Altruism in eWOM: Propensity to Write Reviews on Hotel Experience

Abstract

This research tests the relationship between aspects of customer influenceability at the time of booking a hotel with the propensity to write a review in electronic word-of-mouth communication. A valid sample of 739 online questionnaires was obtained. An Exploratory Factor Analysis was conducted in order to reduce the dimensions of the two critical variables, and a measurement model was built. Then a Path analysis was carried out. The novelty of this research lies in measuring the evolution from being a passive eWOM reader to a proactive eWOM writer. Results indicate a relationship between being influenced by reading reviews and the propensity to write reviews. The most important underlying motivation to write a review is altruistic. Managers should try to identify the most responsive customers and encourage them to write reviews on altruistic grounds. This study effectively validated the impact of being responsive to reading reviews on the inclination to, in turn, write them. Findings contribute to the evolving research landscape in eWOM within the hospitality and tourism sector, offering practical insights for industry practitioners to formulate more effective strategies in soliciting and managing customer reviews.

1. Introduction

An increasing number of scientific studies highlight the influence of the Online Travel Agencies (OTAs) on hotel bookings and the electronic word-of-mouth (eWOM) contained in them as a major factor in the booking [1,2]. This is due to the fact that, nowadays, in the hotel sector, consumers first come into contact with the offer through the web and the online environment. This contact in advance of arrival was not possible in pre-internet times, as there was only the possibility of looking at some print pictures in a physical medium, e.g., leaflet, brochure or travel guide. The online environment is so critical nowadays that some studies conclude that hotels without online reviews are much less likely to be considered and chosen [3]. The lack of knowledge about actual product features and the physical inability to evaluate them makes eWOM one of the main channels consumers use to make their booking decisions [4] due to the accessibility it provides to a large number of reviews [5]. The perceived risk in unfamiliar destinations and accommodation also makes eWOM an effective tool in the trust and adoption of online information [6]. Reimer & Benkenstein [7] argue that e-WOM is a more credible and accessible information mechanism than traditional advertising, although it can also have negative consequences such as loss of credibility and skepticism among customers.

The fact that potential customers cannot physically evaluate the product features of the hotel, makes them look for reliable information and cues about the quality of alternative options, including images of the hotels, reviews and their membership to some quality club or platform with an independent ranking. Some customers are more inclined to read reviews than others and within those who read them, some value specific aspects of the reviews over other aspects.

The aim of this paper is to analyze factors influencing the traveler’s propensity to share travel experiences online, including reviews in OTAs (e.g., Booking, TripAdvisor), Google reviews and other Social Media platforms (e.g., Facebook, Instagram, Pinterest). Some authors [8] examined the effects of different motivations on the frequency of writing online reviews, focusing on altruistic and egoistic motivations. Their results suggest that customers are more willing to help other customers than to help the hotel or the restaurant. Mathwick and Mosteller [9] examine the factors that motivate online reviewer engagement, which is an attitude with varying degrees of altruistic and egoistic market-helping motives coupled with an individual’s intrinsic motivation to review when certain needs are fulfilled. They find three reviewer types with varying degrees of altruistic and egoistic motivations. Chen & Huang [10] explore how the characteristics of reviews and the reviewers might shape their reviewing frequency and continuity.

Magno et al. [8] studied guests’ attitudinal and behavioral reactions to received solicitations to write a review, a widespread strategy of some hotels, which is about soliciting their guests to write online reviews. This is a controversial technique, as it may produce a increase in number of reviews but it may also irritate a significant number of customers, specially if they are asked to write positive reviews. Gössling et al. [11] documented the increasing popularity of new strategies to manipulate online reviews.

The authors discern a research gap in the examination of the complex dynamics surrounding the influenceability of electronic Word-of-Mouth (eWOM) in the context of hotel bookings. Specifically, they identify an unexplored area regarding the interplay between the persuasiveness of eWOM, the underlying motivations that drive individuals to articulate their opinions through reviews, and the consequential association with the likelihood of actually composing a review. This is the journey from being a passive eWOM reader to a proactive eWOM writer (provider). This research gap underscores the need for a comprehensive investigation into how the influence of eWOM during the hotel booking process contributes to the motivations behind user-generated reviews and, subsequently, how these motivations influence the propensity of individuals to engage in the act of writing a review. Addressing this gap not only enhances our understanding of the intricate relationships involved but also holds significant implications for businesses seeking to leverage and manage online reviews in the hospitality industry. Hence the goal is to measure the relationship between the influence of reviews and the aspect that most influences them and the propensity to write reviews. Three aspects of the reviews are considered in the study, i.e., volume and percentage of positive reviews, quality of the reviews’ content and valence of the rating (score). The research questions are what the extent of the relationship is between the influenceability of eWOM in the context of hotel bookings and the underlying motivations to write reviews; and secondly, what is the nature of the relationship between these motivations and the ultimate propensity to write reviews?

