Length of Stay and Sustainability: Evidence from the Schist Villages Network (SVN) in Portugal

sustainability Article Length of Stay and Sustainability: Evidence from the Schist Villages Network (SVN) in Portugal Fidel Martínez-Roget 1,* , JoséAlberto Moutela 2and XoséA. Rodríguez 3 1Department of Applied Economics, University of Santiago de Compostela, Avda. Xoán XIII, s/n, 15782 Santiago de Compostela, Spain 2Doctoral Programme in Economics and Business, University of Santiago de Compostela, Avda. Xoán XIII, s/n, 15782 Santiago de Compostela, Spain; [email protected] 3Department of Quantitative Economics, University of Santiago de Compostela, Avda. Xoán XXIII, s/n, 15782 Santiago de Compostela, Spain; [email protected] *Correspondence: [email protected]; Tel.: +34-881-811-51 Received: 13 April 2020; Accepted: 11 May 2020; Published: 14 May 2020   Abstract: There is evidence that increasing the Length of Stay (LOS) contributes to the sustainability of destinations. This paper analyzes LOS determinants in the Schist Villages Network (SVN) in Portugal, where almost half of the visitors are same-day visitors. Given that these visitors generate more environmental impacts, an increase in LOS would help guide tourism towards more sustainable standards. The analysis was performed using several Heckman selection models. Unlike the findings of previous studies, LOS depended mainly on the characteristics of the trip and the destination—not on visitor profile—and can therefore be considered a highly specific form of tourism. Comparatively, motivations associated with rest, residents’ hospitality, and local produce consumption encourage visitors to extend their stay, thereby boosting tourism sustainability. The SVN is also shown to have various cultural and heritage resources that are currently highly valued. However, the leisure offer is insufficient to increase LOS. Furthermore, conventional advertising was not shown to be effective in raising LOS compared with word-of-mouth (WOM). The paper concludes that measures should be adopted to increase the local produce offer, retain population, maintain traditional village activities, and attract visitors from nearby regions. These actions could increase overnight stays and LOS, contributing to more sustainable development. Keywords: length of stay; sustainability; rural areas; motivation; Heckman models 1. Introduction Tourism has become one of the world’s most dynamic industries. International tourist arrivals have practically tripled in the last two decades, rising from 531 million in 1995 to 1401 million in 2018 [ 1 ]. Projections indicate that this growth will continue in the future, reaching 1800 million international arrivals by 2030 [ 2 ]. It is clear that tourism plays a prominent role in the economy of many regions and countries, generating positive impacts, but it is also evident that, more recently, the focus has been on its negative impacts, especially as a consequence of massification, the congestion of public space, the increasing housing rentals, the rise in CO 2 emissions, and the degradation of certain natural and heritage resources [3–9]. In Europe, the expansion of tourism has occurred mainly in cities, towns, and suburbs, with rural areas accounting for only a third of the total overnight stays in the European Union between 2012 and 2018 [ 10 ]. However, a number of studies have highlighted the potential of tourism as a driving force for rural areas and its contributions to sustainable economic development and residents’ well-being. Specifically, tourism is highlighted as a useful instrument for the diversification of the Sustainability 2020,12, 4025; doi:10.3390/su12104025 www.mdpi.com/journal/sustainability Sustainability 2020,12, 4025 2 of 19 productive structure, generating income, and creating jobs, in many cases for women, which is of particular relevance in an environment where female participation in the labor market remains low. Furthermore, the creation of this type of employment contributes to reducing emigration and the depopulation of the rural environment. The role of tourism in retaining certain essential services that would disappear in its absence has also been highlighted [11–16]. In contrast, other studies have drawn attention to the limitations of tourism as an instrument for diversification and regional development [ 17 ]. Others have qualified the dimension of its impacts, which may be limited by questions such as infrastructures, accessibility, a lack of attractions, the sizes of establishments, the greater weight of domestic demand, or the low number of activities carried out by tourists in rural areas [ 18 – 24 ]. These limitations are an obstacle to the economic viability of businesses and condition the role of tourism as a strategy for revitalizing the rural environment. Taking into account the dichotomy between the positive and negative effects and the limitations of tourism in rural areas, the ideal solution would be to find specific formulas for sustainable tourism in rural areas. A number of studies have indicated that sustainable tourism formulas are effective potential tools for generating income and rejuvenating traditional industries in rural areas [ 25 – 27 ]. Sustainability implies maximizing profit whilst minimizing environmental and social costs. Therefore, it means achieving the highest net benefits derived from tourism. Greater sustainability will boost the well-being of society while, at the same time, reducing environmental damage. In addition, sustainability will not only benefit destinations and the environment, but it will also be positive for visitors’ own experiences and satisfaction levels [28]. We consider that the evolution of two variables contributes to a greater net impact of tourism in rural areas: The reduction in the percentage of same-day visitors and a longer length of stay (LOS). The higher the percentage of same-day visitors, the fewer the positive impacts and the greater the negative ones will be. It is clear that same-day visitors spend less than tourists and exert more pressure on the resources of natural areas. Firstly, accommodation is one of the most important expenditures of tourists in the destination, and secondly, non-overnight visitors tend to stay in central locations and visit major tourism attractions, putting pressure on resources [29–31]. Increasing LOS is as important in generating additional benefits as it is in reducing the negative effects of tourism [ 32 ]. Firstly, the increase in LOS will alleviate the overcrowding of certain areas while offering opportunities for others that would be left out of tourism