Development and validation of a new measure of travel destination personality

The main aim of this study was to develop a new scale for measuring travel destination personality – the Destination Personality Scale (DPS). It was assumed that DPS will confirm the applicability of the five-factor structure of the original Brand Personality Scale (BPS) by Aaker (1997) when more appropriate adjectives for the description of travel destination personality are generated. Results confirmed the five-factor structure of DPS, with dimensions Excitement, Sincerity, Competence, Ruggedness, and Sophistication. The final version of DPS consists of 24 items and subscales showed good construct (convergent and discriminant) and criterion validity, as well as reliability.


Highlights:
• The main aim was to develop a new scale for measuring travel destination personality -Destination Personality Scale (DPS).• The study was based on dimensions of the Aaker's Brand Personality Scale (BPS).• Adjectives that are more appropriate for a description of travel destination personality were selected.• DPS contains 24 destination-specific items and measures five dimensions in line with BPS.
People often attribute personality traits to inanimate and inhumane objects (Aaker, 1996;Solomon, 1999).For example, in many commercials, cars are described as sophisticated, PC is described as intelligent, and destinations as friendly or romantic, etc.This tendency is called anthropomorphism.In marketing, attributing human characteristics to different products is called the brand personality.The personality of the brand is defined as "a set of human characteristics that can be linked with the brand" (Aaker, 1997, p. 347).It should be also noted that personality traits of the brand are not strictly related to the physical characteristics of the brand and may be the result of many factors such as advertising, price, stereotypes, etc. (Sirgy, 1985).
Destinations are nowadays regarded as brands and the brand aspect of attributing human characteristics to them is called destination personality.Thus, Ekinci and Hosany (2006, p. 127) define destination personality as "the set of personality traits associated with a destination".The major benefit of a wellestablished destination personality is that it allows consumers to differentiate the destination from competition (Aaker, 1996).Brand personality also allows the development of a strong emotional connection between the brand and consumers, resulting in the creation of loyalty, trust and greater tourist satisfaction (Fournier, 1998).
The most popular general scale for measuring brand personality, created to be applicable for all brands, is the Brand Personality Scale (BPS) developed by Jennifer Aaker (1997).Originally, BPS is based on Big Five human personality structure, consisting of 42 items that measure five dimensions: Sincerity, Excitement, Competence, Sophistication, and Ruggedness.Although, it was created to be a universal scale for brand personality, in case of travel destinations, as specific brands, this scale in its original form, is not fully applicable.There are two main reasons to explain this: 1) BPS is mainly used to measure brand personality of tangible products, while travel destinations offer both tangible goods and intangible services (Xie & Lee, 2013) and 2) the items of BPS are inadequate to describe a travel destination.The previous studies of destination personality suggest that there is a problem in determining the exact number of dimensions and exact items which should be used in describing destination personality (e.g., Ekinci & Hosany, 2006;Kim & Lehto, 2013;Chon & Psau, 2013;Papadimitrou et al., 2014;Jovanović, 2014).The cause of this problem could lie in the fact that researchers are mostly using modified or shortened BPS which makes it difficult to compare the results (see Usakli & Baloglu, 2011 and Table 1).The main reason for applying shortened and/or modified BPS could be the fact that not all items in the original BPS are applicable to travel destinations, as travel destinations do not have the same basic characteristics as other products and service brands used for trait generation in the original scale.Also, some authors developed their own scales or used different methodologies (projective techniques or interviews) for measuring destination personality, which resulted in different personality dimensions and different scales which are not comparable.Some authors used the original BPS, but the original factor structure in most cases was not confirmed (e.g., Kim & Lehto, 2013;Jovanović, 2014).We believe that the reason for this may not be inadequate dimensions of destination personality, but inadequate items that represent these dimensions.This indicates that further research is required to determine specific traits of destination personality and to test the applicability of original BPS five-factor solutions to travel destinations, in the case when items are adapted to describe a travel destination as a specific brand.The main assumption of this research is that the original model of Aaker (1997) with five dimensions is applicable to travel destination personality, but BPS items are inadequate for measuring it.Authors who used BPS and found it useful for measuring destination personality concluded that it may not fully represent the destination personality (Chon & Psau, 2013;Papadimitrou et al., 2014;Kim & Lehto, 2013;Xie & Lee, 2013).For the same reasons, several authors (Chen & Phou, 2013;Kim & Lehto, 2013;Hultman et al., 2015) suggested the development of a more valid, reliable and generalizable destination personality scale for the use in tourism research and this study intends to address this issue.Thus, the main aim of this research was to develop a new destination personality scale which measures five dimensions from the original BPS model, but with items suitable for the description of destination personality.In order to achieve this, three studies on different study samples have been conducted -preliminary study 1 (item generation), study 2 (exploratory factor analysis) and study 3 (confirmatory factor analysis).The new scale is thus supposed to overcome the main issues and questions that have arisen with the use of Aaker's (1997) BPS: 1) whether all five dimensions of this scale can be applied to a travel destination as a specific brand and 2) whether the items of this scale are indeed related to the personality of the destination.
Thus, it is assumed that the scale we developed based on Aaker's model should be a better measure of travel destination personality and that the selected items are related to personality traits which are easily attributed to a travel destination.

