1. Introduction

The credibility of consumer-generated media (CGM) indicates the credibility users assign to the messages other people post on websites regarding their experiences with a product, service, or brand. It is difficult to estimate the credibility of the source of information in e-WOM communications given that they are usually written by anonymous sources that are not related to the recipient (Filieri et al., 2015). Existing research that has sought to understand CGM-based traveler behavior has focused on aspects such as participation in online communities (Wang & Fesenmaier, 2004), the implications of CGM on travel decisions (Cox et al., 2009), the various forms of user-generated content (Chen et al., 2014), the perceived usefulness of CGM content (Neijens et al., 2011), the trustworthiness of the CGM source (Filieri et al., 2015), and information and regulatory influence (Book & Tanford, 2020).

While these studies explain traveler behavior in light of the information found on digital platforms, the perception travelers have of the credibility of CGM deserves further attention. This research aims to develop a measurement scale for the perception of the credibility of CGM through the research question: How can the credibility of the information published in CGM be measured? To do so, a scale that takes into account the increasing use of websites to make travel decisions is presented, as travelers require the sources they use to be perceived as credible and trustworthy. (Girish et al., 2021; Book & Tanford, 2020; Filieri et al., 2015; Ayeh, 2015).

The recipient's perception of a digital media message is systematically studied through a consumer-generated media (CGM) credibility scale. This concept of credibility is consistent with the Credibility Theory, which studies the factors associated with the credibility of the news media (Hovland et al., 1953; Sternthal, et al., 1978). Experience, information quality, information amount and reputation of the medium are used as sub-dimensions of the CGM credibility scale.

This article is divided into the following sections: first, a literature review of the main constructs; second, the methodology used in the qualitative and quantitative studies; third, the obtained results; and finally, the academic and practical implications, as well as future research approaches.

2. Literature review

2.1 A credibility approach for consumer-generated media (CGM)

CGM are online information sources used by consumers to share information about products, services, and topics with each other (Correia-Loureiro et al., 2020; Litving & Dowling, 2018; Xiang & Gretzel, 2010). The development of the internet and the ability of users to create and publish content has led to active online communities that provide useful information about products and services (Ganzaroli et al., 2017). Social networks are key sources of information that play a leading role in helping travelers decide their travel destinations (Litvin & Dowling, 2016; Xiang & Gretzel, 2010).

The factors associated with users’ acceptance of technologies have enabled improvements in the credibility of CGM, aiding vacation or travel planning (Ayeh, 2015). Credibility has become an effective tool for persuading individuals (Newell & Goldsmith, 2001); greater interaction between the user and the company increases the messages’ credibility and creates a feeling of belonging, therefore generating a better corporate reputation (Barreda et al., 2016; Eberle et al., 2013).

The connection between CGM and credibility is illustrated in the Source Credibility Theory (Hovland et al., 1953), which has been used in marketing and communication research to compare the support and credibility of different media (Johnson & Kaye, 2009; Pornpitakpan, 2004). Lowry et al. (2014) explain that the Source Credibility Theory facilitates the understanding of why some sites are more credible than others and details how the degree of persuasion depends on the credibility of its source.

The Source Credibility Theory has been successfully applied to online communication methods (Jang et al., 2020; Ayeh, 2015; Lowry et al., 2014; Flanagin & Metzger, 2007; Fogg, 2003) to examine the relationship between credibility and information usefulness, information system acceptance, website reliability, perceived experience, and other aspects that may be relevant for travelers who access CGM content. Moreover, various studies have used the Source Credibility Theory to analyze the effects on consumer outcomes, such as their attitude towards a message, purchase intentions, or the adoption of information (Jang et al., 2020; Kim & Kim, 2013; Li, 2013).

To summarize, the credibility of the source positively influences people’s motivation to seek information and entertainment, and their inclination to continue to use these sources, which is why travelers are increasing their use of CGM to share information (Filieri et al., 2015; Hur et al., 2017). The crucial role of online media today requires a conceptualization that accounts for the notion of credibility in CGM, which is provided in this article through the development of a measurement scale based on the principles of the Source Credibility Theory.

