Visitors’ satisfaction in Dubai and pre-trip destination image

Understanding the factors that influence tourists’ satisfaction and the pre-trip destination image of potential visitors is particularly important for policy makers and tourism marketers. The objective of this study is twofold; first to assess the satisfaction level of tourists who have visited Dubai and further explore the factors that shape it and associate with it. Second, to assess the intention to visit Dubai according to the pre-trip destination image of potential tourists. This empirical study relies on a unique sample of 210 participants from all over the world in the year 2017. Several demographic characteristics as well as variables related to the trip process and the city attributes are collected and explored in order to document any relationship between the two groups. The major findings of the ordered logit analysis demonstrate that the city attributes are the most significant contributors to tourists’ satisfaction and to non visitors’ intention to visit Dubai. Trip factors and demographic characteristics also play a significant role only for the group of visitors. The overall satisfaction is what creates loyalty and drives tourists to repeat their visit.


Introduction
Travel and tourism is one of the most important activities worldwide that generates billions of income each year.According to the World Travel and Tourism Council (2018a), the total contribution of travel and tourism to GDP in 2017 was USD 8,272.3bn (10.4% of GDP) and this is expected to rise by 4% to USD 12,450.1bn (11.7% of GDP) in 2018.In total, Travel and Tourism in the United Arab Emirates generated AED154.1bn(USD 41,950.5mn),accounted for 11.3% of GDP in 2017 and is expected to accelerate to 4.9% in 2018, and 10.6% in 2028 (World Travel and Tourism Council, 2018b).The statistics note that the number of international tourists visiting the UAE grew by 6.5% in 2017 compared to 2016 (Gulf News, 2018).According to the Dubai Statistics Center (2018), in 2017 the number of tourists who visited Dubai increased by 6% relatively to the previous year (from 14,900,000 to 15,790,000).
The earliest works on image and businesses are from Boulding (1956) and Martineau (1958) leading to an increase in image research (Balmer, 2009).They suggested that human behaviour relies more on the image people have rather than on an objective reality.According to Pike (2016), image and consumer behaviour research builds on Thomas's theorem: "What is defined or perceived by people is real in its consequences" (Thomas andThomas, 1928, p. 572, in Patton, 2002).Since the beginning of the 1970s the concept of image has been applied to tourism destinations.Hunt (1971) in his PhD dissertation, Gunn (1972) and Mayo (1973) were the first authors to demonstrate the relevance of destination images for tourism.Martens and Reiser (2017) stated that the image of a destination is the sum of all perceptions, impressions, feelings and attitudes a tourist [or a potential tourist] has in his/her mind with respect to a destination.In his review, Chon (1990) found the impact of destination image on the destination choice process to be a prevalent topic in tourism literature.Pike (2002) identified tourism destination images as one of the most frequently researched tourism topics.Research on tourism destination image increased especially in the 1990s alongside the growing corresponding attention of destination marketers (Tasci et al., 2007).Unanimously, tourism researchers agree on the importance of tourism destination images for the destination choice process and, thus, for the relevant marketing efforts (eg.Baloglu and McCleary, 1999;Echtner and Ritchie, 1993;Fakeye and Crompton, 1991;Pike, 2008).
The particular challenge complicating the conceptualisation of the tourism destination image formation process is that everything happens in the mind of potential individual tourists (Reynolds, 1965;Fakeye and Crompton, 1991).The latter hampers the identification and conceptualisation of the process, which according to Pike (2016) is as a black box process and not fully comprehended.Several tourism researchers (eg.Erickson et al., 1984;Anand et al., 1988;Gartner, 1993;Stern and Krakover, 1993;Llodrà-Riera et al., 2015;Pike, 2016) based on the conceptualisation of attitude by Fishbein and Ajzen (1975) in order to understand the image formation process as a three step process following the information input.This process starts with the cognitive image, the knowledge, and beliefs somebody has of one place, based on which the affective image, feelings and attitudes towards a destination are developed (Stylidis et al., 2017).Therefore, the affective reaction leads to a conative reaction, apparent in willing to travel.
