1 Introduction

Residents’ perception of tourism is crucial in tourism development, as it influences its hospitality, development, and sustainability. Tourism growth stimulates community building, inspires new businesses, attracts investments, and improves the supply of services by the public sector. Furthermore, it is important to enhance countries’ images in relations with other countries and to lead the fight against racism and intolerance. Considering the above-listed benefits, tourism development is usually synonymous with economic expansion. Many researchers are focusing on studying the tourism impact in terms of residents’ perceptions. The increasing attention to this topic is considered relevant because, despite all the benefits, tourism has potentially negative effects on host communities. Therefore, the need arises to evaluate it in order to manage the tourism growth in local sites (Ko and Stewart, 2002; Lankford and Howard, 1994). Native communities often are not engaged in tourism decision-making processes and consider their areas exploited in a non-sustainable way. In order to give them a voice, the question arises as to how residents' perceptions can be assessed and how the resulting measurement can be used in decision-making processes.

Sustainability in tourism has been a topic of discussion for several years. The UNWTO has defined sustainable tourism as ‘tourism that takes full account of its current and future economic, social and environmental impacts, addressing the needs of visitors, the industry, the environment and host communities’ and, additionally, they emphasise that an equilibrium between environmental, economic, and social-cultural dimensions is essential to ensure lasting sustainability. An essential element in achieving this balance is the participation of local communities in policy-making processes (Garcìa et al. 2015). The paper is focused on the city of Matera area and aims to understand the residents’ perceptions of tourism concerning the three above dimensions, economic, environmental, and sociocultural and how these perceptions impact the destination development process. Tourism development in Matera is expected to touch the everyday life of those living in the city and influence their perceptions of the impacts of tourism in their communities. The growing number of tourists increases the use of natural resources, generates socio-economic impacts and causes stress on facilities. Tourists increasingly use services created to meet the needs of citizens. In addition, online accommodation platforms are becoming more popular and this could increase the number of tourist accommodations and decrease homes available for long-term rental, creating resentment among inhabitants. Tourism can be considered sustainable if it possible to find a balance between residents’ communities and tourist needs. This can be reached engaging citizens, managing the tourist flows, decreasing seasonality and creating services diversification. It is important to understand residents’ approach towards tourism to guarantee the development of successful sustainable tourism strategies.

2 Theoretical framework and hypotheses

In this section, we define tourism impact from a theoretical perspective and define a conceptual model outlining the determining factors and the relationship between them and tourism impact.

Many studies on the impact of tourism have focused on the economic, socio-cultural and environmental aspects, defining positive and negative aspects, so that residents perceive that tourism growth brings benefits and costs to their community (Díaz and Gutiérrez 2010; Gursoy and Rutherford 2004). Residents' perceptions capture the interest of many researchers, primarily in the USA (Kim et al. 2013; Andereck et al. 2007) and Japan (Miyakuni 2012; Zamani-Farahani and Musa 2008). Most researchers focus on the impacts of tourism affecting the three above dimensions, economic, socio-cultural and environmental (Almeida-García et al. 2015; Choi and Kim 2013; Long and Kayat 2011). Social and cultural aspects are jointly measured using a single socio-cultural variable (Fredline et al. 2003; King et al. 1993). The study of residents' perceptions of tourism is gaining importance (Rasoolimanesh et al. 2017). In popular tourist destinations, rising arrivals influence residents' daily lives and change their life quality (Kim et al. 2013). Considering the mentioned effects, the hypothesis is that residents' perceptions hinder tourism development (Rasoolimanesh et al. 2017; Andereck et al. 2007; Látková and Vogt 2012). Residents are the direct players in sustainable tourism development (Lundberg 2011) because they interact with tourists, and their hospitality is influenced by the perceptions they originate. Comprehending residents' perceptions of tourism impact provides crucial information for people who play a valuable role in tourism strategies and services implementation (Brida et al. 2010).

As theoretical framework, the model implemented by Doxey (1975) seeks to deploy or forecast the residents' attitudes as a function of the tourist number increasing in a destination. Doxey associates the development of tourism with the locals’ feelings that evolve. At first, they feel euphoria and empathy, but later they can turn into annoyance and antagonism. Setting away the criticisms and extensions of these models, some researchers assume that tourism growth, as measured by the increase in the number of tourists, is associated with tourism-related reactions.

