This Is My Hometown! The Role of Place Attachment, Congruity, and Self-Expressiveness on Residents’ Intention to Share a Place Brand Message Online

Abstract

In a gradually more hypercompetitive global tourism arena, destination marketers are increasingly seeking effective ways to promote places through appealing place brand messages. Local residents can be valuable ambassadors for the place, as well as co-creators of place-related brand communication. However, research focusing on place advertising from the residents’ perspective is scant. To address this gap, this work identifies three main antecedents of residents’ intention to share a place brand message online: place attachment, place ad–brand congruity, and self-expressiveness. A model is developed and tested on a sample of current and former residents of a Swedish city. Structural equation model results support that these antecedents have a positive influence on current as well as former residents’ intention to share a place brand ad online. Place attachment also has a significant indirect influence on intention to share, with place ad–brand congruity and self-expressiveness as partial mediators. Furthermore, place ad–brand congruity partially mediates the relationship between place attachment and self-expressiveness.

Competition between places and destinations to attract tourists and other stakeholders is increasing (Braun, Eshuis, and Klijn 2014) at a fast pace in a gradually more hypercompetitive arena (Mariani et al. 2014). This has led to widespread adoption of branding and marketing strategies among places of all sizes (Cleave et al. 2017). For instance, cities like New York and Amsterdam have embarked on several successful place marketing initiatives to improve their appeal and attractiveness for tourists as well as for residents. However, fads and trends that make or break new place brands can be very dynamic, making reactive marketing strategies quite ineffective. Accordingly, destination marketers and destination marketing organizations (DMOs) have developed an awareness that marketing, branding, and advertising activities are not only critical for the survival, resilience, and success of a place but also that branding campaigns need to be continuously updated to help destinations keep a competitive edge.

Along with the increased focus on place marketing and branding, there has been a rapid growth of scientific research related to these topics (Vuignier 2017). The aim of place branding is to create, influence, and reinforce an image in the minds of external and internal target audiences through association of the place, including stories about it (Warnaby and Medway 2013; Zenker and Braun 2010; 2017); thus, advertising plays a key role. Place branding can enable different stakeholder groups to interact with one single brand, as well as assist communications by working as a tool for ensuring that communications are consistent and integrated (Merrilees, Miller, and Herington 2012).

Residents constitute a stakeholder group that is considered to be essential to the competitiveness of tourism destinations (Uchinaka, Yoganathan, and Osburg 2019). They possess thorough direct knowledge of the place (Zenker, Braun, and Petersen 2017) and can act as trustworthy and credible sources of information regarding the place. Furthermore, as they are part of the place as such, they will affect visitors’ experiences (Chen, Dwyer, and Firth 2014; Freire 2009; Zenker, Braun, and Petersen 2017). Hence, residents can become important ambassadors for the place, as well as co-creators of place-related brand communication (Braun, Kavaratzis, and Zenker 2013; Strandberg and Styvén 2019). Still, there is a lack of research focusing on place branding from the perspective of residents (Stylidis 2020; Zenker, Braun, and Petersen 2017). In particular, a better understanding of residents’ role as place brand ambassadors is called for, especially in digital and social media contexts (Stylidis 2020; Uchinaka, Yoganathan, and Osburg 2019) where research has focused mostly on visitors so far (Mariani, Di Felice, and Mura 2016).

As residents develop bonds with their place, the concept of place attachment—in other words, an affective bond to a particular geographic area—works as an antecedent to behavioral outcomes such as electronic word of mouth (eWOM) (Chen and Šegota 2016). Scholars have noted that moving away from a place to which one is deeply attached does not mean that this attachment ceases to exist (Andéhn, Hietanen, and Lucarelli 2020; Casey 1993). Therefore, in addition to those currently residing in a place, its former residents could also be considered a relevant target group who could potentially be engaged in spreading eWOM. Following the fundamental tenets of advertising, brand messages should correspond to the knowledge and experience of the target group—in this case, current and former residents—in order to be perceived as relevant, comprehensible, and appealing (Halkias and Kokkinaki 2014; Keller 1993). That is, it should be congruent with the person’s perception of the place and with his or her self-image (Sirgy 1982). Thus, in line with identity theory, sharing place brand messages publicly could allow individuals to use the symbolic value of the brand to express their self-concepts (Taylor, Strutton, and Thompson 2012). This suggests that the level of self-expressiveness of a place brand message is likely to influence residents’ intention to share the message online (Strandberg and Styvén 2019; Taylor, Strutton, and Thompson 2012).

Extant studies have not examined in depth the role played by residents’ perceptions on their intention to share a place ad in an online context, despite the increasing attention that marketing and tourism scholars have paid to residents versus visitors in supporting tourism and pro-tourism behaviors (e.g., Shen, Geng, and Su 2019; Stylidis, Shani, and Belhassen 2017). This is surprising, as consumers today search, benchmark, and choose places and destinations using online channels, and current as well as former residents can play a critical role as online brand ambassadors. To bridge this research gap, the current study aims at assessing the influence of place attachment, perceived place ad–brand congruity, and self-expressiveness on current and former residents’ intention to share a place brand message online.

This study is important and timely as it (1) contributes to advance the nascent research stream on the role of residents as ambassadors in place marketing; (2) combines attachment theory (Altman and Low 1992), congruity theory (Osgood and Tannenbaum 1955), and identity theory (Sirgy 1982) in developing and testing a conceptual model explaining residents’ intention to share a promotional ad online; and (3) generates multiple practical implications for destination and place marketers to shape cost-effective online marketing and advertising campaigns in a period of drastic promotional budget cuts for DMOs (Hays, Page, and Buhalis 2013).