Measuring Brand Personality
Brand personality is mainly measured using idiographic or general scales (Helgeson & Supphellen, 2004).Idiographic scales are qualitative and are generally implemented first, to determine which personality traits can be attributed to the selected brand.Therefore, they contain only those features that are relevant for a given brand.General scales are quantitative and can be used for measurement of all types of brands, as the already mentioned BPS (Aaker, 1997).Originally, BPS consists of 42 items that measure five dimensions: Sincerity (e.g., honest, real), Excitement (daring, cool), Competence (reliable, corporate), Sophistication (glamorous, charming), and Ruggedness (masculine, outdoorsy).Three dimensions -Sincerity, Excitement, and Competence are associated with the personality dimensions of the Big Five model -Agreeableness, Extraversion, and Conscientiousness.Sophistication and Ruggedness are different from any personality traits and are considered specific personality characteristic of a brand.

Measuring Travel Destination Personality
Although BPS should represent the general/universal measure of brand personality, not limited to the specific type of brand, many studies indicated that BPS is not appropriate for measuring various types of brands and that it must be adjusted and modified (Austin et al., 2003;Venable, Rose, Bush, & Gilbert, 2005;Herbst & Merz, 2011;Usakli & Baloglu, 2011;Matzler et al., 2016).BPS has also found its application in the tourism field and various authors have encountered the same problem with its generalizability.For example, Ekinci and Hosany (2006) were the first who applied BPS to measure destination personality.Specifically, they tested the original 42-item BPS scale, but finally applied the reduced scale of 27 items, as these 27 items were chosen by 70% of the pretest respondents to describe travel destinations (the reduction was made in order to achieve content validity).Instead of five original dimensions, Ekinci and Hosany (2006) extracted three dimensions: Sincerity, Excitement, and Conviviality (as a new dimension and specific to destinations), with the total of 12 items retained.In their next studies, the three-factor solution was confirmed (Hosany et al., 2006;Ekinci et al., 2007).After that, many authors (Murphy, Moscardo, & Benckendorff, 2007;Pit et al., 2007;Sahin & Baloglu, 2009;Usakli & Baloglu, 2011) applied BPS to various travel destinations and obtained a different number of dimensions with quite a different factor structure.Only two studies (Pit et al., 2007;Sahin & Baloglu, 2009) confirmed that the original five-factor structure of BPS can be applied to travel destinations, but with a different number of items.The summary of studies that explore destination personality structure published in 2009 or before is shown in Usakli and Baloglu (2011), while in Table 1, newer findings are presented covering the period from 2011 to 2018.As presented in Table 1 and in the overview of previous research shown in Usakli & Baloglu (2011), some of the studies used reduced BPS, some extracted different dimensions, but by analyzing all of the available studies, it seems that BPS cannot be applied to travel destinations in its original form.The analysis of mentioned studies shows that Sincerity and Excitement are dimensions which were most frequently extracted, while Ruggedness was excluded from most of the previous studies.
The review of the available literature indicates that there are two main issues which should be considered regarding the application of this scale to travel destinations.The first issue is whether this scale can be applied to travel destinations as a specific type of brand.In line with that, some authors (Austin et al., 2003;Venable et al., 2005;Herbst & Merz, 2011;Usakli & Baloglu, 2011;Matzler et al., 2016) argue that BPS is not generalizable when applied to specific brand categories, which is also the case with travel destinations.Authors also consider that not all brands can be described with the same personality traits, due to their specific nature and representation in the consumers' mind.Travel destinations could be considered a product, but they are a very complex one as they contain both tangible elements (i.e., buildings, beaches, flora, and fauna) and intangible ones (i.e., service, hospitality) (Smith, 1994).Therefore, they obviously cannot be described with the same personality traits used for typical consumer brands.For example, in the case of Levis' jeans