2.2 Sub-dimensions of credibility in CGM posted by traveling users

Theorists describe credibility as a multi-dimensional construct (Ohanian, 1990), although single-item measures are treated as one-dimensional (MacKenzie & Lutz, 1989). Therefore, using a reliable and valid scale facilitates standardization and encourages new research (Newell & Goldsmith, 2001).

In order to identify the dimension of credibility, studies on CGM in the area of hospitality and tourism must considered. Given that social networks are the main platforms travelers use to share their experiences online (Xiang & Gretzel, 2010), these considerations are essential for travel planning (Yoo & Gretzel, 2011). Existing research examines the perception of credibility of CGM amongst travelers, as it analyses how the perceived credibility is affected by information valence and source identity in the travel services that are being reviewed in CGM (Kusumasondjaja et al., 2012).

This entails the identification of the role of the following factors perceived by travelers: ease of use, reliability, enjoyment, and integrity. The perceived expertise factor is also considered (Ayeh, 2015) when making travel decisions using reviews from CGM (Ayeh et al., 2013). Additionally, information quality, website quality, satisfaction, and user experience increase the perceived credibility of CGM (Filiery et al., 2015). The source and quality of online reviews, blogs, and texts have a significant effect on the credibility of CGM (Chakraborty & Bhat, 2018) and the influence of brand loyalty is seen in reviews published on CGM platforms (Litvin & Dowling, 2018).

The dimensions and the credibility scales used in previous studies are similar to the methodological design of the scale used in this article. Credibility research has measured and conceptualized the scales according to source credibility, advertising credibility, ad content credibility, and media credibility (Soh et al., 2007). Therefore, the scales used to create the CGM credibility scale for travelers were based on the existing scales of subcategories of source and media credibility; both of these concepts are related to CGM as they all are online information sources that provide user reviews.

According to the procedure carried out in step two of the methodology, the following credibility dimensions were obtained for CGM posted by travelers:

Media Reputation: Blogs where information is posted are considered a form of CGM that can influence a company’s reputation, particularly blogs in which the company’s employees publish information about the business, which increases the number of reviewing users (Agag & El-Masry, 2017; Aggarwal et al., 2012). CGM users search for information on websites with comments from people who have had similar experiences regarding travel and travel-related matters (Kwok et al., 2017). Reputation directly affects the users’ perceived risk of using a hosting service (Sun, 2014).

Information Amount: The amount of information seems to have a significant impact on travelers’ approval of online reviews (Filieri, 2015; Filieri & McLeay, 2014). Travelers need to gather enough information before making decisions. In addition, tourism-related products include different services such as accommodation, transport, restaurants, and entertainment, which makes travel planning complex (Filieri et al., 2015).

Information Quality: People have emotional responses based on the quality of the information they read, which is one of the inputs that enables them to better relate to a brand (Gu et al., 2007; Kim & Johnson, 2016). Furthermore, to enhance the quality of the information, websites should focus on encouraging users who post comments to provide objective and constructive feedback about a product or service (Salehi-Esfahani et al., 2016). If the content is reliable, a user can form a distinct attitude towards the tourism product/service (Sparks et al., 2013).

Experience: The human experience is based on users’ level of satisfaction with the tourist services offered by different companies in the industry. The level of satisfaction serves as an evaluation whereby clients test the services received at the site (Hunt, 1975). Experience includes the paradigms of expectation and reputation, which are manifested in the tourists’ level of satisfaction, specifically in the way it influences the choice of a destination, the consumption of products, and the decision to return to a given destination (Yoon & Uysal, 2005).

3. Methodology

3.1 Sample and procedure

A scale to measure credibility was developed following the process established by Churchill (1979). The development consisted of four steps: specifying the construct, analyzing the items, data cleansing, and confirming the scale.

