Research Online Research

This study aimed to develop and test a conceptual model of the antecedents of residents ’ attitudinal behavior to support future film tourism in a popular tourism destination. These antecedents include: socio-cultural impacts of existing tourism, place attachment, external perceived realism of media programs, and perceived importance of residents ’ involvement. A total of 416 survey questionnaires were completed from residents living in the province of Seville, Spain. Partial least squares-structural equation modelling (PLS-SEM) was used to examine the relationship between these factors and the residents ’ support for future film tourism (RSFFT). The results indicate that place attachment and perceived importance of residents ’ involvement are the most significant predictors. Also, it confirms that the importance of residents ’ external perceived realism of movies and TV series shot in their locale is a key predictor of the RSFFT.

One of the most explored antecedents of tourism support has been residents' perceptions of tourism impacts that consist of economic, socio-cultural, and environmental benefits and drawbacks. However, previous research has been skewed toward economic impacts since tourism has been predominantly used as a means for economic (re) development. The limited research with particular attention to sociocultural impacts of tourism on host communities (Eyisi, Lee, & Trees, 2021;Jaafar et al., 2017) is surprising because "socio-cultural impacts are the 'people impacts' of tourism" (Glasson, Godfrey, & Goodey, 1995, as cited in Mbaiwa, 2005. That said, these impacts influence how tourism transforms residents' daily lives, values and identities. Therefore, the ability to maintain tolerable levels of socio-cultural impacts on the host communities can be a deciding factor in the sustainability of tourism (Nunkoo & Ramkissoon, 2007).
Previous research has also examined the influence of residents' place attachment (Ganji, Johnson, & Sadeghian, 2021;Stylidis, 2020) and involvement in tourism planning and development (Gannon, Rasoolimanesh, & Taheri, 2021;Rasoolimanesh, Ringle, Jaafar, & Ramayah, 2017) on tourism support. However, as discussed in the next section of literature review, the inconclusive results justify the need to further explore these relationships, particularly in the context of film tourism where these relationships have been barely analyzed with the only exception (i.e., Kim, Kim, & Oh, 2017).
Moreover, previous film tourism studies have anecdotally documented residents' concerns about media representation that generated film tourism flows, and concluded that residents do not appreciate, for example, stereotypical and idealized representations of their place of residence (Chen & Mele, 2017;Mordue, 2009). However, no previous study has directly measured the importance that residents give to the match between the actual world and the portrayals circulated by the media production, theorized as 'external perceived realism' (Bilandzic & Busselle, 2011). This is ironic, given that residents are the ones who have to "live with the consequences of what image has been portrayed, and to whom the image has been portrayed" (Heitmann, 2010, p. 39).
All in all, there is still a lack of in-depth understanding of the underlying mechanisms and structures through which local residents lead toward the RSFFT in their locale. Moreover, previous film tourism research has been primarily conducted in non-mature film tourism destinations (Beeton, 2016;Croy et al., 2018;Kim & Park, 2021;Law, Bunnell, & Ong, 2007;Mordue, 2009).
Having acknowledged these critical limitations, the aim of the current study is twofold. Theoretically, this is the first study that introduces the importance of external perceived realism (Bilandzic & Busselle, 2011) of media programs to the field of tourism studies to investigate the extent to which it affects residents' support for future film tourism (RSFFT) and the importance of residents' involvement in its planning and development. Furthermore, the study aims to empirically examine the close, yet complex, structural relationships that a set of antecedents (namely, perceived socio-cultural impacts of existing tourism, residents' place attachment, perceived importance of their involvement, external perceived realism of media programs) have on the RSFFT in a popular tourism destination. Practically, the study sheds light on identifying the factors relevant stakeholders should pay attention to in order to facilitate and maximize residents' support for film tourism development.

Perceived socio-cultural impacts of tourism and Residents' support for future film tourism
Residents' perceptions of tourism impacts have been extensively used as an antecedent to support for tourism development in various contexts (Çelik & Rasoolimanesh, 2021;Nunkoo & Ramkissoon, 2011;Qin, Shen, Ye, & Zhou, 2021;Rasoolimanesh et al., 2017;Sharpley, 2014;Stylidis, 2016;Woosnam, 2012). Previous studies have primarily relied on the triple bottom line approach, which distinguishes between residents' perceptions of the environmental, economic, and sociocultural effects of tourism.
Loosely adopting the Social Exchange Theory, such studies collectively conclude that the support of residents for tourism development depends on equity scores evaluated in terms of the perceived costs and benefits, which is suggestive of the notion that this support tends to be withdrawn if the former outweighs the latter (Kim & Park, 2021;Sharpley, 2014). Furthermore, the triple bottom line approach allows reaching a rounded understanding of residents' perceptions of tourism impacts, given that the approach reveals specific insights into how each pillar of perceived tourism impacts influences residents' support for tourism development.
Despite these contributions, research with particular focus on sociocultural impacts of tourism is relatively limited (Eyisi et al., 2021;Jaafar et al., 2017), although maintaining socio-cultural impacts on the host community at tolerable levels is a determinant of tourism sustainability (Nunkoo & Ramkissoon, 2007). In this vein, prior studies have empirically confirmed the significant role that perceived positive socio-cultural impacts have on residents' support for tourism development, whereas negative socio-cultural effects negatively impact residents' support for tourism (Ahadian, 2013;Chen, 2001;Dyer, Gursoy, Sharma, & Carter, 2007;Stylidis, Biran, Sit, & Szivas, 2014).
In the context of film tourism, drawbacks such as the overcommercialization of local culture and heritage and conflict between community members created a skeptical view on future support for film tourism (Beeton, 2016;Kim & Park, 2021;Mordue, 2009;Thelen et al., 2020;Yoon, Kim, & Kim, 2015).
The following is thus hypothesized:

Perceived socio-cultural impacts of tourism and perceived importance of Residents' involvement
Residents' perceptions of socio-cultural impacts of tourism can also influence the perceived importance of residents' involvement in new tourism initiative. For example, Jaafar et al. (2017) confirmed that residents of Langkawi Island, Malaysia, with positive socio-cultural perceptions of tourism impact are willing to participate in new or further tourism development, while they equally perceived the importance of their involvement in future tourism planning and development to reduce negative tourism impacts. Similarly, prior film tourism research collectively suggested that socio-cultural positive impacts such as community pride (Connell, 2005;Croy et al., 2018;Kim, Suri, & Park, 2018;Mordue, 2009;Park, 2018;Thelen et al., 2020) can lead to the enthusiasm of local residents for their possible engagement and/or involvement in future film tourism planning and development, considering the importance of their involvement.
Drawing on the above, the following is hypothesized: Hypothesis 3. The perceived positive socio-cultural impacts of existing tourism have a positive effect on the perceived importance of residents' involvement in future film tourism planning and development.
Hypothesis 4. The perceived negative socio-cultural impacts of existing tourism have a positive effect on the perceived importance of residents' involvement in future film tourism planning and development.

Perceived importance of Residents' involvement and Residents' support for future film tourism
The relationship between perceived importance of residents' involvement in future film tourism planning and development and their support for future film tourism (RSFFT), however, has received limited scholarly attention. This may be because local authorities have frequently failed to produce any substantial film tourism plans or development programs (Kim et al., 2018), due to the often sporadic and sudden influxes of film tourists. Consequently, reactions and planning are commonly opportunistic and ad-hoc (Beeton, 2016;Connell, 2005;Heitmann, 2010;Kim et al., 2017;Mordue, 2009;Yoon et al., 2015). Moreover, even if programs and plans for film tourism development are in place, residents are either not invited to play an active role (Yoon et al., 2015) or are dissatisfied with any discussions that are organized (see Mordue, 2009).
Regardless of the limited empirical support for a hypothesized relationship between perceived importance of residents' involvement and residents' support for future film tourism (RSFFT), the study on Iran's historical cities by Gannon et al. (2021) and the work of Rasoolimanesh et al. (2017) on Malaysian world heritage sites confirmed that community participation in decision making has a positive effect on the backing for future tourism development in broader contexts. Consequently, it is hypothesized that: Hypothesis 5. The perceived importance of residents' involvement in future film tourism planning and development is positively related to residents' support for future film tourism.
2.4. Place attachment, Residents' support for future film tourism, and perceived importance of Residents' involvement Place attachment, the relationship between an individual and a place, has been used to explore residents' support for tourism development (Eusébio, Vieira, & Lima, 2018;Gursoy, Chi, & Dyer, 2010;Jurowski, Uysal, & Williams, 1997;Wang & Chen, 2015). However, prior studies have arrived at inconclusive results. While some have showed that place attachment has little significant effect on residents' support for tourism development (Vargas-Sánchez, do Valle, da Costa Mendes, & Silva, 2015;Stylidis, 2020), others have concluded that highly attached residents are more willing to support tourism development than their less-attached peers (Eusébio et al., 2018;Gursoy & Rutherford, 2004;Lee, 2013;Stylidis, 2018).
In the context of film tourism, Kim et al. (2017) on Korean towns in remote and peripheral regions that experience film tourism seemed to support that people with a stronger attachment to their place of residence are more likely to express stronger support for further development of film tourism.
Place attachment has been also used in tourism studies to analyze the importance of residents' involvement in their local community and its surroundings (Manzo & Perkins, 2006). In their research into residents' place-related behaviors, Chen and Dwyer, 2018 conclude that "the more strongly residents are attached to a place, the more willing they are to forsake personal resources" (e.g., time, financial) to preserve their relationship with it. This willingness is consistent with the argument of Plunkett, Phillips, and Kocaoglu (2018), who postulate that those who are more attached to their community may have additional motivations for not only continuing to live in that place, but also for protecting what they have and making improvements to it. Accordingly, Hypotheses 6 and 7 are theorized within the context of film tourism: Hypothesis 6. Place attachment has a positive effect on residents' support for future film tourism.
Hypothesis 7. Place attachment has a positive effect on perceived importance of residents' involvement in future film tourism planning and development.
