Examining Perceived and Projected Destination Image: A Social Media Content Analysis

: Destination image has been extensively studied in tourism and marketing, but the questions surrounding the discrepancy between the projected (perceptions from the National Tourism Organizations) and perceived destination image (perceptions from tourists) as well as how the discrepancy may inﬂuence sustainable experience remain unclear. Poor understanding of the discrepancy may cause tourist confusion and misuse of resources. The aim of this study is to empirically investigate if the perceived (by tourists) and projected (by NTOs) destination image are signiﬁcantly different in both cognitive and affective aspects. Through a comprehensive social media content analysis of the NTO-generated and tourist-generated-contents (TGC), the current study identiﬁes numerous gaps between the projected and perceived destination image, which offers some important theoretical and practical implications on destination management and marketing.


Introduction
Nowadays, information technology has been integrated into each field of tourism management. For example, new technologies, such as face recognition, social applications, digital payments, and service robots, have been widely in tourism services. For tourism organizations, information technology has changed from a traditional tourism marketing tool to a tool of knowledge creation and innovation [1]. Especially with the rapid development of mobile and internet technology, the open data and shared social knowledge provided by social media have not only laid a foundation for tourism innovations, but also influenced the communication between tourism organizations and tourists. Tourists can now create and share their travel experiences on social media platforms (e.g., Facebook, Flicker, Microblogs, and Wechat), and subsequently, tourism destination image has been constantly co-created through both the traveler-generated content (TGC) and contents created by National Tourism Organizations (NTOs) and Destination Marketing Organizations (DMOs) on social media [2].
The concept of destination image was first proposed by Crompton (1979) [3], which has at present become one of the most discussed topics in the field of tourism and marketing [4]. Destination image refers to the sum of beliefs, ideas, and impressions that a person has of a destination [3]. Most prior studies have shown that the destination image has an important influence on tourists' decision-making, satisfaction, word-of-mouth, willingness of recommendation, and willingness of revisiting [5,6]. The findings of prior destination image studies can provide important decision-making references for NTOs and DMOs to improve their destination marketing strategies [7]. From the supply-demand perspective, Kotler et al. (1993) [8] proposed two categories of destination image: Projected and perceived destination image. Projected destination image refers to the image that a destination intends to create in the mind of tourists through tourism marketing activities [9]. Perceived destination image, however, refers to the understanding and impression of the destination formed by tourists [10]. Achieving a good level of congruency between the projected and perceived destination image has become a key objective for NTOs and DMOs. This congruency can help national tourism organizations assess if their projected destination image has been accepted by their targeted visitors [11]. In other words, reducing any incongruency between the projected and perceived destination image will not only help retain any established tourism market [12], but also attract or develop new tourist markets [13].
Despite the importance of this issue [14], most destination image studies have examined only projected destination image [15], while few have examined projected destination image [16], with even fewer that have looked at the discrepancy of these two types of destination image [17][18][19]. Having said this, a handful of studies have attempted to compare perceived online destination image with projected offline destination image [20], whilst no reported study has directly compared the projected and perceived destination image using social media content analysis.
As one of the most famous tourism destinations in the world, Australia attracts a significant number of international tourists. Australia has rich natural tourism resources, known as the "tourism paradise in Oceania", with great potential for sustainable development. Tourism has become one of Australian pillar industries. Australian NTOs works with partners, such as online travel platforms, airlines, traditional media, opinion leaders, and social media, to promote the destination image of Australia. According to the Australian Bureau of Statistics (ABS), China has become Australia's largest tourist market in 2018. Australian NTOs has actively used their official Chinese website and social media (e.g., microblog, WeChat, Xiaohongshu, Tik Tok) to strengthen their online interactions and communications with Chinese tourists. However, are these marketing strategies successful? Is there any discrepancy between the destination image that Australia NGOs attempt to project and that perceived by international tourists, such as the Chinese tourists?
The current study, therefore, will attempt to fulfil the following objectives: (1) To examine the cognitive and affective attributes of both perceived and projected image based on TGC and NTO-generated contents; (2) to examine the differences between the perceived and projected destination image from the cognitive and affective aspects based on the TGCs and NTO-generated contents; and (3) to discuss the potential impacts these differences may have on sustainable tourism development. In the current context of the global COVID-19 pandemic, it is of even greater importance that these discrepancies be well understood and incorporated into Australia's marketing and promotion strategies in order to achieve a sustainable tourism development post the pandemic.