development [ 33 ]. Secondly, the increase in LOS allows for slow consumption that encourages the creation of jobs and the conservation of local products and resources, thus contributing to sustainability [ 34 , 35 ]. Finally, from an environmental point of view, higher LOS implies a lower impact on the environment. Following [ 36 ], it must also be remembered that greater LOS implies carrying fewer visitors and, consequently, fewer transport infrastructure requirements, lower land needs, lower energy consumption, and lower greenhouse gas emissions [ 37 , 38 ]. Therefore, one of the main goals of tourism management agents is to design policies that encourage greater LOS. Despite the relevance of these two factors in boosting the net impacts of tourism, we were unable to locate any papers that analyze the determinants of LOS in rural areas. To fill this gap, our empirical analysis focuses on the Schist Villages Network (SVN), a set of 27 villages located in 19 municipalities in the center of Portugal. It is an experience developed with the help of an EU community-based initiative known as the Schist Villages Network Program, which, supported by the participation of various stakeholders, has recovered the architectural and cultural heritage, generated income, and created jobs. The main goals of this paper are to identify the LOS in the case of SVN and analyze whether the factors underlying the decision to stay overnight and those affecting the number of overnight stays differ. In this study, we use Heckman selection models that allow for the differentiation between the factors determining the decision to stay overnight and those affecting the number of overnight stays. This methodology also enables us to distinguish between same-day visitors and tourists. This paper aims to contribute to understanding of the SVN’s tourism characteristics from the perspective of LOS as well as the factors behind the decision to stay overnight and those affecting the number of Sustainability 2020,12, 4025 3 of 19 overnight stays. Determining whether these factors are the same is crucial because the economic and environmental impacts of tourists and same-day visitors are very different. Furthermore, a joint analysis of both decisions is necessary, since the number of overnight stays depends on the initial decision of whether to stay overnight or not. Finally, clarifying the drivers of both decisions allows for the formulation of policies aimed at increasing the relative weight of tourists with respect to same-day visitors, thereby increasing the sustainability of tourism. The conclusions drawn here can be extended to other rural tourist destinations characterized by a large number of same-day visitors and short average tourist stays. 2. Length of Stay: Importance and Determinants There is a broad consensus in scientific literature regarding the importance of LOS in tourism management. Although some studies have found that average daily spending is higher in short stays [ 39 , 40 ] the average spending per trip is heavily influenced by the LOS [ 41 , 42 ]. In addition, the LOS conditions the environmental impacts of tourism. The use of fossil fuels and the resulting emissions are the main environmental problem related to tourism [ 43 , 44 ]. Indeed, transport is the activity that impacts most negatively on the environment [ 37 ]. If LOS is reduced, more visitors will be required to maintain revenue, which will mean a greater use of transportation and, therefore, an increase in greenhouse gas emissions. Conversely, an increase in LOS will increase economic benefits without increasing environmental damage [ 31 , 38 , 45 ]. At the same time, an increase in LOS will allow tourists to abandon the usual circuits in the destination, thereby benefiting more peripheral locations [33]. The decision whether or not to stay overnight has important consequences in terms of economic or environmental impacts. When comparing same-day visitors and tourists, the latter spend more because accommodation expenditure is a key component of overall tourist spending in a destination [ 46 ]. In addition, same-day visitors have a limited time and, therefore, they focus on the main attractions [ 47 ], which, as Reference [ 36 ] (p. 2089) noted, makes it difficult “to distribute tourism’s benefits to a greater number of stakeholders, specifically in more peripheral regions”. This means that the socioeconomic impacts are fewer in the case of same-day visitors, and the negative impacts on natural and heritage resources increase. In recent years, evidence has pointed to a downward trend in LOS [ 36 , 48 , 49 ]. Taking into account this trend and the importance of LOS for destinations, a significant number of studies have focused on the study of LOS and its determining factors. As examples, we will cite some of the most relevant works that have been published in the last decade. In research focused on low-cost tourism in Catalonia (Spain), Reference [ 50 ] used duration models to analyze the determinants of LOS, identifying as significant factors nationality, age, education, type of occupation, type of accommodation, season of travel, and destination area. In turn, Reference [ 51 ] analyzed LOS in Spain with a random-parameter logit model. Their empirical results indicated that income, household size, age, distance, motivations, size of the city of origin, and available days are significant factors. In addition to some determinants previously referred to, such as age and education, Reference [ 29 ], using duration models, identified other significant determinants, such as the type of hotel, the prime motivation for trip, events, climate, the beach, and hospitality, when they analyzed LOS in the golf tourism segment in the Algarve (Portugal). Reference [ 52 ] used a truncated Poisson regression to analyze LOS in the Balearic Islands (Spain) and found familiarity with the destination, the number of tourist trips, and the number of people in the group to be significant determinants. In an analysis of LOS in Madagascar, Reference [ 53 ] used duration models and found that nature, gastronomy, the sea, life style, the physical appearance of the population, and security were relevant factors. When analyzing factors influencing the LOS in cultural tourism in Bolzano (Italy) using a zero-truncated negative binomial model, Reference [ 54 ] identified the costs associated with the journey to be a key element. The analysis of LOS in senior tourism conducted by [ 55 ], based on a negative binomial model, revealed that factors such as visiting friends and relatives, accommodation type, lone travel, and certain activities