Therefore, we can have two types of effects linked to the impact of tourism. Considering the benefits, tourism can increase job opportunities for residents (Andereck and Nyaupane 2011; Diedrich and García 2009), raise the life quality of locals, and preserve monuments, artistic heritage and local culture (Andereck et al. 2005). Adverse impacts, on the other hand, include environmental, economic and socio-cultural aspects, such as overuse of public and leisure facilities, traffic congestion and crowded parking areas (Sheldon and Abenoja 2001); significant ecological damage and significant rise in wastage and pollution (Andereck et al. 2005; McGehee and Andereck 2004). These tourism-related disadvantages may lead residents to create negative perceptions of tourism. Interest in cons has increased since the 1980s (Jurowski and Gursoy 2004; Lankford 1994; Liu and Var 1986; Long et al. 1990; Nunkoo and Gursoy 2012) because they can be an obstacle for touristic destinations in their development and sustainability (Ap 1992; Butler 1980; Diedrich and García 2009). Residents' antagonistic behaviour against tourists could restrict the growth of tourism; instead, a friendly approach could be beneficial for tourism expansion. In most cases, tourists hesitate to travel to places they do not feel welcome (Yoon et al. 1999). Consequently, residents' approval is crucial for developing tourism (Butler 1980; Dyer et al. 2007).

Academics consider important to include residents monitoring at an early stage in the development of tourism strategies (Liu et al. 1987). Analysis of locals’ attitudes is essential to evaluate their perceptions, and this should be integrated with tourism planning. Understanding residents' perceptions can lead to decisions that minimise the disadvantages and maximise the positive impacts (Garcìa et al. 2015). Residents' contribution to decision-making can facilitate the creation of positive attitudes regarding the tourism (Robson and Robson 1996). Local policy initiatives on sustainability and tourism development can be more effective if local communities could voice their needs and if they could gain more economic, social and environmental returns (D'Amore 1983; Marien and Pizan 2005).

Researchers have conducted studies on the impacts of tourism, considering the pros and cons of effects on economic, socio-cultural and environmental (Almeida García et al. 2015; Stylidis et al. 2014). Benefits for economic and socio-cultural effects of tourism growth are new job opportunities, improved incomes, living standards and public services, and accessibility to commodities and facilities (Frent 2016). Tourism also induces disadvantages for residents. It is possible to find an increase in the cost of living (Almeida García et al. 2015), pollution of air and water, deterioration of nature, overcrowding, excessive noise, criminality, and problems with cultural identities and value systems (Jaafar et al. 2017); all these adverse effects often cause residents to oppose tourism development in the long term. The Triple Bottom Line (TBL) represents the core of these notions, and John Elkington (1994) added two bottom lines, the socio-cultural and the environmental. According to the above, it is necessary to balance the three dimensions to reach sustainability (Žak 2015; Udeh and Akporien 2016).

A bottom-up approach (Aguiñaga et al. 2017) involving local people will generate a self-sustainable community considering economic, socio-cultural and environmental factors in alignment with their needs. Consequently, considering residents' perceptions of tourism development becomes important when embracing the Triple Bottom Line approach, prompting them to co-create their idea of sustainable tourism. No study has yet investigated the tourism impact in Matera; understanding the attitudes of Matera residents concerning the tourism effects would allow an understanding of the state of tourism in this destination. These insights could help design strategies to improve the awareness and acceptance of tourists among the residents, improving hospitality, development and sustainability of tourism in Matera. Our pioneering study examines the effects of residents' socio-demographic attributes on tourism impacts, proposing residents profiles based on their perception of the effects of tourism on the local population and socio-cultural life. The study aims to investigate how residents' perceptions on tourism are influenced by their attributes to propose profiles of residents in a tourist destination and propose improvements in the management and planning of this destination. The most relevant socio-demographic variables identified by the research are gender, age, income level, years of residence, proximity to the tourist area, proximity to the city centre, and membership of associations.

From the arguments discussed previously, the following research hypotheses are formulated:

H1:

Residents manifest a significant perception of the economic impact derived from tourist activity on the Matera development Process.

H1a:

The perceived effect of the economic impact on Matera development process is significantly different by residents’ sociodemographic profiles.

