A Study on the Tourism-Related Information Consumption Process of Tourists on Social Networking Sites

Abstract

The current study was conducted to examine the consumption process of tourists through the SIPS model as they experienced tourism-related information and products on social networking sites. Data was collected online from Koreans who have experience in using social networking sites, and a total of 479 responses were used for the data analysis. The statistical package for social sciences 23 and analysis of moment structures 23 were used to evaluate stability, consistency for the measurement items, and to perform structural equation modeling to test hypotheses. There were three main results that emerged from the study. First, three dimensions of sympathy (emotional sympathy, content sympathy, and sympathy for situations and publishers) were extracted. Second, among the dimensions of sympathy, only “content sympathy” showed a significant impact on identification while all dimensions of sympathy significantly affected tourism participation intention. It indicated that tourists are willing to participate in tourism activities immediately if they sympathize with the information obtained from social media without going through the identification process, which is the opposite result of what the SIPS model argued. Lastly, travel content production experience has been shown to have a moderating effect in the relationship between identification and tourism participation intention. The current study will contribute to understanding tourists’ consumption process of tourism-related information on social networking sites and to establish efficient marketing strategies.

1. Introduction

Social networking services (SNS) are now recognized as a powerful communication channel and have deeply permeated into the lives of consumers. The number of SNS users worldwide increased from 970 million in 2010 to 4.48 billion in 2021 [1]. Statistics also showed that out of 4.48 billion SNS users, 99% access websites or apps through their mobile phone. It also indicated that individuals have at least eight social media accounts. The platform that consumers can choose varies, e.g., blogs, Instagram, YouTube, and Facebook [1]. Individuals use SNS when finding tourism-related information. One of the main causes of this phenomenon is higher perceived reliability in consumer opinions than traditional tourism information sources. Reliability matters especially in the tourism industry, because tourists have a lack of direct experience with tourism products. That is, it is difficult to evaluate their quality and benefits before products are consumed by tourists. Therefore, word of mouth (WOM) from tourists who have already experienced tourism products has been considered as one of the most important sources of information for travel planning [2]. Thus, more tourists create their holidays using social media to find information and make plans rather than going through the packaged tour method [3].

In addition, it has been reported that the information tourists post, such as reviews or recommendations on SNS, actually affects tourists’ decision-making [4,5]. The point to note here is how tourists go through the information they encounter through SNS to make final decisions. Obviously, tourists are exposed to a lot of tourism-related information on various platforms. Some of the information may be overlooked, and some information may be the driving force behind tourists’ consumption of the tourism product. Examining how tourism-related information and contents leads tourists to final consumption of a tourism product will play a major role in understanding behavior of tourists on SNS and future tourism marketing.

In Japan, several models explaining the process of consumers purchasing products have been introduced [6], and among them, Dentsu Inc. proposed the SIPS model in 2011, as previous models failed to sufficiently explain the recognition and emotion process of consumers who live with social media. The SIPS model includes the following progression of SNS use: Sympathize → Identify → Participate → Share and spread. What the model suggests is that sympathy with the philosophy or information propagated by individuals or companies serves as the beginning of consumer behavior. The motivation for behavior is born through consumer sympathy and identification, and it leads to participation. That is, through initial motivation, consumers use search engines or social media to find additional information and take specific actions (including behavior without purchase) [7]. The SIPS model emphasizes sympathy, as it is one of the most important key values that drives consumers in the social media era [8,9]. The literature has also consistently verified the effectiveness and importance of sympathy in consumer behavior studies. Kim, Geo, and Nam [10] identified sympathy with consumers as more important than persuasion when verifying the effectiveness of marketing communication tools. Ishak, Khalid, and Sulaiman [11] demonstrated that individuals with a strong sense of religious affiliation were likely to participate in long-term boycotts due to their religious responsibilities and universal brotherhood obligations, demonstrating the importance of universal sympathy in terms of Muslim consumerism.

Tourism can be perceived as an information-intensive field, providing comprehensive and necessary services for tourists. According to Buhalis and Amaranggana [12], the tourism industry is related to extensive and dynamic information retrieval, as tourists not only collect a lot of information through the Internet and make choices, but also compare their choices that they have made in different ways. It has become a daily routine for tourists to search and acquire travel information through SNS and internet searches and share their travel experiences with their acquaintances [13]. As SNS occupies our lives, research on the role of SNS in tourism literature has gradually increased. Different issues that have been studied include the effects of SNS on travel experiences [14,15], the effect of tourists’ emotions on e-WOM [16], the effect of social media on destination image [5,17], complaining practices on social media [18], and blogs in tourism [19,20,21]. Given the growing dependence on social media as a means of effective marketing and communication to attract customers [22], understanding tourist behavior through the application of the SIPS model can have many implications, but discussion on such matters is insufficient. Therefore, the current study aims to validate the SIPS model in the context of tourism to fully understand what makes tourists move in the SNS era, what content they develop sympathy with, and whether they intend to travel to a place where they sympathize. In other words, in this study, the significance of sympathy (S), identification (I), and participation (P) in the SIPS model presented by Dentsu [9] will be validated through empirical analysis. This will verify the important role of sympathy in tourism consumer behavior studies and provide basic standards to establish efficient marketing strategies to promote tourism attractions or tourism products through SNS for potential tourists.