were all significant. In an analysis of the LOS in Brazil with Sustainability 2020,12, 4025 4 of 19 duration models, Reference [ 56 ] identified the following significant determinants: Multi-destination travel, travel purpose, transport mode, type of accommodation, group size, previous visits, season, expenditure, and type of destination. Reference [ 57 ] analyzed LOS in the segment of student tourism and found booking time, daily trip costs, the tourism and trip motivations, the gender, and the trip month to be significant factors. He used the ordinary least squares model, the zero-truncated negative binomial model, and the Weibull survival model. In a project focused on international tourism in south-west Norway, Reference [ 31 ] explored the aspects influencing planning LOS using a binary logistic regression, finding that perceptions of the time ‘needed’ for desired activities, limited holiday leave, and budgets were the most important aspects. Finally, Reference [ 30 ] analyzed the factors influencing LOS in the city of Santiago de Compostela (Spain) using five alternative Heckman selection models. The results of their work show that different factors related to tourists, travel, and the destination itself significantly determine LOS. More specifically, they found that foreign visitors traveling for business or MICE (Meetings, Incentives, Conferences and Exhibitions) reasons are the most likely to stay for longer periods. Although there are many studies analyzing the tourism demand in rural areas [ 58 – 60 ], including some that focus on the determining factors of spending [ 61 ], we failed to find any that focus exclusively on analyzing LOS determinants. For this reason, we consider that our work is pioneering in its nature and that the results shed light on LOS determinants in rural areas. 3. Materials and Methods 3.1. Study Area Description The Schist Villages Network (SVN) is situated in the center of Portugal (Figure 1). Supported by several public programs and initiatives, the SVN project originated in 2000. The project has evolved over the years, and the network currently comprises 27 villages. One of the strengths of the SVN area is its natural and historical environment with ample opportunities for use as a tourist resource, which caters to the growing demand in the rural tourism segment. This demand, which, in many cases, originates in cities, seeks leisure in contact with nature, the local population, and the endogenous territorial resources. The SVN allows the tourists to enjoy nature and discover the region’s cultural traditions. Sustainability 2020, 12, x FOR PEER REVIEW 5 of 20 Figure 1. Location of the scope of study. 3.2. Questionnaire Survey and Data The survey instrument consisted of a questionnaire divided into three main sections. Section I consisted of a set of questions designed to identify visitors’ previous experiences in the SVN and how they heard about the Network. Section II included a set of questions intended to assess the SVN’s prior image as well as visitors’ motivations and satisfaction levels. Section III included a miscellaneous set of questions covering aspects such as the tourists’ profile, the trip characteristics, and the destination characteristics. The survey instrument was revised and finalized in accordance with feedback from a pilot sample of 15 visitors to the SVN. The final questionnaire included two types of questions. On the one hand, a nominal scale was used for the questions related to the visitor’s profile, the travel characteristics, and destination characteristics. On the other hand, in the questions related to previous image, motivation, and satisfaction, the respondents were asked to rate their evaluations on a fivepoint Likert-type scale. In this study, previous image and satisfaction were assessed using a single-item method, and the respondents were asked to rate their previous image and satisfaction with the activities and overall travel experience on a five-point Likert scale (1 = very dissatisfied and 5 = very satisfied). For motivation, 13 items were rated on a five-point Likert scale, where 1 = most disagree and 5 = most agree. An exploratory factor analysis (EFA) was conducted to identify the underlying dimensions of visitors’ motivation, and the arithmetic means of items included within each factor obtained were used in the subsequent analysis. It is widely accepted that motivation is one of the central elements in both the choice process and tourist behavior at the destination [68]. There is a broad consensus that there are both "push motivation" and "pull motivation" factors [69,70]. Push motivation is related to internal elements, typical of tourists, and includes factors such as relaxation, quest for knowledge, and family togetherness. Pull factors depend on the destination’s appeal and include elements such as the range of activities on offer, natural or cultural attractions, security, or accessibility [68,71,72]. SPSS (ver. 20.0) was used for the exploratory factor analysis (EFA). A four-factor solution with eigenvalues greater than one and a factor loading above 0.5, accounting for 66.63% of the total variance, was chosen. Factor loadings of variables varied from 0.61 to 0.85. This study therefore meets the criteria suggested by [73] (pp. 122–129) (factor loadings equal to or above 0.50, eigenvalues equal to or above 1.0, and results of the factor analysis explaining at least 60% of the total variance) [27]. North The Schist Villages (SVN), Portugal Figure 1. Location of the scope of study. The SVN is an example of collaboration between the various stakeholders: The administration, private companies, and the local population. The Network’s action lines include the revalorization of cultural heritage and the use of built architectural heritage and endogenous resources [ 62 ]. In this sense, houses were rehabilitated and infrastructures were built, fostering the socioeconomic fabric and a renewal of the arts and crafts. A differentiated tourism brand was also created for the Network based Sustainability 2020,12, 4025 5 of 19 on the authentic values of the endogenous resources as intangible cultural heritage and deeply rooted customs [63–66]. A total of 70 establishments make up the network’s accommodation offer, totaling nearly 1000 beds. Tourists’ demand for regional gastronomy is covered by a group of 13 restaurants that are members of the Network, which also includes 18 “Schist Villages Shops”, where a wide range of local produce and traditional handicrafts can be acquired. Moreover, the Network territory features around 700 km of walking trails through magnificent country and mountain landscapes, as well as 900 km of bicycle routes, with different levels of difficulty supported by six Mountain Bikes (MTB) centers. The region has 50 river beaches where bathing is possible in summer [ 67 ]. The villages offer a varied recreation program throughout the year, tailored to the specific characteristics of each season. There is also space for activities that encourage tourists’ contact with culture, local heritage, and the authenticity of the population and local customs. 