The first step entailed conducting a literature review to identify different measures, perspectives, and ideas about the Credibility scale. The selection of articles published in high-impact scientific journals was based on the presentation by Briner and Walshe (2015) in which several articles available on the Web of Science (WoS) were identified. Consequently, the topic of interest was limited to studies published in the main tourism and hospitality magazines (Gursoy & Sandstrom, 2016).

The search yielded 552 articles: 15 on consumer-generated media (CGM), 13 on credibility scales, and 524 on reputation. The 524 reputation articles were filtered by most cited in the field, resulting in 34 articles. The base result was 62 articles that were reviewed in-depth for psychometric scales that measured subjects’ perception of the credibility of the information, thus seeking to obtain valid constructs with which to develop the measurement scale. Then, following the classification proposed by Soh et al. (2007) to measure credibility, scales were selected if they belonged to the categories source credibility or media credibility.

3.2 Experts

An in-depth interview is an information gathering technique based on oral interactions between an interviewer and an interviewee, which enables a dialogue to be built based on a research problem (Bozeman, 2003; Goldman, 1962). According to Churchill (1979), a scale shows a concept that can be compared, and in order to do that, a list composed of elements determined by the literature must be developed. It is then followed by in-depth interviews.

The present research study took this step to increase the validity of the scale. The interviews were conducted in 2018 and ten were selected by intentional open sampling (Sandelowski & Barroso, 2003). They were considered saturated after the eleventh interview, when the answers given were found to be comparable to the ten previous interviews. Table 1 shows the item selection process.

3.3 Identified and categorized dimensions

The process of identification and classification of the information was carried out employing a basic thematic analysis, which consisted of an analysis of the content found in traditional and digital media to quantify the number and type of advertisements (Andreú, 2002; Bardin, 1996). The analysis of individual interviews focused on establishing the points of consensus in the participants’ responses. Although this investigation consisted of twelve qualitative reports, it was found that after the ninth report no new relevant information was introduced (Sandelowski, 1995; Sandelowski & Barroso, 2003).

Afterward, the 80 selected items were presented to the interviewees. The experts then indicated which items would allow for a better measurement of the credibility of CGM. The items of a scale can be adjusted to the context of the field of study (Choi et al., 2016). Based on the methodology proposed by Liu et al. (2017), 16 credibility items were obtained, which were identified and categorized into four dimensions: environmental reputation, information quality, information amount, and experience. Each dimension was composed of four items. It is important to highlight that the items regarding information amount were proposed by the experts (Bardin, 1996). Based on their recommendations, only the items they most agreed on were used.

3.4 Questionnaire and CFA

The analysis unit includes users who employ CGM platforms and use the internet to acquire or purchase tourism-related products, hotel reservations, and other services offered on these platforms. Table 2 shows the technical data framework of the study.

Based on the results of the qualitative component of this study, a 16-item scale is proposed. This scale is used to measure credibility for users who access CGM. The items are of the reflective type in association with the following dimensions: 1. media reputation, 2. information amount, 3. information quality, and 4. experience. The items included in information amount were the result of the recommendations provided by the experts during the in-depth interviews conducted in the qualitative component of this research.

The results of the confirmatory factor analysis are presented; the number of factors, the level of relationship or independence between them, and the relationship between the factors and the variables are determined (Long, 1983; Stapleton, 1997; Petrick, 2002). The exploratory factor analysis does not provide evidence for the one-dimensionality of the essential measures in the development of the scale (Bush et al., 1990). Therefore, confirmatory factor analysis can be used to group variables into one factor or into the required number of factors (Hair et al., 1999). Table 3 displays the results obtained in the factor analysis.