2. Conceptual Framework and Hypotheses

2.1. Tourism Information and Travel Content on Social Networking Sites

Social media is defined as a web-based application where individuals or organizations create, share, and exchange data, content, information, and ideas [23]. A distinct feature of SNS tools is that they promote dynamic interactions between consumers (e.g., tourists) and companies (e.g., travel agencies) on internet places [18]. This interaction fosters socio-cultural capital by generating information and opinions that influence travel consumption choices [24,25]. SNS tourism information refers to digitalized information provided through the network system for tourists [26]. Typical characteristics of SNS include social presence for self-expression, information and content sharing, communication with others, and relationship formation [27]. For tourists, SNS is used for different reasons, such as obtaining travel related information, providing travel tips, social connections, and having fun by sharing their tourism experiences with others [13,25,27]. For example, individuals use integrated tourism websites that are combinations of several websites, and they use these websites not only to obtain tourism-related information but also to plan their trips by booking hotels and flights [28]. Leung, Law, and Lee [29] indicated that during the planning stages, tourists can be influenced by the sharing of information on traveler-generated contents and among them, traveler-generated reviews about tourism attractions and accommodations have become an important information source for tourists [30].

Tourists’ blogs, one of the key tourist-generated content channels that provides tourism-related information, is also a platform for exchanging information and experiences [19]. The number of monitored blogs has reached about 300 million as of 2021, with over 7 million blog posts published every day [14], up from three million in 2004 [19]. Travel blogs are a global phenomenon that has attracted public attention and are part of the tourist experience [19]. Tourism destination marketing organizations and tourism companies use blogs as part of their business strategies [20]. Li, Lin, Tsai, and Wang [17] investigated Chinese tourists’ destination image formation of Taiwan by analyzing contents of 1033 blog articles. Both positive and negative impressions of five categories (e.g., attraction, shopping, food and beverage, accommodation, and transportation) were extracted from the blog articles.

SNS enables tourists to create travel-related content and share them with others [31]. Travel and tourism-related content is a tourism-related information product that produces, processes, and delivers information that consists mainly of photos, texts, and videos [32,33]. Tourists produce, distribute, and share such various tourism related contents based on their tourism experiences on SNS. Generally, consumers on social media are divided into five categories depending on their level of participation: spectators, who simply read content produced by others; joiners, who join a specific group and build a relationship with others; collectors, who gather and share information; critics, who evaluate information; and creators, who actively produce and distribute information [34]. It was said that about 90% of consumers just watch the content distributed by the brand, and only 1% of consumers regularly and 9% occasionally reproduce the content [35]. That is, what determines the success or failure of the marketing strategy is whether consumers can be encouraged to engage in activities beyond being “collectors.” [35]. Eventually, the focus of a social media marketing strategy shifts to how to make a spectator into a creator. Creators who create content, curators that select what consumers want, and social influencers with many followers play an important role in increasing consumer participation. Therefore, in this study, it was analyzed whether the experience of the production of travel content has a moderating effect in identifying the impact of the sympathy and identification of SNS tourism information on tourism participation intention.

2.2. SIPS Model

There are several models that describe consumer behavior, and over time, the models have changed to reflect the trends of the times. The following models describe customers’ purchasing behavior: AID (attention, interest, desire), AIDA (attention, interest, desire, and action), AIDAS (attention, interest, desire, action, and satisfaction), and AIDMA (attention, interest, desire, memory, and action) [6,7,36]. More recently, Dentsu Inc. proposed the AISAS model by modifying the AIDMA, which describes five stages of consumer behavior: “attention”, “interest”, “search”, “action”, and “share”.

This model can be seen as the first model that describes consumer behavior of online product purchases [37]. Consumers post information about products after purchasing on blogs and social media, and this information helps other consumers make decisions about the products. Although the AISAS model captures the purchasing behavior of online consumers, it has been observed that this model doesn’t fully explain online consumers’ purchase behavior in relation with consumers’ recognition and their emotional process when purchasing a product [7].