Consequently, the current study hypothesizes that place attachment is a multidimensional, higher-order latent factor consisting of four firstorder elements/dimensions: place identity (the symbolic and emotional attachment an individual has with the place they inhabit), place dependence (the capacity of a place to facilitate residents' living experiences), bonds with friends and family (the social connection with relatives and friends living in the same locality), and place familiarity (the knowledge of a locale and any pleasant memories associated with it).
2.5. External perceived realism of media programs, Residents' support for future film tourism, and perceived importance of Residents' involvement External perceived realism has been defined as an audience's perception of an existing match between a narrative constructed in a media program and the actual world (Bilandzic & Busselle, 2011). According to Bilandzic and Busselle (2011), external perceived realism is infringed when the narratives presented in media representations are inconsistent with the experiences and expectations of audiences with and about the world in reality. This can occur, for example, when residents as audiences perceive that the places where they live are misrepresented in media products.
Earlier studies have described well-documented community concerns about the misrepresentation of their places, cultures and/or peoples within the processes of media production (Beeton, 2016;Chen & Mele, 2017;Kim & Butler, 2015;Póvoa, Reijnders, & Martens, 2019;Thelen et al., 2020). Kim and Butler (2015), for example, describe how a small, sleepy, remote South Australian town became a film tourism hotspot after the release of the movie Snowtown (2011). The influx of tourists caused various social, spatial, and psychological conflicts between, on the one hand, the tourists' quest to experience the infamous murder-focused film as a consumable social construction of reality, and, on the other, the residents' everyday reality, which had little relevance to the story of Australia's worst-ever serial killings portrayed in the movie.
Occasionally, residents cooperate to challenge media representations. For instance, after the release of the gangster movie Monga (2010), some residents and community activists in Taiwan organized a photography exhibition featuring local voices and visions, with tourists invited to explore the actual location in a more realistic way (Chen & Mele, 2017). This was in opposition to the way in which city authorities initially celebrated the anticipated economic impact of the film on the region.
Notwithstanding the role of (fictional) narratives in explaining the expectations of film tourists (Beeton, 2016), their satisfaction with and intention to visit an area (Kim, 2012), and place identities (Reijnders, 2020), little is known about, first, the importance that residents give to external perceived realism when their places of residence are used as backgrounds and/or foregrounds in media productions and, second, the impact that these perceptions may have on residents' involvement in and their support for future film tourism planning and development. Acknowledging this, two additional hypotheses have been formulated for this study: Hypothesis 8. External perceived realism of their places of residence in media programs has a positive effect on residents' support for future film tourism.

Hypothesis 9. External perceived realism of their places of residence in media programs has a positive effect on perceived importance of residents' involvement in future film tourism planning and development.
Finally, previous studies have concluded that residents' relationship with the tourism sector influence their support for future tourism development, indicating that those that obtained direct advantages from this industry adopt more supportive attitudes toward it (e.g., Erul et al., 2020;McGehee & Andereck, 2004;Oviedo-Garcia et al., 2008). Therefore, in the present study the relationship with the tourism sector (i.e., personally working in tourism, having close friends/relatives working in tourism, and neither working themselves nor having close friends/relatives working in tourism) will be used as a control variable to explore the effect it has on the RSFFT. Figure 1 illustrates the conceptual model tested in this study, with place attachment again represented as a higher-order construct reflecting the first-order factors/dimensions of place identity, place dependence, friends and family bonding, and place familiarity. The previously hypothesized relationships between the constructs are further represented via the directional arrows in Fig. 1.

Research context
The research context focuses on the province of Seville in Spain, the second most visited country worldwide prior to COVID-19 pandemic, attracting 83 million arrivals (UNWTO, 2019). This Spanish province was chosen for two main reasons. First, the province of Seville has been the filming location of multiple national and international media productions like Game of Thrones, Star Wars and Lawrence of Arabia. Film tourism has been considered a viable tourism market and thus actively promoted by the Andalucía Film Commission via their project entitled "Andalucía Destino de Cine" that was launched in 2018. Second, the capital of the province of Seville, also named Sevilla, is the third-most visited city in Spain; and its region, Andalusia, is the second-most visited Spanish region in, 2019 (INE, 2019). In sum, the study aims to develop more rigorous theoretical and empirical foundations to improve current knowledge on the antecedents of future film tourism support from the residents' perspectives in a popular tourism destination.