H2:

Residents manifest a significant perception of the environmental impact derived from tourist activity on the Matera development Process.

H2a:

The perceived effect of the environmental impact on Matera development process is significantly different by residents’ sociodemographic profiles.

H3:

Residents manifest a significant perception of the sociocultural impact derived from tourist activity on the Matera development Process.

H3a:

The perceived effect of sociocultural impact on Matera development process is significantly different by residents’ sociodemographic profiles.

3 The analysis of an italian site: the city of matera

Matera is an Italian town of 60.459 inhabitants,Footnote 1 capital of the homonymous province and second largest city in Basilicata by population,Footnote 2 as well as the largest municipality in the area of BasilicataFootnote 3 and the sixteenth in Italy.

The great thrust that accelerated the process of rehabilitation and redevelopment of the old part of Matera was given by UNESCO, which in 1993 declared the Sassi as a World Heritage Site. Meanwhile the city grows, no longer looks to its past with contempt but with respect and awareness of the potential that the territory can and must offer. The Sassi become repopulated and become fertile ground for museums, "cave houses", restaurants, b & bs, luxury hotels, spas and lots of artisan shops that find inspiration here. In the same caves where only fifty years ago people died of malaria and starvation, today tourists from all over the world come to admire the rock churches, the numerous exhibitions, to stay, to taste a typical dish or to buy a souvenir of tuff or terracotta.

In 2008 the city of the Sassi, thanks to the Matera 2019 Committee, embarked on the candidacy path for the European Capital of Culture in 2019. The candidacy involved the drafting of a dossier to be examined by the judging commission. Matera, representing the entire Basilicata Region, first entered the short list, the 6 finalists, along with Cagliari, Lecce, Perugia-Assisi, Siena and Ravenna, and then on 17 October 2014 with 7 out of 13 preferences is designated European Capital of Culture for 2019.

The city of Matera since the beginning of the twentieth century has always seen a slight but steady growth of the population, except for a significant increase in the second post-war period, due in part to an always positive natural balance and partly to the depopulation of the small neighboring towns.

3.1 Tourism impact in Matera

From the data collected by the ATP Basilicata in the following tables in the last 20 years there has been a constant increase both in terms of the accommodation capacity and the arrivals and presence of Italians and foreigners. Significant increases have occurred since 2008, the year in which Matera embarked on the candidacy path for the European Capital of Culture in 2019 and from 2014, the year in which Matera was proclaimed European Capital of Culture 2019. In particular, it can be noted as, with regard to the movement of customers, the dominant sector is that of foreigners.

In Basilicata, in 2018, it is the city of Matera, which records the largest increase in tourist flows, with 344,813 arrivals (+ 22.5%) and 547,532 presences (+ 22.3%), followed by the Costa Jonica with 255,537 arrivals (+ 18.2%) and 1,281,873 presences (+ 0.9%).

The tourism impact on Matera can be measured both on the supply side and on the demand side.

On the demand side, thanks to the Tourist Intensity Index (TIR = ratio between overnight stay and resident population) we can measure the importance that tourism has for Matera and the pressure it exerts on the territory. In fact, if we consider the data for 2014 and 2018, we can see that in 2014 the TIR = 4.05 and in 2018 it is TIR = 9.06, highlighting how over the years tourism has become increasingly important. If we analyze the data in detail, considering the overnight stays of Italian and foreign tourists, we can see that while for Italians the ratio remains constant (TIR 2014 = 2.64; TIR 2018 = 2.54), for foreigners the situation changes considerably (TIR 2014 = 1.41; TIR 2018 = 6.53), highlighting once again the fact that there is the dominant sector.

On the supply side, the tourism pressure on Matera is highlighted thanks to the Tourist Function Index (FI = ratio between the number of available beds and the resident population). If you always consider the data for 2014 (FI = 0.05) and 2018 (FI = 0.1), you can see how tourism is constantly growing, pushing accommodation activities to grow and expand.