However, as Kim and Chen (2016) argue, the affective reaction is inseparably linked to the belief or cognition and not a mere consequence of cognition.Cronin et al. (2000) and Clemes et al. (2011) in their research have proven that behavioral intention is influenced to a large extend by the customers' satisfaction which, in turn, is influenced by their experience (Chen and Chen, 2010;Zins, 2002).Using the concept of schemas, i.e. the mental structures of the classified knowledge and information in the human mind, Kim and Chen (2016) developed a new tourism destination image formation model, according to which associations or attributes on the destination are collected in these schemas and connected to the destination through memory nodes.
Further, Gallarza et al. (2002, p. 59) mention that "attitudes towards tourism can be a significant component of the destination image formation process."Besides the guests or potential tourists' culture, the culture of the host population at a destination can influence the tourism destination image.Henderson (2006) mentions that the United Arab Emirates has a very important strategic location providing a bridge to connect Europe with the Indian subcontinent as well as the Far East and Africa.Reisinger (2009) writes about Emiratis as being known as very welcoming and hospitable hosts to people of all cultures.Apart from experiencing the culture at the destination, the image of a tourism destination may be influenced by interacting with people from the Middle East in the tourists' country of residence.In the relevant vein of literature, products are found to influence the image of countries (Lee et al., 2016;Katsumata and Song, 2016).The same holds for the perceptions of products with respect to tourism destination images (Elliot et al., 2010;Lee and Lockshin, 2012).
The image of a tourism destination differs between potential first-time visitors and potential repeat visitors (Fakeye and Crompton, 1991;Giraldi and Cesareo, 2014;Kim and Morrsion, 2005;Liu et al., 2012).Accordingly, Kim and Chen (2016) utilise the concepts of confirmation and disconfirmation leading to a congruity or incongruity with the pre-trip destination schemas.This builds a foundation to assess how the destination image changes through visitation.Lin et al. (2012) found potential first-time visitors to have problems stating emotional connections towards destinations, where the decision of a potential repeat visitor relies on (Gunn, 1972;Baloglu and McCleary 1999).Although Martens and Reiser (2017) concluded that research on potential first-time visitors should focus on cognitive images, research may not exclude possible emotional connections.According to Martín-Santana et al. (2017), there are no studies evaluating how several factors can influence the image gap, i.e. the perception change before and after visiting a destination, while a positive gap can lead to greater levels of satisfaction.In addition, different types of non-visitors may have different types of image formation processes (Cherifi et al., 2014).
Customer satisfaction has attracted much attention in the literature, as Cronin and Brady (2000) indicated, due to its potential influence on the tourists' behavior.As Hoest and Knie- Anderson (2004) underpinned, there are many approaches of consumer satisfaction which is affected by three antecedents, i.e. the perceived quality, the perceived value and expectation (Anderson et al., 1994), while the tourists' characteristics are also significant factors of their satisfaction (Huh, 2002).Yi (1990) stated that satisfaction is a judgment a customer makes following a service encounter in which goods and/or services are exchanged.Tribe and Snaith (2008) defined tourists' satisfaction as the degree to which a tourist's assessment of the attributes of that destination exceeds their expectations for those attributes.Baker and Crompton (2000) indicated that satisfaction is a personal experience which is related to the nitration between personal expectation and actual receipt.
While there are many studies analyzing the tourists' satisfaction, little has been done regarding satisfaction in the United Arab Emirates and more specifically in Dubai.The majority of the aforementioned studies regarding tourist satisfaction investigate the relationship between factors that can play an important role in being satisfied or dissatisfied after visiting one destination.This research explores the factors that contribute to higher levels of tourist satisfaction and documents the factors that correlate positively and negatively with the level of pre-trip and post-trip satisfaction with respect to the destination of Dubai, bringing evidence for one of the most attractive destinations where government policies and marketers efforts can focus on.