the perception of ruggedness could refer to the product structure and its physical characteristics, thus being described with traits such as tough, rugged etc., while in the case of travel destination it could refer to how open that destination is for tourists, how resistant it is to tourism degradation and how (un)touched it is.So, in the case of a travel destination as a specific brand, this dimension could be described with completely different items -mysterious, free, strong, etc.This means that, although brand personality dimensions isolated by Aaker (1997) could be applied to different types of brands, items describing these dimensions should be adapted to the specific type of brand in order to give personality dimension the meaningful context.Using the original items to describe travel destinations as a specific brand might be the reason why, in the majority of previous studies, the original BPS scale was significantly reduced in terms of both the number of items and dimensions.Thus, unlike previous papers that used new items and isolated new dimensions that better describe travel destination i.e., conviviality (Hosany et al., 2006), tranquility (Kumar & Nayak, 2018), vibrancy (Usakli & Baloglu, 2011), and so on, we assume that the original five dimensions defined by Aaker (1997) could be confirmed in the context of travel destinations.In order to achieve this, items of each dimension should be adapted to describe travel destinations.We believe that the original dimensions of brand personality are adequate, comprehensive and meaningful enough to describe this concept.However, specific items should be generated.In light of this, this study aims to develop such a destination personality scale, with items suitable for travel destinations.
The second issue is whether the items of this scale are indeed related to the personality of the brand, in this case, a travel destination.For example, Azoulay and Kapferer (2003) explored whether this scale actually measures the personality of the brand, and concluded that BPS measures dimensions of brand identity (intellectual abilities, gender, and social class) with a personality only as one of its segments.This study states that BPS measures the personification of the brand, rather than brand personality.For example, items such as western or outdoorsy are not describing the personality of the brand, and in that context, authors consider that the new scale should avoid the generation of such items.This was controlled in Study 1 which included the generation of itemsdestination personality traits and excluded items that do not describe personality.
Besides these main issues, there is one more which includes the applicability of BPS items in specific cultures.For example, Jovanović (2014) and Jovanović et al. (2017) concluded that some of the brand personality items change their meaning when translated into Serbian.For example, item western was originally part of the dimension Ruggedness (Aaker, 1997), but when translated into Serbian, this item obtains a different connotation and is more adequate to describe the Competence dimension.In Eastern European countries western is seen as something progressive and well developed.Also, in the study of Jovanović (2014), items masculine and feminine are part of the same dimension -Sophistication, while in the original model of Aaker (BPS, 1997), masculine was part of the Ruggedness dimension and feminine was part of Sophistication.Jovanović (2014) argues that in more liberal societies there are fewer differences between a man and a woman, so these terms might not have significantly different meanings.This means that masculine does not have to be associated with raw, resilient, while feminine does not have to be associated with sophisticated and glamorous.Chon and Psau (2013) also excluded feminine and western as problematic items that are not descriptive enough of the extracted dimensions.This indicates that there should be caution in the use of these terms as they can have different meanings in different cultures and thus should be avoided.All these results indicate the cultural sensitivity of certain aspects of the brand personality.Therefore, it is especially important to create a destination personality scale that will be culturally invariant, as tourists who evaluate the destination personality are often international and coming from different cultures.Special attention was given to this issue during the creation of a new scale as such culture-specific items were avoided.