Thus, the Dentsu Inc. Open Innovation Laboratory proposed a new model, namely SIPS in 2011. According to Dentsu, the SIPS model, a new consumer behavior model for the SNS era, describes the online consumers’ purchase behavior process as follows: Sympathize → Identify → Participate → Share and spread. The SIPS model explains that empathetic information or philosophy is spread by individuals or companies and initiates consumer behavior. Consumers who sympathize with the information go through the process of identifying the information in several ways, which leads to participation and sharing, including behavior without purchase [7]. This process has similarities to the AISAS model introduced earlier. The AISAS model indicates that consumers’ interest is aroused by a particular product or information, which makes them search for information. It then leads to participation in purchasing the product and sharing their opinion. However, the SIPS model’s new proposal in this process is said to be more important than anything else as consumers first sympathize with information when they encounter it on SNS. The SIPS model not only presents new stimuli and transitions in the company’s marketing communication activities, but also reflects changes in society and media and consumer behavior, which can be applied to relevant field studies to gain more insight. However, previous studies using this model to understand online consumer purchasing behavior have not yet received much attention, except some domestic studies.

In Korea, some studies were conducted to examine the behavior of online consumers through the SIPS model. Ahn [38] used the Dentsu model to create criteria for evaluating SNS activities of several well-known companies and verified the SIPS model as a necessary standard for SNS management of companies. Four companies, Hyundai Motor, SKT, KT, and LGU+, were used to analyze how consumers go through the stages of empathy, confirmation, participation, and sharing information on SNS channels operated by each company (e.g., Facebook, Twitter, blogs). Overall, LGU+ showed the highest rating. It was also noted that companies currently operate SNS channels, and efforts to continuously communicate with users with long-term goals and strategies are needed rather than to increase followers in the short term. Shin and Yoon [39] confirmed “authenticity” as an antecedent variable of “sympathy,” a key factor in the SIPS model, and significant relationships between types of authenticity and consumers’ participation levels were found. Yoon and Son [40] focused on the sympathy and participation factors of the SIPS model by comparing the winning works of the Cannes International Advertising Festival and found that it is highly likely to obtain consumer sympathy through “public interest.” Byun, Toon, and Jo [41], Goh and Lee [42], and Jeon [43] verified that sympathy affects purchase intentions, which in turn leads to intentions to recommend, as claimed by the SIPS model, thus describing the advertising effect in the SNS environment. These studies have shown the effectiveness of the SIPS model as a measurement tool to understand online consumers’ behavior. Yet, such research to date has failed to examine the consumption process for tourism-related products and services among tourists using SNS. Therefore, the SIPS model is considered as a viable framework for consumer behavior research in the SNS era and was used as the basis for the current research model.

2.3. Sympathy, Identification, and Tourism Participation Intention

First, “sympathy” has been used along with “empathy” to describe similar notions, but professionals in consumer and advertising fields believe it to be preferable to use the term “sympathy” [9,44]. Sympathy has been defined as “an emotional response stemming from another’s emotional state or condition and it consists of feelings of sorrow or concern for another’s welfare” [45] (p. 92). This point of view regarding sympathy is supported by Escalas and Sternm [46] (2003), indicating that sympathy stems from the observer’s perspective of being conscious of other people’s feelings. However, the SIPS model suggests that “sympathy” can be more comprehensive. Dentsu [7] grouped other items under the concept of sympathy, namely cognitive consent (e.g., know/be), emotional sympathy (e.g., laugh/cry), information value (e.g., beneficial/to be known), and compassion (e.g., respectful/to help). Therefore, based on the concepts of Dentsu [7], this study defines sympathy as a comprehensive concept that includes cognitive/information factors and value/emotional/behavioral factors, which are based on direct and indirect experiences with SNS. Sympathy can be measured differently depending on how it is conceptualized, and the sub-dimensions associated with it.

Table 1. Measurement of sympathy.

Researchers	Sympathy Scale	Dimensions
Ahn [38]	Sympathy Evaluation Model of companies	Information provision method, amount of information and quality of information (originality, differentiation, suitability)
Davis [47,48]; Nam [49]; Soh [50]	Interpersonal Response Index (IRI), 25 items on a 5-point Likert-type scale	Cognitive + emotional elements 1. Take a look 2. Imagination 3. Sympathetic attention 4. Personal pain
Dentsu [7]	N/A	Authenticity, reliability, benefit (useful), convenience, joy of sharing (sense of bond, sympathy)
Feshbach [51]	Three mixed dimensions	Cognitive, emotional, behavioral factor
Jeon [43]	Sympathy Measurement Scale	(Participants) Sharing feelings, participating in my work, imagination, understanding emotions
Mehrabian and Epstein [52]	Emotional Empathy Scale, 33 items on a 9-point Likert-type scale	Sensitivity, awareness of feelings toward strangers, extreme emotional responsiveness, tendency of sympathy, negative emotion
Soh [50]	Consumer Sympathy Response Index, 11 items	Sympathetic understanding, identification, vicarious feelings

Note: developed by researcher.

shows measurements of sympathy developed by domestic and foreign researchers. Through a comprehensive literature review, thirteen measurement items were developed in this study to evaluate tourism activities, and sympathy had four dimensions: sympathy for publishers, cognitive sympathy, emotional sympathy, and behavioral factors.