The novelty of our study lies in, first, investigating an important question for tourism policy implications for Dubai -there is no prior study in this subject matter.This paper purports to evaluate the degree of Dubai's visitors' satisfaction and at the same time assess the pre-visit satisfaction of individuals, as revealed by their intention to visit Dubai.Second, we include in our analysis a variety of variables, divided in trip factors, city factors and demographic factors, which simultaneously have not been explored so far in other existing related studies.Third, with our econometric approach (ordered logit model) we were able to assess the effect of the covariates on different classes of tourist satisfaction and intention to visit (1 for not satisfied, 2 for satisfied and 3 for very satisfied; 1 for no, 2 for maybe yes and 3 for definitely yes, respectively).Our results demonstrate the important impact of several city factors for both groups of participants, while for the group of visitors, several trip factors and demographic characteristics are also significant.This would allow us to derive more detailed conclusions and propose more concrete suggestions.
The remaining of the paper is organized as follows: Section 2 presents our framework of analysis, data and the model for both groups of participants.Sections 3 and 4 present and discuss our findings, respectively.Finally, Section 5 concludes.

Methodology
This section discusses the survey data and presents the selection of the estimation method.

Data
This empirical analysis relies on web-based data obtained from a sample of 150 individuals for the year 2017, using the Convenience Sampling Technique, i.e. a non-probability sampling technique where the subjects are selected just because of their convenient accessibility and proximity to the researcher; therefore, the subjects are selected just because they are easiest to recruit for the study.More than 150 persons have participated to this research.Nevertheless, for robustness reasons, we exclude questionnaires with limited data.Therefore, our final data set consists of 120 observations.The participants were requested to answer various questions about their satisfaction with respect to their visit to Dubai.The dependent variable, tourist satisfaction, is defined as the overall satisfaction of a certain individual after visiting Dubai.A slightly moderated questionnaire was addressed to individuals who haven't visited Dubai so far, in order to assess all critical factors that may play an important role in their intention to visit Dubai or not (Plan_Visit is the dependent variable).The questionnaire was available online and the corresponding data set consists of 90 observations.Both groups of participants, those who have visited Dubai and those who did not, were asked to provide information about their demographic characteristics (demographic variables) and with respect to their travel (trip factors).Additionally, they were asked to evaluate several factors with respect to the city of Dubai (city factors).Finally, the visitors were invited to evaluate their overall satisfaction, their willingness to visit Dubai once again and their intention to recommend it as a travel destination; the non-visitors were invited to state whether they plan to visit Dubai or not.
The set of trip factors (set T) includes variables that capture when the visitors travel to Dubai The set of city factors (set C) includes variables that assess Dubai's attractiveness as a tourist destination; i.e. the city's safety, the number of attractions, the level of cost, the public transportation system, the whether its people are friendly or not, the entertainment and nightlife, and the facilities available for tourists.All variables corresponding to the city's attractiveness were measured on a five-grade scale.The set C comprises the same variables for both groups of participants.A variable corresponding to the average perception with respect to Dubai as a tourist destination was constructed based on the average grade that both groups of participants gave to the aforementioned seven questions.The latter was used interchangeably with set C and did not alter our results.
A number of demographic factors (set D) were also requested, such as Gender, Age, Income, Marital Status, Occupation, and Country.Gender takes the value of 0 for male and 1 for female; Age consists of 5 intervals and takes the value of 1 for <30, 2 for 31-40, 3 for 41-50, 4 for 51-60 and 5 for >61 years old; Income takes the value of 1 for low, 2 for medium and 3 for high; Marital Status takes the value of 0 for not married and 1 for married; Occupation represents the employment status and is 0 for unemployed and 1 otherwise; Country indicates the location of residence and is 0 for rest of the world, 1 for Europe and 2 for Middle East.The set D comprises the same variables for both groups of participants.In the initial set D several other variables were included that did not appear in our final specifications due to limited data, such as Education and Nationality.
Table 1, below, provides the definition of each variable and their corresponding coding.