Dimensions of Destination Image
Destination image is believed to be a multi-dimensional construct. Baloglu and Brinberg (1997) [21] proposed two important dimensions: Cognitive and affective. The cognitive destination image refers to the rational evaluation of the characteristics or attributes of the destination held by tourists [22], while the affective destination image refers to the affective response of tourists to various attributes and characteristics of the destination [23]. Cognitive and affective dimensions may be also interrelated [21]. For example, some studies found cognitive destination image is an antecedent of the affective destination image [24]. Some early studies focused on the cognitive dimension (e.g., Golledge, 1987 [25]), while others focused on the affective dimension (e.g., Hanyu, 1993 [26]). More and more recent studies have covered both the cognitive and affective dimension [27][28][29][30][31]. In addition, Woosnam et al. (2020) empirically tested the relationship between the cognitive, affective, and conative dimension [32]. However, as noted earlier, no empirical study has ever examined both dimensions when comparing the projected and perceived destination. As such, it is hoped that the current study can offer some important theoretical and practical implications.
Furthermore, there may be some sub-dimensions underneath the cognitive and affective dimension [33]. Take cognitive as an example, Chen et al. (2002) claimed the cognitive dimension included two sub-dimensions, namely tourism activity and tourist attractions [34], while Bonn et al. (2005) proposed three cognitive sub-dimensions: Environment, atmosphere, and service [35]. Over the years, numerous studies (e.g., Beerli and Martin, 2004;Choi et al., 2007;Chen et al., 2008;Yang et al., 2012;Kim and Lehto, 2013;Ji and Wall 2015;Önder and Marchiori, 2017) have proposed different sub-dimensions (or components) of either the cognitive or the affective dimension; however, most scholars in general agree that the number of sub-dimensions underneath cognitive or affective dimensions is contingent on the context (e.g., which destination and which market) [10,11,17,[36][37][38][39]. While acknowledging these prior studies, the current study attempts to further explore the sub-dimensions of cognitive and affective destination image based on both TGC and NTO-generated contents text.

Comparation between the Perceived and Projected Destination Image
As discussed earlier, prior studies of destination image have mainly focused on the perceived destination image, in other words, from the tourist perspectives [15]. However, a good understanding of the projected destination image created by NTOs is also vitally important for the tourism market [10,16], because any potential discrepancy (or gap) between the projected and perceived destination image could potentially lead to significant resource waste. This is why assessing congruence between the projected and perceived destination images becomes an urgent task for NTOs worldwide [40].
Despite of the importance, studies (particularly empirical) on projected and perceived destination images are scarce, not mentioning any that has utilized social media. In recent years, a handful of relevant studies (e.g., Marine-Roig and Clavé, 2016; Meneghello and Montaguti, 2016) have primarily been based on quantitative analyses of data obtained through visitor surveys or through official tourism websites [41,42]. The current study, however, will be the first attempt that adopts social media data (i.e., perceived destination image from TGC, such as travel blogs and online reviews, and projected destination image from NTO-generated data, such as microblogs).

Destination Image, Sustainable Experience, and Sustainable Tourism
Being sustainable is one of the key goals for tourism [43]. Many nations, such as Australia, have focused on sustainable tourism, which refer to a "tourism that takes full account of its current and future economic, social and environmental impacts, addressing the needs of visitors, the industry, the environment, and host communities" [44]. This "sustainable" focus would have been emphasized by their NTOs through communicating with their key target markets (i.e., through projected destination image); however, is the focus of sustainable tourism been really understood by international tourists? The question is probably yet to be answered. Sustainable tourism may be reflected from multiple aspects and one of them is sustainable experience, which is a means of enhancing destination sustainability [45].
Events, Tourism, and Hospitality. Routledge. Sustainable experience is a multidimensional construct and interacting with the natural and cultural environment of a destination are two key dimensions [46]. Destination image and experience are closely linked [47]. For example, for those who have visited a destination before, the perceived destination image is largely built on tourists' actual experiences. As to tourists who have not been to the destination, the perceived destination image may also be influenced by other tourists' experiences (e.g., through sharing via social media). In other words, if sustainable experience could be shown from both perceived and projected destination image, there might be a good level of congruency between NGOs and international tourists, which might subsequently post a positive influence on the destination's sustainable tourism development.

Destination Selection
Australia is selected as the tourism destination examined in the current study. Australia lies between the South Pacific and the Indian Ocean, and its territory consists of the Australian mainland, Tasmania and other nearby islands. It is the only country that occupies an entire continent in the world and has the longest coastline in the world. With its rich tourism resources, unique natural environment and well-developed infrastructure, Australia has become one of the most famous nature-based tourism destinations in the world and attracts a significant number of international tourists [48]. From June 2018 to June 2019, Australian tourism revenue reached A $60.8 billion, accounted for 3.1% of Australian GDP [49].
With the development of globalization and increasing disposable income, China has become an important source of tourists in the international tourism market, and the number of Chinese outbound tourists is continuously on the rise [49]. As such, there is an urgent need for NTOs (or DMOs) and other tourism marketers to better understand the Chinese outbound market [50]. Despite the importance, there is still insufficient knowledge on tourists' preferences, interests, and destination choice behaviors [51][52][53].
China is Australia's largest international tourist market. According to Australian Bureau of Statistics (ABS), China has surpassed New Zealand and became the largest tourists' market of Australia in 2018. Up to August 2019, the number of Chinese tourists to Australia reached 1.5 million. The number of Chinese tourists to Australia has grown rapidly from 358,000 in 2008 to 1.445 million in 2018 (see Figure 1). Chinese tourists to Australia have increased by more than a million since 2009, with an annual growth rate of 15 percent. China is also the top spending source of overseas visitors to Australia. By the first quarter of 2018, the annual spending of Chinese tourists in Australia reached around 11.5 billion Australian dollars, accounting for 27% of the total expenditure of overseas tourists in Australia [50]. destination image is largely built on tourists' actual experiences. As to tourists who have not been to the destination, the perceived destination image may also be influenced by other tourists' experiences (e.g., through sharing via social media). In other words, if sustainable experience could be shown from both perceived and projected destination image, there might be a good level of congruency between NGOs and international tourists, which might subsequently post a positive influence on the destination's sustainable tourism development.

