCI = 0.174 and ULCI = 0.430), so H4b is not rejected. In summary, PI significantly mediated the effects of DA on RI and WOM recommendations.

Table 5.

Bootstrapping of the mediating effects of PI.

Hypothesis	Effect	p	LLCI	ULCI	Results
H4a: DA → PI → RI	0.660	***	0.082	0.144	Not rejected
H4b: DA → PI → WOM	0.553	***	0.174	0.430	Not rejected

Note: DA: destination attractiveness; PI: place identity; RI: revisit intentions; WOM: word-of-mouth recommendations; LLCI: lower limit confidence interval; ULCI: upper limit confidence interval; CI: confidence interval. The level of confidence for all CIs is 95%; the number of bootstrap samples for percentile bootstrap CIs is 2000.

*p < 0.05.

**p < 0.01.

***p < 0.001.

Hierarchical regression

Hierarchical regression was conducted using IBM SPSS 22.0 to examine the moderating effects of the golf tourists’ disposable income on the relationship between DA, PI, and RI. Table 6 presents the results of the hierarchical regression analysis of the moderating effects of disposable income. Model 1 analyzed the relationship between DA and RI, and its explained variance was significant (R² = 0.278, p < 0.001). Model 2 added the moderator of disposable income, which significantly increased the explanatory power of the model (F change = 401.449, p < 0.001). The regression coefficient in model 2 was also significant (β = 0.849, p < 0.001), showing that the golf tourists’ disposable income has a positive moderating effect on the relationship between DA and RI. Thus, hypothesis 5a is not rejected.

Table 6.

Hierarchical regression of the moderating effects of disposable income.

	Model 1	Model 2 (with DI)	Model 3	Model 4 (with DI)	Result
H5a: DA → RI	0.531***	0.849***			Not rejected
H5b: PI→RI			0.104***	0.770***	Not rejected
R ²	0.278	0.901	0.418	0.831
F change		401.449		407.085

Note: DI: disposable income; DA: destination attractiveness; PI: place identity; RI: revisit intentions.

***p < 0.001.

Furthermore, model 3 was applied to test the relationship between PI and RI and its explained variance was significant (R² = 0.418, p < 0.001). The moderator of disposable income was added in model 4, leading to a significant increase of the explanatory power of the model (F change = 407.085, p < 0.001). The regression coefficient in model 4 was significant (β = 0.770, p < 0.001), indicating that the golf tourists’ disposable income has a positive moderating effect on the relationship between PI and RI. Thus, hypothesis 5b is not rejected.

Robustness check

To test the robustness of the results above, subsamples were employed. In light of the variety of golf tourists’ experience, the respondents were divided into two groups: 239 respondents who had less golf tourism experience (≤2 times) and 353 with more golf experience (≥3 times). Table 7 presents that the results of regression, mediation, and moderation among the variables are robust.

Table 7.

Empirical results of robustness check with the subsamples.

	Less experienced subsample			More experienced subsample
Hypothesis	Estimate	SE	p	Estimate	SE	p
H1a: DA → PI	0.195	0.131	*	0.188	0.134	*
H1b: DA → RI	0.676	0.145	***	0.801	0.186	***
H1c: DA → WOM	0.193	0.135	*	0.202	0.145	*
H2a: PI → RI	0.445	0.129	**	0.613	0.144	**
H2b: PI → WOM	0.416	0.139	**	0.469	0.128	***
H3: RI → WOM	0.893	0.126	***	0.952	0.120	***
H4a: DA → PI → RI	0.418	0.181	***	0.535	0.203	***
H4b: DA → PI → WOM	0.455	0.189	***	0.382	0.171	**
H5a: DA → RI (DI)	0.427	0.127	***	0.799	0.220	***
H5b: PI → RI (DI)	0.267	0.123	**	0.675	0.144	***
Number of observations		239			353

Note: DA: destination attractiveness; PI: place identity; RI: revisit intentions; WOM: word-of-mouth recommendations; DI: disposable income; SE: standard error.

*p < 0.05.