Theoretical Background

Attachment theory (Altman and Low 1992), congruity theory (Osgood and Tannenbaum 1955), and identity theory (Sirgy 1982) constitute the theoretical underpinnings of the research model in this study, depicted in Figure 1. More specifically, attachment theory was used to devise the place attachment construct (Lewicka 2008); congruity theory was the basis for the place ad–brand congruity construct (Halkias and Kokkinaki 2014; Meyers-Levy and Tybout 1989); and identity theory was utilized to develop the self-expressiveness construct (Sirgy, Lee, and Yu 2018).

Figure 1. Research Model.

The following sections explain the theoretical background to the proposed model and develop the hypotheses to be tested.

Place Attachment and Place Branding

Bonding, in the form of meaningful connections with others, is central to human experience (Scannell and Gifford 2014). Attachment theory “focuses on person-to-person bonding and proposes that an innate psychological system regulates proximity to an ‘attachment figure’” (Scannell and Gifford 2014, p. 23). However, most people also develop bonds with places (Altman and Low 1992; Lewicka 2011).

Literature in place attachment encompasses different definitions and operationalizations of the concept. While some definitions include both affective, cognitive, and behavioral components (Scannell and Gifford 2010), the most essential and most frequently measured is the affective or emotional component (Lewicka 2010). Place attachment is thus commonly understood as emotional ties that people develop with a particular geographic area (Hidalgo and Hernández 2001; Lewicka 2010; Morgan 2010). As Altman and Low (1992) point out, the word attachment emphasizes affect. Feelings for a place can range from disinterest to fondness and devotion, to powerful attachment that makes people willing to defend the place (Seamon 2014). According to Lewicka (2011), place attachment might build independently of time of residence but has been shown to materialize not only for current and former residents but also for visitors to a place.

“Place” itself can be generally defined as a space that has been given meaning through personal, group, or cultural processes (Altman and Low 1992; Lewicka 2011). Therefore, places as an object of study differ in scale and can be, for example, a home, neighborhood, city, region, country, or even a whole continent (Lewicka 2011). Hidalgo and Hernández (2001) found that the strength of place attachment was greater for the home and city levels than for the neighborhood level. While interest in environmental psychology research most often has been on the neighborhood level (Lewicka 2010), place branding research has tended to focus on cities. Place branding as a field of academic study has developed over the past 20 to 30 years (Gertner 2011). More recently, there appears to be an increasing focus on stakeholders’ role and engagement in place branding (Kavaratzis and Dennis 2018). Systematic and meta-analytic literature reviews highlight the importance of residents in place branding and point to specific knowledge gaps when it comes to understanding the roles of residents in the place branding process; in particular in a digital context (for recent reviews, see Acharya and Rahman 2016; Stylidis 2020; Vuignier 2017).

As local residents constitute a key stakeholder group for places that also are tourist destinations (Stylidis 2018), their perceptions and attitudes need to be solicited and assessed (Nunkoo and Gursoy 2012; Stylidis 2018). Kavaratzis (2017) argues that by taking a participatory approach, the place brand will appear “closer” to the residents, as it is based on the reality of the place. This is important, because residents with higher identification levels toward a place are more likely to experience stronger place attachment (Zenker and Petersen 2014). In turn, this can have positive outcomes, such as a will to remain in or close to the place, expressions of pride and love (George and George 2004; Scannell and Gifford 2010), and positive word of mouth (WOM) (Chen, Dwyer, and Firth 2014). Therefore, to stimulate WOM behaviors, place brand messages (advertising) should likely be congruent with how residents perceive the place and reflect their self-concept. Indeed, this type of organic information coming from residents is perceived as more credible than induced sources such as official tourist brochures (Palmer, Koenig-Lewis, and Medi Jones 2013). The following sections discuss place ad–brand congruity, self-expressiveness, and positive eWOM in the form of intention to share a place brand message online.

Congruity between Place Advertising and Place Brand Image

In psychology studies, the underlying and core assumption of congruity theory (Osgood and Tannenbaum 1955) is that individuals tend to express positive attitudes toward a focal object when they perceive it to be consistent with their beliefs and schemata (i.e., cognitive frameworks that individuals develop to organize, categorize, and interpret information; Bartlett 1932) about the self and the external environment. The self, as described in identity theory, is composed of “the meanings that persons attach to the multiple roles they typically play in highly differentiated contemporary societies” (Stryker and Burke 2000, p. 284).

Advertising scholars have developed the concept of ad–brand congruity that can be defined as the extent to which advertising information conforms to consumers’ expectations about a brand, based on consumers’ previous beliefs, schemata, and knowledge of the brand (Meyers-Levy and Tybout 1989). While some advertising researchers have found that a moderate degree of incongruity between ad and brand brings about an improvement of the attitude toward the ad and the brand (Dahlén et al. 2005; Halkias and Kokkinaki 2014), it has also been shown that ad messages that create even a low dissonance can translate into confusion and negative consumer responses (Loken 2006). Following the basic tenets of congruity theory, this is because individuals attempt to minimize any inconsistency between their own beliefs and their perceptions of the object (in this case, an advertisement) (Ul Islam, Rahman, and Hollebeek 2018).

In the context of this study, we apply congruity theory and specifically ad–brand congruity to place branding. Hence, place ad–brand congruity is defined here as the fit between the image of the place transmitted in advertising (a place brand message) and the residents’ existing image of the place. Today, advertising occurs in and across a multitude of channels—not necessarily paid media—and the receivers are often active participants, seekers, and even co-creators of the communication (Dahlen and Rosengren 2016). This is largely the case for place brands. Therefore, the broad definition of advertising as “brand-initiated communication intent on impacting people” (Dahlen and Rosengren 2016, p. 343) is most relevant for the context of this study. As Lane (2000) points out, advertising can invoke the brand’s image or schema and the associations that differentiate the brand. The image of a place, defined as the sum of beliefs, ideas, and impressions that people have of the place (Kotler and Gertner 2002), is perceived, felt, and processed by individuals.

