Residentsʼ attitudes and behavioural support for tourism in host communities

ABSTRACT In our study of the “perceptions–attitudes–behaviours” sequence, we explain how resident perceptions of tourism’s impacts on host communities influence not only their attitudes towards tourism, but also their attitudes towards tourists – i.e. a new variable that has recently been introduced in literature. Moreover, we introduce the behavioural support, a concept that denotes a higher level of involvement and engagement of residents with their communities in comparison with the traditional attitudinal support. Our results indicate: (1) the residents’ perception of the positive economic and cultural impacts of tourism are the main variables influencing their attitudes towards tourism and tourists; and (2) both types of attitudes influence on behavioural support for tourism in host communities.


Introduction
Residents are usually conceived of as a keystone of tourism development strategies in host communities (Easterling, 2004;Sharpley, 2014). Particularly, Ribeiro, Pinto, Silva, and Woosnam (2017) emphasize the need of attending to the needs and concerns of residents to achieve a sustainable tourism development in host communities. Their study mainly bases this argument on the fact that residents will demonstrate a behaviour that either supports or does not support tourism depending on their perceptions of the positive and negative impacts of tourism on their communities. For their part, Zhang, Inbakaran, and Jackson (2006) have previously already established the relevance of managing these perceptions in order to achieve a more harmonious relationship between tourists and residents, which is considered the central element of tourism (Sharpley, 2014;Yu & Lee, 2014). More recently, Palmer, Koenig-Lewis, and Medi Jones (2013) have emphasized that residents that are more involved with their communities have a key role as "ambassadors" since they promote incoming tourism through positive word-of-mouth, which is a very relevant information source for tourists (Litvin, Goldsmith, & Pan, 2008).
Past research has paid special attention to resident perceptions of the economic, sociocultural, and environmental impacts of tourism on host communities (see for example, Sinclair, Gursoy, & Vieregge (2015), Vargas, Plaza, & Porra (2009), and Wang & Xu (2015)) with the purpose of better understanding the nature and formation of this critical issue to be managed in the tourism development strategies of regions and countries. In particular, these perceptions are considered the main driver of resident attitudes towards tourism (Vargas, Oom, Da Costa Mendes, & Silva, 2015). However, it is necessary to indicate that most previous research has studied resident perceptions and attitudes as overlapping concepts or even synonymous ones (Sharpley, 2014;Sinclair et al., 2015). In our study, following Vargas et al. (2015), we distinguish between residents' specific perceptions with regard to the positive and negative impacts of tourism activities on their communities and, as a consequence of these perceptions, their overall attitudes concerning the tourism phenomenon. Particularly, attitude is a psychological variable that is considered a more consistent disposition than perceptions. In our context, it specifically refers to the resident's enduring predisposition toward tourism in host communities (Ap, 1992).
In this context, our study aims to contribute to the literature by developing a model of resident attitudes that includes not only the attitudes towards tourism, but also the attitudes towards tourists. In particular, most studies have focused on the attitudes towards tourism, which are mainly cognitive in nature, but not on the feelings or affective attitudes towards tourists (Woosnam, Norman, & Ying, 2009). In contrast, these relations (mostly known as "host-guest interactions") have been widely examined in previous studies on destination image and satisfaction (see for example Bianchi (2016), Herrero, San Martín, García-de Los Salmones, & Collado (2016)) although they adopted the touristsʼ point of view, not the residentsʼ. Thus, our study aims to explain how both the attitudes towards tourism in host communities and the attitudes towards tourists are formed. In line with Woosnam and Norman (2010), this approach is likely to lead to a more comprehensive understanding of the residents' responses to the tourism phenomenon.