**p < 0.01.

***p < 0.001.

Discussion

The present study first examined the relationship between golf tourism DA, PI, golf tourists’ RI, and their WOM recommendations and then further tested the mediating effect of PI and the moderating effect of disposable income.

The empirical results reveal that DA significantly increases PI, RI, and WOM recommendations. Thus, H1a, H1b, and H1c are not rejected, suggesting that the greater the golf tourists’ perception of DA, the stronger are their PI, RI, and WOM recommendations. It is likely that those golf tourism destinations which are attractive would enhance tourists’ PI and also attract a large volume of golf tourists who intend to recommend those destinations. This is in line with previous findings about the positive influence of the attractiveness of natural scenery, culture, and traditional buildings on PI (Cheng et al., 2013). It was further found that PI significantly enhances golf tourists’ RI and their WOM recommendations. Thus, H2a and H2b are not rejected. It is worth noting that golf tourists’ RI are likely to increase their WOM recommendations. These findings reflect that in the tourists’ decision-making process, PI plays a role in creating an emotional bonding between a tourist and a certain destination (Stylos et al., 2017).

Furthermore, the empirical results also suggest the mediation role of golf tourists’ PI in the relationship between DA, RI, and WOM recommendations. Thus, H4a and H4b are not rejected. This means that DA affects behavioral intentions (i.e. RI and WOM recommendations) via PI in golf tourism. These findings contribute to the golf tourism literature regarding the mediating role of PI. As evidenced in the literature, Song et al. (2017) found that place attachment (PI and place dependence) has a significant effect in mediating the relationship between destination image and golf tourists’ RI, and, in turn, place attachment was found to mediate the relationship between DA and environmentally responsible behavior (Cheng et al., 2013).

These results reveal that disposable income was the moderation variable which most positively affected DA and RI, PI, and RI. Thus, H5b and H5a are not rejected. One reason for these results is that golf is a luxury leisure activity, which requires comparatively more disposable income. It also shows that a high level of disposable income is a factor, which is positively correlated with tourists’ demand in the tourism economy (Papatheodorou and Pappas, 2017).

Conclusion

Theoretical implications

The present study has several important theoretical implications. It contributes to the theoretical foundation of golf tourism research by establishing both the construct validity and the reliability of measures for DA, PI, RI, and WOM recommendations in the field of golf tourism.

As one of the first attempts, this article has contributed to extending the research framework of golf tourists’ behavioral intentions by examining the moderating effects of golf tourists’ disposable income. Furthermore, this study shows that disposable income exerts a positive effect on golf tourists’ RI and WOM recommendations. Therefore, tourists’ disposable income should be reconsidered in the golf tourism research, in particular with respect to the economic context. Moreover, the analysis of the mediating effects of PI fills a gap in golf tourism research. The findings shed light on the relationship between DA, PI, RI, WOM recommendations, and disposable income, which goes further than the previous studies. The results of this study may attract the attention of scholars in general tourism research to the specific field of golf tourism, particularly in the emerging Asian markets.

Managerial implications

This study also has some important managerial implications for golf tourism management, industry practitioners, and government authorities at both national and regional levels. First, golf tourism developers and local residents should make an effort to stress the excellent natural environment of Hainan Island, thereby enhancing PI. The findings also reveal that DA plays a critical role in forming strong PI. To improve the attractiveness of Hainan Island, tourism practitioners should cooperate with local residents and provide unique cultural activities. In the long run, they may develop their own local innovation compared with other places (Pardey et al., 2014, 2016, 2018). For example, tourism developers could host golf-related events such as the Hainan Golf Open. Also, the local government could enhance the development by, for instance, improving its policy-making based on the investment evaluation (Hurley et al., 2014, 2016).