Second, the SIPS model claims that consumers who have received information go through the process of identification after sympathizing with the contents. That said, sympathetic information, benefits or utilization of products, information about the company, and what friends and professionals said were being checked through the internet or other media. The identification dimensions presented by prior studies included fairness and reciprocity [53]; transparency, profitability, availability, and value perspective [7]; authenticity [54]; and permission to disclose and share information [38]. Based on prior research, this study developed seven measurement items for the identification variable, such as “accuracy,” “authenticity,” “reliability,” and “permission to share information.” Third, SIPS noted that consumers enter the purchasing stage after going through the “sympathy,” “identification,” and “participation” process. It means that consumers of social media may spread their experiences or information to acquaintances through likes, comments, and retweets (sharing), if they sympathize with the information in the post, leading to their purchases or visits. As a result, this behavior can be said to be “participating” in marketing, public relations, or purchasing activities [38]. In the SIPS model, the aforementioned activities (likes, comments, retweets, etc.) are recognized as consumer participation behaviors without necessarily purchasing or visiting a destination [55]. The literature has supported the point of view that sympathy had a positive effect on purchase intention, which leads to recommendation intention as claimed by the SIPS model in the SNS environment [41,42,43]. In this study, the intention of participating in tourism activities was measured to examine the effects of sympathy and identification of tourism-related information based on the results of prior research [41,42,56]. More recently, Nam, Lee, and Min [55] found that tourists go through the process of empathy, identification, participation, and sharing information of tourism-related contents on SNS. Based on the study findings [38,39,40,41,55], the following research models (see Figure 1) and hypotheses were established.

Hypothesis 1:

Sympathetic tourism information in SNS will have a significant positive impact on identification.

Hypothesis 2:

Identification of tourism information in SNS will have a significant positive impact on tourism participation intention.

Hypothesis 3:

Sympathetic tourism information in SNS will have a positive effect on tourism participation intention.

Hypothesis 4:

Travel content production experience will have a moderating effect on the relationship between identification and tourism participation intention on SNS.

3. Methodology

3.1. Measurement

Measurement variables were derived through prior research, and research questions were developed after modification to suit the research purpose. The questionnaire included items measuring research concepts (i.e., sympathy, identification, tourism participation intention). Sympathy consists of 13 measurement items, drawn from previous studies [38,43,49,50,51,57,58], with 4 sub-dimensions: ‘publisher sympathy’, ‘cognitive’, ‘emotional’, and ‘behavioral elements’. To measure identification, a total of seven items were developed based on the previous research [7,34,38,39,50]. In addition, a total of three items were drawn from previous research [41,42,56] to measure tourism participation intention, including visiting or recommending tourism attractions, revisiting, or recommending information of tourism attractions received through SNS to others participating. Travel content production experience is the act of producing and posting tourism-related contents on social media. In order to measure the moderating effect of travel content production experience, two questions were composed of the measurement, including content production status and number of productions drawn from previous research [8,33,34]. All of those measurement items were measured on a 5-point Likert-type scale (where 1 = strongly disagree, 3 = neutral, and 5 = strongly agree). The questionnaire also included items asking respondents’ demographic information (e.g., gender, age, education level, income level) as well as traveling characteristics (e.g., travel length, times of travel, and companions). Once the questionnaire was developed, it was pilot tested with 30 students as well as professionals in the tourism field to test ease of understanding, appropriateness, and ease of response to the questions. A final questionnaire was developed after reflecting on the pilot test results.

3.2. Data Collection

The population for data is Koreans who have experience in using SNS and the sampling frame consisted of Koreans who are over 18 years old, students, university faculty members, and professionals in tourism fields living in Gangwon Province and Seoul metropolitan city. The questionnaire was developed using Google’s questionnaire program, and delivered to the respondents online after generating a survey link. Data was collected from 1 October 2015 for two weeks. A total of 503 responses were obtained, and 458 to 479 responses were used for the final analysis. The reason that different numbers of responses were used for the data analysis is because they were used in the analysis after removing the missing values for each variable (for respondents’ general/ travel characteristics). Upon closer examination of responses, 24 questionnaires either were not completed or included ‘straight-line’ responses. With the removal of such questionnaires, this left 479 questionnaires viable for hypotheses testing.

3.3. Data Analysis

Prior to running the main analysis, data was screened for univariate and multivariate outliers by checking z-Scores and Mahalanobis distance, and the data was checked again to identify any missing data; however, no missing data was identified. Such procedures helped to contribute to achieving more accurate statistical results. SPSS 23.0 (The Statistical Package for Social Sciences) was used for descriptive analyses to provide profiles of the respondents and measurement items. Internal consistency and construct validity were evaluated by applying the Cronbach’s α (alpha) test and confirmatory factor analysis. Structural equation modeling (SEM), using the AMOS 23 program, was applied to test the research hypotheses.