Destination Selection
Australia is selected as the tourism destination examined in the current study. Australia lies between the South Pacific and the Indian Ocean, and its territory consists of the Australian mainland, Tasmania and other nearby islands. It is the only country that occupies an entire continent in the world and has the longest coastline in the world. With its rich tourism resources, unique natural environment and well-developed infrastructure, Australia has become one of the most famous nature-based tourism destinations in the world and attracts a significant number of international tourists [48]. From June 2018 to June 2019, Australian tourism revenue reached A $60.8 billion, accounted for 3.1% of Australian GDP [49].
With the development of globalization and increasing disposable income, China has become an important source of tourists in the international tourism market, and the number of Chinese outbound tourists is continuously on the rise [49]. As such, there is an urgent need for NTOs (or DMOs) and other tourism marketers to better understand the Chinese outbound market [50]. Despite the importance, there is still insufficient knowledge on tourists' preferences, interests, and destination choice behaviors [51][52][53].
China is Australia's largest international tourist market. According to Australian Bureau of Statistics (ABS), China has surpassed New Zealand and became the largest tourists' market of Australia in 2018. Up to August 2019, the number of Chinese tourists to Australia reached 1.5 million. The number of Chinese tourists to Australia has grown rapidly from 358,000 in 2008 to 1.445 million in 2018 (see Figure 1). Chinese tourists to Australia have increased by more than a million since 2009, with an annual growth rate of 15 percent. China is also the top spending source of overseas visitors to Australia. By the first quarter of 2018, the annual spending of Chinese tourists in Australia reached around 11.5 billion Australian dollars, accounting for 27% of the total expenditure of overseas tourists in Australia [50]. As noted in earlier sections, Australian NTOs has been actively promoting the destination image of Australia through a variety of channels and platforms in China (e.g., online travel platforms, airlines, outdoor advertising, and opinion leaders), which directly As noted in earlier sections, Australian NTOs has been actively promoting the destination image of Australia through a variety of channels and platforms in China (e.g., online travel platforms, airlines, outdoor advertising, and opinion leaders), which directly communicate with Chinese tourists. For example, Australian NTOs has opened the Chinese official website (www.australia.cn (accessed on 12 February 2021)), the Chinese official accounts on microblog, WeChat, Xiaohongshu, Tik Tok, and other social media platforms to introduce Australian tourism experience in the form of text, pictures, and videos. Australian NTOs have also used social media to release online interactive contents and held livestreaming

Content Analysis
Content analysis is a research method appropriate for any objective analysis of explicit content [54]. This method can provide an in-depth knowledge and understanding of the content of communication [55]. Most existing studies on destination image have used questionnaire surveys, which is prone to bias [51] and cannot fully reflect destination image and its unique attributes. Comparatively, the textual content analysis method can avoid the subjective influence of the investigators as much as possible [28], and the types and numbers of the groups involved are more comprehensive, which can more effectively reflect the social destination image and the unique attributes perceived by tourists. Accordingly, this method has been applied separately in perceived destination image studies [56][57][58] and projected destination image studies [59]. The datasets in these prior studies were mostly from single online travel blogs (or reviews) texts or official tourism websites texts (e.g., Choi et al.,2007;Mak, Athena, 2017), while ignoring the texts from social media of the official tourism organizations (NTOs and DMOs). It is also worthy to note that there are relatively fewer studies that examined destination image through content analysis and from a comparative perspective [60].
Different from prior research, the current study attempts to collect the textual data on social media from TGCs as well as NTO-generated contents related to Australia and compare any differences between the perceived and projected destination image with the aid of the content analysis. As a software of content analysis, CATPA has been widely adopted in destination image studies (e. g., Mak, Athena, 2017 [2]), but it is only used to analyze English textual data. Instead, ROST CM 6 software is a free social computing software used for content mining, text analysis, and knowledge processing in order to assist humanistic and social science research in China. This software is often used to analyze Chinese textual data in destination image studies. Hence, ROST CMS6 software was adopted by the current study to analyze both the TGC and the NTO-generated textual data.