The findings of this study could help golf course practitioners to better understand the determinants of golf tourists’ RI and WOM recommendations. Golf clubs at the destination could create some market segmentation based on customers’ disposable income and focus on their target markets. For instance, golf has its own game features, such as handicap scoring, which allows participants of different skills to play together and provides opportunities for business partners to extend their collaboration. Therefore, launching pricing promotion packages or a membership system could stimulate golf tourists’ behavioral intentions. In addition, travel agents may also diversify golf tourism products and marketing strategies to accommodate the various needs of targeted golf tourists. The promotion of travel programs is likely to improve the reputation of Hainan Island and enhance its exposure, thereby attracting the attention of a greater number of tourists. In addition, it is worth noting that the emotional response of tourists to the physical environment and site-supported activities should be considered as an important factor in the promotion of golf tourism.

Limitations and future research

This study is based on an investigation of Chinese golf tourists, which has expanded the conceptual research framework for the study of golf tourism and also provided practical guidance to the golf tourism industry. As a case study, sampling bias may occur (Zainal, 2007) because this research collected data only from Chinese samples. Although the empirical models show robust results, the samples in this study are Chinese golf tourists, which may have biased the findings as a result of Chinese modesty or Asian politeness (Spencer-Oatey and Ng, 2001).

It has been suggested that the covariance-based SEM method has some limitations regarding the omitted variables, assumptions, and sensitive paths (Anderson and Gerbing, 1988; Cliff, 1983). In the light of this, the use of a Bayesian SEM (with prior information) could provide advantages in empirical tourism research (Assaf et al., 2018). As such, future studies could examine more samples from different markets to capture the influence of respondent heterogeneity by using more general statistical analytics such as component analysis, regression, and market segmentation (Engida et al., 2018). Furthermore, the influence of disposable income on consumer preferences needs to be further explored (Zhang et al., 2017), to satisfy the increasingly demanding and heterogeneous golf tourists. In addition, future studies may include regional differences to further examine the influence of characteristics among various areas (Rao, 2018; Rao et al., 2019). Future studies may also consider the possible effects of various types of golf tourists’ motivation, such as those of business golf tourists and leisure golf tourists, as their needs might be different.

Supplemental material

Supplemental Material, supplement_material - Mediating and moderating effects in golf tourism: Evidence from Hainan Island

Supplemental Material, supplement_material for Mediating and moderating effects in golf tourism: Evidence from Hainan Island by Huimin Song, Jamie M Chen and Yibin Chen in Tourism Economics

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Jamie M Chen

Yibin Chen

Supplemental material

Supplemental material for this article is available online.

References

Akinci

Aksoy

(2019) The impact of service recovery evaluation on word-of-mouth intention: a moderated mediation model of overall satisfaction, household income and gender. Tourism Management Perspectives 31: 184–194.

Anderson

Gerbing

(1988) Structural equation modeling in practice: a review and recommended two-step approach. Psychological Bulletin 103(3): 411.

Assaf

Tsionas

(2018) The time has come: toward Bayesian SEM estimation in tourism research. Tourism Management 64: 98–109.

Bacon

Sauer

Young

(1995) Composite reliability in structural equations modeling. Educational and Psychological Measurement 55(3): 394–406.

Bárcena-Martín

Rodríguez-Fernández

Borrego-Domínguez

(2017) Golf, supply and demand: the influence of economic factors. Tourism Economics 23(6): 1220–1234.

Cheng

Huang

(2013) The influence of place attachment on the relationship between destination attractiveness and environmentally responsible behavior for island tourism in Penghu, Taiwan. Journal of Sustainable Tourism 21(8): 1166–1187.

Cliff

(1983) Some cautions concerning the application of causal modeling methods. Multivariate Behavioral Research 18(1): 115–126.

Correia

Barros

Silvestre

(2007) Golf tourism repeat choice behaviour in the Algarve: a mixed logit approach. Tourism Economics 13(1): 111–127.

Dredge

Jenkins

(2003) Destination place identity and regional tourism policy. Tourism Geographies 5(4): 383–407.

10.

Engida

Rao

Berentsen

, et al.(2018) Measuring corporate sustainability performance—–the case of European food and beverage companies. Journal of Cleaner Production 195: 734–743.

11.

Field

(2009) Discovering Statistics Using SPSS, 3 rd ed. London: SAGE.

12.

Fornell

Larcker

(1981) Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research 18(1): 39–50.

13.