4. Results

4.1. General Characteristics of the Respondents

Table 2. Demographic profile of respondents.

	Frequency (n)	Percentage (%)
Gender
Female	282	58.4
Male	200	41.4
Total	483	100.0
Age groups
10s	94	20.5
20s	255	55.6
30s	21	4.6
40s	32	7.0
Over 50	57	12.4
Total	459	100.0
Education
In university	331	68.5
Bachelor’s degree	86	17.8
Postgraduate (master’s or doctorate)	62	12.8
Total	483	100.0
Monthly income ($US)
<1000	306	64.8
1000–2500	79	10.4
2500–4000	49	10.4
Over 4000	38	8.1
SNS usage frequency
Once per day	367	76.6
Once per week	59	12.3
Once per month	18	3.8
No use	35	7.3
Total	479	100.0

shows the general characteristics of the respondents. The demographic profile of the respondents indicated that 58.4% were female, and half of the respondents (55.6%) were in their twenties. The majority of participants (68.5%) were in a college or university. In terms of monthly household income, for most participants (64.8%) it was under $1000. Most participants (76.6%) used SNS daily and the types of SNS that respondents enjoyed using were shown to be Instagram (42.2%) and Facebook (41.9%).

Table 3. Characteristics of travel behaviors.

	Frequency (n)	Percentage (%)
Frequency of domestic travel over the past year
1–2 times	185	38.7
3–4 times	165	34.5
More than 5 times	98	20.5
Never	30	6.3
Total	478	100.0
Frequency of international travel over the past year
1–2 times	214	44.7
3–4 times	45	9.4
More than 5 times	10	2.1
Never	210	43.8
Total	479	100.0
Companion
Alone	20	4.4
Family/relatives	134	29.6
Friend	291	64.4
Colleague	7	1.5
Total	452	100.0
Length (day)
1	58	12.3
2–5	353	74.6
3–9	18	8.9
10–19	42	5.4
Over 20	2	0.4
Total	473	100.0

shows the characteristics of the travel experiences of respondents. Most of the respondents (about 73%) traveled domestically over the past year, ranging from one to four times and international travel was mostly reported as once or twice (44.7%) per year. When the respondents traveled, about 64% said they were with friends, and more than half said they had traveled for about two to five days.

4.2. Exploratory Factor Analysis

Exploratory factor analysis (EFA) with Varimax rotation was conducted to identify underlying dimensions of all the constructs, including sympathy, identification, and tourism behavioral intention. In order to determine the appropriateness of EFA, the Kaiser-Meyer-Olkin (KMO) and Bartlett’s test of sphericity were examined for the testing variables (see

Table 4. Exploratory factor analysis results.

Variables and Items	Factor Loading	Eigen Value	Variance Explained	Cron Bach’s α
Sympathy Emotional sympathy (“I sympathize with…”)
tourism-related content that contains interesting information.	0.847
tourism-related content that contains the information I want.	0.834
tourism-related content that contains attractive information.	0.801	5.15	39.61%	0.855
informative/useful tourism-related content.	0.719
the emotional tourism-related content.	0.596
Content sympathy
When I see tourism-related content, I feel like I’m actually in that place.	0.823	2.02	15.55%	0.844
When I see tourism-related content, I feel like the person in the content is me.	0.818
Tourism-related content makes me forget daily lives.	0.800
When I look at tourism-related content, I feel like I’m doing what I’ve experienced in the past again.	0.742
Sympathy for situations and publishers
I understand the content related to tourism well.	0.775
I understand how the person who appeared in the tourism-related content felt.	0.746	1.12	8.60%	0.722
I sympathize more with celebrities when they post tourism-related content.	0.634
I sympathize more with friends when they post tourism-related content.	0.591
KMO = 0.87, Bartlett χ2 = 2718.37 (df = 78, p < 0.01)
Identification
I check the accuracy/ authenticity/reliability of content with the number of “likes.”	0.931
I check the accuracy/ authenticity/reliability of content with the number of “followers.”	0.872	2.30	76.61%	0.846
I check the accuracy/ authenticity/reliability of content with the number of “stars.”	0.839
KMO = 0.70, Bartlett χ2 = 643.50 (df = 3, p < 0.01)
Tourism participation intention
I am willing to purchase/visit tourism-related.	0.889
products/services/tourist sites introduced in the content.	0.867
I will recommend these tourism products/services to people I know.	0.841	2.27	75.54%	0.837
I am willing to re-purchase/visit tourism-related products/services/tourist sites introduced in the contents.
KMO = 0.072, Bartlett χ2 = 578.49 (df = 3, p < 0.01)