An upsurge of interest in the conceptualization and operationalization of place brand image in tourism took place in the early 1990s, with Echtner and Ritchie’s (1991) landmark contribution looking at the meaning and measurement of destination image. These authors suggest that destination image can be defined as “not only the perceptions of individual destination attributes but also the holistic impression made by the destination” (Echtner and Ritchie 1991, p. 8). Several scholars (e.g., Baloglu and McCleary 1999; Beerli and Martín 2004a; Beerli and Martín 2004b) have emphasized that the image of a place is a combination of cognitive and affective components. The cognitive dimension of place image involves perceptions of relatively tangible attributes, such as shopping facilities, accommodations, cultural activities, scenery, and so on (Beerli and Martín 2004a; Stylidis, Sit, and Biran 2016), while the affective dimension relates to the person’s feelings toward the place, such as perceiving it as exciting or pleasant (Beerli and Martín 2004a). However, the overall image is greater than the sum of its parts (Bigné Alcañiz, García, and Blas 2005).

Over the past two decades, some travel and tourism researchers have started to study place and destination image from the perspective of residents rather than tourists (e.g., Henkel et al. 2006; Stylidis, Sit, and Biran 2016). Residents play a key role in all of the relevant components of a place brand and are increasingly involved in place brand communication (Braun, Kavaratzis, and Zenker 2013; Strandberg and Styvén 2019). Due to residents’ attachment to and knowledge of the place, they might disagree with a simplified place brand message, which could result in lower brand identification and less favorable behavior (Zenker, Braun, and Petersen 2017). Therefore, the message should correspond to the knowledge and experience of the target group to be perceived as relevant, comprehensible, and appealing (Halkias and Kokkinaki 2014; Keller 1993). We argue that the higher residents’ attachment to the physical place and the meaning ascribed to it (Lee, Kyle, and Scott 2012), the more they will be capable to translate their place identity, dependence, and social bonding (Ramkissoon and Mavondo 2015) into a congruent representation of the place in promotional communications. In other words, the more the residents are attached to their place, the better will be the match between the image of the place transmitted through an ad (place brand message) and the person’s existing image of the place. Thus, we hypothesize:

H1: Place attachment positively influences perceived congruity between a place brand message and the perceived image of the place.

Self-Expressiveness

Beliefs about the self are termed as “self-concept” (Sirgy 1986), which consists of self-image (actual or ideal) and social self-image (actual or ideal) (Sirgy 1982; 1985). Brands whose images are perceived by consumers as congruent (i.e., consistent) with their self-images will be more likely preferred and positively evaluated by consumers (Loken 2006; Sirgy 1982). Building on identity theory (Sirgy 1982), the self-expressiveness of a brand message can be described as the extent to which recipients perceive that it supports and enacts their self-concept and will be recognized publicly as such (Taylor, Strutton, and Thompson 2012). Previous research has shown that products and brands can help consumers self-express either because the products/brands are themselves self-expressive or because they are associated with a desired prototypical consumer group (Karjaluoto, Munnukka, and Kiuru 2016).

Within the travel and tourism environment, the antecedents and consequences of self-expressiveness have been examined almost exclusively in relation to tourists rather than residents (e.g., Bosnjak et al. 2016). However, it could be argued that the self-expressiveness of a place brand message might be a way for residents to differentiate their place from other places. This could be either because they perceive their place brand as different compared to other places or because they feel that the place brand expresses the emotional ties that they have developed with their place of residence (Lewicka 2010). As place brands embody a highly symbolic role (Aaker 1996; Matzler et al. 2016), the self-expressiveness of a place brand message will likely allow residents to better communicate their self-concepts, their pride to be residents of the place, and the uniqueness of the place where they reside. In parallel, residents’ attachment to and knowledge of the place is likely to help them better communicate their self-concepts and ultimately enhance the self-expressiveness of a place brand message. In other words, the more that residents are attached to their place, the stronger they will perceive the self-expressiveness of an ad (place brand message). Hence:

H2: Place attachment has a positive effect on the self-expressiveness of a place brand message.

Self–brand congruity refers to the match between actual self-image and product or brand image (Sirgy 1985). In tourism settings, this is often measured as the match between the tourist’s self-image and the image of visitors at the destination, or destination personality (Sirgy, Lee, and Yu 2018). Based on self-congruity theory, Taylor, Strutton, and Thompson (2012) propose that greater self–brand congruity increases the self-expressiveness of an eWOM message and find support for this relationship in a study of viral videos for consumer brands. Drawing from several research streams concerned with self-identity, Thorbjørnsen, Pedersen, and Nysveen (2007) state that people prefer brands associated with traits that are congruent with their own, and they use these brands more or less consciously to express their own identity and values. Along the same line of reasoning, in the context of our study, we suggest that the more congruent a place brand message is with residents’ own beliefs, ideas, and impressions of the place (i.e., image), the more the message will be perceived as self-expressive. Accordingly, we hypothesize:

H3: Perceived congruity between a place brand message and the perceived image of the place has a positive effect on the self-expressiveness of the place brand message.

Intention to Share

Consumers commonly engage in WOM because they want to help others (Hennig-Thurau et al. 2004) or share social information (Alexandrov, Lilly, and Babakus 2013). While traditional WOM typically occurs face-to-face, eWOM entails one-to-many and many-to-many communication, sometimes reaching millions of people by spreading virally online (Jeong and Jang 2011). Indeed, eWOM can take place through channels such as virtual communities, websites, product review sites, e-mail, and social media (Litvin, Goldsmith, and Pan 2018). In this case, “consumers” can be tourists as well as residents who are engaged in spreading eWOM about places and destinations.

The communication of advertising messages by means of websites and videos is a specific form of eWOM (Taylor, Strutton, and Thompson 2012). In particular, videos today represent a powerful online advertising tool, as streaming video technologies have become extremely popular, making videos and visual content key for social media such as YouTube, Facebook, and Instagram. Videos are deployed as a tool for viral marketing to be conceived as a social process whereby individuals share and disseminate marketing information originally developed and transmitted by marketers to trigger positive WOM (Hinz et al. 2011; Hsieh, Hsieh, and Tang 2012).