It can be observed that the comparative fit index (CFI) indicates a good fit to the model at 0.968, higher than the threshold value of 0.95 (Bentler & Bonett, 1980). The Tucker-Lewis index (TLI) compares the adjustment by degrees of freedom of the proposed and unaccepted models; this index showed a value of 0.956, which indicates a good fit, as it is higher than 0.9 (Kohli et al., 1993). The fit of the model in research related to the development of scales is divided into three steps: (a) Analysis of the standardized parameters and their standard errors (Brown, 2015; Byrne, 2001); (b) an index adjustment (Brown, 2015), and (c) verifying that the absolute value of the standardized residual covariance is greater than 2.0.

Three model fitting iterations were carried out to achieve the final model. Items were eliminated as a consequence of their relatively low loads (.01). Through the covariance of the errors, it was identified that the data fit the model (Byrne, 2001). Three of the factors showed significant covariance (<.05).

3.5 Scale Validation

The validation in the second group sought to meet three objectives: test the scale with a larger sample of credibility of CGM; validate the structure of all factors in a more general population; and demonstrate convergence, concurrence, discrimination, and validity in the final scale (Table 4 presents the covariance results).

3.6 Data collection and sampling

The elements of the previous final model were tested with all the elements of comprehensibility of the validation study to analyze if the factor showed different results with a sample of non-students. A total of 22 items were tested. A second-order model was tested to consider the credibility of CGM. The theory represented by the subdimensions of the four constructs was applied to ensure content validity. It was suggested that the items related to the IQ and IA scales be eliminated due to their high standardized residual covariance. An even weaker model was shown in the second 20-item model. Due to poor standardized residual covariance values for IQ5, IQ6, and IA6, which were reported by MI statistics, these elements were eliminated.

Furthermore, the first item related to IQ was defined in two different factors and was therefore eliminated. The items IA2, IA3, and MR4 were leading to an overemphasis of that consideration. E4 was removed from the scale due to the wording of the items E1 and E3. The element and factor loads are shown in Figure 1. All the item loads are significant (p <0.001) above 0.5. Moreover, the four factors provided loads (p <.001) in the higher order of the CGM credibility construct. Information amount had a load of 0.96, experience of 0.91, information quality of 0.85, and media reputation of 0.92. With CFI = 0.96, TLI = 0.95, RMSEA = 0.06, CFI = 0.70, and SRMR = 0.06, the general fit of the model was good (Brown, 2015). Information quality in the CGM Credibility scale was high (α = .88); it would not be possible to improve Information quality by eliminating elements. In the CGM Credibility scale, Information quality had a CR of .94 (Raykov, 1997). It seems that this 16-item scale is a good representation of the credibility of CGM.

3.7 Discriminant Validity

The discriminant legitimacy confirms that the value "1" was not within the confidence intervals for each pair of configurations; therefore, the confidence intervals are indifferent (Anderson and Gerbing, 1988). Additionally, in each scale, the mean variance extracted was greater than the variance shared with the eliminated structure, confirming the precision of the identification.

In Table 5, a summary of the items for the constructs used in the research is presented. This provides an overview of the interrelations between the items of the four constructs. The correlations are also included.

3.8 Convergent Validity

The OLS method enables the estimation of the parameters in a linear regression model. Mackenzie et al. (2011) state that convergence is used to analyze the relationship between constructs that are meant to be in a similar nomological network. Therefore, an all-concept regression method was developed to verify CGM Credibility relationships within their nomological connections. Each factor has a positive relationship with CGM Credibility, and the regression confirms this contribution (see Figure 1).

3.9 Concurrent Validity

Concurrent validity of the CGM Credibility scale was performed by showing the relationship between the developed construct and the outcome variable, such as commitment (Cheng, 1990), which is a construct used in the fields of business and tourism. If tourists perceive CGM as credible, it can influence their decision-making process and it will increase the likelihood of them booking hotels recommended on the platform, thus enhancing their engagement (Tsai & Men, 2013). A linear regression model using CGM Credibility as the dependent variable and engagement as the independent variable was developed and confirmed (R2 = .13, β = .31, B = .16, SE = 0.02, t(1) = 3.62, p < .001).