3.2. Questionnaire Survey and Data The survey instrument consisted of a questionnaire divided into three main sections. Section 1 consisted of a set of questions designed to identify visitors’ previous experiences in the SVN and how they heard about the Network. Section 2included a set of questions intended to assess the SVN’s prior image as well as visitors’ motivations and satisfaction levels. Section 3included a miscellaneous set of questions covering aspects such as the tourists’ profile, the trip characteristics, and the destination characteristics. The survey instrument was revised and finalized in accordance with feedback from a pilot sample of 15 visitors to the SVN. The final questionnaire included two types of questions. On the one hand, a nominal scale was used for the questions related to the visitor’s profile, the travel characteristics, and destination characteristics. On the other hand, in the questions related to previous image, motivation, and satisfaction, the respondents were asked to rate their evaluations on a five-point Likert-type scale. In this study, previous image and satisfaction were assessed using a single-item method, and the respondents were asked to rate their previous image and satisfaction with the activities and overall travel experience on a five-point Likert scale (1 =very dissatisfied and 5 =very satisfied). For motivation, 13 items were rated on a five-point Likert scale, where 1 =most disagree and 5 =most agree. An exploratory factor analysis (EFA) was conducted to identify the underlying dimensions of visitors’ motivation, and the arithmetic means of items included within each factor obtained were used in the subsequent analysis. It is widely accepted that motivation is one of the central elements in both the choice process and tourist behavior at the destination [ 68 ]. There is a broad consensus that there are both "push motivation" and "pull motivation" factors [ 69 , 70 ]. Push motivation is related to internal elements, typical of tourists, and includes factors such as relaxation, quest for knowledge, and family togetherness. Pull factors depend on the destination’s appeal and include elements such as the range of activities on offer, natural or cultural attractions, security, or accessibility [68,71,72]. SPSS (ver. 20.0) was used for the exploratory factor analysis (EFA). A four-factor solution with eigenvalues greater than one and a factor loading above 0.5, accounting for 66.63% of the total variance, was chosen. Factor loadings of variables varied from 0.61 to 0.85. This study therefore meets the criteria suggested by [ 73 ] (pp. 122–129) (factor loadings equal to or above 0.50, eigenvalues equal to or above 1.0, and results of the factor analysis explaining at least 60% of the total variance) [ 27 ]. Cronbach’s α for the four dimensions ranged from 0.68 to 0.79, indicating high internal consistency and a generally agreed upon lower limit of 0.60 for research at the exploratory stage [ 74 ]. The results show a wide variety of motivations for visitors in rural areas. As can be seen in Table 1, factor 1 exhibited most of the variance (32.61%) with a reliability coefficient of 0.79. This factor incorporated four items of motivation (security and comfort; rest, leisure, and relaxation; local people’s affectivity and friendliness; and local operators’ availability and Sustainability 2020,12, 4025 6 of 19 friendliness). The proportion of the total variance explained for this factor leads us to conclude that among rural visitors of the SVN, relaxation is a central motivational differentiation factor. The results coincide with the work of [ 75 ] for Korea. This factor is related to visitors’ internal emotional factors, reflecting a strong push motivation, but also a pull motivation related to residents’ hospitality. We have named this factor “relaxation and hospitality”. Factor 2 accounted for approximately 13% of the variance in data and incorporates four items related to the natural, architectural, heritage, and cultural attractions of the destination. We identified it as “natural and cultural attractions”. Factor 3 accounted for 11.41% of the variance and included leisure and sporting activities, sightseeing tours, regional cuisine, and catering diversity. This factor was therefore termed “sports and activities”. Finally, factor 4, identified as “local produce”,accounted for approximately 9.61% of the variance in the data. This factor reflects a strong push motivation that reflects visitors’ preference for consuming local produce during their trips and their willingness to pay for them. Table 1. Factor analysis of motivations of Schist Villages Network (SVN) visitors. Factor or Items Fact. Load. Eigenval. Var. Explain. (%) Cum. Var. Explain. (%) Cronbach α Relaxation and hospitality (Mot1)4.24 32.61 32.61 0.79 Security and comfort 0.84 Rest, leisure, relaxation 0.75 Local people’s affectivity and friendliness 0.69 Local operators’ availability and friendliness 0.63 Natural and cultural attractions (Mot2)1.69 13.01 45.61 0.77 Contact with nature 0.85 Natural and landscape resources 0.79 Traditional conserved architecture 0.72 Cultural and traditional heritage 0.61 Sports and activities (Mot3)1.48 11.41 57.02 0.68 Program of leisure and sporting activities 0.80 Sightseeing tours 0.70 Regional cuisine and catering diversity 0.62 Local produce (Mot4)1.25 9.61 66.63 0.71 I am willing to pay more for local produce 0.84 I prefer to consume local produce when I travel 0.82 The survey included questions to determine whether or not the visitor had stayed overnight and, in case he/she had, the number of nights. With these answers, we defined two alternative dependent variables: length1, a binary variable indicating whether or not the visitor stayed overnight, and length2, a discrete numerical variable reflecting the number of nights the trip included. In relation to the determinants of the LOS at the destination, the covariates included in this study were classified into three categories: a. Visitor-based variables: Age, gender, education, occupation, and monthly income. b. Travel-based variables: Motivation; visitor