Measurements
The scales used to measure the constructs proposed in the present study were based on measures drawn from the previous relevant literature. Place attachment was measured using a total of 13 items, with 4 items each used to measure 'place identity' and 'place dependence' taken from Kyle et al. (2004) and Lee (2011), 2 items used to measure 'friends/family bonding' taken from Raymond et al. (2010) and Taima and Asami (2019), as well as 3 items used to measure 'place familiarity' taken from Hammitt et al. (2006).
The perceived positive and negative socio-cultural impacts of existing tourism were measured using 8 and 4 items, respectively, with all previous items taken from existing studies by Yoon et al. (2015) and Jaafar et al. (2017). The perceived importance of residents' involvement in future film tourism planning and development was measured using 3 items adapted to this study's particular research context from Choi and Sirakaya (2005). They were initially used to measure community participation as part of a scale aimed at assessing residents' attitudes toward future sustainable tourism development.
The external perceived realism was measured using a newly developed single item inspired from Bilandzic and Busselle (2011) with the following description: "It is said that a film or TV series is realistic when that which appears on the screen is adjusted to the real world. This applies to, for example, the way in which scenes, people, culture, etc. are represented. Considering this, how important is it for you that in a movie/series filmed in Seville (province) this location is portrayed in a realistic manner?". Finally, residents' attitudinal behavior to support future film tourism (RSFFT) was explored using 4 items, adapted from Rasoolimanesh et al. (2017) for support for tourism. All proposed items were measured using a 5-point Likert scale, where the answers ranged from 1 = Strongly Disagree and 5 = Strongly Agree for all items except the importance of external perceived realism which in this case ranged from 1 = not at all important and 5 = very important in this case.
The items used to measure each construct are summarized in Table  A1 (see Appendix including a full survey questionnaire), along with their means, standard deviations, coefficient of variation (CV) (the ratio of the standard deviation to the mean), as well as their skewness and kurtosis scores and the normality results from the Kolmogorov-Smirnov test. In particular, the standard deviations for all items were relatively small, ranging between 0.81 and 1.34 and with the corresponding CV below the threshold of 1 for all items (see Hair, Anderson, Babin, & Black, 2010). This suggests the absence of outliers in the data, which is expected when using a Likert scale, as in this case all the data is restricted between 1 (strongly disagree/not at all important) and 5 (strongly agree/very important) and cannot fall outside this interval (Leung, 2011).
However, the results of the normality test for all variables as well as skewness results for some of the variables (i.e., absolute value of skewness; in this case >2) showed further deviation from normality, which is also an expected result with Likert scales because most of the responses tend to fall either to the left or right of the neutral point, skewing the dataset (see Leung, 2011). This lack of normality was accounted for in the model analysis using partial least squares structural equation modelling (PLS-SEM) rather than covariance-based structural equation modelling (CB-SEM) as explained next (see Chin, 2010), whereas the former helps minimize possible errors due to lack of normality in this case. In addition, Common Method Variance (CMV) was examined using Harman's one-factor test (see Podsakoff, MacKenzie, Lee, & Podsakoff, 2003) with the results indicating that more than one factor accounted for the majority of co-variation in our data, and with no single factor accounting for >50% of the co-variation in this case, thus suggesting that common method bias is not a concern in this case.
Finally, the instrument was originally created in English and doubletranslated into Spanish by two Spanish native speakers. Furthermore, and to double-check the accuracy of the translation, a back-translation was conducted by a professional translator. Moreover, prior to the main survey, the questionnaire was pilot-tested in July 2020 on 30 people living in Seville. Of them 25 were reached through Dynata's panel, whilst the other five were recruited through snowball sampling by the first author of the paper. The answers from those 30 respondents were used to uphold the unidimensionality, consistency, and clarity of the adopted constructs and to make slight wording changes to the survey instruments prior to conducting the main survey.

Sample size and data collection
The survey was administered online, and respondents were selected by means of quota sampling considering age, gender, length of residence, and relationship with the tourism sector (i.e., personally working in tourism, having close friends/relatives working in tourism, and neither working themselves nor having close friends/relatives working in tourism). To calculate the percentages of each quota, the statistical data from Junta de Andalucía . The screening question was introduced to appropriately approach and sample respondents who were familiar with media productions filmed in their places of residence and thus could better reflect upon: (1) the importance of external perceived realism and (2) the attitudinal behavior to support film tourism development as two key constructs of the present study. It is also noteworthy that the 416 questionnaires successfully passed a quality check which consisted of removing questionnaires that contained straight liners or were completed in <33% of the median duration.
Of the respondents 98.3% (n = 409) were recruited from an online survey panel managed by the company Dynata and the other 1.7% (n = 7) came from Qualtrics. The decision to distribute the same survey via Qualtrics was made to recruit more people working in the tourism sector, given that the employment in the tourism sector had drastically dropped during the first summer of the COVID-19 pandemic and thus the research team had difficulties in recruiting this particular group of respondents, when the data collection took place between 19 August to 21 October 2020. Therefore, the link of Qualtrics was sent via email to multiple tourism-related companies and organizations such as the reginal tourism offices. Almost half 53% (N = 220) were women and 47% (N = 196) were men, with 38 as the average age of the respondents (minimum age = 18, maximum age = 64, and S.D. = 12). Moreover, 45% of the respondents held a university degree, 22% had a technical/professional degree, 31% had a secondary/high school education, and 2% had only a primary school education. Additionally, respondents have been living in the province of Seville for an average of 29 years (minimum residence of one year, maximum residence of 64 and S.D. = 16).
Finally, 52% (N = 215) of respondents were involved in tourism in their city/town, either directly (i.e., personally working in tourism) or indirectly (i.e., having close friends/relatives working in tourism), as opposed to 48% (N = 201) who were not involved in tourism (i.e., neither working themselves nor having close friends/relatives working in tourism). This latter variable of relationship with tourism sector was further used as control variable to test for the potential statistical effect of residents' involvement in the tourism sector on the model results, particularly with reference to residents' support for future film tourism (RSFFT) in order to determine if respondents' level of support for film tourism remained the same regardless of whether respondents were involved in tourism or not. Table 1 summarizes the respondents' characteristics.

