Valuing forest park attributes by giving consideration to the tourist satisfaction

Abstract

Despite the economic valuation of forest park resources has triggered a great deal of interest to the research community and park management, up to now no attention has been paid to the effect of psychological factors such as tourist satisfaction, which could potentially play a considerable role in contributing to the park resource valuation. This article attempts to fill this niche by carrying out a study on the economic valuation of the Xian-Ren-Tai National Forest Park in China, taking consideration of the tourist satisfaction. In the process, the choice experiment survey was used for data collection. Then, both conditional logit (CL) and random parameter logit models were used in parameter estimations to examine the factors that could potentially impact on the tourists’ willingness to pay. Results indicate that the tourists attach the greatest importance to the park’s natural environment and traffic conditions, and yet the level of tourists’ satisfaction on their tourism experience plays a big part in explaining the valuation results. Tourists who uphold a high level of satisfaction tend to pay more for the park attribute improvement than those with less satisfaction ones. The estimated compensating variation of the overall current park attributes has reached ¥184.95 per person per trip, of which 12.83% shall be attributable to the tourist satisfaction.

Keywords

choice experiment recreation value tourist satisfaction willingness to pay

Introduction

In the tourism economic research, a nonmarket valuation study on a green space and participating in nature-based outdoor recreational activity, such as a forest park tour, has long been an important research subject and generated a great deal of interest to both research practitioners and park management authorities. One of the important reasons for that is the fact that the associated research findings have been able to make some imperative contribution to the natural resource policymakings, tourism product designs, and an overall park management (Lópezmosquera and Sánchez, 2011; Lusk and Schroeder, 2004; Mitchell and Carson, 1989). The estimated figures of willingness to pay (WTP) for different types of natural attractions are being frequently referred to as useful information for economic benefit estimation of various types of natural resources and tourism attractions (Clem and Clevo, 2006). As to the relation between WTP and tourist satisfaction under the tourism context, Harrison (1992) made an interesting point by saying that “to some extent, the respondents’ expression on WTP for a certain natural resource or tourism activity is similar to the situation where visitors spend money for buying satisfaction from recreational consumption.” In this sense, if one’s actual experience of a recreational service is better than his or her expectation, then his or her assessment of the park should be marked as “satisfactory and valuable” (Barbara, 1997). Early research on the tourists’ WTP for tourism attraction was mainly focused on the role played by the socioeconomic characteristics (Rahimi, 2011), and less attention was paid to those non-socioeconomic characteristics such as psychological factors. Specifically, no research has addressed the issue being associated with the effect of tourists’ perception toward a recreational activity, which could presumably impact on an economic valuation. According to Spash et al. (2009), standard socioeconomic explanatory variables are far inferior to those of social psychology and philosophy and that these factors offer a better understanding of the motives behind responses to contingent valuation. Therefore, it is reasonable to assume that to a certain degree, the valuation of environmental resources is related to users’ perception such as the degree of satisfaction of their travel experience.

In this study, we used the choice experiment method (CEM) for data collection and took the Xian-Ren-Tai (XRT) National Forest Park (NFP) in China as the study area to investigate how the recreational value of the NFP could be affected by tourists’ perception on their recreational experience such as the level of satisfaction. In the process, the six recreational attributes, including the domains of both natural resources and management, were considered. Although in the recent literature there were several studies adopting the CEM to investigate tourists’ WTP for some important tourism attractions, none of them gave due consideration to the factor of tourist satisfaction. Our study has three objectives: First, we want to identify those critical park attributes that could potentially make contribution to the level of tourist WTP for the XRT NFP in China. Second, we are interested in knowing to what extent the economic value of the park attributes is affected by the level of tourist satisfaction. Lastly, the revealed economic value information of the park attributes involving natural resources and management is used to address both policy and managerial implications for the nature-based recreational site management in general and for the XRT NFP in specific.

The rest of the article is organized as follows. The second section conducts literature reviews focusing on nonmarket valuation methods and factors that are able to make contribution to the tourists’ WTP. The third section introduces the model framework. The fourth section presents the study area and describes the survey designs and data collection. The fifth section presents the model results and the sixth section concludes the article.

Literature review

Methodology of valuing recreational benefits

With the fast development of the recreational industry, accurately monetizing a specific recreational resource or activity remains a challenging task due to its public good nature where the authentic market transaction data are usually absent. So far, nobody has found a universally accepted method and mechanism that can be effectively used in monetizing tourism resources in a systematic manner; although on the academic research front, there are two dominant methods that have been emerged and broadly adopted in recreation valuation for a long period of time. One is the so-called revealed preference (RP) method and the other is the stated preference (SP) method (Davis, 1963). Each of them presents some advantages and disadvantages. The RP method, such as the travel cost method, is deemed to be a more reliable value discovery method, simply because it seemingly mimics a real market situation where an economic transaction activity takes place and from which the economic value of the considered resource or recreation activity can be referred (Latinopoulos, 2014). However, the travel cost method’s strength is limited to the extent that it can only be used in estimating the use value of nonmarket resources being considered and leaves a short in assessing nonuse value of the considered resource (Loomis and Walsh, 1997). To make up this short and others presented in the RP method, the SP method was developed and used as an important tool for nonmarket resource valuation, especially in the field of monetizing the environmental goods (Arrow et al., 1993; Jin et al., 2006; León et al., 2014). Most commonly used SP methods are the contingent valuation method (CVM) and the CEM (Armbrecht, 2014; Choi et al., 2010). The CVM is a value-eliciting mechanism through a man-made market environment where the consumer’s WTP for being evaluated public good is elicited (Bennett and Blaney, 2015).

Davis (1963) is the first researcher to carry out a study on assessing the economic value of recreational activities such as hunting and camping in the Maine state of the United States. Thereafter, economic value assessing on recreation and esthetic attributes has been broadly popularized all over the world. Firoozan et al. (2012) investigated the economic value of NFPs in Iran and found that on average visitors’ WTP for the park’s recreational resources was 8216 Qatari Rial. Although a large body of literature has documented both travel cost method and CVM studies over the last half century and a great deal of success has been achieved in the area of nonmarket resource valuation in terms of both the theory buildup and empirical study, controversy and criticism remain regarding the accuracy and reliability of those prevailed valuation methods. With respect to the CVM, it commonly involves the following three concerns: the first one is its hypothetical nature of the survey questionnaire design; the second is the strategic bias for being potentially begotten from some of the survey participants who are prone to uphold some strong views on valuing resource questions, causing distorted valuation results (Akter et al., 2008; Carson, 2012; Haab et al., 2013); and the third is survey protest behavior being commonly burst during the survey process and the situation may become worse in developing countries due to the less environmental sentiment exhibited among low-income people (Lo and Jim, 2015).

The CEM is another SP method where a survey respondent is given multiple choices and asked to make trade-off decisions between the choice sets. The information is used to estimate marginal utility for a specific attribute and then the economic value can be revealed based on the relative importance of the attributes revealed by survey participants (King and Mazzotta, 2000). One of the primary advantages of the CEM is that it can be used in valuing multiple recreation resource attributes, with each having different levels of quality in one-time survey effort. This method is more cognitively burdensome for participants because the same questions are being asked multiple times (Petrolia et al., 2014); however, a good design can help to decrease the survey respondent burden to some degree (Carson and Louviere, 2011). Under the CEM, the survey participants are invoked to making choices among different attribute combination choice sets with each choice set attached to a specific level of WTP (Hoyos, 2010; Lusk and Schroeder, 2004). The CEM excels over CVM in terms of evaluating work efficiency and mitigating the problem of strategic bias and protest responses (Adamowicz et al., 2014; Hanley et al., 1998). Thus, in recent years, the choice experiment survey method has gained importance in nonmarket valuation studies. For instance, Hasanbasri et al. (2015) studied how to optimize a latent class of the traction of a city park in Kuala Lumpur, Malaysia, using CEM. It was found that the most prominent feature of the city park was leisure and recreational facilities. On the same line, Giergiczny et al. (2015) carried out a study to investigate the public choices on various forest park attribute combinations in Poland, which found that the number of tree species was the most important attribute to attract visitors. The authors suggested that protecting biodiversity and managing targeted recreational value should be the park authority’s first priority.