origin, which defines the distance to the region of SVN (distance); the duration of the trip to the SVN in hours (duration); whether the visitor was travelling as part of a family group or not (family); if they were repeating their visit (repeat); and the number of villages visited during the trip (visits). c. Destination-based variables: How the visitor heard about the destination (promotion); region in which the village is located (region); physical morphology of the region (mountain); village population size (size); existence of stores selling local produce, Schist Villages Shops, loja in Portuguese (shop); village specialization in tourism (tourist activity); the previous image of the SVN (image); satisfaction with activities (activities); and overall satisfaction with the visit (satisfaction). The target population comprised visitors to the SVN located in central Portugal. The empirical study was conducted between July and November 2017. After screening the responses and removing unusable questionnaires, 427 valid questionnaires were obtained. There was a sample error of 4.7% considering a maximal variance with p =q=0.5 and a level of confidence of 95%. Sustainability 2020,12, 4025 7 of 19 3.3. Empirical Analysis LOS is the factor that differentiates a same-day visitor from a tourist in a specific destination. The analysis of LOS determinants is therefore of key importance. In order to estimate the relationship between LOS and its determinants, a specific econometric model is required. The following multiple linear regression model is a suitable starting point: lengthi=X0 iβ+εi(1) where length i is the LOS for the ith visitor (i=1, 2, . . . N), X0 i denotes a vector of determinants (personal and travel characteristics, destination attributes, and other determinants), β is the vector of parameters to be estimated, and εiis the disturbance. The standard regression model (1) is not suitable for our objectives, since it does not allow us to differentiate between a same-day visitor and a tourist. Furthermore, it fails to take into account that the decision regarding the number of overnight stays is nested in the decision to stay overnight (the former depends on the latter), nor does it allow for any variations between the factors influencing the decision to stay overnight or not and those with the potential to influence the number of overnight stays (LOS). A more flexible and attractive model is therefore necessary to take all three aspects into account. From an econometric point of view, a Heckman selection model is a feasible alternative [ 76 , 77 ]. The Heckman selection model was used for the first time to estimate LOS by [ 30 ]. This model consists of two equations: A selection equation to determine whether an individual stays overnight or not: length1∗ i=X0 1iβ1+ε1iwhere length1i=(1if length1∗ i>0 0if length1∗ i≤0(2) and an outcome equation to estimate the number of overnight stays: length2∗ i=X0 2i β2+ε2i where length2i=(length2∗ iif length1∗ i>0 Not observed if length1∗ i≤0(3) It is commonly assumed that ε1i and ε2i have a bivariate normal distribution: ε1i ε2i!≈NID 0 0!, σ2 1σ12 σ12 σ2 2!! (4) where σ2 1=1, σ12 =ρ12σ2, and ρ12 is the correlation coefficient between the two disturbances. The equation in (2) is, in fact, a standard probit model, describing the choice of staying overnight or not. The choice to stay overnight is affected by variables in X 1i with coefficients β1 . The Equation (3) describes the number of overnight stays as a function of the variables in X 2i with coefficients β2 . In principle, the variables in X 1i and X 2i can be different, and the signs and magnitudes of the β coefficients may differ across the two equations [78]. According to [ 79 ], this model might be estimated by maximum likelihood, but due to the difficulty in maximizing its likelihood function (from an empirical point of view, this method often involves difficulties regarding the convergence of the problem [ 80 ]), in our empirical work, a two-step procedure proposed by [ 77 ], often called the Heckman two-step method (HTSM) or Heckit method, was carried out. This method is computationally simpler and provides consistent estimators, making it one of the most commonly used procedures in empirical micro-econometric work [ 78 ]. Stage one of the HTSM consists of estimating the Equation (2) by maximum likelihood as an independent probit model to determine the decision to stay overnight or not (the estimation is made using the entire sample). Then, a vector of inverse Mills ratios is generated from the parameter estimates λi=φX0 1iβ1/ΦX0 1iβ1 , where φ and Φ are the standard normal density and cumulative distribution function, respectively. Sustainability 2020,12, 4025 8 of 19 In the second stage, and using only the data for overnight visitors (tourists), the least squares regression of length2ion X 0 2i2and λis then computed, yielding consistent estimates of β2and σ12: length2i=X0 2iβ2+σ12λi+νi(5) where νi is the disturbance in this equation. According to [ 78 ], a crucial parameter in (5) is covariance ( σ12 ). Only if σ12 =ρ12 = 0 could we simply estimate Equation (3) by OLS (Ordinary Least Squares) and ignore Equation (2). A simple selection bias in the OLS estimator arises if σ12 ,0. Another key empirical question in the use of HTSM is the identification of the model (identification of β2 in the second step). According to [ 78 ], for this method to work properly in practice, it is necessary to introduce exclusion restrictions in the model (that is, variables existent in X1i are excluded from X2i ). In the case where X1i and X2i are identical, the model is only identified through the fact that λi is a nonlinear function, and the working of HTSM will depend on the variation of λi in the sample (for example, if λi does not vary much in the sample, then this term can be effectively approximated by a linear function of X1i , which can introduce severe collinearity among the regressors in Equation (5) [ 80 ]); hence, the importance of additionally introducing exclusion restrictions in order to solve the identification problem. However, there is no practical rule to specify which variables are most appropriate in order to establish the corresponding exclusion restrictions. An alternative is to use those variables that have a significant effect on the probability of staying overnight (as good predictors of length 1 ∗ i , they would be included in X1i ), but do not have an effect on the number of overnight stays (not associated with length 2 i , and therefore not included in X2i ). This alternative for the HTSM to be operative at an empirical level is in line with the proposal made by [30,81], among others. Taking into account that there are no natural variable candidates to establish the exclusion restrictions [ 78 ] and, therefore, the results may be sensitive to the chosen specification, our modeling proposal consists of several alternative empirical models (as a consequence of establishing several alternative exclusion restrictions) to compare the results obtained. If the results from the different models are similar, they can be considered to be more robust and meaningful. 