Data analysis
The partial least squares structural equation modelling (PLS-SEM) using XL-STAT software version 2020 (Addinsoft., 2020) was used to analyze the model presented in this study (Fig. 1), with PLS-SEM in this case centering around two steps: 1) upholding the outer part of the model through confirmatory factor analysis (CFA) to confirm the convergent and discriminant validity of the various constructs assumed in the model, namely place attachment and its dimensions (i.e. place identity, dependence, bonding, and familiarity), as well as the positive and negative socio-cultural impacts of existing tourism, perceived importance of residents' involvement, external perceived realism of media programs, and RSFFT constructs; and 2) examining the inner (structural) part of the model through regression-type path analysis to validate the causal relationships hypothesized between the model constructs (see Chin, 2010).
For this study, PLS-SEM was used rather than the traditional covariance-based structural equation modelling (CB-SEM), due to 1) data deviation from normality as demonstrated through the normality test and skewness/kurtosis measures, as mentioned earlier; and 2) the higher-order nature of the place attachment construct, whereby one of the place attachment dimensions (i.e., bonding) has two indicators. As such, PLS-SEM in this case helps overcome both issues (lack of normality and presence of higher/multidimensional construct) by minimizing model errors and allowing the model to converge to stable results, respectively, which would have not been the case if CB-SEM had been used instead (Do Valle & Assaker, 2016).

Partial least squares analysis
PLS-SEM analysis was run on the full dataset of the unstandardized data, using mode A (reflective scheme) for all model constructs (including the higher-order construct of place attachment) and using the centroid method to estimate the regression coefficients of the inner model (Chin, 2010).

Outer part results
Results of the confirmatory factor analysis (CFA) first upheld the convergent validity of all constructs, as loadings for each construct including the higher-order placement attachment one on its first order dimensions of place identity, dependence, bonding, and familiarity exceeded or were close to 0.7 (see Table 2), suggesting that all items (or dimensions, in the case of the higher-order place attachment) successfully measured the underlying constructs to which they were assigned (Hair Jr, Hult, Ringle, & Sarstedt, 2016). The only exceptions were: 1) "item 1" of the negative socio-cultural impacts of existing tourism, 2) "item 4" of RSFFT, and 3) the "bonding" dimension of the higher order place attachment construct, showing loadings of 0.409, 0.587 and 0.566, respectively (see Table 2).
Nevertheless, all three loadings were still greater than the 0.4 threshold suggested by Henseler, Ringle, and Sinkovics (2009), with the bootstrap test showing high significance levels for all three loadings (bootstrap-based empirical 95% confidence interval does not include zero; see Table 2). This further supports convergent validity and the retention of those three items within their underlying constructs  (Henseler et al., 2009). Moreover, the average variance extracted (AVE) scores for all constructs (the ones with the three items with low loadings above) exceeded the required 0.5 threshold. This result, as well as the fact that composite reliability (i.e., Cronbach's alpha /Goldstein rho) for all constructs exceeded 0.6 (see Malhotra & Dash, 2011), indicate that >50% of the variation in the items used to measure each construct can be captured by that construct, providing further support for using all of the items as initially assumed in measuring their respective underlying constructs.
The results also help uphold discriminant validity, as the square roots of AVE scores for each construct were greater than the correlation coefficients of that construct with every other construct of the model (Do Valle & Assaker, 2016) (see Table 3), suggesting that each construct had the strongest relationships with its own indicators, further supporting the measurement of each construct and its use as a standalone/distinct one in the model. It is worth noting that the higher place attachment construct was not included in Table 3, as it is surrogated for by its dimensions. This is the common method of presenting discriminant results in SEM in the presence of a higher order construct (see Hair Jr et al., 2016).