Overall findings of this work provide relevant contribution to the growing body of research in eWOM in the hospitality and tourism sector and provide valuable assistance to practitioners in formulating better strategies to garner customer reviews.

The rest of the article is organized as follows: Section 2 is divided in six subheadings. The first five subheadings present a review of the literature that supports the theoretical framework for eWOM and the set of hypotheses are derived from related literature. The sixth subheading comprises the Conceptual Model. In Section 3 we show the foundations of the Methodology used to develop our model, based on Path Analysis over a set of questions which had significant relationships. In Section 4 we present the results of the model, showing that overall the model goodness of fit indicators are adequate and the particular relationships are highly significant. In Section 5 implications are discussed, presenting the limitations and future work, and finally Section 6 presents the conclusions.

2. Literature Review and Hypotheses Development

According to the academic literature on the buying process [12,13] the post-purchase stage is becoming increasingly important for the consumer and for marketing management and, in this sense, the motivations and implications of writing reviews that take place after the purchase process are critical. After a thorough review of scientific literature, Bore et al. [1] consider eWOM as the most influential source of information when consumers are making purchasing decisions regarding hotel offerings.

2.1. eWOM and Digital Communication

In the digital strategy of brands there are different types of media content, according to the source. The POEM (Paid, Owned and Earned Media) model is widely used in the digital marketing communications context, where eWOM represents an outstanding strategy to achieve earned media; contents that manage to spread through networks without the use of an advertising strategy and are more likely to gain trust and engagement from other customers [14,15,16].

On the other hand, the reputation of a website can be interpreted as an indication of quality, so that it offers greater credibility to the information published on it. This way brands can compensate for the risk perceived by consumers by encouraging a positive eWOM on sites with an established positive reputation to enhance their effects [17].

Llorente-Barroso et al. [18] mention the POEM method in their study of the impact of social networks on brand strategy, and the importance of the use of eWOM to achieve the coveted social earned media. Wu et al. [19] study the hybrid review format (text reviews accompanied by pictures) and find that hybrid reviews were perceived as more helpful than word-only reviews, explaining this effect by the perceived amount of information.

2.2. Relationship between eWOM and Hotel Booking Intention and Most Influential Aspects of Hotel Reviews

Consumers use eWOM to reduce uncertainty about their choice, but, also and very significantly, reading reviews increases the likelihood of the purchase decision [3,20]. Leong et al. [21] confirm the importance of eWOM and source credibility, among other factors, as influencing the booking decision.

The influence of eWOM on factors such as trust and attitude towards the facility has been studied and tested by several authors. Some authors like [20,22] establish a relationship between positive opinions and consumer trust towards the offer, especially if these include numerical ratings. Ladhari & Michaud [23] also find the effectiveness of positive reviews on attitude, quality perception and trust towards hotels. The eWOM also ensures correct choice and reduces uncertainty [24]. Some authors [25] recently studied the importance of the dispersion of numerical ratings, finding that dispersion negatively affects the booking decision. Formally, we propose the following hypothesis:

Hypothesis 1 (H1).

The level of influenceability by reading reviews has a positive effect on the influential factors when reading reviews.

2.3. Writing Online Reviews and Its Underlying Motivations

So far, we have reviewed and analyzed the literature on the evident influence of e-wom on consumer decision-making in the hotel industry, however, taking into account the enormous influence that positive e-wom exerts on consumer choice (and therefore, on the economic effects in the hotel industry), and the low rate of users who are willing to write their reviews [24], it seems appropriate to delve into the reasons why they do so.