In addition, it should also be highlighted that most studies (see for example Garau, Díaz, & Gutierrez-Taño (2014), Vargas et al. (2015), Almeida, Peláez, Balbuena, & Cortés (2016), and Rasoolimanesh, Roldán, Jaafar, & Ramayah (2017)) have focused on examining how resident perceptions and attitudes are influenced by: (1) extrinsic variables, such as seasonality or tourism development; and (2) intrinsic variables, such as demographic characteristics or social identity. However, much more research is needed to examine how perceptions and attitudes influence the subsequent responses or behaviours of residents (Sharpley, 2014). In this sense, our study also contributes to the literature by considering the "perceptions-attitudes-behavioural support" sequence in our model of resident attitudes. More concretely, unlike most studies that use the concept of support in an attitudinal sense (see for example, Látková & Vogt (2012), Nunkoo & Ramkissoon (2011), Nunkoo & So (2016, and Rasoolimanesh, Roldán, et al. (2017)) this study goes further by examining the support for tourism using a behavioural approach. In particular, we consider that the residents' participation in tourism activities in their communities, as well as their recommendations about them to other people, are two key behavioural indicators of support for tourism.

Literature review
We provide some background on the topic under investigation and conceptualize the variables of our theoretical model: resident perceptions of the positive and negative impacts of tourism, attitudes towards tourism and tourists, and behavioural support for tourism in host communities. In addition, we propose the causal relationships among these variables, thus formulating the hypotheses to be tested in our study.

Residents' perceptions and attitudes towards tourism
As observed in Table 1, three main types of impacts of tourism are identified in literature: economic, sociocultural, and environmental impacts. Residents will form their perceptions of the tourism phenomenon in terms of economic benefits/costs (for example, employment opportunities versus undesirable business activities), sociocultural effects (for example, cultural identity versus delinquency), and environmental impacts (for example, environmental awareness versus traffic congestion).
Subsequently, residents will form their attitudes towards tourism based on their perceptions of the economic, sociocultural, and environmental impacts of tourism on host communities. The theoretical framework most referenced in past research to explain this link between resident perceptions and attitudes is the Social Exchange Theory (Dyer, Gursoy, Sharma, & Carter, 2007;Gursoy & Kendall, 2006;Látková & Vogt, 2012;Vargas et al., 2015). According to this theory, tourists and residents engage in social exchange with the aim of achieving satisfactory outcomes (Sharpley, 2008). Overall, if residents perceive that the positive impacts of tourism in terms of economic, sociocultural, and environmental benefits are greater than the costs or negative impacts, then they will have a positive attitude towards tourism development in their communities (Sharpley, 2014). Based on this theory, numerous studies provide empirical evidence of the influence of resident perceptions of tourism's impacts on their attitudes on tourism or their attitudinal support (see for example, Dyer et al. (2007), Nunkoo & Gursoy (2012), Vargas et al. (2015), Wang & Xu (2015), Yoon, Gursoy, & Chen (2001)). According to the above-mentioned theory, and the evidence available in past research, we establish a link between resident perceptions of tourism's impacts and attitudes towards tourism. Most studies (see for example, Nunkoo and Gursoy (2012), Vargas et al. (2015), and Nunkoo & So (2016)) establish only two types of effects, namely the effects of positive and negative impacts on resident attitudes. In contrast, our study separately examines the six categories of tourism's impacts, namely economic, sociocultural, and environmental impacts, both positive and negative, with the aim of exploring their differential effects on attitudes, thus reinforcing the explanatory ability of our model. In this sense, it is widely established that resident perceptions of tourism's impacts vary depending on the type of tourism or the stage of tourism development of each host community (Bestard & Nada, 2007;Sheldon & Abenoja, 2001;Upchurch & Teivane, 2000;Vargas, Plaza, & Porra, 2009). Consequently, we consider it necessary to develop a model that disaggregates tourism's impacts and is able to explain in detail, in the context of any host community, a complex reality like resident perceptions and attitudes. In addition, this approach is in line with some previous studies in psychology (see for example Cacioppo & Berntson (1994) and Huppert & Whittington (2003)) which methodologically question the combination of distinct perceived variables into summary dimensions. Thus, we establish the following hypotheses (Hs): H1: The residents' perceptions of the positive economic impacts of tourism positively influence their attitudes towards tourism development in their communities.
H2: The residents' perceptions of the negative economic impacts of tourism negatively influence their attitudes towards tourism development in their communities.
H3: The residents' perceptions of the positive sociocultural impacts of tourism positively influence their attitudes towards tourism development in their communities.
H4: The residents' perceptions of the negative sociocultural impacts of tourism negatively influence their attitudes towards tourism development in their communities.
H5: The residents' perceptions of the positive environmental impacts of tourism positively influence their attitudes towards tourism development in their communities.