Inner part results
Results of the path analysis further demonstrated that a sizeable part of the variance of the endogenous construct (i.e., residents' support for future film tourism) could be explained by the model (Fig. 2). In particular, the R 2 result associated with the 'residents' support for future film tourism' construct was found to be equal to 30%, which is above the minimum threshold of 20% proposed by Chin (2010), thereby supporting the prediction validity of the model. Moreover, the Stone-Geisser Q2 values for all indicators of the endogenous 'residents' support for future film tourism' construct, computed using blindfolding procedures, were all >0, suggesting that the model is also an effective predictor for explaining the endogenous construct indicators (the four items of RSFFT), and providing further support for the nomological validity and prediction power of the proposed model (Hair Jr et al., 2016).
Additionally, the path coefficients among the constructs were examined. As PLS-SEM makes no distributional assumptions, bootstrapping with 1000 iterations of resampling was used to generate the regression coefficients for the different hypothesized relationships in this case (Chin, 2010) (Fig. 2), with the results revealing that the positive socio-cultural impacts of existing tourism have a positive significant effect on both residents' support for future film tourism and perceived importance of residents' involvement in future film tourism planning and development (reg. Coeffs. std. = 0.182 and 0.181, respectively), thus supporting Hypotheses 1 and 3. The negative socio-cultural impacts of existing tourism, on the other hand, has both negative and positive significant effects on the RSFFT (reg. Coeff. std. = − 0.105) and the perceived importance of residents' involvement (reg. Coeff. std. = 0.233), respectively, which support Hypotheses 2 and 4. Moreover, the perceived importance of residents' involvement in future film tourism planning and development has a significant positive effect on the RSFFT (reg. Coeff. std. = 0.224), further supporting Hypothesis 5. Finally, both place attachment and external perceived realism have significant positive effects on the perceived importance of residents' involvement in future film tourism planning and development (reg. Coeffs. std. = 0.152 and 0.182, respectively) and the RSFFT (reg. Coeffs. std. = 0.270 and 0.138, respectively), additionally supporting Hypotheses 6-9.
Lastly but not least, relationship with the tourism sector (i.e., personally working in tourism, having close friends/relatives working in tourism, and neither working themselves nor having close friends/relatives working in tourism) which was used as a control variable in the model was found to have a non-significant effect on the RSFFT (reg. Coeff. Std. = 0.040), suggesting that the model results, and the residents' support for future film tourism (RSFFT) specifically, are not affected by whether respondents were involved in the tourism sector in their town/ city or not. Table 4 also summarizes the direct, indirect and total effects of the model constructs on the RSFFT with these and the above results further explained in detail in the discussion section below.