Another advantage of CEM embodies that it can be used in estimating both use value and nonuse value (Hoyos et al., 2012; Juutinen et al., 2014; Oleson et al., 2015; Subade and Francisco, 2014). Certainly, the CEM has been used in valuing various recreational activities such as bicycling, horse riding, and nature watching (Christie et al., 2007). In general, the economic valuation study on NFPs has been concentrated on three aspects: (i) valuing a specific attractive spot within a park; (ii) valuing individual recreational activity such as mountain climbing, hiking, backpacking, bird watching, and hunting; and (iii) valuing an entire park site. However, most of those studies were conducted in other countries, particularly in developed countries (D’Amato et al., 2016). So far, no such studies have been done in China.

Could the tourist’s perception contribute to park recreational values?

A large body of the literature reported that personal perceptions have a significant effect on the level of recreational benefits being accrued to tourism travelers. Arrow et al. (1993) were the first to introduce affective elements into valuation study on the environmental damages caused due to oil leaking. In the field of recreational economic research, it has been well recognized that there exists the complex relationship between tourist attitudes and their preferences (Kahneman et al., 1999). Along the same line, Layard (2006) pointed out that the public economic theory must be updated because the current theory is paralyzed on providing adequate explanation for the human well-being evolvement over the contemporary era. Yet some of the economic welfare theory seems to conflict with the modern, developed psychological theory. Sawe (2017) adopted a method that combined the neuroimaging data with people’s explicit preference records to examine the relationship between people’s affection and their actual money expenditure. It is anticipated that the burgeoning neuroeconomics will perhaps be able to reshape the interpretation for the role that could be played by psychological factors in valuing environmental resources (Genevsky et al., 2013; Knutson and Greer, 2008; Loewenstein, 2000; Venkatraman et al., 2015).

Similarly, there was a small but a growing number of studies that introduced psychological and philosophical factors into CVM studies in order to expand evaluating research tool box (Jin and Kyle, 2011). Several studies have concluded that the affection factor tends to play a more important role in predicting WTP accuracy than those recognizable factors (Reynisdottir et al., 2008). Since tourism satisfaction is the result of affection and recognition, it is believed to play a considerable role in economic valuation studies (Luo et al., 2011; Manning, 2003).

Tourist satisfaction is closely related to tourist perception or subjective feeling (Manning, 2003). However, a person’s perception and feeling has a broad connotation intertwining with psychology, sociology, economic, management, and so on. In an operational level, it links to resources, social environment, management conditions, and so on (Jarvis et al., 2016; Kozak, 2001), which in turn affects tourist behaviors (Baker and Crompton, 2000; Chen and Chen, 2010; Jabarin and Damhoureyeh, 2006; Jim and Chen, 2006; Lópezmosquera and Sánchez, 2011).

In a nutshell, considerable research efforts have been made in valuing recreational resources and activities. However, most of them were focused on identifying factors that are critical in contributing to the value accrued to the objects being evaluated such as social demographic factors (Choi, 2013; Reynisdottir et al., 2008). Few research studies gave consideration for the role that tourists’ affection or perception could play in park resource valuation. To fill out this gap, this study attempts to make another park resource valuation study by taking into consideration the tourist psychological traits.

Model framework

Much of the recent work in valuing different types of environmental goods is based on the random utility maximization (RUM) (Champ et al., 2003). The RUM model assumes that a utility function consists of both systematic (v) and random components (Louviere et al., 2001; Train, 2003) and can be generally presented in equation (1)

U_{i j} = v_{i} (X_{j}, β) + ε_{i j} = {β_{i}}^{′} X_{i j} + ε_{i j} (i = 1, \dots \dots, n; j = 1, \dots \dots, J)

where U_ij is the true but unobservable indirect utility of tourist i as a result of the one who chooses an alternative j, $X_{i j} = {(P_{i j}, X_{i j 1,} X_{i j 2,} \dots X_{i j N})}^{'}$ is a vector of park attributes that are related to the alternative j, P_j is the entrance price, X_jn indicates the remained attributes of n of alternative j, $β_{i} = {(α_{i}, β_{i 1,} X_{i 2,} \dots X_{i N})}^{'}$ is a vector of coefficients for the park attributes to be estimated, and ε_ij is a random error term with zero mean.

This RUM provides the theoretical foundation for constructing an empirical model based on the tourist’s choices among the competing choice alternatives. The choice problem asks tourists to choose the most preferred selection from a given choice set. The probability that tourist i will choose alternative j from a choice set can be expressed as in equation (2):

{Prob}_{i j} = Prob (U_{i j} > U_{i k}) = Prob (v_{i j} + ε_{i j} > v_{i k} + ε_{i k}) (j \neq k, \forall k \in C_{i})

where C contains all of the alternatives in the choice set. By following a routine assumption, the random errors in the RUM are independently and identically distributed following a type I extreme value distribution. Then, the predicted probability a tourist chooses alternative j can be estimated using equation (3):

{Prob}_{i} (j | C) = \frac{exp ({β_{i}}^{′} X_{i j})}{\sum_{k \in C_{i}} exp ({β_{i}}^{′} X_{i k})}

which results in the CL model being developed by Mcfadden (1974).

However, the CL model is required to fulfill the following three restrictions: (i) the preference structure is homogeneous over respondents (Hensher et al., 2005), (ii) choices conform to be independent from irrelevant alternatives assumption (Louviere et al., 2001), and (iii) all errors have the same scale parameter. One alternative for relaxing these restrictions is to assume that the parameters in equation (1) are not fixed, but to be random and follow a predetermined distributional form. Then, the heterogeneity in the sample can be captured by estimating the mean and variance of the random parameter distribution (Wuepper, 2017). This approach is referred to as random parameter logit (RPL) model (Hensher and Greene, 2003; Train, 1999). Similar to equation (1), the RPL model can be expressed using equation (4):

U_{i j} = {β_{h}}^{′} X_{i j} + ε_{i j} (i = 1, \dots \dots, n; j = 1, \dots \dots, J)

where $β_{h} = β_{m} + v_{i}$ and v_i∼N(0, $\sum β_{m}$ ), β_m are the mean values of the estimated parameters associated with each individual attribute, and v_i denotes the standard deviation of each parameter. Following Train (1999, 2003), let’s integrate the unconditional probability over the values of β_h to result in equation (5)

{Prob}_{i} (j | C) = \int \frac{exp ({β_{h}}^{′} X_{i j})}{\sum_{k} exp ({β_{h}}^{′} X_{i k})} f (β) d β

As to β’s distribution, it could have many flexible forms such as normal, lognormal, or triangular. According to Ek and Persson’s (2014) work, the results do not differ greatly across the alternative distribution assumptions (e.g. lognormal vs. triangular). Therefore, in our study, all the attribute parameters to be estimated except for the entrance price are characterized as random parameters with a normal distribution under the RPL specification.

It should be noted that the tourist satisfaction factor with a particular interest in this article cannot be directly examined in the abovementioned models because the variable does not vary across different alternatives. As such, in order to examine the effect of tourist satisfaction on WTP, we need to set up an interacted term by multiplying tourist satisfaction variable with different recreational attributes. By doing so, we can purposefully create a condition of more tourists’ heterogeneity (Champ et al., 2003). Then, the estimated attribute parameter changes could be attributed to the effect of tourist satisfaction. Additionally, it is perhaps essential to create an alternative-specific constant (ASC) for the status quo option to capture the utility associated with that option. Therefore, the ASC was included in the specification to better estimate the change in utility associated with choosing the alternatives different from the status quo (Bennett and Adamowicz, 1999; Louviere et al., 2001).