3.2 Data collection and method

The empirical phase of the research was carried out by means of a survey on residents of the Matera destination. Data collection was conducted using a questionnaire submitted between May and July 2019, in the city of Matera. The questionnaire was designed considering the literature review regarding resident’s attitudes and perceptions of tourism impacts (Gursoy and Rutherford 2004; Johnson et al. 1994; Ko and Stewart 2002; Liu and Var 1986; Williams and Lawson 2002). Particularly, starting from the Ap and Crompton (1998) scale consists of a belief component asking respondents to rate the level of change associated with 35 items and an evaluative component asking residents to rate their level of like or dislike for each item. We have verified social-cultural, economic, and environmental domains. The questionnaire consisted of two sections, with the first capturing socio-demographics features, the second section captured perceptions of tourism’s impact. Because of the variety of attributes available in the literature, selection was based on common items applicable to all destinations, such as improvements to infrastructures and services. Similarly, attributes inappropriate to the context of Matera were excluded, such as nightlife. Items were measured using a five-point Likert-type scale anchored by strongly disagree and strongly agree, with 3 as a neutral point, to capture residents’ perceptions of the four domains of tourism impacts.

The latent variables included several items, taken from existing studies on the perception of impacts on the three mentioned dimensions (Brunt and Courtney 1999; Gursoy et al. 2010; Kim et al. 2013; Látková and Vogt 2012).. Moreover, to avoid Common Method Biass (CBM) problems, the questionnaire has followed the procedural recommendations by Podsakoff et al. (2012) to be designed.

Our sample, selected from a convenience sampling, consists of 250 residents of the Matera destination. The analysed sample well replicates the demographic characteristic (gender and age) of Matera population.

Size requirements were determined by considering the sample size recommendation with a statistical power of 80% (Green, 1991). Our study sample of 250 residents of the Matera, fell between a small-effect size (sample size requirement = 481) and a medium-effect size (sample size requirement = 66). According to Cohen (Cohen 1988), this solution appeared satisfactory for a study in social sciences.

Table 1 presents the respondents profile from Matera city. The results show that 44% were male and 56% female. The respondents were divided into three age groups, 16–20 years (8%), 21–40 (43%) and 41 years and above (49%). 43% of respondent earn less than euros 15,000 euros per year, 39% earn between 15,001–28000 euros and 18% earn more than 28,000 euros. 78% of residents live close to the City Center and 57% of residents' homes are close to the tourist activity. Lastly, 31% of the Matera residents are member of a sociocultural or environmental association.

Table 1 Profile of Residents
Full size table

To investigate the phenomena, we propose a structural equation modelling approach in which the variable “Tourism impact” depends on some factors such as “Economic Impact”, “Socio-cultural Impact”, “Environmental Impact” that influence the overall experience of residents. A good knowledge about those relationships and a better understanding about how those relationships are influenced by different subgroups regarding the sociodemographic profile of respondents are important for public administrators in their management choices. In this vein, The Partial Least Square technique, a composition-based structural equation modelling approach, was used to test the hypotheses formulated (Rigdon 2016). Composite measurement models are constructs defined as weighted linear combinations of their indicators (Nitzl and Chin 2017). This design type is suitable for behavioural constructs in which the markers do not reflect or cause the construct, but they compose it (Henseler 2017). Hence, the PLS-SEM model was chosen for the characteristics of the latent variables (LVs). In the current study, constructs are referred to as composites, so to provide consistent and unbiased estimates a composite-based method such as PLS should be used (Sarsdedt et al. 2016; Rigdon et al. 2017). Composites can be defined as either Composite in mode A or Composites in mode B depending on weather their MVs are highly correlated or not. Thus, Composites in mode A uses correlation weights and Composites in mode B use regression weights (Faul et al. 2009)  to compose the construct respectively (Becker et al. 2013). Due to the nature of all LVs, they have been defined as Composite Mode A. Furthermore, the collinearity statistics (VIF) for most of MVs’ VIF being bigger that 3 advices for not defined LVs as Composites. Secondly, the statistical properties of PLS-PM produce robust evaluations for data with normal and extreme non-normal distributions (skewness and/or kurtosis). The skewness scores of the latent variable indicators are not more than -1.31, revealing that the level of skewness is not severe and the PLS-SEM is fit for model estimation (Hair et al. 2018). Thirdly, PLS-SEM was also employed because a non-parametric SEM approach is appropriate to perform a Multigroup analysis (Hair et al. 2018; 2018; Henseler et al. 2016a) as to test differences in the relationships implicated in the research model according to different groups of Materas’ residents. Lastly, PLS-SEM allows the challenging of constructing a global evaluation of a destination development through composites indicators as in our study. One of the main advantages of using the PLS-SEM is that it gives two weights: one to measure the impact of each indicator on its composite indicator, and the other to measure the influence of the composite indicator on the complex indicator (Trinchera and Russolillo 2010). Thus, providing a complex indicator for destination tourist development, allows destination marketing organizations (DMOs) to assess whether the overall destination development process is improving based on residents’ perception. Smart-PLS 3.2.7 software (Ringle et al. 2015) was used to assess the research model and to perform Multigroup analysis (MGA).