In the context of place advertising, messages that are consistent with place residents’ held image of the place will, on one hand, reveal residents’ positive experience of the place brand and, on the other hand, trigger social processes such as sharing social information of the place (Baumeister, Zhang, and Vohs 2004). Thereby, they help others gain knowledge of the place (Bilgihan et al. 2016) and show concerns for others (Hennig-Thurau et al. 2004). Research suggests that residents often act as ambassadors of their place, which can have a substantial influence on tourists’ image formation and decision making (Stylidis 2018). This type of organic information is perceived as more credible than induced sources such as official tourist brochures (Palmer, Koenig-Lewis, and Medi Jones 2013). In particular, by sharing content about the place in various social media platforms, local residents can be influential in shaping the image of tourist destinations (Palmer, Koenig-Lewis, and Medi Jones 2013; Stylidis 2018).

Based on the preceding discussion, we hypothesize that high levels of place ad–brand congruity would result in stronger intention to spread positive eWOM in the form of sharing the place brand message online. The advertisement (i.e., place brand message) is in this case a streaming video. According to Oh, Baek, and Ahn (2017), eWOM through video-sharing social media represents an underresearched area. Eckler and Bolls (2011) point out that viral videos are potentially even more engaging than traditional advertising in terms of stimulating complex and intense motivational and emotional processes. Therefore, we offer hypothesis 4:

H4: Perceived congruity between a place brand message and the perceived image of the place positively influences intention to share the message online.

Self-expressiveness in the context of place advertising might reflect the need of residents to feel good about themselves and seek positive evaluations from others—a concept known in psychology as self-enhancement (Jones 1973). This might translate into a positive impact on intention and frequency (Hennig-Thurau et al. 2004) to share the advertisement through eWOM. Empirical research has found that eWOM on general social media such as Facebook is affected by self-expressiveness (Wallace, Buil, and de Chernatony 2014).

In advertising contexts, scholars have detected a positive relationship between self-expressiveness of an advertising message (such as a viral video) and the probability that it will be shared (Taylor, Strutton, and Thompson 2012). In line with identity theory, sharing the message publicly allows individuals to deploy the brand’s symbolic value to communicate their self-concepts (Taylor, Strutton, and Thompson 2012). In addition, research suggests that self-expressiveness has a particularly strong influence on behavioral intention for products and services that are symbolic and/or are consumed in public settings (Thorbjørnsen, Pedersen, and Nysveen 2007), both of which are accurate for places and destinations. Similarly, place branding research has found that identification with a place brand positively influences promotion and advocacy of the place (Stokburger-Sauer 2011). Residents of a place are likely to engage in positive eWOM communication about the place because their identity is directly related to it. Summing up the previous research, we hypothesize:

H5: The self-expressiveness of a place brand message has a positive effect on the intention to share the message online.

Extant research in tourism and travel has emphasized that both place attachment and identification are conducive to positive WOM for places allowing people to express their self-identity (Simpson and Siguaw 2008). When considering residents, this could be explained in terms of self-enhancement (Jones 1973) as they might want to feel good about themselves and may seek positive evaluations of themselves and their place from others. Moreover, residents might be looking for self-affirmation to maintain the integrity of their self-image (Steele 1988) and by analogy the image of their place. From an empirical point of view, place attachment has been found to positively influence tourists’ intention to recommend a place (Prayag and Ryan 2011) and residents’ positive WOM about a place (Chen and Dwyer 2018; Chen, Dwyer, and Firth 2014; Zenker and Rütter 2014). Thus, the deeper the bonds of residents to their place, the stronger their intention to share a positive advertising message online about the place that makes them feel good about themselves. Accordingly, we hypothesize:

H6: Place attachment positively influences intention to share a place brand message online.

Method

Design and Procedure

The study was conducted as an online survey incorporating a promotional video that was developed by a local Swedish DMO as part of a long-term branding strategy of the place as a destination to visit during different seasons. The place is predominantly visited by tourists interested in its climate and nature, as well as by individuals based in the region. When answering the survey, participants had to watch the video, which is about 1.5 minutes long, to be able to keep answering the questions. The video was embedded into the questionnaire so that respondents watched it within the survey platform. It is otherwise available on YouTube and thus subject to potential eWOM and viral spread (Hsieh, Hsieh, and Tang 2012).

To increase validity and reliability, we relied on extant scales when developing the questionnaire. All items can be found in the Online Appendix. Affective place attachment was measured by a seven-item scale adapted from Lewicka (2008), while a scale of five items from Taylor, Strutton, and Thompson (2012) was used to capture self-expressiveness. For the dependent construct, intention to share, a three-item scale was employed (Hsieh, Hsieh, and Tang 2012). Likert-type 7-point scales were used for these three constructs.

Place ad–brand congruity was measured using a single item adapted from Lane (2000), with responses on a 7-point scale anchored by Does not fit at all (0) and Fits very well (6). As stated in the theoretical background, place ad–brand congruity is in this study defined as the fit between the image of the place transmitted in advertising (a place brand message) and the residents’ existing image of the place. The latter refers to the overall image, which cannot be adequately measured through an average or sum of attribute scores (Bigné Alcañiz, García, and Blas 2005). Therefore, single-item measurements are commonly used to capture the overall image of a place (e.g., Baloglu and McCleary 1999; Beerli and Martín 2004b; Bigné Alcañiz, García, and Blas 2005; Stylidis, Shani, and Belhassen 2017). Consequently, we argue that also the place ad–brand congruity variable could be measured through a single item.

Prior to releasing the survey, we performed a qualitative pretest on a small sample of consumers, place marketing practitioners, and researchers to increase content and face validity of the instrument. This led to minimal amendment of a few items.