Study area and data collection

Study area

The XRT NFP in China, established in December 2002, is located at the central area of the Liaodong Peninsula in the northeastern China. It is a large topographically changing location of 2931 ha and about 17 km from Anshan city which is boasted of as China’s first steel industrial city (see Figure 1). The park is of particular famous due to its ancient pine, characteristic crags, and numerous species of flora and fauna. With a high altitude of the park land, XRT NFP holds the best position for viewing the sunrise in the early morning and offers a bird view of all outstretched surrounding areas. For example, standing at the highest point of the park, one is able to overlook the Bohai Sea area hundred miles away. That is why the XRT is deemed as one of the best locations in the region for sightseeing, landscape viewing, mountain climbing, bird watching, hiking, and so on. The site has been able to generate tremendous recreation benefits for coming visitors and the local community. Along with a fast tourism development over the last decades, the number of visitors has surged substantially to this park, causing considerable concerns about the overuse of the park resources. As a result, the park management has set up a limitation on the number of tourist entrances to the park at a level of 15,000 persons/day.

Figure 1.

Geographical coordinate of the Xian-Ren-Tai National Forest Park.

The CE survey design

Our survey questionnaire comprises of the following four parts of content. The first part is about tourists’ travel cost, number of trips taken, their travel purposes, and so on. The second part is to ask about tourists’ general assessment of their travel satisfaction. The third part is about the choice experiment survey that includes various hypothetical changes on the quality of the park resource and management choices in the sets. The final part requests tourists’ socioeconomic information.

The survey question of tourist satisfaction was based on “The China tourist satisfaction evaluation system” measured by the China Tourist Satisfaction Index, which was authorized by China National Tourism Administration in 2012. The Tourist Satisfaction Index is characterized by five categories: a flat-out satisfaction, loyalty, demand, expectation, and recommend intention. Each is measured on a five-point scale, where “1” ranks the most dissatisfied and 5 the highest satisfaction (see Table 1 for details), and other numbers are ranked between 1 and 5.

Table 1.

Measuring tourist satisfactions and their validity tests.

Tourist satisfaction scale		Score
In general, I am satisfied with this trip. This trip meets my demand. The travel meets my expectation. If there is an opportunity, I would choose to revisit this place. I would like to recommend the park to friends and relatives.		Five-point Likert-type scale: 1 “completely agree” 5 “completely disagree”
Number of items		5
Cronbach’s α coefficient		0.905
Kaiser–Meyer–Olkin measure		0.851
Bartlett’s test of sphericity	Approximately χ²	4176.075
	df.	10
	Sig.	0.000

In the CE design, a challenging task is to select the attributes and their categorizations, that is, setting up different attribute levels (Leitham et al., 2000). In the preparation phase, two focus group discussions were held in order to identify the key attributes in the XRT NFP. The first group discussion was administrated with 18 participators selected from the tourists at the site. Of which, there were 8 men and 10 women with ages ranging from 20 to 55. These survey participants were asked to identify their mostly preferable attributes while visiting the park. The discussion identified 10 attributes, including forest coverage, cleanliness, density of tourists, and environmental quality. The second focus group involved three professors and five graduate students who are deemed as research experts in the tourism management field. The main purpose of this second group discussion was to select the most important attributes from the 10 attributes identified from the first focus group discussion in order to narrow down the number of attributes to be considered in the final questionnaire design. This is a necessary step for ensuring a manageable factorial permutation of the choice sets. Besides, the literature review was also made use for the attribute’s ranking level design (Lyu, 2017; Wang et al., 2014). Finally, only six attributes were selected for the CE survey questions (see Table 2).

Table 2.

Specifications of the recreational attributes of XRT NFP.

Attribute	Attribute description	Attribute level	Type of variable	Variable name
Vegetation	The vegetation coverage	70%	0,1	Veget*
		80%	0,1	Veget^Better
		90%	0,1	Veget^Best
Infrastructure	Washroom, restaurant, road, and so on	Less developed	0,1	Infra*
		Partially developed	0,1	Infra^Better
		Well developed	0,1	Infra^Best
Rubbish	Number of pieces in a visible scope	>10	0,1	Rubbish^More
		3–10	0,1	Rubbish*
		<3	0,1	Rubbish^Less
Crowding	Number of people observed in a visible scope	<10; 20; 35*; 50; >60	Continuous	Crowding
Travel time	From downtown to XRT	>3 h	0,1	Time*
		1–3 h	0,1	Time^Less
		<1 h	0,1	Time^Least
Entrance price	Admission fee charges	¥30*; ¥35; ¥40; ¥50; ¥80	Continuous	Price

Note: XRT: Xian-Ren-Tai; NFP: National Forest Park.

* Baseline status.

A total of 2025 combinations of attributes were derived from a full factorial permutation procedure ( $3^{4} \times 5^{2} = 2025)$ . Obviously, this number of choice sets was too large to be feasible for questionnaire design and yet prohibitive cooperative burden from tourists, thus will discourage survey participators’ interest of participating in the survey interview (Carlsson and Martinsson, 2008; Ek and Persson, 2014). For this reason, an orthogonal experiment design procedure was executed (Louviere et al., 2001), which ended up with a total of 27 candidate combinations.¹ Thus, eventually three blocks of nine choice sets were made with each set including the present attribute combination, that is, the status quo. In the survey process, each respondent was asked to select a choice set from randomly assigned choice set scenarios. To help participants better differentiate one choice set from another, a few photo pictures were attached to related attribute content in the survey forms (see Figure 2).

Figure 2.

An exemplary choice set attached with visualized photo pictures.

Data collection and descriptive statistics

Before the formal tourist survey process was implemented, a statistical pretest was made through asking 54 people selected from the school campus in order to ensure whether the choice sets and the related questions presented in the questionnaire are comprehensible, meaningful, and manageable by survey participants, to test whether the photo pictures used in the survey form are appropriate, and to better serve the survey purposes. The pretests showed that the nine choice sets and associated questions are reasonable. Thereby, the formal field survey was conducted over the period of May 1–10, 2017 during China’s International Labor Day holiday. A total of 365 questionnaires were distributed, of which 328 were recovered, namely an 89% of effective responding rate. With four choice sets reviewed by each interviewee, this implies 1312 observations in total.

Table 3 summarizes the descriptive statistics of the survey sample. As shown in Table 3, the ratio of male and female is approximately equal. Most tourists’ age ranges from 41 to 60 accounting for 41%, of which 67% received high school education and 63% married with children. In terms of household income, more than 85% of the respondents earned income within the range of ¥40,000 to ¥100,000. The life satisfaction variable is phrased as “All things being considered, how satisfied are you with your life?” This variable was set up as an ordinal variable with 1 = totally dissatisfied with life and 5 = totally satisfied with life. In the sample, 39% of tourists were satisfied with their life.

Table 3.

Socioeconomic characteristics of the sampled tourists.

Variable	Characteristics	N	%	Variable	Characteristics	N	%
Gender	Male	179	54	Marital status	Unmarried	98	30
Gender	Female	149	45		Married (children = 0)	22	7
					Married (children > 0)	208	63
Age	18–24	63	19	Education	Junior high school	30	9
	25–40	113	34		High school	65	20
	41–60	134	41		Undergraduate	220	67
	61 or more	18	5		Graduate and above	13	4
HH income (per year; ¥’000)	≤¥40	137	42	Life satisfaction	Completely dissatisfied	7	2
	¥40–100	139	43		Dissatisfied	12	4
	¥100–200	44	13		Neither dissatisfied nor satisfied	90	27
	¥200 or more	8	2		Satisfied	127	39
					Completely satisfied	92	28

Note: HH: household.

Empirical results

The results from maximum likelihood estimation of both CL and RPL models and two extended versions (i.e. with added interaction terms) are presented in Table 4. For making comparison of the model performance between CL and RPL, a likelihood ratio test (pseudo-R²) was performed. The results suggest that the RPL model is preferred over the CL model, and yet the RPL model with the added interaction terms outperforms the RPL model with no interaction terms (α ≤ 0.01).

Table 4.

Summary of CL and RPL models’ results.