The PLS-PM is defined by two series of linear equations known as the inner (or structural) model and the outer (or measuring) model. The structural model details the relationships between VLs, while the measurement model identifies the relationships between a VL and its manifest variables (MVs). A PLS path modelling (PLS-PM) is analysed and evaluated in two steps: (1) assessment of the measuring model; (2) assessment of the structural model. This study evaluated the measurement model by assessing the reliability and validity of the composites indicators (Hair et al. 2017). The structural model was estimated by examining the importance of the path coefficients and by assessing the R2 as an indication of the explanatory power and the Stone-Geisser test (Q2) to shows the predictive relevance (Henseler et al. 2009) of the research model. Furthermore, the value of the standardized room mean square residual (SRMR) was obtained as a goodness of fit indicator to detect for model misspecification (Henseler et al. 2015, 2016a). The nonparametric method Henseler’s MGA (Henseler et al. 2009) was used to performance the multigroup analysis. Furthermore, prior to the MGA analysis, measurement invariance was also evaluated using MICOM (Hair et al. 2018; Henseler et al. 2016a, b) analysis implemented in SmartPLS software.

3.3 Results

Table 2 shows loadings above 0.7 for all the Composites Mode A indicators. Composite Reliabilities (CRs) are over 0.9, and Average Variances Extracted (AVEs) are over 0.4. Therefore, Composite Mode A meets the internal consistency and convergent validity requirements. Table 3 shows that Composite Mode A satisfy discriminant validity criteria according to HTMT90. None of the confidence intervals contains value one, as pointed out from HTMT inference tests, and this highlights the difference between each construct (Henseler et al. 2015).

Table 2 Measurement model for Mode A composites: loadings, construct reliability and convergent validity
Full size table
Table 3 Discriminant validity
Full size table

Table 4 shows the path coefficients and their significance. The hypothesis test has been conducted through the bootstrapping procedure with 5,000 subsamples to generate the t-statistics and the p-value. (Ali et al. 2016).

Table 4 Structural Model Estimates
Full size table

The results from Table 4 support the three effects on Tourism Impact resident’s perception (βECO = 0.166, p-value = 0.025), thus supporting hypotheses H1. Environmental Impact’s resident perception influence significantly Tourism impact’s resident perception (βENV = 0.220, p-value = 0.003) which supports hypothesis H2. Socio-cultural impact has a significant influence on Tourism Impact’s resident perception (βSOCU = 0.474, p-value = 0.000) thus supporting hypothesis H3 (Fig. 1).

Fig. 1
figure 1

The Triple Bottom Line

Full size image

From Table 4 we observe that the model presents both a good explanatory power (R2 = 0.625) and a high predictive power (Q2 = 0.456). Lastly, a SRMR value of 0.97 is considered to be acceptable for a proper fit since that value is within the thresholds that are considered satisfactory (SMR = 0.10 and a more conservative threshold, SMR = 0.08).

The analysis of the invariance of the measures across different groups is necessary when using PLS-SEM for group comparison. To do this, the procedure called “measurement invariance of composite models “(MICOM) is carried out before performing a multigroup analyses in PLS-SEM (Hair et al. 2018) and it consists of three steps hierarchically interrelated (Henseler et al. 2016): (1) configural invariance, (2) compositional invariance, and (3) the equality of composite mean values and variances. Step 1 is not included in SmartPLS since it does not involve statistical test. The criteria are fulfilled under the following assumptions: identical indictors of composite across the groups; identical data treatments; identical algorithm settings or optimization criteria across groups. In step 2, SmartPLS return permutation-based confidence intervals. Through them, it is possible to determine if the correlation between the two group composite scores is significantly lower than one (null hypothesis H0:c = 1). If H0 is not rejected, the composite score does not diverge much in the two groups, and there is compositional invariance. In step 3, permutation-based confidence intervals for the mean values and the variances make it possible to assess if composites' mean values and variances differ across groups.