). For sympathy, as the results indicate in Table 4, the KMO value was 0.88 with Bartlett χ2 = 2718.37 (df = 78, p < 0.01). The correlation matrix ranged from 0.85 (tourism-related content makes you forget reality) to 0.92 (I sympathize with emotional tourism-related content), which indicates the construct was appropriate for the factor analysis. On the other hand, the principal component analysis of sympathy resulted in 3 factors with an eigenvalue of 1 or higher, and the total variance explained was 63.75%. The three dimensions were named as “Emotional sympathy” (eigenvalue = 5.15, total variance explained = 39.61%), “Content sympathy” (eigenvalue = 2.02, total variance explained = 15.55%), and “Sympathy for situations and publishers” (eigenvalue = 1.12, total variance explained = 8.60%). Cronbach alpha coefficients were examined to evaluate the reliability of the extracted dimensions. It was 0.86, 0.84, and 0.72 respectively.

For the identification variable, only one factor was extracted by the factor analysis with an eigenvalue greater than 1 (eigenvalue = 2.30, total variance explained = 76.61). The KMO value was 0.70 with Bartlett χ2 = 643.50 (df = 3, p < 0.01). The correlation matrix ranged from 0.65 (I check the accuracy/authenticity/reliability of contents with the number of “likes”) to 0.77 (I check the accuracy/authenticity/reliability of contents with the number of “stars”) with a Cronbach alpha coefficient of 0.85.

The principal component analysis of tourism participation intention was extracted as one single factor with an eigenvalue greater than 1 (eigenvalue = 2.27, total variance explained = 75.54) and the results showed that the KMO value was 0.72 with Bartlett χ2 = 578.49 (df = 3, p= 0.000). The correlation matrix ranged from 0.68 (I am willing to purchase/visit tourism-related products/services/tourist sites introduced in the contents) to 0.76 (I will recommend these tourism products/services to people I know) with a Cronbach alpha coefficient of 0.84.

4.3. Confirmatory Factor analysis (CFA)

Confirmatory factor analysis (CFA) was conducted to evaluate the appropriateness of the measurement model (see

Table 5. Confirmatory factor analysis results.

Variables and Items	Standarized Loading	S.E.	C.R.	CR	AVE
Sympathy Emotional sympathy (“I sympathize with…”)
tourism-related content that contains interesting information.	0.799
tourism-related content that contains the information I want.	0.827	0.056	19.453
tourism-related content that contains attractive information.	0.771	0.056	17.900	0.715	0.656
informative/useful tourism-related content.	0.772	0.058	17.927
the emotional tourism-related content.	0.566	0.072	12.453
Content sympathy
When I see tourism-related content, I feel like I’m actually in that place.	0.806
When I see tourism-related content, I feel like the person in the content is me.	0.796	0.057	17.724
Tourism-related content makes me forget daily lives.	0.714	0.059	15.790	0.606	0.539
When I look at tourism-related content, I feel like I’m doing what I’ve experienced in the past again.	0.722	0.055	15.996
Sympathy for situations and publishers
I understand the content related to tourism well.	0.735
I understand how the person who appeared in the tourism-related content felt.	0.739	0.078	13.799	0.572	0.516
I sympathize more with celebrities when they post tourism-related content.	0.522	0.083	10.155
I sympathize more with friends when they post tourism-related content.	0.5622	0.073	10.907
Identification
I check the accuracy/ authenticity/reliability of content with the number of “likes.”	0.901
I check the accuracy/ authenticity/reliability of content with the number of “followers.”	0.800	0.050	18.649	0.647	0.600
I check the accuracy/ authenticity/reliability of content with the number of “stars.”	0.727	0.048	16.994
Tourism participation intention
I am willing to purchase/visit tourism-related.
products/services/tourist sites introduced in the content.	0.848
I will recommend these tourism products/services to people I know.	0.780	0.054	18.138	0.751	0.706
I am willing to re-purchase/visit tourism-related products/services/tourist sites introduced in the contents.	0.761	0.055	17.664

Note: p < 0.001.

). The results of CFA showed acceptable model fit χ2(142) = 396.38, p < 0.05, comparative fit index (CFI) = 0.94, normed fit index (NFI) = 0.91, goodness of fit index (GFI) = 0.92, incremental fit index (IFI) = 0.94, root mean square error of approximation (RMSEA) = 0.06, and root mean square residual (RMR) = 0.05. Also, as shown in Table 5, all constructs were highly reflected by their respective observed variables based on standardized loadings estimates (>0.50), which displayed an adequate construct validity [59]. Also, the average variance extracted (AVE) for all constructs ranged from 0.52 to 0.71, which is greater than 0.5, reflecting an acceptable convergent validity. Construct reliability for all the constructs was either above or close to acceptable thresholds of the recommended composite reliability (>0.60) [60], ranging from 0.57 to 0.75. Even though “Sympathy for situations and publishers” showed lower reliability compared to other constructs, it was retained because the research was the first to develop the SIPS scales in the context of tourism and therefore was exploratory in nature) [59].