Attributes and interactions	CL model			RPL model
	(1)	(2)	(3)	(4)		(5)		(6)
	Mean	Mean	Mean	Mean	Standard deviation	Mean	Standard deviation	Mean	Standard deviation
ASC	1.168*** (0.192)	1.200*** (0.197)	1.234*** (0.202)	2.774*** (0.681)	—	4.288*** (1.756)	—	3.829** (1.229)	—
Price	−0.025*** (0.007)	−0.023*** (0.007)	−0.021*** (0.007)	−0.055*** (0.019)	—	−0.052** (0.027)	—	−0.044** (0.025)	—
Veget^Better	0.547*** (0.180)	0.884*** (0.252)	0.900*** (0.253)	0.313* (0.486)	−3.364*** (1.193)	0.404 (0.868)	5.891** (3.261)	0.408 (1.022)	6.350** (2.712)
Veget^Best	0.525** (0.214)	0.819** (0.347)	0.887** (0.359)	0.993** (0.523)	2.289* (1.446)	1.745 (1.073)	3.454** (1.578)	1.933 (1.015)	3.917** (1.493)
Infra^Better	0.242 (0.247)	0.528 (0.307)	0.635 (0.317)	0.885 (0.579)	0.416* (1.007)	1.780* (0.959)	0.917^* (1.232)	1.823* (0.882)	1.945* (1.102)
Infra^Best	0.511* (0.280)	0.704** (0.306)	0.782** (0.313)	0.503 (0.821)	4.200*** (1.454)	1.138 (1.111)	4.728** (3.259)	1.759 (1.094)	3.036** (1.359)
Rubbish^more	−1.545*** (0.280)	−1.798*** (0.207)	−1.934*** (0.465)	−4.969*** (1.509)	4.625*** (1.692)	−6.095*** (2.095)	4.772*** (2.381)	−5.227** (2.823)	5.419*** (2.642)
Rubbish^Less	−0.082 (0.200)	−0.003 (0.207)	−0.048 (0.209)	0.852 (0.492)	−0.305 (0.841)	0.512 (0.633)	−0.636 (1.355)	0.208 (0.574)	−0.488 (1.423)
Crowding	−0.295*** (0.073)	−0.331*** (0.085)	−0.319*** (0.088)	−0.344*** (0.145)	0.080 (0.497)	−0.665** (0.332)	0.866 (0.624)	−0.492** (0.211)	0.203 (1.457)
Time^Less	0.290* (0.180)	0.483** (0.201)	0.501** (0.205)	0.621* (0.352)	−0.102 (0.575)	0.918 (0.611)	−0.351 (1.388)	0.654 (0.448)	0.234 (0.851)
Time^Least	0.960*** (0.234)	1.074*** (0.245)	1.090*** (0.251)	1.974*** (0.543)	−0.993 (1.298)	3.200*** (1.463)	−0.849 (1.265)	2.405*** (0.804)	−1.110 (1.425)
Education × Rubbish^More	—	−0.674* (0.722)	−0.753* (0.760)	—	—	−0.998* (0.605)	—	−1.358** (0.715)	—
Education × Rubbish^Less	—	−0.650 (0.380)	−0.679 (0.386)	—		−0.623 (0.457)	—	−0.702 (0.566)	—
Education × Crowding	—	0.105 (0.102)	0.115 (0.105)	—	—	0.656 665 (0.325)	—	0.449** (0.292)	—
Income × Rubbish^More	—	−0.054** (0.731)	−0.061** (0.842)	—	—	−0.253 (0.452)	—	−0.442** (0.625)	—
Income × Rubbish^Less		0.169 (0.430)	0.345 (0.455)	—	—	0.512 (0.446)	—	0.336 (0.105)	—
Income × Crowding	—	−0.096*** (0.039)	−0.097*** (0.039)	—	—	−0.285 (0.172)	—	−0.231** (0.129)	—
Tourist satisfaction × Veget ^Better	—	—	0.021 (0.197)	—	—	—	—	0.006 (0.725)	—
Tourist satisfaction × Veget ^Best	—	—	−0.056 (0.257)	—	—	—	—	−0.399 (0.805)	—
Tourist satisfaction × Infra^Better	—	—	0.396 (0.276)	—	—	—	—	1.048 (0.875)	—
Tourist satisfaction × Infra^Best	—	—	0.454* (0.276)	—	—	—	—	1.500 (1.026)	—
Tourist satisfaction × Rubbish^More	—	—	−0.699** (0.361)	—	—	—	—	−3.081** (1.652)	—
Tourist satisfaction × Rubbish^Less	—	—	0.033 (0.211)	—	—	—	—	0.078 (0.538)	—
Tourist satisfaction × Crowding	—	—	−0.127* (0.073)	—	—	—	—	−0.552** (0.276)	—
Tourist satisfaction × Time^Less	—	—	0.376* (0.194)	—	—	—	—	0.889* (0.559)	—
Tourist satisfaction × Time^Least	—	—	0.445* (0.247)	—	—	——	—	0.738* (0.616)	—
Log-likelihood	−384.208	−375.591	−362.697	−371.234		−353.561		−346.745
McFadden pseudo R²	0.153	0.171	0.197	0.162		0.175		0.199
Number of observations	1312	1312	1312	1312		1312		1312
Prob > χ²	0.000	0.000	0.000	0.002		0.000		0.000

Note: CL: conditional logit; RPL: random parameter logit; ASC: alternative-specific constant. Standard errors are in parenthesis.

***0.01 Level of statistical significance

**0.05 Level of statistical significance.

*0.1 Level of statistical significance.

Attributes contributed to the tourists’ choices

As shown in Table 4, under both CL and RPL models, the ASC parameters are positive and significantly different from zero (α ≤ 0.01). The positive sign signifies that choosing all the alternatives other than the one of the status quo is able to improve tourists’ indirect utility (i.e. choosing the alternative of the status quo would be lowering the indirect utility to be potentially accrued to tourists), which suggests that the designated alternatives of upgraded park attributes are indicative of creating more value for the tourists at XRT NFP.

For details, let’s use CL model results (see column 2 in Table 4) to delineate the structure for tourists’ preference. As expected, all estimated coefficients of the recreational attributes are statistically significant except for “Infrastructure better” level and “Rubbish less” level. The negative coefficient of the “Entrance price” attribute reflects the fact that the tourists’ purchasing behavior is consistent with the law of demand, that is, a higher entrance admission cost shall have a negative effect on tourists’ utility. Positive coefficient signs of both “Vegetation” attribute level and “Travel time” attribute level suggest that tourists prefer more green land coverage over less land vegetation coverage and benefit from a better traffic accessibility to the park. Conversely, a negative coefficient being estimated for the “Crowding” variable implies that tourists strongly dislike noisy recreational environment. This may be used to explain why most tourists prefer to have a lower level of congestion in the parks. It is also clear that tourists prefer the best level of “Infrastructure” such as adequate convenient facilities being available at site. It is interesting to note that the “More rubbish” status results in a statistically negative coefficient, whereas “Less rubbish” is not statistically significant at conventional levels. This may suggest that a threshold might exist in terms of the amount of rubbish present, that is, it is not acceptable for a tourist to experience rubbish amount over this limit. Nevertheless, it also implies that tourists do not care for any further rubbish reduction from this level of threshold.

For the RPL model (columns 5 and 6), the overall estimated coefficients are very similar to those generated by CL models because they all exhibit the same signs of the estimated coefficients. However, there exist some differences between them. For instance, under the RPL model, the coefficient of the “best” level of infrastructure variable is positive but not significant. It might be probably due to the fact that most visitors take a 1-day trip to the XRT park, thus they are much less sensitive to the changes of the infrastructure conditions. With respect to standard deviations of the random parameter estimates, five of nine standard deviations are statistically significant under the RPL model with interaction terms included. This result confirms that some heterogeneity is embedded into tourists’ preferences on the designed park attribute combinations.

In order to analyze the effect of tourists’ personal characteristic on their choice decisions, we further estimate both CL and RPL models by incorporating interaction terms of the park attributes and tourists’ sociodemographic characteristics (columns 3, 7, and 8).² The results show that tourists with different levels of education seem to value “Crowding” differently. Someone with a better education level tends to uphold much negative view toward crowded and noisy conditions than those with less education. As noted by Vandermersch and Mathijs (2004), better educated people have a wider selection on tourism goods and services. Likewise, households with better incomes appear to have lower tolerant threshold to park “Crowding.” Along the same line, according to the estimated result of “Rubbish^More,” higher income tourists tend to have a much less willfulness of accepting undesirable park conditions.