4. Results 4.1. Descriptive Analysis From the statistics summary reported in Table 2, we can obtain the socio-demographic characteristics and an approximate travel profile of visitors to the SVN. Around 48% of the 427 respondents were same-day visitors. The average LOS of the 52% who stay overnight in the SVN is around 5.8 days. The number of men and women is similar, and a large majority of visitors (more than 73%) were between 20 and 49 years old. Over half had university studies, 70% were employed, and only 9% had a monthly income of more than 3000 euros. Most of the visitors came from two areas of Portugal. Thirty-seven percent (distance1) came from the center of Portugal, the region closest to the villages, and 38% came from the south of Portugal, Lisbon, and the Portuguese islands. The average duration of the journey to the SVN was 4.3 h, although it ranged from 0.5 to 28 h. Approximately 68% of visitors had travelled to the SVN region previously. The average number of villages visited was five, with those located in Serra da Lous á and Z ê zere receiving the highest number of visitors. Most of the visitors heard about the SVN from friends or relatives, had a fairly favorable previous image of the SVN, and expressed a high degree of satisfaction with the visit. Sustainability 2020,12, 4025 9 of 19 Table 2. Descriptive statistics summary. Variable Description Mean Std. Dev. Min Max Dependent variables Length1 Dummy variable equal to one if the visitor stays overnight and zero otherwise 0.520 0.500 0 1 Length2 Variable indicating the number of overnight stays 5.788 5.236 1 21 Visitor-based variables Age Age1Under 20 years =1, other =0 (reference category) 0.066 0.248 0 1 Age2Between 20–34 years =1, other =0 0.349 0.477 0 1 Age3Between 35–49 years =1, other =0 0.377 0.485 0 1 Age4Over 49 years =1, other =0 0.208 0.407 0 1 Gender Gender (1 =male, 0 =female) 0.515 0.500 0 1 Education Primary Primary education =1, other =0 0.070 0.256 0 1 Secondary Secondary education=1, other =0 0.415 0.493 0 1 University University education =1, other =0 (reference category) 0.515 0.500 0 1 Occupation Occupation1 Employed in the primary, secondary, or tertiary sectors =1, other =0 (reference category) 0.700 0.459 0 1 Occupation2Inactive or unemployed =1, other =0 0.300 0.459 0 1 Income Income1Less 1500 euros =1, other =0 (reference category) 0.539 0.499 0 1 Income2Between 1500–3000 euros =1, other =0 0.372 0.484 0 1 Income3More than 3000 euros =1, other =0 (reference category) 0.089 0.285 0 1 Travel-based variables Duration Variable indicating the number of travelling hours to the SVN 4.331 5.586 0.5 28 Family Family holidays =1, other =0 0.321 0.467 0 1 Distance Distance1 If the visitor comes from Central Portugal = 1, elsewhere =0 (reference category) 0.368 0.483 0 1 Distance2If the visitor comes from North of Portugal and “Grande Porto” =1; elsewhere =00.138 0.345 0 1 Distance3 If the visitor comes from Lisbon, southern Portugal, or the Portuguese islands =1, elsewhere =0 0.375 0.485 0 1 Distance4If the visitor comes from outside of Portugal =1, elsewhere =00.119 0.325 0 1 Motivation Mot1Relaxation and hospitality 3.990 0.729 1.5 5 Mot2Natural and cultural attractions 4.350 0.563 1.8 5 Mot3Sports and activities 3.502 0.785 1 5 Mot4Local products 3.785 0.854 1 5 Repeat If the visitor is repeating their visit =1, otherwise =00.677 0.468 0 1 Visits Number of villages visited during the trip 5.084 3.382 1 18 Destination-based variables Promotion Promotion1 If the visitor got to know SVN as a result “Advertising in Audio–Visual Media” and “Tourist Brochures” =1, other =0 0.185 0.389 0 1 Promotion2 If the visitor heard about the SVN as a result of consultation in travel agencies, internet search, or social media =1, remainder =0 0.199 0.400 0 1 Promotion3 If the visitor heard about the SVN as a result of recommendations from friends and relatives =1, other =0 (reference category) 0.616 0.487 0 1 Sustainability 2020,12, 4025 16 of 19 5. Sequera, J.; Nofre, J. Debates Shaken, not stirred. New debates on touristification and the limits of gentrification. City 2018. [CrossRef] 6. Del Romero, L. Touristification, Sharing Economies and the New Geography of Urban Conflicts. Urban Sci. 2018,2, 104. [CrossRef] 7. Jover, J.; D í az-Parra, I. Gentrification, transnational gentrification and touristification in Seville, Spain. Urban Stud. 2019. [CrossRef] 8. Campos, R.; Sequeira, A. Urban Art touristification: The case of Lisbon. Tour. Stud. 2019 ,20, 182–202. [CrossRef] 9. Andrade, M.J.; Costa, J.P. Touristification of European Port-Cities: Impacts on Local Populations and Cultural Heritage. In European Port Cities in Transition Moving Towards More Sustainable Sea Transport Hubs; Carpenter, A., Lozano, R., Eds.; Springer: Basel, Switzerland, 2020; pp. 187–204. 10. Statistical Office of the European Union (EUROSTAT). Database. Available online: https://ec.europa.eu/ eurostat/data/database (accessed on 9 March 2020). 11. Park, D.-B.; Yoon, Y.-S. Developing sustainable rural tourism evaluation indicators. Int. J. Tour. Res. 2011 ,13, 401–415. [CrossRef] 12. Stetic, S. Specific features of rural tourism destinations management. J. Settl. Spat. Plan. 2012,1, 131–137. 13. Woo, E.; Kim, H.; Uysal, M. Life satisfaction and support for tourism development. Ann. Tour. Res. 2015 ,50, 84–97. [CrossRef] 14. Muresan, I.; Oroian, C.; Harun, R.; Arion, F.; Porutiu, A.; Chiciudean, G.; Todea, A.; Lile, R. Local residents’ attitude toward sustainable rural tourism development. Sustainability 2016,8, 100. [CrossRef] 15. Androshchuk, L.; Chernenko, N. The economic and mathematical analysis of migration of employable population as a factor of national modernization in crisis. East. J. Eur. Stud. 2016,7, 35–47. 16. Roberts, L.; Hall, D.; Morag, M. New Directions in Rural Tourism; Routledge: Abingdon, UK, 2017. 17. Castellano- Á lvarez, F.J.; Del R í o-Rama, M.C.; Á lvarez-Garc í a, J.; Dur á n-S á nchez, A. Limitations of Rural Tourism as an Economic Diversification and Regional Development Instrument. The Case Study of the Region of La Vera. Sustainability 2019,11, 3309. [CrossRef] 18. Yagüe Perales, R.M. Rural tourism in Spain. Ann. Tour. Res. 2002,29, 1101–1110. [CrossRef] 19. Sharpley, R. Rural tourism and the challenge of tourism diversification: The case of Cyprus. Tour. Manag. 