H6: The residents' perceptions of the negative environmental impacts of tourism negatively influence their attitudes towards tourism development in their communities.

Residents' perceptions and attitudes towards tourists
Although the Social Exchange Theory has been widely used for understanding responses given by residents to tourism phenomena in host communities, Andereck, Valentine, Knopf, and Vogt (2005) consider that the above-mentioned framework may provide an incomplete approach to resident attitudes. In this sense, Woosnam et al. (2009) highlight that this theory examines resident attitudes based merely on a cost-benefit perspective or, in other words, in terms of a trade-off between the favourable impacts and negative impacts perceived by residents (Zhang et al., 2006). In addition, some studies using this theory have found mixed results, with no significant relationships between resident perceptions and attitudes (Andereck et al., 2005;McGehee & Andereck, 2004). This fact has led to an alternative approach to resident attitudes that is focused on the interactions between local people and tourists, considered to be the essence of tourism (Sharpley, 2014;Yu & Lee, 2014), in a more emotional and individual sense. In the late 1990s, Lindberg and Johnson (1997) already included residents' pleasure related to the interaction with tourists in their attitudinal model, while Teye, Sönmez, and Sirakaya (2002) found a relationship between the residents' feelings derived of the interaction with tourists and their attitudes. More recently, Woosnam et al. (2009) built the Emotional Solidarity Framework with the purpose of incorporating the feelings towards tourists in the study of residentsʼ attitudes. Particularly, emotional solidarity is defined as the affective connection a person has with another individual based on the emotional closeness and grade of contact (Hammarström, 2005). In the context of tourism, Woosnam et al. (2009) highlight that residents will experience an emotional solidarity with tourists based on interactions and shared beliefs. Later, empirical research conducted by Woosnam (2012) found a positive relationship between emotional solidarity and attitudinal support for tourism, thus validating this theoretical approach.
Thus, despite the fact that most studies (see for example, Stylidis, Biran, Sit, & Szivas (2014), Vargas et al. (2015), Wang & Xu (2015), and Almeida et al. (2016)) have focused on the residents' attitudes towards tourism (cognitive in nature), we consider it necessary to also examine the attitudes towards tourists (affective in nature). By taking as a reference the concept of "affective attitudes towards tourists" proposed by Palmer et al. (2013), our study aims to enrich the literature by establishing a link between resident perceptions of tourism's impacts and their feelings or attitudes towards the tourists. Our statements are based on early studies that use the Social Representations Theory to explain how those perceptions are formed. According to Moscovici (1988), social representations are the mechanisms that individuals use to understand the events and objects around them. In tourism, Zhang et al. (2006) use this theory and establish that residents perceive tourism's impacts through the social representations of the tourism industry in their communities. In turn, representations are developed among residents based on the information published in the media and other external sources, but also on their direct experiences with tourists (Fredline & Faulkner, 2000).
In this context, we consider that resident-tourist interactions in host communities are a key factor in the formation of resident perceptions and attitudes. Particularly, the evaluations of those interactions made by residents will contribute to the formation of their perceptions about the different impacts of tourism in their communities. In addition, following the traditional "perceptions-attitudes" sequence, we postulate that a positive perception of tourism's impacts by residents, which is based to a certain extent on a favourable evaluation of their interactions with tourists, will lead to a positive attitude towards tourists visiting their communities. Thus, we establish the following hypotheses for each category of tourism's impacts considered in this study: H7: The residents' perceptions of the positive economic impacts of tourism positively influence their attitudes towards tourists in their communities.
H8: The residents' perceptions of the negative economic impacts of tourism negatively influence their attitudes towards tourists in their communities.
H9: The residents' perceptions of the positive sociocultural impacts of tourism positively influence their attitudes towards tourists in their communities.
H10: The residents' perceptions of the negative sociocultural impacts of tourism negatively influence their attitudes towards tourists in their communities.
H11: The residents' perceptions of the positive environmental impacts of tourism positively influence their attitudes towards tourists in their communities.
H12: The residents' perceptions of the negative environmental impacts of tourism negatively influence their attitudes towards tourists in their communities.