As being hypothesized, tourist satisfaction would affect their WTP for the park attributes considered, which has been proved by models (3) and (6) (columns 4, 9, and 10) where both CL and RPL models’ performance has improved as the interaction variables of tourist satisfaction and park attributes are considered according to R² in the CL model and pseudo-R² in the RPL model. Specifically, R² surges from 0.171 to 0.197 under the CL model and pseudo-R² is ratcheted up from 0.175 to 0.199 under the RPL model. The results also show that tourists with different levels of satisfaction value attributes differently. For example, with a positive estimated coefficient sign such as “Tourist Satisfaction × Time” and “Tourist Satisfaction × Infra^Best,” it implies that tourists with a high level of trip satisfaction tend to value higher on park attributes than those with lower travel satisfaction. Similarly, a significantly negative coefficient estimated for both variables of “Tourist Satisfaction × Rubbish^More” and “Tourist Satisfaction × Crowding” supports the evidence that less satisfied tourists would have a low WTP for the tourism products. This result echoes Lópezmosquera and Sánchez’s (2011) study, whose report indicates that more satisfied visitors over the period of visiting a green space area impose a positive effect on their WTP.

Preference heterogeneity induced by tourist satisfaction

According to the tourist satisfaction scores, we subdivided the whole survey data sample into two subsamples, that is, low tourist satisfaction group (1≤ satisfaction score ≤4) and high tourist satisfaction group (4< satisfaction score ≤5). To quantify the effect of tourist satisfaction on park attribute in monetary term, the RPL model is reestimated using the two subsamples of data, and the estimated results are presented in Table 5. Similar to the results generated from using the pooled data, the two models perform well (α ≤ 0.05) according to the magnitude of log-likelihood.

Table 5.

Estimated results of RPL model under different satisfaction groups.

Attributes	Low satisfaction group				High satisfaction group
	Mean		Standard deviation		Mean		Standard deviation
	Coefficient	SE	Coefficient	SE	Coefficient	SE	Coefficient	SE
ASC	2.585***	0.720			3.021***	0.916
Price	−0.046***	0.019			−0.036**	0.017
Veget^Better	−0.345	0.681	2.585**	1.342	0.290**	0.833	2.630	2.507
Veget^Best	0.046	0.580	1.506	1.172	0.380**	0.669	3.044**	1.474
Infra^Better	0.634	0.642	0.205	1.056	0.579	0.630	−0.141	1.006
Infra^Best	1.212	0.863	−1.543	1.471	0.716	1.022	1.813	1.526
Rubbish^More	−4.083**	2.034	−4.907**	2.135	−3.802***	1.878	−5.278**	2.576
Rubbish^Less	0.976	0.470	−0.184	0.721	1.192	0.540	−0.251	0.905
Crowding	−0.461**	0.209	0.014	0.543	−0.432**	0.190	−0.051	0.435
Time^Less	0.231	0.486	−0.776	0.886	0.285	0.507	−0.217	0.758
Time^Least	1.295**	0.642	−0.686	0.957	1.427***	0.601	−1.074	1.204
Log-likelihood	−162.062				−198.812
McFadden Pseudo R²	0.162				0.189
Number of observations	596				716
Prob> χ²	0.002				0.006

Note: ASC: alternative-specific constant; SE: standard error.

***0.01 Level of statistical significance

**0.05 Level of statistical significance.

*0.1 Level of statistical significance.

As shown in Table 5, ASC is significant (α ≤ 0.05) under the two models, suggesting that all tourists prefer to change the status quo to a better condition. Neither the high satisfaction group nor the low satisfaction group inclines to raise the park entrance price given the current attribute condition, that is, if the entrance fee rises without coupling with attribute improvement, the level of all tourists’ utility would be decreased. This could be interpreted as a non-satiation behavior where tourists would like to obtain a better recreational experience from the given level of entrance fee payment. For the high satisfaction group (columns 6–9), all significant coefficients, representing the mean value of an individual random parameter, show an identical signal pattern with those resulted from using the pooled data set except for the variable of “Less” travel time. Thus, we can conclude that tourist utility shall be enhanced under the conditions of park attributes improvement, such as lifting up vegetation coverage, suppressing observable rubbish, more convenient traffic, and less park congestion. In comparison, the low tourism satisfaction group shows less interest in increasing the vegetation coverage of the park. A possible reason of this could be that the visitors with low satisfaction might not be very keen to the vegetation coverage of the site.

In sum, it appears that tourists’ preference heterogeneity could be attributed to a degree of their travel satisfaction rather than park attributes, and tourists tend to be more skeptical about their evaluative judgment when they have negative experiences during their travel praxis. More accurate relation analysis on WTP elicited from both high and low satisfaction groups will be conducted in the next section.

Estimating tourists’ WTP

Economic value of the park resources and services is measured by tourists’ WTP, which is conducted by directly calculating the marginal rate of substitution between coefficients of attributes and entrance price variable (Lim et al., 2013). The formula used for the estimation is presented in the following equation

{WTP}_{j} = - \frac{β_{j}}{β_{c}}

where β_j represents the estimated coefficients of each individual non-price park attribute j and β_c denotes the coefficient of the entrance price variable. However, the ratio may lead to unrealistic WTP measures or unidentified distributions if both the price attribute and the other attributes are set to be randomly distributed (Daly et al., 2012). Therefore, a commonly used solution is to hold the price parameter fixed so that the WTP distribution is identical to the distribution of the attribute coefficients (Revelt and Train, 1998, 2000), which simplifies the derivation of WTP’s distribution.³

Table 6 presents the mean WTP results estimated from using the whole sample data and subsample data, as well as their associated 95% confidence intervals that are calculated using the parametric bootstrapping method with 1000 simulated estimates (Krinsky and Robb, 1986). Since the RPL model shows a better fit than the CL model does, only the RPL model’s coefficients are utilized in the WTP estimation here.

Table 6.

WTP estimates under whole data set and subsample data sets representing different satisfaction groups.

	Whole data sample		High satisfaction group		Low satisfaction group
Attributes	WTP	95% Confidence Interval	WTP	95% Confidence Interval	WTP	95% Confidence Interval
Veget^Better	5.724*	[−12.097, 23.545]	7.968**	[−53.064, 37.128]	1.008	[−36.934, 21.919]
Veget^Best	18.156**	[−2.911, 39.224]	10.429**	[−47.116, 26.259]	−7.507	[−23.782, 25.799]
Infra^Better	16.180	[−3.642, 36.003]	15.878	[−20.020, 51.776]	13.772	[−15.384, 42.928]
Infra^Best	9.202	[−18.468, 36.872]	19.648	[−32.234, 71.529]	26.344	[−8.186, 60.874]
Rubbish^More	−72.586***	[−144.799, −0.372]	−104.304***	[−241.954, 33.346]	−88.761**	[−211.607, 16.767]
Rubbish^Less	15.573	[1.504, 29.641]	32.695	[−2.458, 67.849]	21.209	[−1.596, 44.012]
Crowding	−6.289***	[−13.199, 0.619]	−11.860**	[−26.955, 3.235]	−10.011**	[−21.859, 1.838]
Time^Less	11.356*	[−2.148, 24.859]	7.819	[−21.376, 37.013]	5.011	[−16.622, 26.644]
Time^Least	36.098***	[13.182, 59.015]	39.159***	[−6.369, 84.685]	28.143**	[−3.776, 60.063]

Note: WTP: willingness to pay.

***0.01 Level of statistical significance

**0.05 Level of statistical significance.

*0.1 Level of statistical significance.