For the aim of this study, PLS multi-group analyses based on the socio-demographic profiles of Matera residents based on gender, age, income, city proximity, tourist proximity and membership of association have been carried out. The MICOM procedure is applied to the research model and its multi-group analysis for different residents’ profile. In step 1, we have followed the criteria established to ensure the composites’ configural invariances. Table 5 shows the results of applying MICOM procedure, step 2 and step 3, when using gender to conduct a multi-group analysis. For practical reasons, we have omitted the results of MICOM procedure for the rest of socio-demographic profiles. However, results of MICOM for these groups also reveal that the full measurement invariance of the four composites across groups are supported. Next, Henseler’s MGA has been used to compare and test differences in the path coefficients in the structural model between groups.

Table 5 MICOM results: Gender
Full size table

Table 6 and 7 show the structural models across resident’s groups and results of the MGA analysis. For Henseler’s MGA method (Henseler et al. 2009), a p-value lower than 0.05 or higher than 0.95 indicates significant differences between specific path coefficients across two groups at a 5% level of significance.

Table 6 Multigroup analysis. Results of Hypothesis Testing
Full size table
Table 7 Multigroup analysis. Results of Hypothesis Testing by age
Full size table

4 Discussion and conclusions

The results evidence significant differences in the perception of the economic impact from tourist activities on the destination development process between men and women, being that men’s perception bigger than women’s perception. Moreover, significant difference in the perception of the sociocultural impact on the destination development process is found, being the women’s perception higher than men’s perception. No significant differences in environmental impact perception on the destination overall tourism impact by gender is observed. Hence, for gender Hypotheses H1a and H2a are supported and H3a is no supported. No differences in the perception of the economic, environmental and sociocultural impact across residents on the destination overall tourism impact by age group are observed. Thus, Hypotheses H1a, H2a, H3a are not supported when considering residents’ age. However, it is interesting to highlight that the residents with age 21–40 years old perceived significantly the economic impact derived from tourist activities whereas the environmental and sociocultural impact are perceived significantly by the three group of age (less than 21, 21–40, more than41). Focusing on the level of income, this variable has been categorized into three groups: residents with less than 15,000 euros (group 1), residents with income 15,001–28000 euros (group 2) and residents with more than 28,001 euros (group 3). For this variable, significant differences in the perception of the economic impact on the destination development can be found across residents from group 1 and group 3 of income and across residents from group 1 and group 2 of income, being the residents with higher level of income (group 3) those who manifest a significant perception of the economic impact. Likewise, differences and across residents from group 1 and group 3 and from group 1 and 2 of income are observed regarding the perception of sociocultural impact derived from tourist activity on the destination overall impact. Otherwise, the environmental impact is perceived significantly different across residents from group 1 and 3, and across residents from group 2 and 3. In fact, residents with lower level of income (group 1) do not perceive significantly the impact of environmental effect derived from tourism on the destination development. Hypotheses H1a, H2a and H3a are partially confirmed. Being a member of a cultural or/and environmental association lead to significant and positive differences in the perception of the environmental and sociocultural impact on the destination overall tourist impact. However, the belonging or not to an association does not lead to significant differences in the perception of the economic impact. For being a member of an association as resident profile characteristic, Hypotheses H2a and H3a are then supported and H1a is not supported. The proximity of Matera’s residents to the City Centre or tourist activity does not cause significant differences in the perception of the economic and environmental impacts derived from tourism. However, significant differences on the sociocultural impact are observed by City Canter proximity, being those closer to the city Centre with higher perception of the sociocultural impact. Hence, focused on City Centre proximity, the Hypotheses H1a and H2a are not supported and H3a is supported. The proximity of Matera’s residents to the tourist activity does not cause significant differences in the perception of the economic, sociocultural and environmental impacts derived from tourism. Hence, for this characteristic, hypotheses H1a, H2a and H3a are not supported. However, we can observe that those residents who live closer to tourist activity perceive significantly and slightly higher economic, sociocultural and environmental impact than those who live far away from the tourist activity.