Model
The likelihood of a certain tourist being satisfied after his/her visit to Dubai, can be described by an ordered logit model defined as follows: where the endogenous variable Y is the degree of tourist satisfaction and takes values from 1 to 3 (c) in accordance with the aforementioned variable (1 for not satisfied, 2 for satisfied, 3 for very satisfied); F is the standard logistic cumulative distribution function and Χ i is a set of covariates defined as:  The likelihood of an individual's intention to visit Dubai for the first time can also be described by an ordered logit model, defined as follows: where the endogenous variable Y is the individual's intention to visit Dubai and takes values from 1 to 3 (c) in accordance with the aforementioned variable (1 for no, 2 for maybe yes, 3 for definitely yes); F is the standard logistic cumulative distribution function and Χ i is a set of covariates defined as: The selection of the variables in Χ i set can be justified by relevant studies.Various demographic variables such as gender, age, income, occupation are documented in the studies of Master and Prideaux (2000), Kozak (2001), Chen and Tsai (2007), Alegre and Garau (2010) and Jarvis et al. (2016).When it comes to trip factors, such as duration of visit, the overall involvement with the trip, season, and past visits, they are explored in the studies of Alegre and Caldera (2006), Jarvis et al. (2016) and Martín-Santana et al. (2017).Finally, several city attributes, such as cost, infrastructures, safety, number of attractions, are included in a handful of studies (Yoon and Uysal, 2005;Salleh et al., 2013;Sukiman et al., 2013;Jarvis et al., 2016;Martín-Santana et al., 2017).For a comprehensive review, see Stylidis et al. (2017).

Results
Before proceeding presenting the estimates of our model, we first show in Table 2 some descriptive statistics for our two groups of participants (Group A corresponds to individuals that already visited Dubai and Group B to individuals that may plan to visit).The correlations between dependent and independent variables are below 0.5 and not statistically significant.Table 3, below, presents the correlations between all independent variables of set C. As Table 3 shows, the city's safety environment is highly and positively correlated with the number of tourist attractions.In addition, the city's famous entertainment and night life is strongly and positively related with the number of tourist attractions and the good public transportation system (.552 and .519,respectively).
The odds ratios for all specifications are presented in Table 4.One can read the odd ratios as follows: if the odd ratio, a, is bigger than one (a>1), then the probability of a tourist being satisfied from visiting Dubai, i.e.Y it =3 (maximum level of satisfaction), increases by (a-1)*100%, whereas the probability decreases by (1-a)*100%, if the odd ratio is smaller than one (a<1).As Table 4 shows, among the trip factors (T) presented in column (1), only Last_visit and Company have a statistical significant effect on the probability of being satisfied.More specifically, the more recent the period of the last visit of the participant is, the higher the likelihood of the maximum level of his/her satisfaction.The same finding emerges with respect to the Company effect, which is positively related to the tourist satisfaction.Particularly, those who chose to visit Dubai with their families or with their colleagues are about 350% and 230% more satisfied with respect to those who visited Dubai alone, respectively.

Next, column (2) includes only the city factors (C). Results demonstrate all city factors carry
the expected sign with respect to their impact on tourist satisfaction; however, only five out of seven are found to be statistically significant.If the city has a satisfactory safety level, famous entertainment and night life, and satisfactory tourist facilities, then the probability of a tourist being satisfied in increased as the odds ratios indicate.On the other hand, the number of attractions, the cost, the good transportation system and friendly people seem to have a positive effect on tourist satisfaction, but they do not have a statistical significance.
Next, in column (3), among the demographic factors (D), only Gender and Country have a statistical significant effect on the probability of being satisfied.More specifically, gender (being a woman) appears to be associated with tourist satisfaction.A positive relationship is also documented between tourists' satisfaction and their country of living.Particularly, tourists who currently live in Europe and in Middle East have about 600% and 250% probability of being satisfied by visiting Dubai, respectively, than those who currently live in the rest of the world.In contrast, getting older or being unemployed appear to be negatively associated with tourist satisfaction, but with a statistically insignificant effect.