In AMOS one loading, the first items of the variable in this study had to be fixed to 1, thus the C.R. and S.E. cannot be calculated for the item.

Discriminant validity is verified when the correlation between two variables is lower than 0.85 [61]. The results showed that all correlations were lower than 0.85. Therefore, the discriminant validity of the measurement scales was established (

Table 6. Discriminant validity test of measurement model.

	ES	CS	SSP	ID	PI
ES	0.809
CS	0.365 **	0.734
SSP	0.634 **	0.556 **	0.718
ID	0.218 **	0.422 **	0.795 **	0.775
PI	0.647 **	0.476 **	0.374 **	0.253 **	0.840

Note 1: ES = Emotional sympathy, CS = Content sympathy, SSP = Sympathy for situations and publishers, ID = Identification, PI = Participation intention. 2. Bold numbers on the diagonal parentheses are the square root of each construct’s average variance extracted value. 3. ** p < 0.01.

). According to Hair, Ringle, and Sarstedt [60], adequate discriminant validity is demonstrated when the square roots of the AVE for each factor is greater than the correlations between that factor and other factors. As shown in Table 5, because all square roots of the AVE for each variable is greater than the correlations between factors, it can be said that the discriminant validity was also confirmed.

4.4. Testing Hypotheses

As Figure 2 indicated, the model fit of the research model showed an acceptable level χ2(142) = 396.38, p < 0.01, CFI = 0.94, NFI = 0.91, TLI = 0.93, RMR = 0.05, RMSEA = 0.06. Consequently, research hypotheses were tested. H1 proposed that there would be a statistically significant, at α = 0.05 level, positive relationship between sympathy of SNS tourism related information and identification. According to the results, emotional sympathy and sympathy for situations and publishers were not statistically related to identification (β = 0.09, β = 0.10; p < 0.05). Yet, content sympathy showed a statistically positive relationship with identification (β = 0. 34, p < 0.05). Thus, hypothesis 1 was partially supported. H2 indicated that identification would have a significant positive impact on tourism participation intention. The results showed the relationship between identification and tourism participation intention was not significant (β = 0.07, p < 0.05). Thus, Hypothesis 2 was not supported. Hypothesis 3 proposed that sympathetic tourism information in SNS will have a positive effect on tourism participation intention. The results indicated that emotional sympathy (β = 0. 39, p < 0.05), content sympathy (β = 0.13, p < 0.05) and situation and publisher’s sympathy (β = 0.32, p < 0.01) were all statistically significant. Thus, hypothesis 3 was supportedHypothesis 4 proposed that travel content production experience will have a moderating effect on the relationship between identification and tourism participation intention. By taking Bagozzi and Yi’s [62] recommendation, the chi-square difference between the baseline model and nested model was examined to see the moderating effect of invariance by computing for 1 degree of freedom. The results showed that the chi-square of the baseline model was χ2 (284) = 556.12 and the nested model was χ2 (285) = 564.45. A significant chi-square difference was found (Δχ2 = 8.23, Δdf = 1, i < 0.01), thus, supporting the moderating role of travel content production experience (Hypothesis 4).

5. Discussion, Implication and Limitation

5.1. Discussion

The primary purpose of this study is to understand tourists’ tourism product consumption process on SNS in application of the SIPS model presented by Dentsu [7] to provide efficient marketing strategies in the context of tourism industries. The current study confirmed that the relationships of emotional sympathy and sympathy for situations and publishers with identification were not supported. Content sympathy showed a significant positive effect on identification. That is, among the dimensions of sympathy, only ‘content sympathy’ (e.g., forgetting reality, sense of realism, past recollection, self-reflection) had a significant impact on identification. It can be interpreted that tourists go through the process of checking tourism-related information that they sympathize with through social media in various ways. The finding that ‘emotional sympathy’ and ‘sympathy for situations and publishers’ do not have much impact on identification was noteworthy because it was contrary to Dentsu’s [7] claim. According to that model, individuals would go through the process from ‘sympathy’, ‘identification’ and ‘participation’ to ‘shares/spreads’. The current study also found that identification did not influence tourism participation intention. This means that even if tourists check the information that they sympathize with in social media, it does not affect their tourism participation intention. This is also a different result from what the SIPS model indicated. The way tourists identify information acquired on the media was confirmed by checking ‘the number of likes/stars/followers’. Such results are supported by previous studies wherein consumers tend not to trust the information shared by a company, so they prefer to evaluate the quality and value of a product or service through user-generated content rather than the content provided by the company. It has also been confirmed that other customers’ word of mouth and positive opinions represent a more reliable sources of information and play an important role in triggering customer trust in products and services [63,64]. Hypothesis 3 (sympathy → tourism participation intention) was supported, and all three sub-dimensions of sympathy were verified to have a significant impact on tourism participation intention. These results are consistent with the previous research results [41,42,43] and it has been verified that the empathy factor of SNS tourism information exerts the strongest influence. Unlike what the SIPS model indicates, it means that tourists are willing to participate in tourism activities immediately if they sympathize with the information obtained from social media without going through the identification process. This demonstrated the differentiation between the purchase decision model of general consumer products and the purchase decision model of travel/tourism products. It was also confirmed that sympathy for tourism information on social media is a new path (sympathy → tourism participation intention) that directly affects tourism participation intention. In addition, hypothesis 4 (moderating effect of travel content production experience) was supported in this study. This indicates that travel content production experience has been shown to have a moderating effect in the relationship between identification and tourism participation intention. This can be interpreted as the more active the content that is produced by individuals, the more willing they are to participate in tourism activities. These results can be seen as supporting the claim that one collector, critic, and creator rather than 100 subscribers determine the success or failure of social media marketing [8].