2002,23, 233–244. [CrossRef] 20. Irvine, W.; Anderson, A.R. Small tourist firms in rural areas: Agility, vulnerability and survival in the face of crisis. Int. J. Entrep. Behav. Res. 2004,10, 229–246. [CrossRef] 21. Mart í nez-Roget, F.; Rodr í guez, X.A. Occupancy level and productivity in rural tourism establishments: The case of Galicia, Spain. Tour. Econ. 2006,12, 279–289. [CrossRef] 22. Gasc ó n, J. The limitations of community-based tourism as an instrument of development cooperation: The value of the Social Vocation of the Territory concept. J. Sustain. Tour. 2013,21, 716–731. [CrossRef] 23. Neumeier, K.; Pollermann, K. Rural tourism as promoter of rural development—Prospects and limitations: Case study findings from a pilot projectpromoting village tourism. Eur. Countrys. 2014 ,4, 270–296. [CrossRef] 24. Bel, F.; Lacroix, A.; Lyser, S.; Rambonilaza, T.; Turpin, N. Domestic demand for tourism in rural areas: Insights from summer stays in three French regions. Tour. Manag. 2015,46, 562–570. [CrossRef] 25. Lane, B. Sustainable rural tourism strategies: A tool for development and conservation. J. Sustain. Tour. 1994,2, 102–111. [CrossRef] 26. Hall, D.; Richards, G. Tourism and Sustainable Community Development; Routledge: New York, NY, USA, 2003. 27. Cho, H.-S.; Byun, B.; Shin, S. An Examination of the Relationship between Rural Tourists’ Satisfaction, Revisitation and Information Preferences: A Korean Case Study. Sustainability 2014 ,6, 6293–6311. [CrossRef] 28. Iniesta-Bonillo, M.A.; S á nchez-Fern á ndez, R.; Jim é nez-Castillo, D. Sustainability, value, and satisfaction: Model testing and cross-validation in tourist destinations. J. Bus. Res. 2016,69, 5002–5007. [CrossRef] 29. Barros, C.P.; Machado, L.P. The length of stay in tourism. Ann. Tour. Res. 2010,37, 692–706. [CrossRef] 30. Rodr í guez, X.A.; Mart í nez-Roget, F.; Gonz á lez-Murias, P. Length of stay: Evidence from Santiago de Compostela. Ann. Tour. Res. 2018,68, 9–19. [CrossRef] 31. Jacobsen, J.K.S.; Göosling, S.; Dybedald, P.; Skogheim, T.S. Exploring length of stay: International tourism in south-western Norway. J. Hosp. Tour. Manag. 2018,35, 29–35. [CrossRef] 32. Alegre, J.; Pau, L. The length of stay in the demand for tourism. Tour. Manag. 2006 ,27, 1343–1355. [CrossRef] Sustainability 2020,12, 4025 17 of 19 33. Fakeye, P.C.; Crompton, J.L. Image differences between prospective, first-time, and repeat visitors to the Lower Rio Grande Valley. J. Travel Res. 1991,30, 10–16. [CrossRef] 34. Ax, C. Slow consumption for sustainable jobs. In Sustainable Solutions; Charter, M., Tischner, U., Eds.; Greenleaf: Sheffield, UK, 2001; pp. 402–409. 35. Hall, C.M. The contradictions and paradoxes of slow food: Environmental change, sustainability and the conservation of taste. In Slow Tourism. Experiences and Mobilities; Fullagar, S., Markwell, K., Wilson, E., Eds.; Channel View Publications: Bristol, UK; Buffalo, NY, USA; Toronto, ON, Canada, 2012; pp. 53–69. 36. Gössling, S.; Scott, D.; Michael Hall, C. Global trends in length of stay: Implications for destination management and climate change. J. Sustain. Tour. 2018,26, 2087–2101. [CrossRef] 37. Gössling, S.; Peeters, P.; Ceron, J.-P.; Dubois, G.; Patterson, T.; Richardson, R.B. The eco-efficiency of tourism. Ecol. Econ. 2005,54, 417–434. [CrossRef] 38. Gössling, S.; Ring, A.; Dwyer, L.; Andersson, A.-C.; Hall, C.M. Optimizing or maximizing growth? A challenge for sustainable tourism. J. Sustain. Tour. 2016,24, 527–548. [CrossRef] 39. Thrane, C.; Farstad, E. Domestic tourism expenditures: The non-linear effects of length of stay and travel party size. Tour. Manag. 2011,32, 46–52. [CrossRef] 40. Mayer, M.; Vogt, L. Economic effects of tourism and its influencing factors. An overview focusing on the spending determinants of visitors. Z. Tour. 2016,8, 169–198. [CrossRef] 41. Divisekera, S. Economics of tourists’ consumption behavior: Some evidence from Australia. Tour. Manag. 2010,31, 629–636. [CrossRef] 42. Divisekera, S.; Deegan, J. An analysis of consumption behavior of foreign tourists in Ireland. Appl. Econ. 2010,42, 1681–1697. [CrossRef] 43. Gragl, H.; Kokott, J.; Kulessa, M.; Luther, J.; Nuscheler, F.; Sauerborn, R.; Schellnhuber, H.-J.; Schubert, R.; Schulze, E.-D. Climate Protection Strategies for the First Century: Kyoto and Beyond. Special Report; WBGU: Berlin, Germany, 2003. 44. Thomas, C.D.; Cameron, A.; Green, R.E.; Bakkenes, M.; Beaumont, L.J.; Collingham, Y.C.; Erasmus, B.F.N.; Ferreira de Siquira, M.; Grainger, A.; Hannay, L.; et al. Extinction risk from climate change. Nature 2004 ,427, 145–148. [CrossRef] 45. Göossling, S.; Scott, D.; Hall, C.M. Inter-market variability in CO2 emission-intensities in tourism: Implications for destination marketing and carbon management. Tour. Manag. 2015,46, 203–212. [CrossRef] 46. Marrocu, E.; Paci, R.; Zara, A. Micro-economic determinants of tourist expenditure: A quantile regression approach. Tour. Manag. 2015,50, 13–30. [CrossRef] 47. Turkle, S. Reclaiming Conversation; Basic Books: New York, NY, USA, 2015. 48. Salmasi, L.; Celidoni, M.; Procidado, I. Length of Stay: Price and Income Semi-Elasticities at Different Destinations in Italy. Int. J. Tour. Res. 2012,14, 515–530. [CrossRef] 49. Ferrer-Rosell, B.; Mart í nez-Garcia, E.; Coenders, G. Package and no-frills air carriers as moderators of length of stay. Tour. Manag. 2014,42, 114–122. [CrossRef] 50. Mart í nez-Garc í a, E.; Raya, J.M. Length of stay for low-cost tourism. Tour. Manag. 2008 ,29, 1064–1075. [CrossRef] 51. Nicolau, J.L.; M á s, F.J. Simultaneous analysis of whether and how long to go on holidays. Serv. Ind. J. 2009,29, 1077–1092. [CrossRef] 52. Alegre, J.; Mateo, S.; Pau, L. A latent class approach to tourist’s length of stay. Tour. Manag. 2011 ,32, 555–563. [CrossRef] 53. Peypoch, N.; Randriamboarison, R.; Rasoamananjara, F.; Solonandrasana, B. The length of stay of tourists in Madagascar. Tour. Manag. 2012,33, 1230–1235. [CrossRef] 54. Brida, J.G.; Meleddu, M.; Pulina, M. Factors influencing length of stay of cultural tourism. Tour. Econ. 2013,19, 1273–1292. [CrossRef] 55. Al é n, E.; Nicolau, J.L.; Losada, N.; Dom í nguez, T. Determinant factors of senior tourists’ length. Ann. Tour. Res. 2014,49, 19–32. [CrossRef] 56. Santos, G.E.D.O.; Ramos, V.; Rey-Maquiera, J. Length of stay at multiple destinations of tourism trips in Brazil. J. Travel Res. 2015,54, 788–800. [CrossRef] 57. Thrane, C. Students’ summer tourism: Determinants of length of stay (LOS). Tour. Manag. 