Data Collection
With the fast development of mobile Internet, social media has become more favored by tourists because of its interactivity and accessibility. Sina microblog is not only one of the most popular social media platforms in China but also an important platform for foreign NTOs and DMOs to directly interact and communicate with Chinese tourists. The projected destination image portrayed by NTOs and the perceived destination image formed by Chinese tourists have been compared with the research data in this study. The authors select tourists' online travel blogs and comment texts as well as NTOs' microblogs texts as research data. The time span of both texts was from 1 January 2015 to 31 December 2019. The data collection was divided into four steps as follow.
First, to develop a dataset for perceived destination image, the top three Chinese tourism websites, including Ctrip.com, Mafengwo.cn, and Qunar.com, were selected according to the Chinese website traffic ranking provided by Alexa. In this study, "Australia" was used as the key word to search on these tourism websites. A total of 600 travel blogs, which were ranked according the "Hot Travel Blogs", were obtained by using the web crawler software to capture the top 200 travel blogs from each website. At the same time, "Australia" was used as the key word to search for tourism reviews, sorted by "useful number", and a total of 3600 reviews were obtained by using the web crawler software to capture the top 1200 reviews of each website. On this basis, travel blogs and reviews that did not relate to Australia as the only destination or not related to Australia tourism were deleted. Further, travel blogs with less than 200 words and reviews with less than 10 words were deleted. After data cleaning, 418 travel blogs and 2983 reviews were obtained, with a total word count of more than 1.2 million words, which formed the original data of tourists' texts.
Second, to develop a dataset for projected destination image, through the web crawler software tool, this study collected 3872 microblogs from the official microblog of "Tourism Australia", the National Tourism Organization of Australia. Up to December 2019, the "Tourism Australia" microblog account had 1.39 million followers, and it is one of the most direct platforms for information exchanging between Tourism Australia and Chinese tourists. Through data cleaning, contents that were unrelated to Australia as a destination or irrelevant to travel or tourism were deleted. Further, repeated contents were also deleted for clarity purposes. The final dataset of projected destination image (NTO-generated) included 3220 microblogs with a total of 300,000 words.
Third, the TGC and NTO generated data were further cleaned. For example, words, emoticons, links, and other information that were irrelevant to the research topic were removed. Words like "travel" and "tourism" were unified into one form of expression in the two textual data. The textual data processed in accordance with the above procedures were saved as two .txt files.
Last, with the aid of the "ROST CM6" software, categorization of 291 frequently appearing words from the TGC generated data and 286 frequently appearing words from the NTO-generated content data were segmented based on a list of synonyms created by HowNet dictionary. The HowNet Dictionary is developed by CNKI (China National Knowledge Infrastructure), which is one of the most comprehensive Chinese language platforms with a total of 91,016 Chinese words (i.e., including degree level words, negative evaluation words, positive evaluation words, negative affection words, positive affection words, assertion words, and so on).