Sample and Data Collection

As previously described, this study focuses on current and former residents as place brand ambassadors, and the “place” is in this case a city (more formally, a municipality). Therefore, to reach current residents as well as those who had previously lived in the municipality, the survey was distributed through three different channels: (frame 1) at the DMO’s Facebook page; (frame 2) via e-mail invitations to randomly sampled students at the city’s university; and (frame 3) to an online consumer panel which comprised men and women between 16 and 75 years residing in the county (i.e., the larger geographical region surrounding the studied place).

In total, 289 responses had been gathered at the closure of the survey. Four responses were removed due to missing values greater than 15% (Hair et al. 2010). Then, given the focus of the study, we excluded another 72 respondents who were neither current nor former residents of the municipality. Thus, 213 responses were retained for the analysis, of which 107 came from sampling frame 1 (DMO’s Facebook), 43 from frame 2 (students’ e-mails), and 63 from frame 3 (online panel). Among these 213 respondents, 60 were former residents while 153 currently resided in the municipality. Of the total sample, 60% were women. Furthermore, the median age category was 35 to 44 years. With regard to occupation, 43% of the respondents were working, 36% were students, and 14% were retired.

No significant differences in means between the current and former residents on 15 of the 16 items were detected in an independent-samples t test. Item PA7, however, had a slightly lower mean in the former residents group; perhaps not surprising considering the meaning of the statement (“I want to be involved in what is going on in [the place]”). Demographic characteristics were also similar in both groups, although there were more students among the current residents. This is understandable in light of the data collection methods and the fact that many students move away from the city after completing their studies.

The latter was also reflected in the results when comparing means across the three samples (from frames 1, 2, and 3). One-way analysis of variance (ANOVA) indicated that mean values were lower on 14 of the 16 items in the student sample compared to the others. A comparison of time of residency between the three groups showed that respondents from sampling frame 1 (DMO’s Facebook) had lived in the city 24.5 years on average, which was not significantly different from frame 3 (online panel), for which the mean time of residency was 23.5 years. In the student sample, however, the mean was 9.8 years. Further analysis showed that time of residency had a significant positive (Pearson) correlation with 13 of the 16 items. Independent-samples t test of groups created through a median split based on time of residency (removing nine respondents who were exactly at the median value) yielded the same results. While differences in means can reflect natural sociodemographic differences within the combined sample, it is important that the metric structure of the measurement model is sufficiently invariant between the subsamples. Therefore, we tested for invariance before running the structural model. This is described in the section Assessment of Measurement Invariance later in this article.

Measurement Validation

Very few missing values remained after dropping the four respondents with larger parts of missing data. Little’s MCAR test showed that the remaining missing values could be labeled as missing completely at random (χ² = 223.94, p = .898). Nevertheless, as testing the full research model requires the use of structural equation modeling (SEM), we deployed the expectation–maximization (EM) method to substitute missing data in the quantitative variables. Otherwise, it would not have been possible to obtain some of the necessary output (such as modification indices). According to Hair et al. (2010), the EM approach to missing value replacement is desirable as it introduces the least amount of bias into SEMs. Skewness and kurtosis values were well within recommended ranges (−1 to 1 for skewness; −1.5 to 1.5 for kurtosis) (Hair et al. 2010).

We ran a confirmatory factor analysis (CFA) using AMOS Version 25 to assess the constructs in terms of model fit indices, factor loadings, modification indices, and standardized residuals. For estimation purposes, the single-item construct place ad–brand congruity was assumed to have a reliability of .95, corresponding to a factor loading of .90 and an error term of .10 (Anderson and Gerbing 1988; Petrescu 2013). The loading and the error term were fixed to these respective values when specifying the model (Hair et al. 2010). Estimates obtained from the CFA, including factor loadings, modification indices, and standardized residual covariances, were assessed in conjunction with reliability values of the scales. This resulted in dropping three items from the place attachment construct (PA1, PA4, and PA7) due to high modification indices, standardized residual covariances >2, and/or factor loadings below .70. Factor loadings for all retained items in the multi-item scales are provided in the Online Appendix. The final measurement model had a χ² of 114.183 (58 df), χ²/df = 1.969, comparative fit index (CFI) = .980, standardized root mean square residual (SRMR) = .0317, and root mean square error of approximation (RMSEA) = .068 (nonsignificant), suggesting adequate fit between the measurement model and sample data (Byrne 2001; Iacobucci 2010).

Furthermore, as presented in Table 1, all average variances extracted (AVE) were higher than .50 and also higher than the shared variance for all pairs of constructs, indicating discriminant validity (Fornell and Larcker 1981). Internal consistency values of the scales (composite reliability and Cronbach’s α) as well as standardized factor loadings of all retained items exceeded .70, indicating convergent validity (Hair et al. 2010).

Table 1. Reliability, average variances extracted (AVE), and correlation matrix.

	α	CR	AVE	ABC	PA	SE	IS
Ad–brand congruity (ABC)	N/A	.951	.951	.975
Place attachment (PA)	.922	.892	.734	.769	.857
Self-expressiveness (SE)	.946	.951	.795	.657	.679	.892
Intention to share (IS)	.953	.962	.895	.781	.738	.704	.946

Note. CR = composite reliability; N/A = not applicable; square root of AVE on diagonal (in bold).

SEMs normally do not require extensive samples. Hair et al. (2010), as well as Iacobucci (2010), point out that sample-size decisions must be based on several different factors. They suggest it is possible to run well-performing models even with samples of 100, provided that the model has five or fewer constructs, data are normally distributed, item communalities (i.e., squared standardized factor loadings) exceed .60, measures are reliable, and there are at least three items per construct. In this case, the model meets all of these criteria, with the exception of the single-item construct of place ad–brand congruity. Thus, the sample size of 213 is considered adequate to perform the desired analyses.

Assessment of Measurement Invariance

After obtaining a satisfactory measurement model in the total sample, measurement invariance between the different samples (Facebook, students, and online panel) and between current and former residents was assessed by running a multigroup analysis in AMOS (Byrne 2004). Hence, we tested measurement invariance in terms of equality of factor loadings between each pair of groups (Yi and La 2004).