The authors draw into the two-step communication theory applied to marketing, outlined in 1957 by Katz [26], who conducted pioneering research on the role of individual influencers in disseminating communication messages. Emphasizing the significance of perceived prestige, Katz et al. (2017) [27] assert that an opinion leader must be regarded as prestigious to exert influence over others. Katz [27] (1957, pp. 73–75) delineates the defining traits of opinion leaders, highlighting three key attributes. Firstly, opinion leaders possess the capacity to embody specific values. Secondly, they demonstrate expertise in particular domains. Thirdly, they maintain a social network that values their opinions and possesses a discernible size. This theory could be used to explain the influence of online influencers, but also to explain the influence exerted by online opinion platforms such as OTAs platforms and Social Networks, where those key attributes are embodied by the corporate brand (e.g., TripAdvisor, Booking, etc.). In this context, the power to influence is diffused in a larger amount of users. Social media has empowered customers as active players in the value formation process. This is true for the different Social media channels and in particular in OTAs platforms and Social Networks, which we study. Shin et al. [28] studied the helpfulness of hotel reviews looking into the relationship among review ratings and semantic and linguistic review features, finding that the interaction between readability and review rating was not significant. Alwash et al. [29] distinguish two levels of eWOM providing behavior, according to the required engagement: shallow and deep. At the shallow level consumers perform an action of low cognitive effort and require a short time to perform, such as “liking” and sharing existing content. On the other hand, the deep level involves higher cognitive effort and takes longer to perform, like creating or producing content, such as reviews and commenting on others’ eWOM communications [29]. Research by Ismagilova et al. [30] focusses on eWOM providing behavior at both levels. They do a meta-analysis to reconcile conflicting findings of the factors affecting consumers’ intention to engage in eWOM provision. Our research focuses on a deep level of engagement, in particular on providing eWOM hotel reviews, which we consider the most time consuming and most demanding in terms of cognitive effort. Naumann et al. [31] studied the customer engagement for both, positive and negative eWOM, and found involvement to be a strong driver of positive customer engagement at the time of providing eWOM content.

The main eWOM factors studied in literature affecting retailer sales are volume and valence [32]. Results of previous studies are conflicting, Rosario et al. [33] carried out a meta-analysis to examine the effect of eWOM characteristics, volume and valence on sales, finding that eWOM volume has a stronger impact on sales in comparison with eWOM valence. In the hotel sector, the most studied influential aspects of hotel reviews are review valence, review extremity, review volume and review quality [3,34] but the percentage of valence (positive/negative) seems to be the most relevant. This work tries to link the sensitiveness to reviews to the level of influence by the three main factors, volume, valence and quality of opinions, and at the same time clarify if the influenceability by any of those factors can drive eWOM provision. These effects have not been tested before in literature, being part of the novelty of this paper. Hence we propose the following hypotheses:

Hypothesis 2 (H2).

The level of influenceability by reading reviews has a positive effect on the underlying motivations to write a review.

Hypothesis 3 (H3).

The level of influenceability by reading reviews has a positive effect on the propensity to write a review.

Hypothesis 4a (H4a).

The influenceability by volume of opinions has a positive effect on the underlying motivations to write a review.

Hypothesis 4b (H4b).

The influenceability by quality of opinions has a positive effect on the underlying motivations to write a review.

Hypothesis 4c (H4c).

The influenceability by valence of opinions has a positive effect on the underlying motivations to write a review.

Already in 2012, Cheung & Lee [35] analyzed the different triggers for writing online reviews. The motivations studied were selfish (seeking return from others and reputation when writing opinions), collective belonging (they are integrated and feel part of a group), altruistic (not expecting return or personal benefit), principlism (moral motives such as justice), and knowledge self-efficacy (self-judgment ability). They conclude that a sense of belonging has the greatest weight in the tendency to write e-WOM, while reciprocity, moral obligation and self efficacy do not show a clear relationship with e-WOM intention.

Altruistic motives have been suggested as the main triggers to write opinions [7] although they may be somewhat promoted by monetary incentives. Magno et al. [8] study the influence on e-wom writing frequency of altruistic motives (helping other consumers) and add selfish ones (enjoyment and revenge towards companies). They conclude that managers should focus on customers writing reviews to help others, but not to help the company.

Recent studies [13] confirm that service quality positively influences customer satisfaction, but that neither service quality nor customer satisfaction is a guarantee that customers will be encouraged to express their opinion through eWOM. On the other hand, they do find a higher propensity in the case of having experienced positive emotions during the stay.