2016 ,54, 178–184. [CrossRef] 58. Abaladejo Pina, I.P.; D í az Delfa, M.T. Rural tourism demand by type of accommodation. Tourism Manage. 2004,26, 951–959. [CrossRef] Sustainability 2020,12, 4025 18 of 19 59. Mart í nez-Roget, F.; Rodr í guez, X.A. Rural tourism demand in Galicia, Spain. Tour. Econ. 2006 ,12, 21–31. [CrossRef] 60. Eus é bio, C.; Carneiro, M.J.; Kastenholz, E.; Figueiredo, E.; Da Silva, D.S. Who is consuming the countryside? An activity-based segmentation analysis of the domestic rural tourism market in Portugal. J. Hosp. Tour. Manag. 2017,31, 197–210. [CrossRef] 61. Kastenholz, E. Analysing determinants of visitor spending for the rural tourist market in North Portugal. Tour. Econ. 2005,11, 555–569. [CrossRef] 62. Moutela, J.A.; Carreira, V.; Martinez-Roget, F. Authenticity in interior rural areas: A systemic study of stakeholders in the Schist Villages Network. In Authenticity & Tourism: Materialities, Perceptions, Experiences; Rickly, J.M., Vidon, E., Jafari, J., Eds.; Tourism Social Science Series; Emerald Publishing: Bingley, UK, 2018; Volume 24, pp. 109–126. 63. Carvalho, P. A AIBT do Pinhal Interior e as Aldeias do Xisto: Novos caminhos e actores para o desenvolvimento de territórios de baixa densidade en ambientes de montanha. Cadernos Geogr. 2009,28/29, 185–191. 64. Ramos, B.A. Aldeias do Xisto: Um conceito de turismo sustentável inovador. Rev. Tur. Desenv. 2009,11, 103–107. 65. Mart í nez-Roget, F.; Moutela, J.A. Novas motivaç õ es dos turistas: Uma oportunidade para o territ ó rio rural. Análise da “Rede das Aldeias do Xisto” no Centro de Portugal. Rev. Tur. Desenv. 2013,19, 99–110. 66. Mart í nez-Roget, F.; Moutela, J.A.; Est é vez, J.C. Chaves do ê xito do Turismo em Espaço Rural (TER): Evid ê ncias a partir da lealdade turística na Rede das Aldeias do Xisto. Rev. Port. Estud. Reg. 2015,40, 7–8. 67. Agency for the Tourist Development of Schist Villages (ADXTUR). Available online: https://aldeiasdoxisto. pt/entidade/706 (accessed on 11 February 2020). 68. Yoon, Y.; Uysal, M. An examination of the effects of motivation and satisfaction on destination loyalty: A structural model. Tour. Manag. 2005,26, 45–56. [CrossRef] 69. Dann, G.M.S. Anomie, Ego-Enhancement and Tourism. Ann. Tour. Res. 1977,4, 184–194. [CrossRef] 70. Dann, G.M.S. Tourist Motivation: An Appraisal. Ann. Tour. Res. 1981,8, 187–219. [CrossRef] 71. Chi, C.G.Q.; Qu, H. Examining the structural relationships of destination image, tourist satisfaction and destination loyalty: An integrated approach. Tour. Manag. 2008,29, 624–636. [CrossRef] 72. Rid, W.; Ezeuduji, I.O.; Pröbstl-Haider, U. Segmentation by motivation for rural tourism activities in the Gambia. Tour. Manag. 2014,40, 102–116. [CrossRef] 73. Hair, J.H., Jr.; Black, W.C.; Babin, B.J.; Tatham, R.L. Multivariate Data Analysis; Prentice Hall: Upper Saddle River, NJ, USA, 2005. 74. Hair, J.; Money, A.; Page, M.; Samouel, P. Research Methods for Business; John Wiley and Son Ltd.: Chichester, UK, 2007. 75. Park, D.-B.; Yoon, Y.-S. Segmentation by motivation in rural tourism: A Korean case study. Tour. Manag. 2009,30, 99–108. [CrossRef] 76. Heckman, J. The common structure of statistical models of truncation, sample selection and limited dependent variables models and a simple estimator for such models. Ann. Econ. Soc. Measur. 1976,5, 475–492. 77. Heckman, J. Sample selection bias as a specification error. Econometrica 1979,47, 153–161. [CrossRef] 78. Verbeek, M. A Guide to Modern Econometrics, 4rd ed.; Wiley & Sons: Chichester, West Sussex, UK, 2012. 79. Greene, W.H. Econometric Analysis, 3rd ed.; Prentice Hall Inc.: London, UK, 1998. 80. Wooldridge, J.M. Econometric Analysis of Cross Section and Panel Data, 2nd ed.; The MIT Press: Cambridge, UK, 2010. 81. Wynen, J. Explaining travel distance during same-day visits. Tour. Manag. 2013,36, 133–140. [CrossRef] 82. Ashworth, G.; Page, S.J. Urban tourism research: Recent progress and current paradoxes. Tour. Manag. 2011,32, 1–15. [CrossRef] 83. Book in Xisto. Aldeias do Xisto. Available online: https://bookinxisto.com/es (accessed on 3 March 2020). 84. Fleischer, A.; Pizam, A. Tourism constraints among Israeli Seniors. Ann. Tour. Res. 2002 ,29, 106–123. [CrossRef] 85. Valls, J.-F.; Mota, L.; Vieira, S.C.F.; Santos, R. Opportunities for Slow Tourism in Madeira. Sustainability 2019,11, 4534. [CrossRef] 86. Lane, B. Rural tourism: An overview. In The SAGE Handbook of Tourism Studies; Jamal, T., Robinson, M., Eds.; Sage Publications: London, UK, 2009. 87. Farah Atiqah Mohamad Noor, F.A.M.; Nair, V.; Mura, P. Conceptualizing a Framework for Slow Tourism in a Rural Destination in Malaysia. Adv. Sci. Lett. 2015,21, 1185–1188. [CrossRef] Sustainability 2020,12, 4025 19 of 19 88. Hjalager, A.-M.; Kwiatkowski, G.; Larsen, M.Ø. Innovation gaps in Scandinavian rural tourism. Scandin. J. Hosp. Tour. 2018,18, 1–17. [CrossRef] 89. United Nations. Sustainable Development Goals. 2020. Available online: https://www.un.org/ sustainabledevelopment/(accessed on 23 February 2020). 90. Alias, A.; Aziz, A.; Karim, M.S.A.; Isa, S.S. Local food consumption at ecotourism destination. In Adventure and Ecotourism in Malaysia; Mariapan, M., Isa, S.S., Aziz, N.A.A., Lin, E.L.A., Hakeem, K.R., Eds.; Faculty of Forestry: Selangor, Malaysia, 2015; pp. 40–45. 91. Kim, K.-H.; Park, D.-B. Relationships among Perceived Value, Satisfaction, and Loyalty: Community-Based Ecotourism in Korea. J. Travel Tour. Mark. 2017,34, 171–191. [CrossRef] 92. Oppermann, M. Tourism Destination Loyalty. J. Travel Res. 2000,39, 78–84. [CrossRef] 93. OECD. OECD Tourism Trends and Policies; OECDE Publishing: Paris, France, 2018. [CrossRef] 94. Sidali, K.; Kastenholz, E.; Bianchi, R. Food tourism, niche markets and products in rural tourism: Combining the intimacy model and the experience economy as a rural development strategy. J. Sustain. Tour. 2015 ,23, 1179–1197. [CrossRef] 95. Flanigan, S.; Blackstock, K.; Hunter, C. Agritourism from the perspective of providers and visitors: A typology-based study. Tour. Manag. 2014,40, 394–405. [CrossRef] 96. Pulido-Fern á ndez, J.I.; Ward-Perkins, D.; Krukova, O.V. Nature-Based Tourism. In Tourism in Russia: A Management Handbook; Diamanche, F., Andrades, L., Eds.; Emerald Group Publishing Limited: Bingley, UK, 2015; pp. 181–230. 97. Lai, I.K.W.; Hitchcock, M.; Lu, D.; Liu, Y. The Influence of Word of Mouth on Tourism Destination Choice: Tourist–Resident Relationship and Safety Perception among Mainland Chinese Tourists Visiting Macau. Sustainability 2018,10, 2114. [CrossRef] © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).