Semantic Network Analysis
In the current study, the "social network and semantic network analysis" function of ROST CM6 software was used to analyze TGC and NTO textual data, so as to visually reveal the perceived and the projected cognitive destination image of Australia. Semantic network analysis, an effective method for text analysis, is often adopted to construct a network map reflecting the psychological cognition of text publishers through the connections between words, which can reveal the semantic core and internal structure of the text [61]. The nodes in the semantic network diagram represent frequently appearing keywords, and the lines between nodes represent the connections between frequently appearing keywords.
The semantic network diagram of TGC textual data is shown in Figure 2. First, "Australia" is the first-level central node in the semantic network diagram. By using "Australia" as the central node of the cluster, tourists to Australian tourist attractions (e.g., beach, seaside, opera house), tourist and general infrastructure, such as restaurant, hotel, airport, and properties were more prominent than other properties such as history, culture and art, tourist leisure, and recreations (e.g., tourism, visit). Second, nodes such as "Sydney", "airport", and "hotel" are the secondary central nods of the semantic network diagram, and three relatively obvious clusters are formed. This finding indicates that "Sydney" is the most important Australian tourism hub city and that tourists pay special attention to the transport, accommodation, and other tourist infrastructure in Australia. Third, nodes such as "scenic area", "Melbourne", "urban area", and "free" are the threelevel central nodes of the semantic network diagram. This further shows that tourists pay more attention to Australian tourist attractions and location-related environment. The results also reflect that Australian "free" (or promotional) tourism events have left a deep impression on Chinese tourists. level central nodes of the semantic network diagram. This further shows that tourists pay more attention to Australian tourist attractions and location-related environment. The results also reflect that Australian "free" (or promotional) tourism events have left a deep impression on Chinese tourists. The semantic network diagram of NTOs' microblog textual data show three obvious clusters related to the cognitive aspect of the projected destination image (See Figure 3). The first cluster takes "Australia" as the core node, and its related nodes including tourist attractions (e.g., beach, beautiful scenery, koala), tourism environment (e.g., Sydney, Western Australia, Queensland), history, culture and art, tourist leisure and recreation (e.g., guideline, explore), and tourist services and facilities (e.g., restaurants, wine, food). The second cluster takes "tourist office" as the core node, and its related nodes mainly include natural resources, location-related environment, history, culture and art, and catering services (e.g., restaurant, delicious food and wine). The third cluster takes "experience" as the core node, and its related nods mainly include location-related environment (e.g., South Australia, Queensland, Northern Territory), history, culture and art, and catering services (e.g., delicious food, wine, restaurants). This result reveals that NTOs pay more close attention to the dissemination of tourism experience in the aspects of Australian history, culture, and art, as well as catering services. In summary, the cognitive destination image perceived by Chinese tourists can be described as follows: Australia is rich in natural resources and humanistic resources, but tourists have not been deeply impressed by Australian history, culture, and art; It seems Chinese tourists care more about transportation and accommodations than catering. In addition, Chinese tourist have been impressed by the free travel programs in Australia. The cognitive destination image projected by Australian NTOs can be described as: Australia has rich natural resources, superior geographical location, rich history, culture and art, and delicious food, which NGOs hope to provide tourists with a high-quality The semantic network diagram of NTOs' microblog textual data show three obvious clusters related to the cognitive aspect of the projected destination image (See Figure 3). The first cluster takes "Australia" as the core node, and its related nodes including tourist attractions (e.g., beach, beautiful scenery, koala), tourism environment (e.g., Sydney, Western Australia, Queensland), history, culture and art, tourist leisure and recreation (e.g., guideline, explore), and tourist services and facilities (e.g., restaurants, wine, food). The second cluster takes "tourist office" as the core node, and its related nodes mainly include natural resources, location-related environment, history, culture and art, and catering services (e.g., restaurant, delicious food and wine). The third cluster takes "experience" as the core node, and its related nods mainly include location-related environment (e.g., South Australia, Queensland, Northern Territory), history, culture and art, and catering services (e.g., delicious food, wine, restaurants). This result reveals that NTOs pay more close attention to the dissemination of tourism experience in the aspects of Australian history, culture, and art, as well as catering services. level central nodes of the semantic network diagram. This further shows that tourists pay more attention to Australian tourist attractions and location-related environment. The results also reflect that Australian "free" (or promotional) tourism events have left a deep impression on Chinese tourists. The semantic network diagram of NTOs' microblog textual data show three obvious clusters related to the cognitive aspect of the projected destination image (See Figure 3). The first cluster takes "Australia" as the core node, and its related nodes including tourist attractions (e.g., beach, beautiful scenery, koala), tourism environment (e.g., Sydney, Western Australia, Queensland), history, culture and art, tourist leisure and recreation (e.g., guideline, explore), and tourist services and facilities (e.g., restaurants, wine, food). The second cluster takes "tourist office" as the core node, and its related nodes mainly include natural resources, location-related environment, history, culture and art, and catering services (e.g., restaurant, delicious food and wine). The third cluster takes "experience" as the core node, and its related nods mainly include location-related environment (e.g., South Australia, Queensland, Northern Territory), history, culture and art, and catering services (e.g., delicious food, wine, restaurants). This result reveals that NTOs pay more close attention to the dissemination of tourism experience in the aspects of Australian history, culture, and art, as well as catering services. In summary, the cognitive destination image perceived by Chinese tourists can be described as follows: Australia is rich in natural resources and humanistic resources, but tourists have not been deeply impressed by Australian history, culture, and art; It seems Chinese tourists care more about transportation and accommodations than catering. In addition, Chinese tourist have been impressed by the free travel programs in Australia. The cognitive destination image projected by Australian NTOs can be described as: Australia has rich natural resources, superior geographical location, rich history, culture and art, and delicious food, which NGOs hope to provide tourists with a high-quality In summary, the cognitive destination image perceived by Chinese tourists can be described as follows: Australia is rich in natural resources and humanistic resources, but tourists have not been deeply impressed by Australian history, culture, and art; It seems Chinese tourists care more about transportation and accommodations than catering. In addition, Chinese tourist have been impressed by the free travel programs in Australia. The cognitive destination image projected by Australian NTOs can be described as: Australia has rich natural resources, superior geographical location, rich history, culture and art, and delicious food, which NGOs hope to provide tourists with a high-quality travel experience. Results showed some similarities between perceived and projected destination image from the cognitive side, but identified one prominent difference between them, which is experience. NGOs seem to emphasize their destination image on "experience", but tourists do not seem to have that word in mind when they think about Australia as a tourism destination.