Discussions and conclusions
This study aimed to develop and test a conceptual model of the underlying dynamics and complexities of residents' attitudinal behavior to support future film tourism (RSFFT) in a popular tourism destination, namely the province of Seville, Spain. To this end, a set of antecedents of this support was identified theoretically and examined empirically. These comprised: (1) the perceived socio-cultural impacts of existing tourism; (2) the perceived importance of residents' involvement in future film tourism planning and development; (3) residents' place attachment; and (4) the residents' external perceived realism of films and series filmed in their place of residence. All of the path relationships proposed in the model were supported statistically significantly. The key findings are discussed below.
First, the perceived positive socio-cultural impacts of existing tourism had a positive effect on the residents' support for future film tourism (H1), while negative effects had a negative influence (H2). This is consistent with previous studies in general tourism contexts. These (re)confirm that the residents of a tourism destination are more reluctant to support, and are less enthusiastic about, future tourism plans when the drawbacks outweigh the benefits, or when they themselves have had more negative experiences (Nunkoo & Ramkissoon, 2011;Rasoolimanesh et al., 2017;Sharpley, 2014;Stylidis, 2018). More importantly, the positive socio-cultural effects of existing tourism perceived by the local residents involved in the current study had the third-most influence on the extent to which it is supported in future. This support was also affected, both directly and indirectly, by the perceived importance of residents' involvement in future film tourism planning and development. As such, this study confirms the significant role played by previous positive experiences of the impacts of tourism in general when it comes to predicting both the perceived importance of residents' involvement in future film tourism planning and development and residents' support for future tourism such as film tourism in this study. This finding supports the core of Social Exchange Theory (SET) in the context of film tourism.
In comparison, although the residents' determinations of the negative socio-cultural effects of existing tourism had the least influence on their attitudinal behavior to support future film tourism, they had the most impact on the perceived importance of residents' involvement in decisions on the issue. This suggests that the latter is a more significant antecedent of the RSFFT when negative socio-cultural effects of existing tourism are experienced more by local residents.
Second, the residents' perceptions of both the positive and negative socio-cultural impacts of existing tourism in the province had a significant positive effect on the perceived importance of their involvement in the planning and development of future film tourism (H3 and H4). This indicates that, regardless of the socio-cultural costs or benefits of existing tourism, the residents in Seville agreed that their active participation in future film tourism planning and development is both necessary and significant. This finding corresponds with the conclusions reached by Jaafar et al. (2017) in their study on Langkawi Island, Malaysia, which describes the interest of residents in participating in tourism development activities, irrespective of their perceptions of the impacts these will have on their individual lives and the community more generally.
Furthermore, perceived negative socio-cultural effects of existing tourism have a greater influence on the perceived importance of residents' involvement in future film tourism planning and development than positive ones. As it is the first study to examine that relationship in the context of film tourism, there is no comparative previous studies on this finding. However, this corroborates the findings of Jaafar et al. (2017), whose study suggests that residents with a greater perception of the negative effects of tourism are more willing to participate in tourism development activities in order to increase the positive socio-cultural impacts and minimize the negative ones. It is, thus, not unreasonable to argue that residents are interested in using their voices in tourism planning and development process because of the impacts that tourism  has on their immediate lives as both individuals and a community. Third, and related to the above, the study's hypothesized relationship between the perceived importance of residents' involvement and the RSFFT (H5) is corroborated. This aligns with both the conclusions of Gannon et al. (2021) in a tourism study conducted in two historical cities in Iran, as well as the findings of Rasoolimanesh et al. (2017) derived from a study conducted in world-heritage destinations in Malaysia. Although the present study is thought to be the first to explore such a causal relationship in the context of film tourism, it is noteworthy, and thus future studies are invited to re-affirm that the perceived importance of residents' involvement in future film tourism planning and development has the second-most influential effect on the RSFFT than other antededents such as perceived socio-cultural impacts of existing tourism and residents' external perceived realism of media programs.
This empirically and theoretically reinforces the proposition that involving residents in decision-making processes of tourism planning and development is key to the success of film tourism initiatives in future that will both benefit and are accepted by a variety of stakeholders (Beeton, 2016;Heitmann, 2010;Mordue, 2009;Yoon et al., 2015).
Fourth, place attachment has a significant impact on the perceived importance of residents' involvement in future film tourism planning and development (H7). This is consistent with the argument of Chen and Dwyer, 2018, which highlights the willingness of place-attached residents to invest personal resources (e.g., time) to maintain their relationship with that place. The current study found that the residents' place attachment has the most influence, both directly and indirectly, on the RSFFT (H6). These findings give credence to the research by Kim et al. (2017), which asserts that residents with a strong, consistent community attachment -understood as their emotional experiences and bonds with people in their locality -are more likely to support the development of a film tourism destination. This is a particularly important finding, given that the current study is the first attempt to simultaneously test the efficacy of place attachment on the perceived importance of residents' involvement in future film tourism planning and development and residents' support for future film tourism (RSFFT). Consequently, further work is required to refine these relationships and widen the study's applicability. The current study also proposed, and confirmed empirically, the notion that place attachment should be viewed as a multi-dimensional (e.g., place identity, place dependance, friends/family bonding, and place familiarity), not a two-dimensional, construct, even though the simplicity of the latter approach has meant that it has been overwhelmingly adopted in previous research.
Fifth, when examined alongside the place attachment and the sociocultural impacts of existing tourism, it was found that the residents' external perceived realism of movies and TV series filmed in their place of residence also influenced both the RSFFT (H8) and the perceived importance of residents' involvement in future film tourism planning and development (H9). No comparable studies exist since this research is the first to introduce and test the residents' external perceived realism of media programs from a local community perspective. However, such a relationship between residents' external perceived realism and their involvement in future film tourism planning and development could be explained by: (1) some residents' awareness of the potential role played by media representations in raising tourists' expectations of a filming location (Beeton, 2016;Kim & Park, 2021); and (2) residents' interest in contributing to the creation of less stereotypical and more accurate representations of where they live (Póvoa et al., 2019).
Furthermore, the fact that residents are aware of the potential effects of media productions on the public's travel decisions could help researchers to better understand the positive relationship between the residents' external perceived realism and the RSFFT. In this regard, the current study makes a significant contribution to the film tourism literature by incorporating the residents' external perceived realism of media programs, a concept that has been predominantly used in a nontourism context (e.g., Hall & Bracken, 2011). This study confirms that a host community's perspective on the significance of external perceived realism of media productions is of paramount importance when seeking support for future film tourism initiatives. This specific contribution is particularly relevant, as residents have little or no control over how their areas are portrayed in the media, even though these representations can sometimes have undesirable effects on them and their places of residences as well as their wider communities (Beeton, 2016;Thelen et al., 2020;Yoon et al., 2015).
Finally, unlike previous research (e.g., McGehee & Andereck, 2004;Oviedo-Garcia et al., 2008), the relationship that the study's respondents had with the tourism sector and/or the personal benefits currently derived from it made no difference to their willingness to support the future development of film tourism -overall, the residents in the province of Seville were receptive to the development of film tourism, regardless of their occupation and dependence on the tourism sector.