Nam, Baker, Ahmad & Goo [36] examine the motivations for writing e-wom (both positive and negative), and conclude that, in the case of positive e-wom, in addition to the aforementioned altruism, attachment factors (sense of belonging to an online community) and the enjoyment that comes from finding previous interesting opinions are also very relevant. In the case of the negative e-wom, the main triggers that the same authors explain are dissatisfaction with the product or service, and the existence of previous opinions that do not correspond to the reality experienced.

Cheung & Lee [35] point out that altruism refers to serving the public good to benefit one or more others. The motive for the public good can be linked to empathic emotion, being empathy (feelings of sympathy, compassion, tenderness, and the like) a source of altruism.

Therefore, both motivations, altruistic and hedonic, have been widely studied, but to compare which of the two has the most influential effect on writing eWOM, the following hypotheses are proposed:

Hypothesis 5a (H5a).

The altruistic underlying motivations to write a review has a positive effect on the propensity to write a review.

Hypothesis 5b (H5b).

The hedonic underlying motivations to write a review has a positive effect on the propensity to write a review.

The authors also investigated some mediating effects in the propensity to write a review, and to that end added the following hypothesis to the study related to the mediating effect of Q2 on the relationship between Q1 and Q4:

Hypothesis 6 (H6).

The most influential factor from hotel reviews mediates the relationship between the Level of influenceability by opinions at the time of booking a hotel and the Propensity to write a review.

Authors, in their aim of covering a research gap found in the literature review, added the following hypothesis related to the mediating effect of Q3 on the relationship between Q1 and Q4:

Hypothesis 7 (H7).

Underlying motivations to write a review mediates the relationship between the Level of influenceability by opinions at the time of booking a hotel and the Propensity to write a review.

2.4. Sociodemographic Factors: Age and Gender

Sociodemographics have also been studied as explanatory factors of the differences in influence and needs met by eWOM on consumers. Gender has been considered as one of the most influential [37,38,39,40]. Meyers-Levy and Maheswaran [41] notice a higher level of involvement (effort, therefore) of women in the whole process of information acquisition, which is why they are interested in the whole and each of the pieces and compare more thoroughly the information obtained. In contrast, men seem to be more selective [42] when capturing information and focus on those specific details that they find more useful [38] to process the whole and obtain the desired informative result.

In terms of age, Rakhit & Laohavichien [43] also find a greater ease and inclination of Generation Z to use eWOM, and Shome [44] also confirms a greater inclination of millennials to use eWOM versus traditional advertising for hotel and travel bookings.

Nam et al. [36] examine the motivations for writing e-wom (both positive and negative) and establish a relationship between age and positive e-wom. The younger the consumer, the greater the likelihood of positive opinions, possibly because they are used to sharing positive experiences in their social networks. In the same vein, age (Generation Z) explains the tendency of young people to share their opinions and experiences more than other age groups, as they need to be visible to their peers [45].

Khalifa et al. [46] proposed a model to explore the factors motivating Tunisian customers to communicate with eWOM. Their empirical investigation reported that eWOM is significantly linked to customer satisfaction, which is the result of expectation confirmation. In their results, this tendency is moderated by consumers’ altruism and age, also suggesting that females are more likely to communicate eWOM than males do.

Authors also investigated moderating effect of Sociodemographic factors (age and gender) in the present model, and to that end they added the following hypotheses to the study:

Hypothesis 8a (H8a).

Age is a moderating factor of the propensity to write a review.

Hypothesis 8b (H8b).

Gender is a moderating factor of the propensity to write a review.

2.5. Conceptual Model

The authors, after reviewing the available literature, find there is a research gap in measuring the relationship between the level of influenceability by eWOM for booking a hotel and the propensity to write a review. The conceptual model in Figure 1 measures the direct effect of the “level of influenceability by reading reviews” (Q1) on the “propensity to write reviews” (Q4), as well as the indirect effects of Q1 on Q4 through “Most influential factor from hotel reviews” (Q2) and “underlying motivations to write a review” (Q3). A special section is dedicated to the mediating effects of Q2 and Q3. Finally, the authors also study the moderating effect of Sociodemographic factors (Age and Gender). The hypotheses formulated in the previous subheadings are illustrated in the arrows of Figure 1.