Noun Frequency Examination
The frequently appearing words of textual data can reflect the importance of certain words in the entire text [62]. Therefore, the frequently appearing nouns in the textual data can reflect the cognitive destination image of the text publisher. The frequently appearing nouns have been sorted, and 24 place names such as "Australia", "Sydney", "Melbourne", and "Western Australia" were deleted from the NTO-generated content textual data, and 24 place names such as "Sydney", "Melbourne", "Brisbane", "Cairns", "Adelaide", and "Perth" were deleted from the TGC textual data. At the end, this study selected the top 90 frequently appearing nouns in each textual data as analysis data.
These frequently appearing nouns are classified according to cognitive attributes. In order to increase credibility, three research assistants independently classified and encoded the frequently appearing nouns in the two textual data [40], and the level of agreement of the final list was 93 percent (intercoder reliability = 0.93). The results of the coding showed that the cognitive attributes were divided into five dimensions: Tourist attractions, tourism environment, history, culture and art, tourist leisure and recreations, and tourist and general infrastructure. The summarized results are shown in Table 1. The cognitive attributes projected by NTOs and perceived by tourists are largely consistent, covering five sub-dimensions (See Table 1). Some percentage differences were detected. For example, tourists' cognition of Australian tourism is mainly focused on tourist attractions (47.68%), tourist and general infrastructure (25.29%), and tourist leisure and recreations (13.07), while NTOs mainly focus on tourist attractions (38.52%), tourist and general infrastructure (30.48%), tourist leisure and recreations (11.53), and tourism environment (11.48%).
Further differences between NTOs and tourists were detected in the attribution side (i.e., how much it contributes to destination image). The attribution of Australian tourist attractions to perceived destination image (47.68%) is higher than that to the projected destination image attribution (38.52%). The attribution of natural resources to the perceived destination image (27.20%) is lower than that to the projected destination image (31.96%), while the attribution of Humanistic resources to the perceived destination image (20.48%) is much higher than that to the projected destination image (6.56%). These results show that Australian rich tourist attractions do leave a deep impression on Chinese tourists, but the dissemination of Australian Humanistic resources needs to be strengthened.
The attributions of the projected destination image (i.e., the three sub-dimensions of tourism environment, history, culture and art, and tourist and general infrastructure) were all higher than those of the perceived destination image attribution. This result reveals that NTOs should adjust their communications related to these three sub-dimensions in order to improve the perception of tourists in these dimensions. The possible explanations for the differences between perceived destination image and projected destination image on historical, cultural, and artistic attributes may be as follows: Firstly, Australia's rich natural tourism resources are still the main attraction to Chinese tourists [49]. As such, Chinese tourists are still keen on nature-based tourism resources but have low expectations for experiences of history, culture, and art (such as museums, art galleries, opera performances, parties, and aboriginal culture). Secondly, Chinese tourists to Australia often travel in groups through outbound travel agencies who may tend to favor low budgeted travel routes in order to save costs. Activities associated with Australian history, culture, and art may be more expensive than those associated with nature-based resources.
The attribution of tourist and general infrastructure in the perceived destination image (25.29%) is lower than that to the projected destination image attribution (30.48%). In particular, the catering dimension accounts for a much higher percentage in the projected destination image (24.28%) than in the projected destination image attribution (6.71%), while dimensions of accommodation and transportation account for a smaller share in the projected destination image than in the perceived destination image attribution. This shows that the NTOs may overemphasize the attractiveness of Australian "catering" to tourists, but tourists actually care more about "transportation" (9.83%) and "accommodation" (8.75%).

Adjective Frequency Examination
Unlike the cognitive destination, sematic network analysis diagram could not be produced for affective destination image, because the total number of adjectives were far less than that of nouns contained in the two textual data. In other words, the semantic network diagram can only reveal the semantic connection between the nouns in the two textual datasets (See Figures 2 and 3) but cannot show the connection between the adjectives in the two textual data.
In order to capture the affective attributes of the projected and perceived destination image, this study adopted the circumplex model of affect [63] as the category of content analysis for identifying the attributes of the affective destination image. Based on four pairs of affections: "arousing-sleeping", "exciting-gloomy", "pleasant-unpleasant", "relaxingdistressing", this study classified the first 50 frequently appearing adjectives of the two texts' frequently appearing vocabularies. Three research assistants independently checked through all the categorization and coding processes [40], and the level of agreement of the final list was 96% (intercoder reliability = 0.96). The results are shown in Table 2.
"Pleasant" and "exciting" are the two main affections of NTOs and tourists-generated texts (See Table 2). It showed both texts towards Australian tourism were mainly focused on positive affections. However, the percentage (44.23%) of the "exciting" attribute in the TGC's data were higher than that (33.22%) of the NGO's data. This indicates that there may be a gap between the affective dimension of the perceived and projected image; in other words, tourists strongly associated Australia with excitement; however, NGOs have not placed excitement into a prominent place in their communications. Therefore, NTOs will need to strengthen the "exciting" sub-dimension in the affective destination image. In contrary, the projected image seems to have a higher percentage (43.11%) of the "pleasant" attribute as compared to that in the perceived image (39.23%). The "relaxing" attribute in the projected image (13.61%) was slightly higher than that in the perceived image (10.08%). The "relaxing" attribute for both tourists and NTOs were the third most important positive affective just behind "pleasant" and "exciting". This showed that the focus of "Leisure Travel in Australia" projected by Australia's NTOs has been indeed received or accepted by tourists. It should also be noted that the proportions of "distressing" and "unpleasant" in the projected destination image was higher than those of the perceived destination image.
In conclusion, the affective dimension in both perceived image and projected image were dominated by positive affections (98.1% and 95.3%, respectively), while the proportions of negative emotions in either projected or perceived image was very low (1.9% and 4.7%, respectively).

Affection Tendency Examination
Through the semantic analysis of the text contents, the affection tendency of TGC or NTOs can be judged. This study uses ROST CM6 software to analyze the affection tendency of two textual datasets from the three dimensions of positive, neutral, and negative affection. The results are shown in Table 3. Table 3. Comparing the affection tendency between perceived and projected. Results showed that positive affections were the main affection tendencies of NTOs and TGC towards Australia as a tourism destination (See Table 2). The percentage (84.77%) of positive affections perceived by tourists is very similar to the percentage (83.91%) of positive affections projected by NTOs. This finding suggests that Australia, as a pleasant tourism destination, is largely congruent between its perceived and projected destination image. The proportion (4.94%) of neutral affections perceived by tourists was slightly higher than the proportion (2.7%) of neutral affections projected by NTOs. Both datasets detected some minor negative affection. The percentage (10.29%) of the negative affection perceived by tourists appeared to be slightly lower than the percentage (11.58%) of the negative affection projected by NTOs. With a closer look at the TGC text and the NTO text, it can be revealed that the negative affections perceived by tourists mainly stemmed from tourists' complaints about airports and bad weather as well as some sad descriptions of humanistic landscapes, while the negative affections projected by NTOs mainly came from the description of the bleak and beautiful scenery, and the troubles and pressures of modern life. The current dataset cannot show the exact causes of these negative affection, which is certainly an interesting area to explore in future research.