Residents' attitudes and behavioural support for tourism
According to Sharpley (2014), one of the main limitations of past research on residents is that the influence of perceptions and attitudes on subsequent responses or behaviours has rarely been examined. To explain this link, it is fitting to use the theory of reasoned action (TRA) and the theory of planned behaviour (TPB) as references. According to these theories (Ajzen, 1985;Ajzen & Fishbei, 1980), people's behaviour is a function mainly of their attitudes, which are defined by Eagly and Chaiken (1993) as psychological tendencies expressed by evaluating an "object" (in the context under investigation the object would be the host community). These theoretical approaches have been successfully used to predict a wide variety of behaviours based on individual attitudes, among them the protourism behaviours (Lepp, 2007). Nunkoo and Ramkissoon (2010) use TPB to theoretically develop a model linking impacts, attitudes, and support for tourism. These authors postulate that if residents have positive attitudes towards tourism, then they will engage in behaviours supporting the tourism activity in their communities.
Some key behaviours in this field would include the recommendation to other people of the host communities, that is advocacy behaviour (Palmer et al., 2013), as well as participation in tourism activities in their communities, that is self-consumption (O'Shaughnessy & O'Shaughnessy, 2003). According to Andereck and Vogt (2000) and Andriotis (2005), residents' attitudes would determine the extent to which the local communities will engage or not engage in these behaviours that offer support for tourism. Palmer et al. (2013) in particular provided some empirical evidence of a positive relationship between the residents' attitudes towards tourists and their propensity to engage in advocacy to support inward tourism. In line with this, our study postulates that a positive attitude towards tourism and tourists will lead to a more positive behaviour supporting tourism in host communities. Therefore, our purpose is to make advancements in the academic literature by introducing the variable of behavioural support and postulating that: H13: The more positive the residents' attitudes towards tourism, the more positive their behaviour supporting tourism in host communities.
H14: The more positive the residents' attitudes towards tourists, the more positive their behaviour supporting tourism in host communities. Figure 1 illustrates the research hypotheses established in our study.

Methodology
In order to test the hypotheses, empirical research was developed in the region of Cantabria, which is located in the north of Spain. This country is the third destination in the world in terms of international tourists incoming (United Nations World Tourism Organization [UNWTO], 2017). In addition, Spain was ranked as the most competitive destination worldwide in 2016 (World Economic Forum, 2017), thus being a good benchmark for collecting empirical data on resident perceptions and attitudes in tourism. It is necessary to indicate that, despite the fact that Spain is internationally recognized for its sun-and-beach tourism, this activity sector is very heterogeneous depending on each specific region. In particular, while sun-and-beach tourism is predominant on the Mediterranean coast and the Spanish islands, northern and interior Spain show a more heterogeneous tourism offer, with great importance attached to natural and cultural heritage. Specifically, taking as a reference the extrinsic variables influencing residentsʼ perceptions, as established by Sharpley (2014), tourism in the region of Cantabria can be defined as a non-mass and seasonal activity, which is mainly based on the natural resources and cultural heritage of the place. Therefore, the host community under investigation is reasonably free of bias caused by an intensive development of sun-and-beach resorts, with a high potential impact in environmental and sociocultural terms. In addition, visitors are mainly from other Spanish regions, and international tourists only represent a 15-20% of the visitors. In this sense, a non-intensive arrival of foreign tourists also reduces the negative potential bias with regard to the formation of perceptions and attitudes among residents.
The target population of empirical research consisted of residents over 18 years old. In particular, data were collected using a personal questionnaire that included the following: (1) the residents' perceptions of the positive and negative impacts usually related to tourism activity; (2) the residents' attitudes towards tourism and tourists; (3) the residents' behavioural support for tourism in host communities; and (4) the sociodemographic characteristics of respondents. The variables of the theoretical model were all measured using multi-attribute instruments adapted from previous works in order to assure their content validity (Appendix). In particular, residents' perceptions of the positive and negative impacts of tourism were measured taking as an initial reference frame the set of impacts considered in several recent empirical works (Nunkoo & Gursoy (2012), Prayag et al. (2013), Garau et al. (2014), Sinclair et al. (2015), Vargas et al. (2015), and Wang & Xu (2015)). We checked the list of impacts to avoid duplications and, subsequently, three main impacts for each category were selected (i.e. those that better fit the characteristics of the host community under investigation). Attitudes towards tourism were measured with an instrument adapted from the study of Wang, Pfister, and Morais (2006), while attitudes towards tourists were captured based on the instruments developed by Teye et al. (2002) and Palmer et al. (2013). Finally, behavioural support for tourism was measured taking as a reference the concept of residents' participation in tourism, as established by Palmer et al. (2013). It is necessary to indicate that all the items were measured with a 10-point Likert scale in order to ensure the variability of the data.