Fritz

MacKinnon

(2008) A graphical representation of the mediated effect. Behavior Research Methods 40(1): 55–60.

14.

Gholipour Soleimani

Einolahzadeh

(2018) The influence of service quality on revisit intention: the mediating role of WOM and satisfaction (case study: Guilan travel agencies). Cogent Social Sciences 4(1): 1560651.

15.

Hainan Provincial Bureau of Statistics (2018) Available at: http://www.hainan.gov.cn/hainan/tjsj/list_tjsj.shtml.

16.

Hair

Ringle

Sarstedt

(2013) Partial least squares structural equation modeling: rigorous applications, better results and higher acceptance. Long Range Planning 46(1–2): 1–12.

17.

Hall

(1998) Introduction to Tourism: Development, Dimensions and Issues. Melbourne: Longman.

18.

Henkel

Agrusa

, et al. (2006) Thailand as a tourist destination: perceptions of international visitors and Thai residents. Asia Pacific Journal of Tourism Research 11(3): 269–287.

19.

Ritchie

(1993) Measuring destination attractiveness: a contextual approach. Journal of Travel Research 32(2): 25–34.

20.

Hultkrantz

(1995) On determinants of Swedish recreational domestic and outbound travel, 1989–93. Tourism Economics 1(2): 119–145.

21.

Humphreys

Weed

(2014) Golf tourism and the trip decision-making process: the influence of lifestage, negotiation and compromise, and the existence of tiered decision-making units. Leisure Studies 33(1): 75–95.

22.

Hurley

Rao

Pardey

(2014) Re-examining the reported rates of return to food and agricultural research and development. American Journal of Agricultural Economics 96(5): 1492–1504.

23.

Hurley

Rao

Pardey

(2016) Re-examining the reported rates of return to food and agricultural research and development: reply. American Journal of Agricultural Economics 99(3): 827–836.

24.

Hutchinson

Lai

Wang

(2009) Understanding the relationships of quality, value, equity, satisfaction, and behavioral intentions among golf travelers. Tourism Management 30(2): 298–308.

25.

Jorge

Monteiro

(2011) Competitive choice dimensions of golf destinations: a multivariate perceptual mapping analysis. European Journal of Tourism Hospitality and Recreation 2: 29–54.

26.

Jorgensen

Stedman

(2001) Sense of place as an attitude: lakeshore owners attitudes toward their properties. Journal of Environmental Psychology 21(3): 233–248.

27.

Jorgensen

Stedman

(2006) A comparative analysis of predictors of sense of place dimensions: attachment to, dependence on, and identification with lakeshore properties. Journal of Environmental Management 79(3): 316–327.

28.

Kim

Ritchie

(2012) Motivation-based typology: an empirical study of golf tourists. Journal of Hospitality and Tourism Research 36(2): 251–280.

29.

Kyle

Graefe

Manning

, et al. (2003) An examination of the relationship between leisure activity involvement and place attachment among hikers along the Appalachian trail. Journal of Leisure Research 35(3): 249–273.

30.

Lee

Chung

Lee

(2013) Presence in virtual golf simulators: the effects of presence on perceived enjoyment, perceived value, and behavioral intention. New Media & Society 15(6): 930–946.

31.

Leonidou

Coudounaris

Kvasova

, et al. (2015) Drivers and outcomes of green tourist attitudes and behavior: sociodemographic moderating effects. Psychology & Marketing 32(6): 635–650.

32.

MacKinnon

Fairchild

Fritz

(2007) Mediation analysis. Annual Review of Psychology 58: 593–614.

33.

Mikulić

Krešić

Miličević

, et al. (2016) Destination attractiveness drivers among urban hostel tourists: an analysis of frustrators and delighters. International Journal of Tourism Research 18(1): 74–81.

34.

Moital

Dias

Machado

(2013) A cross national study of golf tourists’ satisfaction. Journal of Destination Marketing & Management 2(1): 39–45.

35.

Moore

Graefe

(1994) Attachments to recreation settings: the case of rail-trail users. Leisure Sciences 16(1): 17–31.