The Present Study
The aim of this research is to develop the Destination Personality Scale (DPS) based on the original five dimensions of the Aaker's (1997) model of brand personality, but with items more applicable to travel destinations.The development of DPS includes three studies.The aim of the first preliminary study was to generate items for DPS, the aim of the second study was to explore the factor structure of the DPS, and the aim of the third study was to further validate the final version of DPS by determining its construct and criterion validity, as well as the reliability.To examine construct validity, factor structure was tested, as well as convergent and discriminant validity.Convergent validity, which explains how well observed variables posited to a latent construct converge or share a high proportion of variance was examined by calculating the score of average variance extracted (AVE, see Fornell & Larcker, 1981) for each DPS dimension.AVE can be obtained as a sum of squared loadings divided by the number of items, thus it measures the level of variance captured by a dimension versus the level due to measurement error.Furthermore, correlations with similar constructs, e.g., with cognitive and affective image were calculated.Hosany and colleagues have found that affective destination image is interrelated with sub-scales (sincerity, excitement, and conviviality) of destination personality (Hosany et al., 2006;Hosany, Ekinci, & Uysal, 2007), while Chen and Phou (2013), Xie and Lee (2013) and Ekinci, Sirakaya-Turk, and Baloglu (2007) confirmed the relationship between the cognitive image and destination personality.Discriminant validity provides evidence that each construct can capture its own unique information not obtained from other constructs in the model and where each observed variable is posited to only one construct.It was examined by comparing squared correlations between each pair of dimensions with AVE scores (Kline, 2011).Discriminant validity was confirmed if squared correlations were lower than AVE scores.Criterion validity was tested by examining the relationship of destination personality with tourists' satisfaction with the destination that they have visited, since previous studies showed that destination personality positively affects tourist satisfaction with the visited destination (Chen & Phou, 2013;Hultman et al., 2015).

Participants and Procedure
Participants were 170 (57% female) students of tourism and hospitality from the University of Novi Sad who have traveling experience (M age = 22.34, SD age = 8.76).This sample was only used for the purposes of Study 1.The participants were approached at the Faculty of Sciences, Department of Geography, Tourism and Hotel Management in January, 2015.They were introduced with the purpose of the study and that their participation is anonymous and voluntary.Eliminatory question for participation in the study was whether they have travelled in the last two years.

Instrument
Study 1 was conducted in order to generate items for DPS.Participants were asked to write 3 to 5 adjectives for each dimension of the Brand Personality Scale (Sincerity, Excitement, Competence, Sophistication, and Ruggedness), that, according to their opinion, best describe the personality of destinations in general, not referring to any particular destination.The authors provided the example for each dimension to clarify their meaning in the context of tourist destination (Sincerity -a destination that does not deceive tourists and information about it are true, Excitement -a destination that offers exciting activities and experience, Competence -destination which is competent to provide any type of service, Sophistication -a destination that offers a sophisticated experience, and Ruggedness -destination that has yet to be shaped, that is still unspoiled).

Results: Generation of the Item Pool
After adjectives for each dimension were generated, synonymous items were merged and this resulted in a total of 551 adjectives (Sincerity = 126, Excitement = 91, Competency = 132, Sophistication = 113, and Ruggedness = 89 adjectives).Thereafter, 15 adjectives with the highest frequencies (over 10) were selected for each dimension resulting in a total of 75 adjectives which represent the initial pool of items for the preliminary version of DPS.

Instrument
Since Study 1 was focused on item generation, the first real use of the DPS was in Study 2. In order to determine the factor structure of the DPS principal component analysis was conducted.DPS is comprised of 75 adjectives generated in Study 1 that describe personality attributes of travel destinations.
Participants were asked to imagine the last travel destination they have visited and to rate how much each adjective describes that destination.The ratings were done on a Likert scale from 0 = not clear, 1 = does not describes, to 5 = totally describes.Option not clear was given to provide an answer if participants think that adjective is not appropriate or not clear for describing travel destination.