3. Materials and Methods

3.1. Questionnaire Design

Prior to the creation of the questionnaire, a qualitative study was carried out consisting of two focus groups. The first brought together eight consumers between 18 and 30 years of age, and the second brought together ten consumers between 35 and 60 years of age. The inclusion criteria were individuals who have stayed in hotels at least once in the past year and were regular users of online platforms for reading or writing hotel reviews. Participants were recruited through leveraging personal networks, and discussions were facilitated using the moderation of one of the authors. Sessions were audio-recorded for later analysis. The topics discussed explored the participants’ motivations and experiences related to writing hotel reviews, such as factors influencing their decisions, preferred review content, and motivations for writing reviews. These initial discussions provided additional items for some of the key questions in the survey.

From this qualitative phase, we extracted the aspects to which they gave most importance, and which were helpful in the questionnaire design, as well as the revision of previous research, such as Wang et al. [47] and Hyun and Park [48] but adapted to the context of tourism, as in the research of Ek-Styvén and Foster [45] in which the need for uniqueness (NFU) and opinion leadership (OL) are prioritized. Particularly useful have been the guidelines collected by Wang [49] to answer questions about travelers’ factors for using eWOM and for generating eWOM content in turn.

A pilot test was conducted on 50 consumers of different ages and backgrounds to validate the questionnaire and scales, and, as a result, technical expressions that turned out to be poorly understood by the general population, were also corrected.

The questionnaire was structured in five sections: the first as a presentation and filter so that only those consumers who use the web to search for information on tourist accommodations could be consulted, and to detect on which specific platforms they read reviews. A second section focuses on the reviews consulted, and which aspects are the most relevant when selecting them (both positive and negative). The third section delves into the reasons why users read these reviews and the factors that they consider to have the greatest influence on their choice, such as quantity, type of reviewer, opinion ratings, etc.

A fourth section is dedicated to the core of the study, which are the factors that would promote a greater inclination on the part of the customers to write a review about their experience after their stay. The questionnaire ends with six socio demographic questions that allow the sample to be classified and analyzed according to different criteria.

3.2. Description of Variables in the Questionnaire

Among the many questions on the questionnaire, the authors selected the variables which were relevant for the study. Q1 and Q2 examine the motivations to read reviews. These factors include the volume of reviews, the characteristics of the reviewers, and the ratings expressed, among others. Q3 and Q4 frame the factors that could stimulate an increased propensity among customers to write a review about their experience following their stay. Q2 and Q3 are sets of items, 7 items in each of them, and an Exploratory Factor Analysis (EFA) was carried out to reduce their dimensionality.

Table 1. Variable definition.

Variable Name	Variable Definition
Q1	Level of influenceability by opinions at the time of booking a hotel
Q2	Most influential aspect from hotel reviews
Q3	Underlying motivation to write a review
Q4	Propensity to write a review

shows the variables definition.

3.3. Sample and Data

During the months of September and October 2021, an online self-administered survey was conducted among users of online tourism opinion platforms using a non-probabilistic, convenience sampling method [50]. This type of sampling is appropriate when the limits of the population under study are not known or are very broad [51].

The questionnaire was sent to an initial sample of 200 people in a capital city in Europe, over the age of 20 to 75 years and Internet users, who were asked to redistribute it to their contacts. A total of 788 completed questionnaires were obtained, of which 49 had some missing responses. This reduced the final number of valid surveys to 739.

Among them, 40.9% were male and 59.1% female. Age was concentrated in the ranges 22–30 (34%) and 31–45 (25%). Bachelor’s degree was the most common education level (47%). There were only 70 customers who never wrote reviews and 365 write seldom, the rest write either half of the times, often and always.

Means, standard deviations, skew and kurtosis of questions Q1–Q4 can be found in Supplementary Table S1.

The critical questions in the study are Q2 and Q3. See

Table 2. Items of Question 2: Most influential aspect from hotel reviews.

Items Q2	Item Definition
Q2_1	The number of opinions about it (whether good or bad)
Q2_2	The percentage of positive/negative opinions
Q2_3	Hotel rating (stars, number out of 10...)
Q2_4	The label (e.g., “fabulous”, “highly recommended”, etc.)
Q2_5	The content of the most favorable and least favorable opinions
Q2_6	How well written and substantiated an opinion is.
Q2_7	That includes additional content, such as photos, videos, etc.

and

Table 3. Items of Question 3: Underlying motivation to write a review.