It has been well understood that in a choice experiment model, the results of WTP measures are particularly useful in assessing the relative importance of a park attribute and the trade-off effect between different choosing criteria being used (Blamey et al., 1999; Chen and Chen, 2012). As indicated in Table 6 (columns 2 and 3), the attribute with the strongest effect on tourists’ utility (choice) is “Rubbish” whose negative sign and magnitude suggests that tourists need to be given some form of compensation in order for their preference to be changed from the status quo to accept for “More rubbish.” Previous research also suggested that rubbish may be an important issue to tourism demand and number of visitors being attracted (Siano and Siglioccolo, 2011). The attribute of “Travel time” serves as another important factor for tourists to consider while choosing a forest travel product. The reason is that, against the backdrop of a well-developed transportation system in China, private car driving has become a common choice for most tourists while making a tourism travel. In addition, tourists are willing to pay ¥6.29 for a marginal reduction of the park crowding.

As expected, differences present in terms of visitors’ preference over the park attribute between the two tourist groups. Of all the recreational attributes except for “Infrastructure” being not significant, the high satisfaction group exhibits a higher level of WTP than that of the low satisfaction group.

In the welfare economic analysis, the Hicksian welfare measure of compensating variation (CV) is commonly used in estimating consumers’ well-being for different quality of environmental goods, which is deemed to be equivalent to the WTP measurement. By definition, the CV is the amount of money that must be given to or taken away from a tourist in order to ensure him or her the same level of well-off before and after a change takes place (Champ et al., 2003). Equation (7) is adopted for CV calculation here⁴

CV = \frac{1}{β_{c}} [ln (\sum_{j = 1}^{J} exp V_{j}^{1}) - ln (\sum_{j = 1}^{J} exp V_{j}^{0})]

where V₀ and V₁ are the expressions of utility for the base and altered cases, respectively. Note that if V₀ and V₁ are linear in attributes, and the goal is to evaluate a change in a single attribute, equation (7) reduces to the ratio of the attribute coefficient and the marginal utility of money assuming zero income effect. Using the model results presented in Table 4, the CV is estimated at ¥184.95 per person per trip under the base model (model 3). However, as the influence of tourist satisfaction factor is excluded, the magnitude of CV is reduced to ¥161.22 per person per trip (model 2). As expected, the CV is increased with improved tourist satisfaction. By contrast, the CV for the low satisfaction group is estimated at ¥121.36 per person per trip, whereas it is ¥181.11 per person per trip for the high satisfaction group (Table 5). According to the annual statistical report assembled by the Department of Anshan Tourism Administration, on average, there were 50,000 people who participated in outdoor recreation activities every day, of whom about one-fifth chose to visit the XRT NFP. This could be translated into approximately ¥675.06 million of economic value being accrued to the park annually, of which 12.83% could be attributable to the tourists’ satisfaction factor. Thus, tourist satisfaction on the tour could play a vital role in contributing to the economic value of park resources and activities. Negligence of this psychological factor would most likely result in a downward biased economic value estimation.

Conclusions and implications

Estimating economic value of the tourism resources potentially accrued to the NFPs provides essential information for the park authority and resource policymakers due to its close relevancy to the resource uses and park management. The general uses of the economic valuation data could include but not limit to the park admission fee management, tourism product design, compensation for ecological damages, and so on. However, owing to its public or quasi-public good nature, the true value of the NFP resources is hardly revealed through the market transaction process. As such, the nonmarket valuation technique such as CEM has been widely used in valuing nature-based park resources. However, for many of these studies, researchers have commonly ignored the role that tourists’ satisfaction attitude could play in explaining the valuation outcomes. In this article, we put tourist satisfaction factor into the choice experiment survey design to investigate the effect it could potentially impose on the park attribute values.

To this end, both CL and RPL models are used in parameter estimations. It turns out that the RPL model outperforms the more restrictive CL model. The estimate results lead us to the following conclusions: (i) there are five park attributes which are able to make a statistically significant contribution to the tourists’ choices. In general, the increase of the rubbish amount and a worsening park’s congestion condition impose significantly negative effects on the tourists’ indirect utility. On the contrary, tourists tend to uphold a positive attitude toward the improved attributes of vegetation coverage and shortened travel time. Among those critical NFP attributes, tourists are WTP the most for reducing the number of rubbish on the site. (ii) The result of the RPL model with the interaction term included supports the hypothesis that tourists’ satisfaction plays an essential role in their valuation process and decisive choices. Compared with the high satisfied tourists, the less satisfied tourists exhibit less WTP for the entrance fees, and so does the CV. (iii) The total economic value of those considered park attributes at XRT is estimated to be ¥675.06 million per year. Of which, 12.83% shall be attributable to the factor of tourist satisfaction.

Those analytical results can be used to shed light on several policy and managerial implications for the NFP governance. Firstly, according to the visitors’ WTP, the vast majority of tourists have the desire for further improving the park resources and management from the current conditions. This means that the current park condition in terms of the park attribute combination is not adequate for satisfying tourists’ demand. Specifically, the park visitors display some diversified preferences in terms of the park attribute combinations. The tourists’ strong preferences on the attributes of rubbish reduction and traffic convenience imply that the park management ought to exert efforts on park sanitation work and traffic access improvement. Secondly, the revealed economic value results of the park tourism resources provide essential information for presumptive cost–benefit analysis of the park investment. As long as the benefits accrued to a particular resource attribute or recreation activity are larger than its costs, any management effort will become worthwhile and justifiable from the economics point of view.

Nowadays, the sustainable tourism becomes a buzzword where researchers and park authority tend to establish some sustainable performance indicators based on the monetized costs and benefits of policies. As Liu et al. (2016) noted, sustainable tourism is capable of proving tremendous benefits to the tourists’ experience. However, in the past decades, the rubbish problem has remained a challenge to the sustainable tourism site development in the XRT NFP. One possibly effective way of dealing with the rubbish problem should still be focused on tourist education or periodically carrying out rubbish cleaning up campaign at the park mountain areas. Besides, it might be advisable to design an incentive plan to award those visitors who consciously conduct rubbish collection activity while touring at the site. The next low-hanging fruit perhaps is the congestion control at the park. It seems that there is a conflict between the national holiday system currently implemented in China and its goal of natural resource protection. The former’s goal is to stimulate the general public to engage in more tourism activities during the centrally mandated holiday periods, such as “May 1st Golden Week” and “October 1st Golden Week.” During these holiday seasons in China, billions of people make travels domestically and internationally, imposing a tremendous pressure on the NFP resources and environment. Thus, there is a distinct trade-off between the nationwide tourism motivating plan and the park carrying capacity management. It might be wise and necessary as well for the nation to carry out a reform on the current holiday system for the sake of safeguarding a sustainable nature-based tourism development in China.

One caveat with respect to the amount of the tourists’ WTP for the XRT NFP is that the estimated value information is only to be useful for the short-run policy and management analysis, because it is reasonable to assume that the tourists’ preference pattern is relatively stable within a short period of time. As the time goes by, the pattern of tourists’ preferences over the park attributes may be changed. As a result, any study based on the cross-sectional data analysis such as this research could bring about both heteroscedasticity and consistency problems, which is likely a limitation of doing a choice experiment study like ours. For this reason, the temporal limitation of the SPs must be recognized for a long-term valuation study in order to extract an adaptive tourism resource conservation strategy.

Supplemental material

Appendix_A - Valuing forest park attributes by giving consideration to the tourist satisfaction

Appendix_A for Valuing forest park attributes by giving consideration to the tourist satisfaction by Nannan Kang, Erda Wang and Yang Yu 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) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The research was funded by the National Natural Science Foundation of China (Grant No 71640035).

Supplemental material

Supplemental material for this article is available online.

Notes

References

Adamowicz

Glenk

Meyerhoff

(2014) Choice modelling research in environmental and resource economics. In: Hess

Daly

(eds) Handbook of Choice Modelling. Cheltenham: Edward Elgar, pp. 661–674.

Akter

Bennett

Akhter

(2008) Preference uncertainty in contingent valuation. Ecological Economics 67(3): 345–351. DOI: 10.1016/j.ecolecon.2008.07.009.

Armbrecht

(2014) Use value of cultural experiences: A comparison of contingent valuation and travel cost. Tourism Management 42(3): 141–148. DOI: 10.1016/j.tourman.2013.11.010.