5.2. Theoretical Implications

This study theoretically suggests several implications. First, it attempted to explore the process and tourism participation intention of how tourists purchase tourism products using information obtained through social media based on Dentsu’s SIPS model. As a result, it was found that the consumer’s product purchase process explained by the SIPS model is different from when tourists purchase tourism products through SNS. These results could be used to develop strategies on how to promote tourism products in the tourism industry through SNS in the future. Secondly, the concept of ‘sympathy’ was one of the first applied to the study in the field of tourism to understand tourism participation intentions of tourists using social media. Thirdly, the measurement scales of sympathy and identification factors were developed to investigate tourists’ tourism participation intention in SNS as a foundation. This study, which describes the process of purchasing tourism products of tourists on SNS, helped bridge the gap in research that is lacking in this area. Therefore, starting with this study, follow-up research should be carried out in different settings to validate the scales and results found in this study.

5.3. Practical Implications

This study may also provide local governments and tourism entities with several practical implications. Firstly, local governments will be able to use the sympathy dimensions presented in this study to promote tourist attractions and to produce advertisements for tourist products that tourists sympathize with. In particular, promotion to a specific segment of the tourism market can be used to attract tourists to specific tourism destinations as well as to induce consumption of tourist products by using a measure of sympathy developed in this study. Secondly, considering the result that tourists sympathize with the contents of tourism information on SNS, tourists go through the process of identifying and verifying the information they found on social media before purchasing tourism products or visiting tourism destinations. Thus, the information that is shared should be accurate, reliable, and authentic.

Third, in this study, it was found that sympathy (e.g., emotional sympathy, content sympathy, and sympathy for situations and publishers) directly affects the tourism participation intention. Therefore, in order to encourage tourists to actively engage in tourism activities, information producers who produce tourism products and content should seek ways to maximize tourists’ sympathy by applying sympathy measurement items. In particular, given that emotional sympathy has the greatest influence on tourists’ behavior intention among the three sympathy factors in this study, it is believed that tourism information producers should make tourists more emotionally sympathetic with the tourism content and information.

Fourth, it was found that travel content production experience has been shown to have a moderating effect in the relationship between identification and tourists’ behavioral intention. This indicates the more active the content that is produced by tourism information producers, the more willing they are to participate in tourism activities. Therefore, it would be necessary to encourage active participation of travel content producers (especially collectors, critics, and creators). In addition, it will be necessary to identify the number of users who use the produced tourism information and the preferred social media by age group to check users’ responses from time to time, and efforts should be made to communicate with users more efficiently.

5.4. Limitations

Similar to all studies, our current study also had a few limitations. First, this study may have limitations in the research approach, as it uses the SIPS model, a consumer behavior model of general consumer products to understand how tourists use the information that they obtain on SNS and verify its effectiveness by attempting a psychological approach to tourism behavior research. Therefore, in future studies, a research model can be developed to explain the tourism consumption process of tourists in the field of tourism by reflecting the results of this study. Secondly, in this study, it was attempted to develop a scale of sympathy and identification factors for information obtained by tourists on SNS, but the question still remains whether this measure can be used for all types of SNS. Therefore, future research will require continuous examination and verification of this unexplored area. Third, another limitation of the current study is related to the generalization of this study’s results due to cultural differences. The study was conducted on Korean tourists living in South Korea. Therefore, tourists living in countries with different social or cultural backgrounds from Korea can lead to different research results. Therefore, based on the results of this study, future research on how people from other countries acquire tourism information and participate in tourism activities through the process of verifying information will be needed. Lastly, since there is a lack of international research on the behavior of tourists using SIPS models, most of the prior studies referred to in this study are conducted in South Korea. Therefore, it is hoped that this study will be a foundation to enable future research in developing measures to explain the behavior of tourists who use SNS and live in other countries.