The sampling strategy consisted of two stages and methods; in the first phase, we used a quota method to build the profile of respondents according to the characteristics of the population under investigation. In particular, we used the statistics provided by Cantabrian Institute of Statistics (ICANE) , an observatory of the population in the region of Cantabria, to determine the distribution (or percentages) of residents by gender and age. In a second stage, we used a convenience method to define the geographical areas for data collection. Particularly, we selected the main areas in the region of Cantabria according to the statistics of arrival of tourists provided by ICANE; for example, the municipality of Santander or the Area of Torrelavega were selected for data collection. With this information, the interviewers were distributed in the different geographical areas and were trained to administer the questionnaire without causing bias in the responses, obtaining a valid sample of 619 residents (Table 2 provides the sociodemographic characteristics of the respondents). In relation to the gender and age of respondents, it is important to indicate that the survey sample is close to the target population, thus achieving an adequate level of typological representativeness of our sample.

Results
A structural equation model (SEM) approach was used in order to test our research model. Firstly, the reliability and validity of the measurement scales were checked by means of a confirmatory factor analysis using EQS 6.1 software ( Table 3). The fit criteria indicate the extent to which the factorial model fits the data. In particular, three types of criteria are relevant to assess the model fit for the data: measures of absolute fit (Bentler and Bonett's Nonnormed Fit Index -BBNNFI is very near the recommended valued of 0.90, and Root Mean Square Error of Approximation -RMSEA is below 0.08), measures of incremental fit (Incremental fit index -IFI and Comparative fit index-CFI are above the recommended value of 0.90), and measures of parsimonious fit (normed χ 2 is below 3.0).
The reliability of our measurement scales was evaluated using the Cronbach's alpha, compound reliability, and AVE coefficients (Bagozzi & Yi, 1988) (Table 3). The values of Cronbach's alpha and compound reliability are, in almost every case, clearly above the required minimum values of 0.7 (Hair, Black, Babin, & Anderson, 2010). AVE coefficients are above the recommended value of 0.5 (Hair et al., 2010), except in the cases of "Behavioural support for tourism", "Attitude towards tourism", and "Perception of positive sociocultural impacts". However, it is necessary to indicate that (1) in the last two cases the AVE coefficients are very close to 0.5; and (2) "Behavioural support for tourism" is a construct scarcely studied in the literature, and its scale may require refinement in further research, even though the values obtained for the Cronbach's alpha and composite reliability are reasonably good. Accordingly, we consider that the results obtained for supporting the reliability of our scales are acceptable.
The convergent validity of our scales was also confirmed (see Table 3), since all items are significant to a confidence level of 95% and their standardized lambda coefficients are higher than 0.5 (Steenkamp & Van Trijp, 1991). The discriminant validity of the scales was tested following the procedure proposed by Fornell and Larcker (1981), which requires the comparison of the variance extracted for each pair of constructs (AVE coefficient) with the squared correlation estimated between constructs. Despite the low value of the AVE coefficient for the latent variable "Behavioural support for tourism", the results summarized in Table 4 show that there are no discriminant validity problems as the AVE is higher than the squared correlation between each pair of latent variables. This, together with the reasonably good indicators obtained for convergent validity (i.e. Lambda coefficients) and reliability (i.e. Cronbach's alpha and composite reliability), support the general psychometric properties of the latent factor "Behavioural support for tourism". Accordingly, empirical results support the psychometric properties of the scales used in this research.