Finally, column (4) presents the fully-fledge specification with all trip, city and demographic variables included.As before, the same variables appear to be statistically significant, maintaining the expected sign.In addition, among the trip factors, except the period of last visit and the tourists' company, duration of the visit seems to play an important and statistical significant role with respect to their satisfaction.More specifically, those who visit Dubai for more than one week have more than 300% probability of being satisfied than those who visit for less than one week.Among the city factors, except the aforementioned variables, the number of attractions and the affordable prices seem to also have an important and positive effect.Finally, among the demographic factors, age, income and marital status seem to have also a statistical significance and are associated with tourist satisfaction.More specifically, there is a positive relationship between income and tourist satisfaction, i.e. the higher the income level of a tourist, the higher his/her probability of being satisfied at the maximum level.A positive relationship, significant at a 10% level, is also documented between marital status and tourist satisfaction.
Particularly, those who are married have 125% probability of being satisfied at the maximum level with respect to singles.In contrast, getting older appears to be negatively associated with tourist satisfaction, and the probability of a tourist being satisfied decreases by 36.3% when the participant ages.In sum, estimates do not alter neither in sign, nor in statistical significance across all specifications of Group A, and remain robust.
With respect to the Group B, where only the fully-fledged model is presented, only some city factors seem to play a significant role in forming the willingness of an individual to visit Dubai.
All the variables referring to the city attractiveness are positively associated with the dependent variable and carry the same sign with respect to Group A specifications.Among the city factors, the number of city's attractions, the good public transportation system and the satisfactory tourist facilities are statistically significant at a borderline level of significance (10%), and increase the probability of an individual's planning to visit Dubai about 175%, 95%, and 135%, respectively.
Overall, our findings strongly support that the number of Dubai's attractions and its tourist facilities are the only variables among the city factors set that contribute to tourist satisfaction of an individual after his/her visit to Dubai, and at the same time they are the only important contributors in shaping an individual's willingness to travel there.For the latter, a good public transportation system also plays an important and significant role.With respect to those that already have visited Dubai once, other city factors, such as the safety of the city's environment, the cost level, and the entertainment and its night life, play also a significant and positive role in shaping their satisfaction level.Among the trip factors, the more recent period of last visit, the longer the duration of traveling with friends or colleagues, the higher the probability of being satisfied.The same holds for women, for those who belong to higher income classes and for those who currently live in Europe or Middle East and the opposite holds for the elderly.The reason of their visit and whether they are employed or not play no role at all across all specifications and seem to have no impact on tourist satisfaction.Finally, as diagnostics of bottom part of Table 4 demonstrate, all specifications have a satisfactory fitness.

Discussion
Understanding what shapes tourist satisfaction in a specific country or city is particularly important for policy makers and tourism marketers, as it provides critical information to develop targeted interventions.The same holds for the case of prospect visitors.Zhang et al. (2014) believe that it is critical to recognize how the components of image condition the future behavior of the stakeholders in order to understand the behavioral processes.Based on that, Stylidis et al.
(2017) provide destinations marketers with critical knowledge related to what drives behavioral intentions.Baloglu and McCleary (1999) state that if any potential destination is to experience success in the tourism industry, the development of a positive overall image is a prerequisite.In line with that, Dubai should formulate a positive image since our results demonstrate that the city factors corresponding to the city's attractiveness are important for both groups of participants.
Handszuh (1995) has highlighted that among several reasons that causes a high level of tourists' satisfaction, are the quality of services provided, such as the infrastructure, cleanliness, and security and safety.Salleh et al. (2013) mentioned that among the factors that make tourists to visit a place are the beautiful scenery, customs and culture, hospitality, the quality of food and the friendliness of local people, while Sukiman et al. (2013) concluded that the majority of domestic tourists are satisfied with the accessibility to destinations.Lyssiotou (2000) mentioned that tourists typically visit different destinations having a chance to enjoy different things, such as the climate, culture, wildlife, and whatever else gives them after use satisfaction.As Murphy et al. ( 2007) indicated, higher self-congruity is connected to a higher satisfaction of visitors.