Conclusions
The present study, through a social media analysis, examines Australia's projected and perceived destination images in the Chinese market. It contributes the understanding of destination image in the following aspects.
Through analyzing both the tourist and NTO's text data, the study identifies five major categories (and 12 sub-categories) under the cognitive destination image dimension: (a) Tourist attractions (natural resources, humanistic resources); (b) tourism environment (i.e., natural environment, social environment, location-related environment); (c) history, culture, and art, (d) tourist leisure and recreations, and (e) tourist and general infrastructure (catering, accommodation, transportation). This study also identifies four pairs of emotions (eight types), namely "arousing-sleeping", "exciting-gloomy", "pleasant-unpleasant", and "relaxing-distressing", under the affective destination image dimension.
The current study also identifies some gaps or incongruencies between the projected and perceived destination image in both cognitive and affective dimensions. Firstly, there is a significant difference between projected and perceived destination image in terms of cognitive attributes. Tourists pay more attentions to cognitive attributes such as tourist attractions, while NTOs are more concerned with attributes such as tourism environment, history, culture and art, tourist leisure and recreations, and tourist and general infrastructure. It is also worthy to note that NTOs emphasize Australia's "catering", while tourists are more concerned about the general infrastructure. Secondly, some differences are also detected in affective attributes of the projected and perceived destination image. Tourists appear to focus more on the "exciting" attribute of Australia while NGOs seem to focus more on the "pleasant" aspect. While the negative affection projected by NTOs are slightly higher than those perceived by tourists, the difference was marginal.

Discussion
First, as mentioned above, prior studies have not reached a consensus regarding the attributes (or sub-dimensions) of cognitive destination image [15,33]. This study has identified key attributes of cognitive destination image by encoding the top frequently appearing nouns of TGC textual data and NTO-generated content textual data. Comparatively, this approach (using traveler-generated and NTOs-generated contents) may be more comprehensive in terms of developing concrete and sound attributes of cognitive destination image for both regional or national destination. Beerli and Martin's (2004) research on the attributes of the destination's cognitive image has been generally recognized by tourism scholars [36]. Except for "political and economic factors", the other attributes found in this study are consistent with those of Beerli and Martin (2004). The attributes of cognitive destination image in this study are more comprehensive analyzed (e.g., Bonn et al. (2005) [35] and Yang et al. (2012) [38]), and therefore, provides NTOs with more operational references. In addition, this study detects four pairs of affections (eight categories), which shows that Russell's (1980) circumplex model of affect [64] is appropriate for the analysis of affective destination image.
Second, previous studies on destination image have mostly focused on the cognitive aspect, while paying little attention to the affective aspect. The need of using TGC in destination image studies has been recognized (e.g., Mak, Athena, 2017), while the role of NTOs and DMOs generated content are still neglected. Consequently, the current study adopts two social media datasets and provides a more comprehensive comparison between the perceived and projected destination image. On one hand, findings on cognitive attributes support prior studies, which have found that the perceived destination image seldom coincide with the projected destination image [37,41]. On the other hand, findings on the affective dimension suggested that the projected and perceived destination image might not be significantly different (e.g., Perry, 1978). Cultural differences may provide some explanations for the differences between the projected and perceived destination image. Previous studies have also shown that that people from different cultures backgrounds have different perceptions of destinations [64,65]. Liu et al. (2017) argued that the behavioral characteristics and preferences of Chinese outbound tourists may differ significantly from those of international tourists [51], and the marketing team of NTOs may lack of understanding of the Chinese tourists. All the above reasons may lead to a mismatch between destination image projected by Australian NTOs and destination image perceived by Chinese tourists to Australia.
Third, as discussed earlier, the two major dimensions of sustainable experience are interacting with the natural and the social environment of a destination [47]. The data from NTOs detected not only nature, but also arts and history (which have a strong social element). However, the TGC's data (perceived destination image) appeared to only detect "interacting with nature". This finding suggests that the sustainable experience shown from the perceived destination image is less comprehensive than that shown from the projected destination image. It suggests that some measures should be placed in order to close this gap or discrepancy for a sustainable development purpose.
Last but not least, different from Mak and Athena (2017), who have found that TGC better reflected affective destination image than NTO-generated content [2], the present study finds that these two texts have no obvious advantage or disadvantage in reflecting the affective destination image attribute. This difference may be due to the larger textual datasets used in this study. In Mak and Athena (2017), 93 travel blogs (53,422 words) of TGC textual data and 49 pages (18,027 words) of NTOS-generated textual data were obtained for analysis. While in the current study, 418 Trave blogs and 2983 reviews (1.2 million words) of TGC textual data and 3220 microblogs (300,000 words) of NTOS-generated textual data were obtained for analysis.