Next, the structural model was estimated to test our research hypotheses. The analysis was run using a robust maximum-likelihood estimation procedure, which avoids the problems related to non-normality of data by providing the outputs "robust chi-square statistic" and "robust standard errors". These outputs have been corrected for non-normality (Byrne, 1994) and, consequently, guarantee the validity of the model estimation. The results obtained in the first estimation of the model show that six research hypotheses are not significant. In particular, the residents' attitudes towards tourism are not significantly affected by their perceptions of the negative economic impacts or by their perceptions of positive/negative environmental impacts (H4, H5, and H6 are rejected). In addition, the residents' attitudes towards tourists are not significantly affected by their perceptions of negative economic and sociocultural impacts or by their perceptions of positive environmental impacts (H8, H10, and H11 are rejected). Following the procedure proposed by Hair et al. (2010), we proceeded to the re-specification of the model by eliminating the causal relationships not supported by empirical data. Once the model was re-specified, the goodness-of-fit indices obtained were within the recommended values (Table 5). Specifically, our results indicate that the residents' attitudes towards tourism are positively influenced by their perceptions of the positive economic and sociocultural impacts, and negatively influenced by their perceptions of negative sociocultural impacts (H1, H3, and H4 are confirmed). In addition, our results confirm that the residents' attitudes towards tourists are positively influenced by their perceptions of the positive economic and sociocultural impacts, and negatively influenced by their perceptions of negative environmental impacts (H7, H9, and H12 are confirmed). In addition, it is important to note that residents' behaviours supporting tourism in their communities are influenced in a similar way by their attitudes towards tourism and their attitudes towards tourists (H13 and H14 are confirmed). Finally, it is necessary to indicate that the R 2 statistics for each dependent variable were higher than 0.4 (or very close to this recommended value), thus reflecting that a significant percentage of variation in the dependent variables was explained by the independent variables (see Table 5).

Conclusions
Residentsʼ perceptions of tourism's impacts, and their influence on attitudes towards tourism in host communities, have received a lot of attention in previous research. However, several recent studies demand more effort to examine in greater depth this relevant topic for academics and practitioners. In this sense, our study represents a new step in the study of the tourism phenomenon from the residentsʼ point of view by taking a different and more comprehensive approach. On the one hand, our study generates new knowledge about the chain of effects from the perceptions of tourism's impacts to attitudes in the context of communities hosting non-mass tourism (in contrast to the most studied communities that host mass tourism, mainly based on sun and beach). On the other hand, our study examines not only the relationships between resident perceptions and attitudes, but also the links between resident attitudes and subsequent behaviours. Particularly, our study develops a new model that is based on the "perceptions-attitudes-behaviours" sequence and includes two types of resident attitudes, namely attitudes towards tourism and attitudes towards tourists, and the novel concept of behavioural Table 5. Results for the re-specified structural model. support, namely residentsʼ behaviours supporting tourism in host communities. With regard to the formation of attitudes, our results confirm that the residents' perceptions of the positive economic impacts of tourism are the main variable determining their attitudes towards tourism development in their communities. This finding is in line with Vargas et al. (2015), who establish that the perceptions of these specific impacts have been identified to be the most influential factor on resident attitudes formation. In addition, according to the Social Exchange Theory, residents aim to obtain satisfactory outcomes with tourism developed in their communities. Thus, in a context of an economic crisis which many places experience nowadays, it is logical to think that the perceptions of the economic opportunities that tourism generate in host communities (for example, job opportunities and new businesses) are the most critical factor for residents in the formation of their attitudes towards tourism.
Furthermore, another important contribution of this study is related to the formation of other types of resident attitudes, namely those linked to the tourists, which have received little academic attention in literature. Unlike the study of Palmer et al. (2013), which defines these attitudes as an antecedent to advocacy behaviour, our paper also examines how residentsʼ attitudes towards tourists are influenced by their perceptions of tourism's impacts. In this context, it has been empirically demonstrated that the perception of the positive sociocultural impacts of tourism is the most important variable influencing attitude towards tourists in host communities. This finding is coherent with the development of the Social Representations Theory in tourism. More concretely, it is reasonable to believe that positive sociocultural impacts are the category of tourism's impacts more likely to be conditioned by the interactions between residents and tourists in host communities (a key element in the formation of social representations in tourism). Therefore, the perceptions of these specific impacts would represent a key factor in the formation of attitudes towards tourists. In addition, these attitudes are also affected by the perceptions of both the positive economic impacts and the negative environmental impacts, but to a lesser extent.