For the three sample groups, model fit for the unconstrained three-group measurement model was satisfactory, with χ²/df = 1.695, CFI = .948, SRMR = .078, and RMSEA = .058 (nonsignificant). When constraining all factor loadings in the model to be equal among all three samples, Δχ² was nonsignificant (p = .10), suggesting overall metric invariance. We proceeded to test each construct one by one, constraining factor loadings to be equal across group 1 versus group 2, group 1 versus group 3, and finally group 2 versus group 3. Results of this analysis indicated that full metric invariance was found for the place attachment (PA) and self-expressiveness (SE) constructs between all three groups. Intention to share (IS) was fully invariant between group 1 versus group 2 and group 1 versus group 3, while it was partially invariant between group 2 versus group 3, with two of the three factor loadings being equal between the two samples.

Next, we tested the measurement model for invariance between current versus former residents, following the same procedure. Results showed that Δχ² was nonsignificant (p = .49) at the overall level; nor were any significant differences found at the construct level. Model fit for the unconstrained two-group measurement model was also acceptable: χ²/df = 1.851, CFI = .966, SRMR = .054, and RMSEA = .064 (nonsignificant). Hence, following the recommendations from Hair et al. (2010), the model was found acceptable for making comparisons of relationships between constructs including the different samples and groups.

Common Method Bias

As a final step before proceeding to structural model testing, we addressed the potential threat of common method bias (CMB). First, as multicollinearity can indicate CMB (Kock and Lynn 2012), we ran a linear regression using place attachment, ad–brand congruity, and self-expressiveness as predictors of the criterion variable (intention to share) in order to retrieve collinearity statistics. In this case, variance inflation factors (VIFs) for the tested variables were between 1.897 and 2.180—in other words, well lower than the generally recurrent thresholds of 5 (Kline 1998) or 10 (Hair et al. 2010).

Second, we applied the common latent factor approach, in which all items are permitted to load on their theoretical constructs and on a common latent factor simultaneously, thus uncovering how much of the variance is common between all observed indicators (Podsakoff et al. 2003). In the analysis, the unconstrained common method factor model is compared to a model in which all loadings between items and the common latent factor are constrained to zero (Jaber and Oftedal 2020). Then, we used Gaskin’s chi-square test (available at http://statwiki.kolobkreations.com/) to examine whether the difference between the unconstrained and the fully constrained model was significant. The test returned a p value of < .001, which indicates that CMB might influence the results. Therefore, we retained the latent factor in the structural model, which allowed us to effectively control for CMB (Jaber and Oftedal 2020; Schaller, Patil, and Malhotra 2015). Model fit for this CMB-corrected measurement model was well within recommended levels, with χ²/df = 1.685, CFI = .989, SRMR = .0247, and RMSEA = .057 (nonsignificant).

Results

After validation of the measures, we ran the structural CMB-corrected model in AMOS using maximum likelihood estimation (MLE). Results suggest good fit between the model and data: χ²/df = 1.685, CFI = .989, SRMR = .0247, RMSEA = .057 (nonsignificant). The standardized path coefficient between place attachment and place ad–brand congruity is highly positive and significant (β = .647, p < .001), providing support for hypothesis 1. Place attachment also influences self-expressiveness positively (β = .387, p < .001), as suggested in hypothesis 2. Moreover, the place brand message self-expressiveness appears to become stronger with higher place ad–brand congruity (β = .219, p < .05), indicating support for hypothesis 3. Furthermore, all three independent constructs are positively and significantly related to intention to share the place message online. In this regard, place ad–brand congruity has the strongest influence with β = .393 at p < .001 (hypothesis 4), while the path coefficients of self-expressiveness (hypothesis 5: β = .250, p < .001) and place attachment (hypothesis 6: β = .295, p < .001) are somewhat lower. Figure 2 depicts the results of the structural model.

Figure 2. Structural Model Results.

Overall, the squared multiple correlations of the endogenous constructs indicate that the model explains 62.4% of the variance in intention to share. Moreover, the model explains 41.8% of the variance in place ad–brand congruity and 30.8% of the variance in self-expressiveness.

Summary of Hypothesis Tests

Positive and significant relationships between the constructs indicate support for all hypothesized relationships. The level of perceived congruity between how the place is portrayed in the video (ad) and one’s perception of the place (brand) significantly increases the likelihood that current and former residents will share the video with others online. This influence is further driven by place attachment, which increases place ad–brand congruity. Self-expressiveness, in other words, the extent to which consumers perceive that a place brand message supports and enacts their self-concept and will be recognized publicly as such (Taylor, Strutton, and Thompson 2012), also tends to increase the intention to share the video—even more so when the place attachment of current and former residents is high.

Nonhypothesized Results

Mediation Effects

In addition to the direct relationships tested in the model, we evaluated the mediating role of self-expressiveness and place ad–brand congruity on intention to share, by running the Aroian version of the Sobel test suggested by Baron and Kenny (1986). Results (summarized in Table 2) suggest that place attachment has a significant indirect influence on intention to share, with place ad–brand congruity and self-expressiveness as partial mediators. This indirect effect is stronger than the direct relationship. Furthermore, place ad–brand congruity partly mediates the relationship between place attachment and self-expressiveness. Results also indicate that place ad–brand congruity may have some indirect influence on intention to share with self-expressiveness as a partial mediator, though the test is significant only at a level slightly above 5%.

Table 2. Results of mediation tests.

Relationship	Direct Effect	Indirect Effect	Total Effect	Indirect/Total	Test Statistic	p
PA → ABC → IS	.295	.386	.681	56.7%	4.245	< .001
PA → SE → IS	(Included in effects above)				2.589	< .01
PA → ABC → SE	.387	.142	.529	26.8%	2.123	< .05
ABC → SE → IS	.393	.055	.448	12.3%	1.930	.053

Note. Standardized effects. ABC = ad–brand congruity; IS = intention to share; PA = place attachment; SE = self-expressiveness.