Data Analysis
The principal component analysis has been applied in order to explore structure of destination personality.The decision about the number of components was based on parallel analysis (O'Connor, 2000).Promax rotation was applied because that we expected the extracted components to be correlated (Jovanović et al., 2017).Before conducting the analysis, 19 out of 75 adjectives were excluded from the instrument because they had more than 5% of "not clear" answers (brave, subtle, confident, hardworking, raw, consistent, caring, resourceful, courageous, sophisticated, cuddly, fair, sincere, capable, loyal, determined, competent, meek, direct).The 5% cut off point was applied because it is a standard margin of error used in quantitative research (Reid, 2013).Higher percentage of "not clear" answers could indicate invalid (erroneous) items for describing given destination personality dimension.After this, a total of 56 adjectives were retained.

Results
Representativeness was good (KMO = .85)and Bartlett's sphericity test was significant (χ 2 (1540) = 7041.82,p < .001),which confirmed that data are suitable for the analysis.Based on results of parallel analysis, five dimensions were extracted.From a total of 56 adjectives, 8 adjectives with loadings lower than .40were excluded (real, predictable, natural, peaceful, quiet, inflexible, modest, and unspoiled) and after that, 48 adjectives remained.The parallel analysis confirmed the five-component solution and components were interpreted as Excitement, Sincerity, Competence, Ruggedness, and Sophistication, in accordance with the original BPS model.In order to shorten the scale, seven adjectives that had the highest loading were selected for each dimension.As a result, another 13 adjectives were excluded from the study (competitive, innovative, relaxed, romantic, charismatic, discreet, interesting, exciting, unique, passionate, proud, curious, and joyful).Finally, 35 adjectives were retained (see Table 2).Correlations between components were significant and positive and ranged from .12 (between Ruggedness and Excitement) to .57(between Sincerity and Excitement).

Data Analysis
In order to validate the DPS structure from Study 2, confirmatory factor analysis (CFA) was conducted.Given that the value of multivariate kurtosis was over 7, the robust method was used (Bentler, 2006).To determine the fit of the model, the following indicators were used: Sattora-Bentler χ² (S-Bχ²), which should not be statistically significant and ratio χ2/df, which should be less than 2 (Hoelter, 1983), or less than 3 (Kline, 2015); the square root of the mean squared errors of approximation (RMSEA) and standardized square root of the average of the square residuals (SRMR) which should be less than .08;Bentler Comparative Fit Index (CFI), which should be higher than .90for adequate model fit (Hu & Bentler, 1999).CFA was conducted in EQS 6.1 software.The convergent validity of each dimension was examined in two ways.Firstly, it was examined by calculating the score of the average variance extracted (AVE, see Fornell & Larcker, 1981).A substantial convergent validity is achieved when all item-to-factor loadings are significant and the AVE score is higher than .50within each dimension, but AVE higher than .40 is still acceptable if composite reliability (CR) is higher than .60(Huang, Wang, Wu, & Wang, 2013;Fornell & Larcker, 1981).Discriminant validity was examined by comparing squared correlations between each pair of dimensions with AVE scores (Kline, 2011).If the squared correlations based on total scores are lower that AVE, discriminant validity was confirmed (Fornell-Larcker, 1981;Zaiţ & Bertea, 2011).
For further construct validation, correlations with cognitive and affective images of tourist destination were calculated, and for exploring the criterion validity, prediction of tourists' satisfaction with the destination based on DPS dimensions was tested.

Results
Results of CFA showed that the first model with correlated factors (Model 1, see Table 3) did not show satisfactory fit indices, so it was necessary to make certain changes.Firstly, the Lagrange Multiplier Index (LM) indicated that the items cool, sociable and kind should have loadings on three dimensions and thus they were eliminated from the model in order to avoid cross-loadings (Model 2).In a further improvement of the model fit, all items that did not have a loading above .30(items imaginative, adventurous, tolerant, strong, mysterious, signified, and sophisticated) were removed (Model 3).This resulted in good fit indices (Table 3) and factor loadings (Figure 1).Thus, the final model consisted of 24 items distributed into five dimensions.All dimensions had AVE higher than .40 and CR higher than .60(Table 4) which indicates good convergent validity.Furthermore, correlations with the cognitive and affective image were all significant, except for the correlation between Sophistication and affective image (Table 5).Note.All correlations .18and higher are significant at p < .001.
The range of squared correlations based on total scores is from .07 to .35, which is lower than AVE.Thus, results confirm that DPS dimensions have sufficient discriminant validity.
All five dimensions have good Cronbach's α (≥ .82)considering the small number of items (Table 4).Furthermore, mean values for all dimensions are relatively close and range from 3.51 to 3.81, which means that there is no favoring of any of the dimensions.
Regarding criterion validity, Competence obtained the highest correlation with the satisfaction with the destination, compared to the rest of the dimensions (Table 6).Moreover, prediction of satisfaction was significant (R = .34,R 2 = .11,p < .001) with Competence as the main predictor, followed by Excitement (Table 6). 55 In order to control their effects, sex, age, and education were entered in the first step of hierarchical regression analysis and the same results were obtained.