Arrow

Solow

Portney

. (1993) Report to the national oceanic and atmospheric administration panel on contingent valuation. Federal Register 58: 4601–4614.

Baker

Crompton

(2000) Quality, satisfaction and behavioural intentions. Annals of Tourism Research 27(3): 425–439.

Barbara

(1997) The service profit chain: How leading companies link profit and growth to loyalty, satisfaction, and value. International Journal of Service Industry Management 9(3): 312–313.

Bennett

Adamowicz

(1999) Some Fundamentals of Environmental Choice Modelling. Canberra: School of Economics and Management, University College, The University of New South Wales.

Bennett

Blaney

RJP

(2015) Estimating the benefits of farm animal welfare legislation using the contingent valuation method. Agricultural Economics 29(1): 85–98. DOI: 10.1016/S0169-5150(03)00037-9.

Blamey

Gordon

Chapman

(1999) Choice modelling: Assessing the environmental values of water supply options. Australian Journal of Agricultural and Resource Economics 43(3): 337–357.

10.

Carlsson

Martinsson

(2008) Does it matter when a power outage occurs? — A choice experiment study on the willingness to pay to avoid power outages. Energy Economics 30(3): 1232–1245. DOI: 10.1016/j.eneco.2007.04.001.

11.

Carson

(2012) Contingent valuation: A practical alternative when prices aren’t available. Journal of Economic Perspectives 26(4): 27–42.

12.

Carson

Louviere

(2011) A common nomenclature for stated preference elicitation approaches. Environmental and Resource Economics 49(4): 539–559.

13.

Castellani

Pattitoni

Vici

(2015) Pricing visitor preferences for temporary art exhibitions. Tourism Economics 21: 83–103.

14.

Champ

Boyle

Brown

(2003) A Primer on Nonmarket Valuation. Berlin: Springer Netherlands.

15.

Chen

(2010) Experience quality, perceived value, satisfaction and behavioral intentions for heritage tourists. Tourism Management 31(1): 29–35. DOI: 10.1016/j.tourman.2009.02.008.

16.

Chen

(2012) Research note: Exploring tourists’ stated preferences for heritage tourism services - the case of Tainan city, Taiwan. Tourism Economics 18(2): 457–464.

17.

Choi

(2013) Nonmarket values of major resources in the Korean DMZ areas: A test of distance decay. Ecological Economics 88(7): 97–107. DOI: 10.1016/j.ecolecon.2013.01.014.

18.

Choi

Ritchie

Papandrea

. (2010) Economic valuation of cultural heritage sites: A choice modeling approach. Tourism Management 31(2): 213–220. DOI: 10.1016/j.tourman.2009.02.014.

19.

Christie

Hanley

Hynes

(2007) Valuing enhancements to forest recreation using choice experiment and contingent behaviour methods. Journal of Forest Economics 13(2): 75–102. DOI: 10.1016/j.jfe.2007.02.005.

20.

Clem

Clevo

(2006) Information, wildlife valuation, conservation: Experiments and policy. Contemporary Economic Policy 24(1): 144–159. DOI: 10.1093/cep/byj014.

21.

Daly

Hess

Train

(2012) Assuring finite moments for willingness to pay in random coefficient models. Transportation 39(1): 19–31. DOI: 10.1007/s11116-011-9331-3.

22.

D’Amato

Rekola

. (2016) Monetary valuation of forest ecosystem services in China: A literature review and identification of future research needs. Ecological Economics 121: 75–84. DOI: 10.1016/j.ecolecon.2015.11.009.

23.

Davis

(1963) Recreation planning as an economic problem. Nature Resources Journal 3(2): 239–249.

24.

Persson

(2014) Wind farms — Where and how to place them? A choice experiment approach to measure consumer preferences for characteristics of wind farm establishments in Sweden. Ecological Economics 105(9): 193–203. DOI: 10.1016/j.ecolecon.2014.06.001.

25.

Ferrini

Scarpa

(2007) Designs with a priori information for nonmarket valuation with choice experiments: A Monte Carlo study. Journal of Environmental Economics and Management 53(3): 342–363.

26.

Firoozan

Abed

Bahmanpour

. (2012) Estimated recreational value of Lahijan forest using by contingent valuation method. Journal of Agricultural and Biological Science. Available at: http://www.cabdirect.org/abstracts/20123393087.html (accessed 6 November 2017).

27.

Genevsky

Västfjäll

Slovic

. (2013) Neural underpinnings of the identifiable victim effect: Affect shifts preferences for giving. Journal of Neuroscience the Official Journal of the Society for Neuroscience 33(43): 17188.

28.

Giergiczny

Czajkowski

Żylicz

. (2015) Choice experiment assessment of public preferences for forest structural attributes. Ecological Economics 119: 8–23. DOI: 10.1016/j.ecolecon.2015.07.032.

29.

Haab

Interis

Petrolia

. (2013) From hopeless to curious? Thoughts on Hausman’s “dubious to hopeless” critique of contingent valuation. Working Papers 35(4): 593–612.

30.

Hanley

Wright

Adamowicz

(1998) Using choice experiments to value the environment. Environmental and Resource Economics 11(3–4): 413–428. DOI: 10.1023/A:1008287310583.

31.

Harrison

(1992) Valuing public goods with the contingent valuation method: A critique of Kahneman and Knetsch. Journal of Environmental Economics and Management 23(23): 248–257. DOI: 10.1016/0095-0696(92)90003-F.

32.

Hasanbasri

Karim

Bakar

(2015) Willingness to pay for recreational attributes of public parks: A choice experiment approach. Singapore Economic Review 60(05): 1550048-1–1550048-18. DOI: 10.1142/S0217590815500484.

33.

Hensher

Greene

(2003) The mixed logit model: The state of practice. Transportation 30(2): 133–176.

34.

Hensher

Rose

Greene

(2005) Applied Choice Analysis: A Primer. Cambridge: Cambridge University Press.

35.

Hoyos

(2010) The state of the art of environmental valuation with discrete choice experiments. Ecological Economics 69(8): 1595–1603. DOI: 10.1016/j.ecolecon.2010.04.011.

36.

Hoyos

Mariel

Pascual

. (2012) Valuing a Natura 2000 network site to inform land use options using a discrete choice experiment: An illustration from the Basque Country. Journal of Forest Economics 18(4): 329–344. DOI: 10.1016/j.jfe.2012.05.002.

37.

Jabarin

Damhoureyeh

(2006) Estimating the recreational benefits of Dibeen national park in Jordan using contingent valuation and travel cost methods. Pakistan Journal of Biological Sciences 9(12): 2198–2206.

38.

Jarvis

Stoeckl

Liu

(2016) The impact of economic, social and environmental factors on trip satisfaction and the likelihood of visitors returning. Tourism Management 52: 1–18.

39.

Jim

Chen

(2006) Recreation-amenity use and contingent valuation of urban greenspaces in Guangzhou, China. Landscape and Urban Planning 75(1): 81–96. DOI: 10.1016/j.landurbplan.2004.08.008.

40.

Jin

Wang

Ran

(2006) Comparison of contingent valuation and choice experiment in solid waste management programs in Macao. Ecological Economics 57(3): 430–441. DOI: 10.1016/j.ecolecon.2005.04.020.

41.

Jin

Kyle

(2011) Fairness of prices, user fee policy and willingness to pay among visitors to a national forest. Tourism Management 32(5): 1038–1046. DOI: 10.1016/j.tourman.2010.08.016.

42.

Juutinen

Kosenius

Ovaskainen

(2014) Estimating the benefits of recreation-oriented management in state-owned commercial forests in Finland: A choice experiment. Journal of Forest Economics 20(4): 396–412. DOI: 10.1016/j.jfe.2014.10.003.

43.

Kahneman

Ritov

Schkade

(1999) Economic preferences or attitude expressions? An analysis of dollar responses to public issues. Journal of Risk & Uncertainty 19(1/3): 203–235. DOI: 10.1023/A:1007835629236.

44.

King

Mazzotta

(2000) Ecosystem valuation. Available at: http://www.ecosystemvaluation.org/default.htm (accessed 20 November 2017).