Items Q3	Item Definition
Q3_1	Make a complaint about something I didn’t like
Q3_2	Congratulations for the good experience
Q3_3	Offer information to other customers to improve their choice and experience
Q3_4	To state that what I received exceeded my expectations
Q3_5	Provide feedback to the hotel on any area for improvement
Q3_6	Receive some reward in return (improve my rating as a reviewer, increase my number of followers, etc.)
Q3_7	Tell about my experience in Social Networks (Facebook, Instagram, Twitter, etc.)

for the list of items of Q2 and Q3 respectively. In order to assess the reliability of these scales, we calculated Cronbach’s alpha coefficient, which measures the internal consistency of a set of items, on a standardized 0 to 1 scale. The values obtained (0.723 for Q2 and 0.729 for Q3) can be considered as acceptable [52]. Quantitative assessment is not typically employed to evaluate validity. Instead, the evaluation involves a meticulous comparison of the measurement method with the conceptual definition of the construct.

An exploratory factor analysis (EFA) was carried out in questions 2 and 3. As a result, three factors were considered in both questions.

We conducted Bartlett’s Test of Sphericity to make sure that the correlation matrix of the items in Question 2 diverges significantly from the identity matrix. The obtained p-value (lower than 0.001) indicates that a data reduction technique is suitable to use. Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy (0.794) shows that these items are suitable for conducting factor analysis. Three factors were chosen, retaining 69.95% of variability and explaining all seven items (communalities between 0.53 and 0.77). Results can be found in Supplementary Tables S2–S5. Finally, interpretation of the factors was done after performing a varimax rotation (see

Table 4. Factor Analysis Q2: Correlations between items and factors component.

	1	2	3
Q2.1	0.861	0.147	0.055
Q2.2	0.807	0.193	0.243
Q2.3	0.296	0.024	0.770
Q2.4	0.022	0.220	0.830
Q2.5	0.421	0.526	0.282
Q2.6	0.163	0.789	0.004
Q2.7	0.085	0.733	0.167

Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. Rotation converged in 5 iterations. Bold represents the highest correlations for each factor.

). As a consequence, three influenceability dimensions on reading eWOM reviews unfolded from the survey responses: volume (Q2_F1), quality (Q2_F2) and valence (Q2_F3). These factors are also mentioned in the literature review [3,32].

Similar results were obtained for Q3, which are shown in

Table 5. Factor Analysis Q3: Correlations between items and factors component.

	1	2	3
Q3.1	0.183	−0.009	0.962
Q3.2	0.805	−0.030	0.150
Q3.3	0.810	0.169	0.128
Q3.4	0.840	0.056	−0.022
Q3.5	0.611	0.307	0.270
Q3.6	−0.034	0.884	0.089
Q3.7	0.263	0.810	−0.093

Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. Rotation converged in 4 iterations. Bold represents the highest correlations for each factor.

: p-value lower than 0.001 for Bartlett’s Test of Sphericity and a value of 0.763 for KMO measure of sampling adequacy, which is also considered acceptable. Three factors retain 72.91% of variability and all communalities were between 0.54 and 0.79, with a high value of 0.96 for Q3_F1. Results can be found in Supplementary Tables S6–S9. Again, three underlying motivation dimensions for writing a review on a hotel experience unfolded from the survey responses: altruistic (Q3_F1), hedonic (Q3_F2) and conflicted (Q3_F3). Some other factors mentioned in the literature review do not unfold in the EFA (i.e., selfishness or belonging). This may be due to a masking effect, as they are included in the main factors (e.g., Selfishness as part of the Hedonic factor; Belonging as part of the Altruistic factor).

3.4. Use of Path Analysis on eWOM Research

Path Analysis is widely used and accepted in the marketing areas: see for example [53]. More specifically, we can find it recently applied in the field of hospitality and eWOM. Uslu [54] has used it to relate the relationship between this form of communication, service quality, customer satisfaction and behavioral intention. Andriani et al. (2019) analyzed and explained how much influence eWOM has towards visit intention with destination image as mediating variable. Zhao et al. [55] studied eWOM and consumer purchase intentions in social e-commerce.

Path Analysis is an extension of multiple regression. It allows for the analysis of complicated models where there are several final dependent variables and “chains” of influence, as in our dataset.