Moderation Effects

Finally, we tested for possible differences between current and former residents in the structural relationships of the CMB-corrected model by running a multi-group analysis in AMOS. This means fitting the model to one group’s data while assuming it to be the same in the second group (Iacobucci 2010). Each path coefficient was constrained to be equal for both groups (current versus former residents). A significant change in χ² indicates that the path coefficient differs between the two tested groups, which suggests that group membership has a moderating influence on the relationship between constructs. In this case, the changes in χ² were very small for all the paths in the model (.105 − 1.718) and none was significant (p = .190−.812), suggesting that the hypothesized relationships were similar for current as well as former residents.

Conclusions

Theoretical Implications

Drawing from attachment theory (Altman and Low 1992), congruity theory (Osgood and Tannenbaum 1955), and identity theory (Sirgy 1982), this study has shed new light on the role of place attachment, place ad–brand congruity, and self-expressiveness on residents’ intention to share a place brand message in the form of an online streaming video.

First, while previous research has studied the congruity or incongruity between advertising and brand in a consumer products context, this study is to our knowledge the first to adapt the concept to place branding from the perspectives of local residents. Thus, we define place ad–brand congruity as the fit between the image of the place transmitted in advertising (a place brand message) and the residents’ existing image of the place. In contrast to some previous advertising studies that have found positive effects of incongruity between ad and brand (Dahlén et al. 2005; Halkias and Kokkinaki 2014), this study suggests that there are several positive outcomes of a strong match between the way the place is communicated in the ad and residents’ perception of the place. This is in line with congruity theory and with other research which has shown that negative receiver responses can result from advertising that creates dissonance (Loken 2006).

Compared to product brands, place brands are much more complex and intangible, embodying a multitude of elements and stakeholders (Hultman, Yeboah-Banin, and Formaniuk 2016). Considering these particularities, it is likely more difficult to capture and communicate the “essence” of a place to the extent that its residents feel that the message reflects their own beliefs, ideas, and impressions. However, when this is done successfully, the perceived congruence may also reinforce residents’ self-concepts. This is consistent with research in self-identity, which proposes that people use brands that are congruent with themselves to express their own identity and values (Thorbjørnsen, Pedersen, and Nysveen 2007).

Moreover, the more similar an object is to the self, the stronger the tie to the self and the feelings of value and attachment (Sirgy 1982). The results of this study suggest that affective place attachment has a positive influence on place ad–brand congruity as well as on self-expressiveness of the place brand message. This implies that strong emotional bonds to the place increase the likelihood that residents perceive place brand messages as confirming their own image of the place and as communicating their self-concepts accurately—for example, by instilling pride. Advertising thus has the potential to augment positive emotions that current as well as former residents feel about their place. Once developed, place attachment does not necessarily cease when a person moves away (Andéhn, Hietanen, and Lucarelli 2020; Casey 1993). This study adds to the body of knowledge in advertising and place branding by including the perspectives of both current and former residents.

Previous research has pointed to residents as ambassadors of their place, which can be effectively achieved by sharing content about the place in social media (e.g., Stylidis 2018). Thereby, they can influence tourists’ image formation and decision making, such as visitation intentions (Palmer, Koenig-Lewis, and Medi Jones 2013; Stylidis 2018). This study found that positive eWOM, in the form of current and former residents’ intentions to share a place brand video online, was amplified by place ad–brand congruity, self-expressiveness of the place brand message, and place attachment. Hence, aspects of self-enhancement (Jones 1973) as well as emotional bonds to the advertised place are likely to increase the tendency to spread place brand messages. The place brand’s highly symbolic value allows residents to articulate and communicate their self-concepts and to seek positive evaluations of themselves and their place from others. Identification with a place intensifies attachment to it (Zenker and Petersen 2014), and the feeling of pride and being part of “my” place fosters a will to show it to others.

Overall, place branding is still relatively young as a field of research, and there are many challenges related to the complexity of places as branded entities and the multiplicity of stakeholder groups (Kavaratzis and Dennis 2018). There is also a recently increasing focus on stakeholders’ role and engagement in place branding (Kavaratzis and Dennis 2018). This study, therefore, contributes to the understanding of a key stakeholder group—current as well as former local residents—and points to their importance in creating and communicating the place brand.

Practical Implications

A rich set of managerial implications for tourism destination policymakers, tourism destination marketers, and professional advertisers are derived from the study.

Our findings suggest that tourism destination policymakers should increasingly recognize residents—current as well as former—as a relevant stakeholder group. This means involving them more intensively and proactively in promoting their destinations as this might translate into more effective advertising that can improve host–guest balance (Jamal and Getz 1995) and enhance destination competitiveness (Crouch and Ritchie 2012; Ritchie and Crouch 2003). As such, destination tourism policies should be designed to incentivize DMOs to listen more proactively to both residents and hosting communities when defining the destination’s development and planning goals (Morrison 2019). This might demand the development of public–private partnerships in order to set up and maintain effective relationships between the public sector (destination policymakers and local administrators) and the private sector, including enterprises and hosting communities (Mariani, Di Felice, and Mura 2016). Overall, policymakers should internalize the idea that place advertising embedding residents can be relevant to sustain tourism through a more “organic” (Uchinaka, Yoganathan, and Osburg 2019) and likely cost-effective form of place marketing in which residents are the primary sources of marketing communication and active proponents of the place brand. This might ultimately translate into improved and effective local community-driven economic planning and development policies.

Tourism destination marketers should allocate a higher share of their resources and efforts to design more inclusive integrated destination marketing communication plans (Morrison 2019) whereby residents should be incentivized to participate in an increasing number of communication activities, including advertising. For instance, best managerial practices in relation to the aforementioned point have been developed by the DMOs of the Swedish city of Gothenburg (Göteborg & Co) and the Spanish city of Barcelona (BarcelonaTurisme). These cities have prioritized the quality of life of residents, which has likely translated into high levels of place attachment, as DMOs have been able to involve residents and the local community in marketing communications.