45.

Knutson

Greer

(2008) Anticipatory affect: Neural correlates and consequences for choice. Philosophical Transactions of the Royal Society B: Biological Sciences 363(1511): 3771–86. DOI: 10.1098/rstb.2008.0155.

46.

Kozak

(2001) Comparative assessment of tourist satisfaction with destinations across two nationalities. Tourism Management 22(4): 391–401. DOI: 10.1016/S0261-5177(00)00064-9.

47.

Krinsky

Robb

(1986) On approximating the statistical properties of elasticities. Review of Economics and Statistics 68(4): 715–719.

48.

Kuhfeld

Tobias

Garratt

(1994) Efficient experimental design with marketing research applications. Journal of Marketing Research 31(4): 545–557.

49.

Laitila

Paulrud

Fredman

. (2008) Anglers’ valuation of water regulation dam removal for the restoration of angling conditions at Storsjö-Kapell. Tourism Economics 14(2): 283–296. Available at: http://www.cabdirect.org/abstracts/20083144145.html (accessed 19 June 2018).

50.

Latinopoulos

(2014) The impact of economic recession on outdoor recreation demand: An application of the travel cost method in Greece. Journal of Environmental Planning and Management 57(2): 254–272.

51.

Layard

(2006) Happiness and public policy: A challenge to the profession. Economic Journal 116(510): C24–C33. DOI: 10.1111/j.1468-0297.2006.01073.x.

52.

Lazari

Anderson

(1994) Designs of discrete choice set experiments for estimating both attribute and availability cross effects. Journal of Marketing Research 31(3): 375–383.

53.

Leitham

Mcquaid

Nelson

(2000) The influence of transport on industrial location choice: a stated preference experiment. Transportation Research Part A Policy and Practice 34(7): 515–535. DOI: 10.1016/S0965-8564(99)00030-0.

54.

León

Araña

Hanemann

. (2014) Heterogeneity and emotions in the valuation of non-use damages caused by oil spills. Ecological Economics 97(3): 129–139. DOI: 10.1016/j.ecolecon.2013.10.010.

55.

Lim

Maynard

. (2013) U.S. consumers’ preference and willingness to pay for country-of-origin-labeled beef steak and food safety enhancements. Canadian Journal of Agricultural Economics/revue Canadienne Dagroeconomie 61(1): 93–118. DOI: 10.1111/j.1744-7976.2012.01260.x.

56.

Liu

Horng

Chou

. (2016) An empirical examination of the form of relationship between sustainable tourism experiences and satisfaction. Asia Pacific Journal of Tourism Research (7): 1–24.

57.

Jim

(2015) Protest response and willingness to pay for culturally significant urban trees: Implications for contingent valuation method. Ecological Economics 114: 58–66. DOI: 10.1016/j.ecolecon.2015.03.012.

58.

Loewenstein

(2000) Emotions in economic theory and economic behavior. American Economic Review 90(2): 426–432.

59.

Loomis

Walsh

(1997) Recreation Economic Decisions: Comparing Benefits and Costs. Pennsylvania: Venture Publishing.

60.

Lópezmosquera

Sánchez

(2011) Emotional and satisfaction benefits to visitors as explanatory factors in the monetary valuation of environmental goods. An application to Periurban green spaces. Land Use Policy 28(1): 151–166. DOI: 10.1016/j.landusepol.2010.05.008.

61.

Louviere

Hensher

Swait

(2001) Stated Choice Methods: Analysis and Application. Cambridge: Cambridge University Press.

62.

Luo

Huang

Cheng

. (2011) Study on the effect of affective factors on tourist experience and satisfaction——taking the performance” impression of liusanjie” in Guilin as an example. Tourism Tribune 26(1): 51–58.

63.

Lusk

Norwood

(2005) Effect of experimental design on choice-based conjoint valuation estimates. American Journal of Agricultural Economics 87(3): 771–785.

64.

Lusk

Schroeder

(2004) Are choice experiments incentive compatible? A test with quality differentiated beef steaks. American Journal of Agricultural Economics 86(2): 467–482.

65.

Lyu

(2017) Which accessible travel products are people with disabilities willing to pay more? A choice experiment. Tourism Management 59: 404–412. DOI: 10.1016/j.tourman.2016.09.002.

66.

Manning

(2003) What to do about crowding and solitude in parks and wilderness? A reply to Stewart and Cole. Journal of Leisure Research 35(1): 107–118.

67.

Mcfadden

(1974) Conditional Logit Analysis of Qualitative Choice Behavior. New York: Academic Press.

68.

Mitchell

Carson

(1989) Using Surveys to Value Public Goods: The Contingent Valuation Method. Washington, DC: Resources for the Future /Johns Hopkins University Press.

69.

Neuts

(2016) An econometric approach to crowding in touristic city centres: Evaluating the utility effect on local residents. Tourism Economics 22(5): 1055–1074.

70.

Oleson

KLL

Barnes

Brander

. (2015) Cultural bequest values for ecosystem service flows among indigenous fishers: A discrete choice experiment validated with mixed methods. Ecological Economics 114: 104–116. DOI: 10.2139/ssrn.2516507.

71.

Petrolia

Interis

Hwang

(2014) America’s Wetland? A national survey of willingness to pay for restoration of Louisiana’s coastal wetlands. Social Science Electronic Publishing 29(1): 17–37. DOI: 10.2139/ssrn.2054867.

72.

Rahimi

(2011) Estimated value of forest recreation parks and identify factors influencing willingness to pay Bonab Arsanjan visitors (tourists). Australian Journal of Basic and Applied Sciences 5(12): 1802–1805.

73.

Revelt

Train

(1998) Mixed logit with repeated choices: Households’ choices of appliance efficiency level. Review of Economics & Statistics 80(4): 647–657.

74.

Revelt

Train

(2000) Customer-specific taste parameters and mixed logit: Households’ choice of electricity supplier. Economics Working Papers 19(9): 934–938.

75.

Reynisdottir

Song

Agrusa

(2008) Willingness to pay entrance fees to natural attractions: An Icelandic case study. Tourism Management 29(6): 1076–1083. DOI: 10.1016/j.tourman.2008.02.016.

76.

Sawe

(2017) Using neuroeconomics to understand environmental valuation. Ecological Economics 135: 1–9. DOI: 10.1016/j.ecolecon.2016.12.018.

77.

Siano

Siglioccolo

(2011) Rubbish emergency impact on in-bound tourism demand and on the number of visitors to museums. International Journal of Culture Tourism and Hospitality Research 5(1): 69–79.

78.

Spash

Urama

Burton

. (2009) Motives behind willingness to pay for improving biodiversity in a water ecosystem: Economics, ethics and social psychology. Ecological Economics 68(4): 955–964. DOI: 10.1016/j.ecolecon.2006.09.013.

79.

Subade

Francisco

(2014) Do non-users value coral reefs?: Economic valuation of conserving Tubbataha Reefs, Philippines. Ecological Economics 102(102): 24–32. DOI: 10.1016/j.ecolecon.2014.03.007.

80.

Train

(1999) Mixed logit models for recreation demand. In: Valuing Recreation and the Environment: Revealed Preference Methods in Theory and Practice. Northampton: Edward Elgar.

81.

Train

(2003) Discrete Choice: Methods with Simulation. Cambridge: Cambridge University Press.

82.

Vandermersch

Mathijs

(2004) Consumer Willingness to Pay for Domestic Milk. Leuven: Katholieke Universiteit Leuven, Centre for Agricultural and Food Economics.

83.

Venkatraman

Dimoka

Pavlou

. (2015) Predicting advertising success beyond traditional measures: New insigh. Journal of Marketing Research 52: 436–452.

84.

Wang

Wei

(2014) Valuing natural and non-natural attributes for a national forest park using a choice experiment method. Tourism Economics 20(6): 1199–1213.

85.

Wuepper

(2017) What is the value of world heritage status for a German national park? A choice experiment from Jasmund, 1 year after inscription. Tourism Economics 23(5): 1114–1123.

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