Discussion
The aim of this research was to develop a new measure of the destination personality -Destination Personality Scale (DPS), which would be based on the original five dimensions of the Aaker's (1997) model of brand personality, but would consist of items more suitable for travel destinations as brands.After the validation process, the final version of DPS contains 24 items/adjectives that describe five dimensions from the original Aaker's (1997) model.Although the five original dimensions of brand personality have been confirmed, DPS offers new items which are solely personality traits, aimed to be devoid of cultural specificities and are more appropriate for describing destination personality.This approach differs from those of previous studies that either isolated new dimensions using modified and shortened BPS scales (e.g., Hosany et al., 2006;Kumar & Nayak, 2018;Usakli & Baloglu, 2011), or developed new scales for destination personality (e.g., Pan et al., 2017;Kumar & Nayak, 2018) with dimensions and items for specific destinations.Thus, one of the main theoretical contributions of this study is that it develops and tests a scale designed to measure destination personality based on the original five dimensions from Aaker (1997) model, but with items specific to travel destination.Thus, proposed DPS also addresses the criticism that Aaker's BPS may not be suitable to study destination personality because some of the dimensions of the traditional BPS are not applicable to travel destinations (Kim & Lehto, 2012;Murphy, Moscardo et al., 2007;Usakli & Baloglu, 2011).Unlike these studies, the current study confirmed that dimensions of BPS are applicable to a travel destination when items are adapted to describe the specific nature of the travel destination.But, even Aaker (1997) emphasizes that BPS might not be appropriate for measuring brand personality in a different cultural context.So, in its original form, BPS is not applicable to travel destinations and is sensitive to the culture from which the tourist originates (Jovanović 2014;Jovanović et al., 2017).Also, in order to avoid measuring brand identity instead of brand personality (Azoulay & Kapferer, 2003), the authors were careful to use only the items that are personality traits.A final contribution of the study is a 24-item five-dimensional scale with dimensions Excitement, Sincerity, Competence, Ruggedness, and Sophistication.DPS dimensions showed good construct and predictive validity and reliability.
Even though the original model of brand personality indicated that five dimensions are independent and uncorrelated (Aaker's, 1997), previous results showed that these dimensions are related (Jovanović, 2014;Jovanović et al., 2017;Murphy et al., 2009).In the final DPS model, correlations between dimensions are low to moderate, indicating that although some dimensions shared similarities, the distinction between them could be made.Furthermore, discriminant validity of the dimensions was confirmed by showing that AVE scores are higher than squared correlations between dimensions.
Results showed that Sincere destinations could be described as benevolent, emotional, open, friendly and convivial, destinations where people feel welcome, cared for and safe, where people could take care of their needs and destinations where everything is known and without hidden intentions.Travel destinations that are perceived as Exciting are seen as energetic, fun, vibrant, dynamic, and destinations where there is always something new going on, with a variety of activities and various forms of entertainment.Competent travel destinations are perceived as reliable, responsible, dedicated, serious, organized, and attentive.Those are travel destinations that take care of their customers, with a focus on their needs and wishes, as well as organized destinations that provide high-quality services to their customers.Sophisticated travel destinations are perceived as smooth, balanced, gentle and charming destination for people with refined taste.Rugged destinations are destinations that are mysterious, spontaneous, free, uncontrollable and resistant, destinations that are still waiting to be discovered, with interesting stories, as well as destinations that are still waiting to emerge on the tourist market.
If we compare items of the original BPS with the new DPS, only 4 items are the same -charming, smooth (Sophistication), sentimental (Sincerity) and reliable (Competence).We could conclude that DPS manages to represent Aaker's (1997) model but contains items that are more suitable for the description of destination personality.For example, BPS included items that are not related to personality traits (for example, technical, up-to-date) and are culture-specific (western).Major differences are observed in the dimension Ruggedness which is now fully modified to describe the personality of a travel destination.In the BPS, Ruggedness destination trait is described as rugged, tough, masculine, western, and outdoorsy.Ruggedness from BPS was shown to be inadequate in describing the personality of a travel destination in a previous study (Jovanović, 2014;Jovanović et al., 2017;Papadimitrou et al., 2014;Usakli & Baloglu, 2011).For example, items rugged, tough, and western are difficult to associate with a travel destination (Jovanović, 2014;Jovanović et al., 2017).Moreover, some authors claimed that not a single item of the original dimension of Ruggedness can be really regarded as a personality trait (Azoulay & Kapferer, 2003) and many previous studies did not extract this factor (Usakli & Baloglu, 2011;Jovanović et al., 2017).Ruggedness in this scale has been given a different connotation in order to overcome previously mentioned issues.