Our analysis also suggests that advertising should capture relevant aspects of the place and the people who live there so that residents feel that it reflects how they view themselves. Therefore, destination marketers are encouraged to develop their integrated destination marketing communication plans after a careful analysis of residents’ perceptions of the place. A positive correlation between place brand advertising campaigns and tourist expenditure has been empirically shown in recent research (Wen, Wu, and Whitfield 2019). Consequently, destination marketers and their respective DMOs should increase the share of their integrated destination marketing communication budget allocated to advertising and ask residents to play an active role in place brand advertising campaigns. Typically, residents display higher levels of place attachment than visitors and this might increase the virality of promotional messages, especially in online contexts as our study suggests. This is in line with previous findings that have found online content based on quizzes leveraging local dialects and residents’ knowledge of the place to increase social media engagement with the destination’s promotional activities (Mariani, Di Felice, and Mura 2016).

Professional advertisers should increasingly recognize that advertising a place and tourism destination brand is different compared to advertising a tangible product. Places not only consist of natural and cultural assets and heritage but are also made by people as well as by the symbolic meanings that revolve around local communities. Advertisers trying to promote places would benefit from involving residents in place advertising and branding activities, as they are active and credible place ambassadors who can help spread advertising messages to a wider audience. Our analysis suggests that place advertisers should consider ways to include not only people who currently reside in the area but also those who have moved away. Former residents still have strong emotional bonds to the place, which implies that they will relate to place brand ads and eventually spread them.

In the digital age, advertisers should look at the most suitable digital channels and media to reach the target audiences. Our study suggests that digital advertising campaigns involving residents are effective as they are conducive to share the digital ad (in the form of an online video). However, a precondition for this to happen is that the message and its execution should be congruent with the image residents have of the place. Ultimately, place advertisers should allocate more time, resources, and efforts to develop viral videos. Content that is spread organically by peers is perceived as more credible than induced sources such as official tourist brochures (Palmer, Koenig-Lewis, and Medi Jones 2013). In addition, according to recent advertising research (Eckler and Bolls 2011), viral videos are able to induce more intense emotional processes than traditional advertising.

Limitations and Future Research

The findings and implications outlined in the preceding sections should be considered in light of some limitations. First, as is often the case in place branding studies, this is a cross-sectional study of a single place, which limits the possibilities to generalize results. Causal effects in the tested model can only be inferred, not proven. Moreover, to reach relevant target groups for the empirical study, a purposive, nonrandom sampling strategy was adopted. While this is common for online surveys, it may raise questions of representativeness. Future research in other cities and countries, relying on larger samples, is needed to validate the proposed model’s relationships. Studies adopting an experimental design, allowing for manipulation of one or more of the independent variables, are also encouraged in order to overcome the potential weaknesses of relying on a single source. In addition, such research could reduce the potential impact of high correlations and overlap found between a minority of the constructs in this study.

Second, the perspective taken in the study was that of current and former residents, as they are key stakeholders and disseminators of eWOM about their city. Although place attachment is most often discussed in relation to residents of a focal place, research has shown that it is a valid concept also for visitors of a place (Lewicka 2011). Therefore, testing the same model from the perspective of visitors, or conducting comparative studies including both residents and visitors of a place, would make valuable contributions to knowledge. Third, we conceptualized place attachment as an affective component, which is the most common way to understand and study the concept (e.g., Lewicka 2010). Some researchers have, however, also included cognitive and behavioral dimensions of place attachment (Scannell and Gifford 2010). It may therefore be interesting to test the proposed model with different conceptualizations of the place attachment construct.

Fourth, the use of a single-item construct to measure place ad–brand congruity could be potentially viewed as a limitation. The practice of single-item measures in SEM is debated, with a number of arguments raised both for and against it (Petrescu 2013). However, several studies have shown that single-item measures are valid for concrete constructs, which are simple and easy to understand (Bergkvist 2015; Hair et al. 2010; Petrescu 2013), for example, attitude toward the ad, attitude toward the brand, liking (Bergkvist 2015; Bergkvist and Rossiter 2009; Hair et al. 2010), and image congruity (Marshall et al. 2008). For such concepts, multiple-item scales may reduce the quality of responses and add very little information over a single-item measure (Drolet and Morrison 2001). Hence, we found a single item to be suitable for the place ad–brand congruity construct as it has these characteristics and because we attempted to make the questionnaire concise to reduce respondent fatigue, undesired response styles, nonresponse, and early break-off (Adigüzel and Wedel 2008).

Finally, concepts related to the self and identity may vary among different cultures (e.g., Inglehart and Baker 2000). Future studies could therefore test the model in other countries and across cultures. Moreover, though the model explains a fair amount of the variance in the dependent variables (e.g., more than two-thirds of intention to share), there are likely other constructs that work as antecedents to residents’ intention to share a place brand message online. A possible avenue for research could be the addition of other self related constructs such as self–brand congruity, value congruity, or social identity expressiveness, as well as testing potential moderating influences of demographics or other characteristics on the measured relationships.

Additional information

Funding

This study is funded by The R&D Fund of the Swedish Tourism and Hospitality Industry.

Notes on contributors

Maria Ek Styvén

Maria Ek Styvén (PhD, Luleå University of Technology) is a professor, Luleå University of Technology, Luleå, Sweden.

Marcello M. Mariani

Marcello M. Mariani (PhD, University of Bologna) is a professor, Henley Business School, University of Reading, Henley on Thames, Oxfordshire, United Kingdom.

Carola Strandberg

Carola Strandberg (MSc, Luleå University of Technology) is a doctoral student, Luleå University of Technology, Luleå, Sweden.