Implications
The findings of this study show that NTOs (or DMOs) should better use social media to project a cognitive and affective destination images of a nation, such as Australia, that are more in line with tourists' expectations.
First, NTOs and DMOs should establish a system for regularly monitoring TGC and NTO-generated contents. Previous studies have shown that, compared with the positive online reviews of tourists, reducing the negative online review of tourists is more important for the enhancement of destination image [66]. Therefore, in addition to regular and systematic review of TGC [67], NTOs and DMOs should also be cautious of the social media contents that are generated by themselves, and pay particular attention to the textual expression of the content (such as landscapes introduction and warning information about tourism). In addition, more and more tourists began to pay attention to the sustainable development of the destination [68], and tourism endowed tourists with more significance. So, NTOs should also try to use positive contents related to the destination's implementation of sustainable tourism actions to project a positive affective destination image.
Second, prior studies have shown that NTOs' efforts may be undermined, if there is a significant difference between NTOs-generated content and TGC [69]. Önder and Marchiori (2017) argued that NTOs and DMOs need to keep information consistency in order to avoid cognitive dissonance among tourists [39]. Therefore, NTOs and DMOs should continue to promote natural tourism resources, meanwhile also strengthening the innovative promotion of humanistic resources on social media. NTOS should explore the destination culture (e.g., Australian Aboriginal culture) and use social media to showcase the destination culture so that tourists can form a preliminary identity with the destination culture before traveling. NTOs should also encourage tourists to communicate with destination residents in the process of tourism, so that tourists can experience the destination culture and cultural resources of the destination can also be inherited.
They should also increase tourism and leisure activities that allow tourists to participate. In addition, the transportation, accommodation, and other service facilities also should be further improved. NTOs and DMOs should not only focus on the key attributes of cognitive destination image, but also develop more positive attributes of the affective destination image expectations [70] so as to establish a unique destination image in tourists' minds [71].
Third, NTOs and DMOs should change the way they communicate with visitors online. Tourists are increasingly turning to Internet tools to obtain information about their destinations [71]. So, firstly, NTOs and DMOs should combine various social media tools in order to enrich the communications with potential and repeat visitors. Secondly, they should also adopt vivid texts, photos, short videos, and other contents to spread the topics of destinations in line with tourists' expectations and evoke their emotional responses. Thirdly, they can consider adopting anthropomorphic communication strategies to reduce the distance between NTOs (DMOs) and tourists and improve the quality of communication on official social media. Lastly, they should use various incentive methods to engage realistic tourists to generate the content related to the destination, in order to influence potential tourists, and achieve the purpose of tourist-NTOs (DMOs) joint participation in the construction of destination image.
Fourth, in the use of social media platforms for destination image management process, NTOs (DMOs) should also support the implementation of sustainable marketing strategy, educate tourists on eco-tourism-related knowledge through social media, and promote tourists to form environmentally responsible behaviors, in order to realize the sustainable development of tourism industry.
Last, this study also provides a tool for NTOs and DMOs to measure the degree of fit between the projected and perceived destination image. NTOs and DMOs can use data mining technology to evaluate the effectiveness of the marketing strategy aimed at a particular target market [24] and provide more scientific decision support for developing a favorable and sustainable destination image.

Limitations and Future Research
The limitations of this study are mainly reflected in the following aspects. First, this study conducted a content analysis by using the TGC (travel blogs and reviews from Ctrip.com (accessed on 12 February 2021), Mafengwo.cn and Qunar.com) and the NTO-generated content (Sina microblogs) on social media. Because of the textual data in the current study was substantial, other social media textual data were not included in this study. Tussyadiah and Fesenmaier (2009) argued that different forms of TGC may reflect different message of destinations [72]. Pan et al. (2014) also argued that photos as a non-verbal communication method may be more effective than words to reveal the perceived destination image attribute. Therefore, data of different forms, such as photos and videos, generated by tourists and NTOs (DMOs) on social media should be collected in future research to further understand the perceived and projected destination image.
Second, TGC textual data used in this study may have bias in their tourist coverage. Tourists who shared travel blogs and reviews on social media are mainly young, educated, and internet literate, while those who may not like to share travel blogs and reviews online are neglected. Therefore, future research should consider conducting a questionnaire survey or in-depth interviews with tourists to Australia with the help of NTOs, DMOs, and outbound travel agencies. Comparing the tourist data collected both online and offline together with the NTOs-generated content data may help deepen the understanding of the differences between the perceived and projected destination image.
Third, TGC and TNOs-generated content textual data in the current study were dated from 2015 to 2019. In fact, existing studies have shown that destination image is evolving [69,73]. In other words, destination image is a dynamic construction process between tourists and NTOs (DMOs). For example, the Covid-19 pandemic has brought dramatic changes to the tourism sector [74], future research may examine the data after Covid-19 to see if the projected or the perceived destination image has been significantly impacted by the pandemic.