36.

Papatheodorou

Pappas

(2017) Economic recession, job vulnerability, and tourism decision making: a qualitative comparative analysis. Journal of Travel Research 56(5): 663–677.

37.

Pardey

Alston

Chan-Kang

, et al. (2018) The shifting structure of agricultural R&D: worldwide investment patterns and payoffs. In From Agriscience to Agribusiness. Cham: Springer, pp. 13–39.

38.

Pardey

Andrade

Hurley

, et al. (2016) Returns to food and agricultural R&D investments in sub-Saharan Africa, 1975–2014. Food Policy 65: 1–8.

39.

Pardey

Chan-Kang

Dehmer

, et al. (2014) Investments in and the economic returns to agricultural and food R&D worldwide. In Agriculture and People. Amsterdam, the Netherlands: Elsevier, pp. 78–97.

40.

Parent

(2013) Handling item-level missing data: simpler is just as good. The Counseling Psychologist 41(4): 568–600.

41.

Pereira

Correia

Schutz

(2015) Towards a taxonomy of a golf-destination brand personality: insights from the Algarve golf industry. Journal of Destination Marketing and Management 4(1): 57–67.

42.

Petrick

Backman

(2002) An examination of the construct of perceived value for the prediction of golf travelers’ intentions to revisit. Journal of Travel Research 41(1): 38–45.

43.

Podsakoff

MacKenzie

Lee

, et al. (2003) Common method biases in behavioral research: a critical review of the literature and recommended remedies. Journal of Applied Psychology 88(5): 879–903.

44.

Pretty

Chipuer

Bramston

(2003) Sense of place amongst adolescents and adults in two rural Australian towns: the discriminating features of place attachment, sense of community and place dependence in relation to place identity. Journal of Environmental Psychology 23(3): 273–287.

45.

Rao

(2018) Land fragmentation with double dividends—the case of Tanzanian agriculture. European Review of Agricultural Economics 46(4): 609–635.

46.

Rao

Hurley

Pardey

(2019) Are agricultural R&D returns declining and development dependent? World Development 122: 27–37.

47.

Sangpikul

(2018) The effects of travel experience dimensions on tourist satisfaction and destination loyalty: the case of an island destination. International Journal of Culture. Tourism and Hospitality Research 12(1): 106–123.

48.

Song

Kim

Yim

(2017) The mediating effect of place attachment on the relationship between golf tourism destination image and revisit intention. Asia Pacific Journal of Tourism Research 22(11): 1182–1193.

49.

Spencer-Oatey

(2001) Reconsidering Chinese modesty: Hong Kong and mainland Chinese evaluative judgements of compliment responses. Journal of Asian Pacific Communication 11(2): 181–201.

50.

Stylos

Bellou

Andronikidis

, et al. (2017) Linking the dots among destination images, place attachment, and revisit intentions: a study among British and Russian tourists. Tourism Management 60: 15–29.

51.

Hsu

Marshall

(2014) Understanding the relationship of service fairness, emotions, trust, and tourist behavioral intentions at a city destination in china. Journal of Travel & Tourism Marketing 31(8): 1018–1038.

52.

Urry

(1990) The Tourist Gaze: Leisure and Travel in Contemporary Societies. London: Sage Publications.

53.

Wang

Fang

Law

(2018) Effect of air quality in the place of origin on outbound tourism demand: disposable income as a moderator. Tourism Management 68: 152–161.

54.

Wang

Chen

(2015) The influence of place identity on perceived tourism impacts. Annals of Tourism Research 52: 16–28.

55.

Yoh

Pedersen

Park

(2006) Sources of information for purchasing golf clubs: personal and non-personal references. International Journal of Sports Marketing and Sponsorship 7(2): 31–41.

56.

Zainal

(2007) Case study as a research method. Jurnal Kemanusiaan bil 5(1): 1–6.

57.

Zhang

Rao

Wang

(2017) Organization, technology and management innovations through acquisition in China’s pork value chains: the case of the Smithfield acquisition by Shuanghui. Food Policy 83: 337–345.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.36 MB