Theoretical and practical implications
In view of the above, the study has important theoretical and practical implications. Indeed, it uses a novel integrative conceptual model to assess the antecedents of the residents' support for future film tourism. It also takes us a step closer to achieving a deeper understanding of the support given by residents for this ever-growing type of global tourism. It also enables examinations of how perceived importance of residents' involvement in future film tourism planning and development can be shaped by place attachment, the perceived sociocultural impacts of existing tourism and the residents' external perceived realism of media programs in the context of film tourism (including fan and media tourism).
Furthermore, the study is the first known attempt to demonstrate: (1) the relationship between the perceived importance of residents' involvement in tourism planning and development and residents' support for future film tourism (RSFFT); and (2) the relationship between residents' external perceived realism of media programs and the RSFFT. Consequently, this study highlights the necessity to go beyond the linearity between perceived benefits and costs of tourism to better understand the complexity of resident's support for future film tourism development. In this vein, the study makes a significant theoretical contribution by confirming the efficacy of importance of residents' involvement and the external perceived realism of media programs in predicting the likelihood that residents will support future film tourism development. It also demonstrates the wider applicability of these findings and the integrated model proposed herein.
Moreover, understanding the extent to which the residents of a destination support and think about their involvement in the planning and development of film tourism is vital to the creation of sustainable policies and regulations on the issue, which has been largely unknown (Kim & Reijnders, 2018).
From a practical point of view, the current study has revealed that, regardless of how the residents of Seville perceive the impact of existing tourism, they place importance to their involvement in the development and planning of future film tourism initiatives.
Importantly, the study's participants who do not work in the tourism sector also expressed an interest in getting involved in the planning and development process of future film tourism, suggesting a genuine desire on the part of the local community to be included in decisions that will have (in)direct effects on their everyday lives, values, and identities. Consequently, it is recommended that municipalities and regional governments seeking to develop and capitalize on film tourism initiatives should design effective platforms or mediums to collaborate with the local residents right from the start of the production-planning stage. This re-confirms the findings of a study in Bali by Kim and Park (2021) relating to the Hollywood movie Eat Pray Love. This approach is particularly important given that the more residents are involved in tourism planning, the more favorable their attitudes are toward tourism development and community change (Easterling, 2005). Furthermore, by acknowledging the interest that some residents have in collaborating with media-production teams (Póvoa et al., 2019), producers could also benefit from establishing an advisory-board service via which the heterogenous resident community would be well represented. The goal would be to reflect critically on how the relevant culture, idiosyncrasies and, above all, places of residence have been portrayed previously in the media (production), and how they could be depicted in the future, given that the residents' external perceived realism as a single item boasts the highest mean score (4.30 out of 5.00) among all constructs and their respective measures adopted in the current study. Indeed, residents' feedback and local knowledge, as evidenced in this study, should never be understood as a constraint on the creative process. Instead, it should be understood as an additional source of important information that media producers could use to create less stereotypical and more accurate representations, which have the potential to be received more favorably and thus supported more positively by local communities.

Limitations and suggestions for further research
Despite the study's practical and theoretical contributions, it nevertheless has some limitations. First, results obtained cannot be generalized since the study is focused on a very specific province and it does not account for any possible differences in results across geographical locations and demographical profiles. Future research could thus replicate the study in other geographical areas and explore how certain demographics (e.g., income) and relationships (e.g., selfidentified film tourists vs. residents who do not engage in this type of tourism) shape residents' support for, and desire to be involved in, film tourism development. This would also enable the study's findings to be re-examined and extrapolated across different geographical areas, respondent profiles and multi-stakeholder perspectives.
Second, future projects conducted in a specific province or region could compare the reactions of residents from rural and urban areas, which may play host to different numbers and types of media productions. This would be valuable since research conducted in a variety of locations could reveal different relationships between the constructs adopted and tested herein.
Third, after newly confirming the impact that residents' external perceived realism has on constructs such as residents' support for future film tourism (RSFFT), subsequent studies should delve deeper into the construct of external perceived realism as used by Green (2004) and Bilandzic and Busselle (2011). This would help identify what specific aspects of external perceived realism are valued by the residents (e.g., dialogues, scenarios, characters). Similarly, future studies should consider aspects related to residents' participation on film tourism development that go beyond the involvement of residents in future film tourism planning and development, for example, by measuring residents' actual experiences in the development of policies for film tourism.
Lastly, the study's data collection began in 2020 during the first summer of the COVID-19 pandemic and took place in region whose GDP significantly depends on the tourism sector (13.1% in 2019, Consejería de Turismo, Regeneración, Justicia y Administración Local). This may therefore have shaped the residents' attitudinal behavior to support future tourism initiatives. Consequently, repeating the study in the same location could help to identify any effects of the extraordinary circumstances on the participants' responses, as well as any changes to their perceptions. Overall, the results of the study can serve as the basis for more comprehensive research that gives voice to resident communities in (potential) film tourism destination.

Funding
This investigation was funded by the European Commission through the H2020 Marie Skłodowska-Curie Actions [Grant number 843473].

Declaration of Competing Interest
None.