Nevertheless, although the aforementioned is associated with the shopping experiences, it is clear enough that high prices can reduce trip satisfaction whereas prices in line with budget or when considered a good value for money increase satisfaction and/or even the intention to visit a destination (Brkic and Dzeko, 2008).The involvement with the trip, the time dedicated to the search for information, and the number of attractions visited influence the change in cognitive image (Martín-Santana et al., 2017).According to the same study, the image of destination may change during or after the visit based on characteristics of the trip, along with the process for secondary information.Pike (2002;2007) identified that 71% of the 262 reviewed destination image studies focus only on the cognitive destination image component by the use of an attribute list for tourism destination image assessment.The cognitive image of a destination may depend on the perception of the larger surrounding area.This depends on the knowledge the potential tourist has of the destination, i.e. the destination image of Zagreb is depending on the overall destination image of the country Croatia (Gartner, 1993).In addition, for the case of Eilat, Israel, the indifferent city image residents have may threat its success as a tourist destination (DiPietro et al., 2007).Accordingly to the latter, the image of for example  (Huntington, 1996;Reisinger, 2009).
Balakrishnan (2016) argues that the UAE needs to be perceived as Emirati as opposed to Arab or Middle Eastern on a global market because of these negative perceptions.In addition, high crime levels, natural disasters or any kind of danger decrease satisfaction (Jarvis et al., 2016).
Given the significance of the overall image in influencing future behavioral intentions as well as the visitors' satisfaction, marketing strategies must be developed to promote that specific component of the destination image (Stylidis et al., 2017).In addition, the overall trip satisfaction is the most important factor which influences tourists to repeat their trip (Alegre and Cladera, 2009).Chi (2012) documented that repeat visitors can be considered as a stable market for a destination and in the form of word of mouth recommendations they can provide free advertisement.According to Assaker and Hallak (2012) and Baker and Crompton (2000), repeat tourists can reduce marketing costs and price sensitivity amongst consumers and, also, increase economic profit (Choo & Petrick, 2014).Although the first time visitors are less likely to return than the repeat ones, they are not willing to repeat their trip if their overall satisfaction level is low (Alegre and Cladera, 2006).Foster (2002) also suggested that the on-going systematic measurement of satisfaction with destinations is a valuable exercise with tangible benefits.
Although behavioral intentions are positively affected by destination image and satisfaction (Chen and Tsai, 2007), several demographic and cultural factors play an important role in travel decision-making (Alegre and Pou, 2002).Furthermore, the likelihood of a tourist to return to a specific location depends on a range of factors, such as age and income.For example, Master and Prideaux (2000) analyzed the influence of several demographic and travel characteristics on different levels of satisfaction.In contrast, Alananzeh et al. ( 2018) document a significant impact on tourist attractions and facilities, and front office services on tourist satisfaction, while they do not prove a significant difference in the impact of satisfaction in favor of gender, age, occupation, educational level and purpose of visit.
Several studies demonstrated so far that there is a strong and positive association between city attributes and tourist satisfaction, while specific factors with respect to the trip itself and the tourists also play a significant role.For the individuals that have never visited a destination, the same holds when expressing their intentions to visit a destination or not.To our knowledge, this is the first attempt in the literature that all aforementioned factors are included in a specification for both groups of participants; therefore we are not able to perform comparisons with existed related studies.Overall, a key factor of tourist satisfaction as well as of behavioral intention is the city attributes, while specific characteristics with respect to the trip or the demographic characteristics seem to play an important and significant role only for the case of visitors.In most studies the majority of the respondents are males, married, their average age is between 35 and 45 years old, they are mostly Europeans and they travel to the end destination for the purpose of vacations, similar findings with those of the current study (Shahrivar, 2012;Forozia et al., 2013;Mohammed et al., 2014;Cong and Dam, 2017).
Furthermore, it is worth to evaluate the corresponding questionnaire as there was no methodology to base upon and, in addition, there were missing data in the initial data set; so possible errors or changes with respect to the way data were collected could influence the results.