Practical Contributions
This research proposed a new measure of destination personality -Destination Personality Scale (DPS), based on the original five dimensions of the Aaker's (1997) BPS, but consisting of items more suitable for travel destinations as brands.Aside from the theoretical contribution, these findings also have practical implications, especially for destination marketers.As the items of the scale are generated specifically for travel destinations, destination marketers can benefit from using this scale for measuring destination personality, which is crucial for effective positioning and differentiation on the market.Previous measurement scales of destination personality based on BPS may not depict the real nature of tourist destinations, so DPS makes it easier to explore specific personality traits of a destination that can be used for promotional purposes.Tourism destinations could be differentiated and positioned based on these personality characteristics.

Limitations and Future Research
Although this research might have surpassed some limitations of previous studies of destination personality, it also has several of its own.First of all, the study was done only on tourists from one country (Serbia), and for the crosscultural validation of the scale, future research should include international samples.Secondly, the sample for Study 3 consisted of participants who evaluated only one type of travel destination (Roman sites in Serbia).Future studies should include tourists who would assess different types of destinations, e.g., sea, mountain.Stability of destination personality should be examined i.e. future research should explore whether the structure of DPS remains the same before and after a trip.Also, since there may also be heterogeneity within a culture (Zhang, Decosta, & McKercher, 2015), future research should analyze socio-demographic variables such as race, ethnicity, education etc., which might affect the perception of the destination personality.

Conclusions
The current study developed a new measure of the destination personality DPS that contains 24 items generated specifically for travel destinations as brands.To achieve this, authors applied a three-study process (items generation, scale validation by exploratory factor analysis and further scale validation by confirmatory factor analysis) on three different samples.A new scale showed satisfactory construct (convergent and discriminant) and predictive validity, as well as reliability.DPS contains new personality traits, aimed to describe the complex nature of travel destinations.Thus, it confirmed that BPS dimensions can be applied to travel destinations when items are generated specifically for travel destination personality.A major improvement of the scale has been done in the context of the dimension Ruggedness as its structure is now modified to describe travel destinations.DPS scale also aims to be independent of cultural specificities (by avoiding items which may be specific for certain cultures), however, this aspect requires further research and cross-cultural validation of the scale.

Study 2 :
Principal Component Analysis of the DPS Participants and Procedure Participants were 180 (60% female) students of tourism and hospitality from the University of Novi Sad (M age = 21.81,SD age = 7.89).The procedure was the same as in the Study 1, and the research was conducted during March, 2015.

Table 1
Destination personality studies and their similarity to the original Brand Personality Scale

Table 2
An excerpt from the pattern matrix of the Destination Personality Scale(N = 180) Note.Loadings lower than .30were omitted.

Table 4
Descriptives and measurement properties of the Destination Personality Scale(N = 502)

Table 5
Convergent validity of the Destination Personality Scale: Relations with cognitive and affective image of the Roman sites on the culture route(N = 502)

Table 6
Criterion validity of the Destination Personality Scale: Relations with satisfaction with