In relation to tourism's impacts that are not found to influence the residents' attitudes (i.e. negative economic impacts and positive environmental impacts), it may be due to two reasons. First, the tourism development model of the region under investigation is not based on a mass tourism model, so the negative economic impacts (usually attributed to conventional tourism) are relatively small and, consequently, residents may not consider them a key element in the formation of their attitudes. Second, one of the main motivations for tourists visiting this Spanish region is the contact with nature, so residents may place some importance on the negative environmental impacts of this type of tourism, but they may not particularly consider the positive ones when they form their attitudes.
Finally, it is also important to emphasize that our paper represents a significant effort to generate new knowledge about the influence of attitudes on subsequent behaviours or responses of residents. Compared with some previous studies such as those by Látková and Vogt (2012), Nunkoo and Ramkissoon (2010), Nunkoo and So (2016), Rasoolimanesh, Ringle, Jaafar, and Ramayah (2017), and Wang and Pfister (2008), our study goes further in analysing these responses and examines the concept of behavioural support for tourism, which represents a more reliable variable to measure the true commitment to tourism than the attitudinal support. Particularly, in line with previous studies that have demonstrated an "attitude-behaviour" link based on the TRA or TPB, our results confirm that residentsʼ attitudes positively influence behaviours supporting tourism in host communities. In particular, resi-dentsʼ behaviours in terms of participation in tourism activities in their communities and recommendation of these places to other people (i.e. behavioural support) are affected, in a similar way, by their attitudes towards tourism and their attitudes towards tourists.

Managerial implications
In a review of previous research, Sharpley (2014) emphasizes the significance of examining not only the perceptions-attitudes relationship, but also the responses or actions of residents from a management perspective. In this sense, we recommend that destination managers develop a variety of social responsibility activities in order to enhance the quality of the relationship with residents and, therefore, reinforce their attitudes and behavioural support for tourism. These activities, which require a close coordination between the public sector and the private sector (Su, Wang, Law, Chen, & Fong, 2017), should be mainly focused on different campaigns that raise awareness of the positive impacts of tourism among local residents. Specifically, destination managers may use both conventional communication (for example, advertising or public relations) to target older people, as well as communication 2.0 (for example, social networks or photo-sharing websites) to target younger people, in order to achieve this goal. In addition, destination social responsibility should pay attention to other issues such as enhancing the physical and cultural well-being of residents, increasing the participation of the local community in decision-making processes, or complying with the different kinds of regulations.
Although our findings indicate that these positive tourism impacts are the most influential factors in the formation of residents' attitudes, destination managers should not forget to minimize the negative impacts typically perceived by residents related to tourism activity. More concretely, managers should pay special attention to the negative sociocultural and environmental impacts in their host communities. With this in mind, the objective should be to involve residents in different activities, such as workshops or meetings in a natural environment, for the purposes of eliminating certain stereotypes surrounding the cultural and environmental damage of tourism activity and/or explaining the measures developed by the communities to reduce undesirable tourist behaviours.

Limitations and further research
Despite the rigorous methodology used in empirical research, this study has several limitations that should be managed in future research. First, the fact that this research focuses on a specific Spanish region could be a constraint in the generalization of our results. In this sense, it would be interesting to replicate empirical research in other communities with different types of tourism or different states of development. Specifically, host communities promoting new products as alternatives to sun-and-beach tourism may be interested in examining how the different types of tourism influence the perceptions and attitudes of residents. Thus, policymakers can develop their tourism development strategies in a more efficient way. Second, the analysis of the "perceptions-attitudesbehaviours" sequence has been conducted for all residents considered in empirical research. It would be very interesting to examine how the model works depending on different variables such as the personal values or sociodemographic characteristics of residents. Finally, some of the measurement scales show values below 0.5 for the AVE coefficient, which might raise some doubts about their reliability. However, the values obtained for Cronbach's alpha and compound reliability support inner reliability of all the scales. This problem is slightly bigger with regard to the dependent variable (i.e. behavioural support for tourism), whose operationalization should be refined by revising the proposed items and adding other behaviours such as volunteering in tourism activities or buying local products in their communities.

Disclosure statement
No potential conflict of interest was reported by the authors.