Nevertheless, for robustness purposes, we try with alternative indices1 with respect to the proposed one and results do not change significantly.Moreover, results remain robust when the average city factor was used interchangeably to city factors set.Finally, there might be several confounding factors that have contributed to these findings.Therefore, further research could focus on a country-level analysis, taking into consideration omitted factors such as individual motivation, attitudes and emotional factors that are not taken into account, and evaluate the use of the questionnaire as well as the possible ceiling effect.Future research should use a multi-item questionnaire in order to explore in-depth the overall destination image of (potential) tourists.In addition, it would be interesting to compare the destination image that visitors have before travel and whether their expectations are fulfilled right after the trip.The same analysis could be performed in repeat tourists as well.

Conclusions
Tourist satisfaction is affected by economic, social and environmental factors which, in turn, is found to affect the likelihood of a tourist returning, creating loyalty in one specific destination (Alexandris et al., 2006).Thus a successful tourism industry except from attracting new tourists, also needs to encourage repeat visits (Jarvis et al., 2016).
The destination image is the most important factor that can influence the intention of an individual to visit one destination and at the same time the satisfaction level after the trip.Given the importance of the destination image, it can be used by destination marketers as a framework for the design of marketing campaigns aiming to enhance the image and word of mouth recommendations of this stakeholder group.
Although it is a general belief that the overall experience at the destination is what causes a greater positive change in the destination image (Smith et al., 2015) and therefore higher levels of satisfaction, the true problem lies on communicating the considerable improvements to the wide public.Internal campaigns and educational programs should be developed and promoted respectively, targeting residents and general public.

(
Last Visit), the length of stay (Duration), the season they chose to travel (Season), the whom they have traveled with and the reason why (Company and Reason, respectively).Last Visit consists of 4 intervals and takes the value of 1 for before 2007, 2 for 2008-2010, 3 for 2011-2013 and 4 for after 2014; Duration is a dummy variable that takes the value 0 if the tourist visited Dubai for less than 1 week, otherwise is 1; Company takes the value of 1 if someone visited Dubai alone and the values 2, 3 and 4 if someone visited Dubai with family, friends and colleagues, respectively; Season takes the value 1 for summer, 2 for spring, 3 for winter and 4 for autumn; Reason takes the value of 1 if someone visited Dubai for any other reason expect of business (value 2) and of vacations (value 3).The set T is slightly different for the non-visitors group, since the variables Last Visit and Company are not included.
where the first five variables represent the trip factors set of variables (set T), the following seven variables represent the city factors set (C) and the remaining six variable represent the demographic factors set (D).Details about the variables' classification are given in the preceding section.
β 16 Country i where the first three variables represent the trip factors set of variables (set T), the following seven variables represent the city factors set (C) and the remaining six variable represent the demographic factors set (D).Details about the variables' classification are given in the preceding section.
Group A refers to the visitors.More specifically, column (1) presents estimates of the model for those who have visited Dubai where only the trip factors (T) are included.Next, column (2) shows estimates of the model, where only the indicators regarding the city's attractiveness (C) are included.Next, column (3) presents estimates of the model, where only the demographic factors (D) are included.Finally, column (4) presents estimates, where the full set of covariates (X) is included.Group B presents the corresponding estimates, where the full set of covariates (X) is included, for individuals who may plan to visit Dubai.

Table 2 : Summary statistics of all variables
Table2shows, the majority of Group A participants are satisfied from their visit to Dubai.They chose to visit Dubai after 2014, during winter, mostly for vacations with their families and they stay less than one week.Further, half of the participants are men, married and belong to middle income class.Finally, the majority of them are Europeans, employed, and between the age of 31 and 50 years old.With respect to Group B, the majority of the participants who are thinking to visit Dubai are also Europeans, under the age of 40 years old, employed, and belong to middle income class.Furthermore, they are thinking to travel for vacations, during winter or spring, and for less than one week.

Table 3 : Correlations between all city factors
* Significance at 5% level of significance.
Dubai, Abu Dhabi, Qatar, Oman or Bahrain in part depends on the overall image of the Middle East region.According to Trimeche et al. (2012) the Middle East is connected with fundamentalism, disrespect of human rights and terrorism.Here a lack of knowledge and cultural differences can lead to stereotyping of Middle-Eastern people amongst Westerners