The authors also studied the analysis of the possible moderating effect of categorical or ordinal variables (such as gender and categorized age). This moderating effect has been investigated using AMOS multi-group analysis, where six age groups were considered. Estimation of the multi-group effects was done using the so-called indirect method (using the Chi-square difference test).

The authors also investigated the potential mediating effects of each of the factors found in Q2 and Q3 on the effect of Q1 over Q4. To this end Path Analysis Mediation Model was used, as in Zhao et al. [53].

4. Results

4.1. Model Development and Estimation

Figure 2 shows the chains of influence of the Path Analysis Model, with its standardized path coefficients and p values.

Table 6. Fitting index table.

Measure	Result
CMIN/DF	2.11
GFI	0.994
CFI	0.98
IFI	0.981
TLI	0.941
RMSEA	0.039
NFI	0.964

shows the main Goodness of Fit (GoF) indicators. All indices have reached a standard level, which shows that the model fit is good.

With regards to the explained variance of the Path model, after estimating the model, we look for the standardized direct and indirect effects in the output, as AMOS does not provide the determination coefficient (R²). These effects represent the pathways through which variables influence each other. We divide the sum of these effects by the total variance of the dependent variable. The value obtained of R² for the final dependent variable (Q4) is 0.5739, meaning that the Path model explains 57.39% of the variance in the dependent variable. This is an acceptable value, taking into account that there are additional variables which could have been included in the model.

As can be seen from

Table 7. Path Analysis results.

Path			Unstandardized Coef.		Standardized Coef.
			b	SE	β	C.R.	p-Value
Q2_F1	<---	Q1	0.51 ***	0.051	0.348	10.088	<0.001
Q2_F2	<---	Q1	0.277 ***	0.053	0.188	5.201	<0.001
Q2_F3	<---	Q1	0.268 ***	0.053	0.182	5.027	<0.001
Q3_F1	<---	Q1	0.139 *	0.055	0.094	2.52	0.012
Q3_F2	<---	Q1	−0.142 *	0.058	−0.096	−2.455	0.014
Q3_F1	<---	Q2_F1	0.165 ***	0.036	0.164	4.53	<0.001
Q3_F2	<---	Q2_F1	0.198 ***	0.038	0.196	5.159	<0.001
Q3_F1	<---	Q2_F2	0.243 ***	0.035	0.242	7.013	<0.001
Q3_F2	<---	Q2_F2	0.142 ***	0.036	0.141	3.888	<0.001
Q3_F1	<---	Q2_F3	0.153 ***	0.035	0.153	4.444	<0.001
Q3_F2	<---	Q2_F3	0.097 **	0.036	0.097	2.664	0.008
Q4	<---	Q1	0.156 **	0.05	0.11	3.131	0.002
Q4	<---	Q3_F2	0.099 **	0.033	0.104	3.014	0.003
Q4	<---	Q3_F1	0.279 ***	0.034	0.292	8.244	<0.001

* p < 0.05. ** p < 0.01. *** p < 0.001.

, the standardized path coefficients of the level of influenceability by opinions at the time of booking a hotel (Q1) on the three factors of influential aspects when reading reviews (Q2) are 0.348, 0.188 and 0.182 respectively, p < 0.001 in all of them; this indicates that the level of influenceability by reading reviews has a positive and significant effect on the influential factors when reading reviews (H1 true).

In order to study the influence of the level of influenceability by reading reviews (Q1) on the underlying motivations to write a review (Q3) we need to focus on the coefficients associated to the two factors which are significant in the model (Q3_F3—Conflicted factor—was not significant). One of them is positive (Q3_F1, 0.094, p = 0.012) while the other one is negative (Q3_F2, −0.096, p = 0.014). Hence, H2 is only partially true.

The standardized path coefficient of the level of influenceability by reading reviews (Q1) on the propensity to write a review (Q4) is 0.110, p = 0.002, which shows that Q1 has a significant positive influence on Q4. Therefore, H3 is supported.

The standardized path coefficients of the influenceability by volume of opinions (Q2_F1) on the two factors of the underlying motivations to write a review (Q3_F1 and Q3_F2) are 0.164 and 0.196 (p < 0.001), which suggests a significant positive effect. Therefore, H4a is supported. Similar results are obtained when studying the effect of the influenceability by quality of opinions (H4b) and valence of opinions (H4c) on the underlying motivations to write a review.

The standardized path coefficient of the altruistic underlying motivations to write a review (Q3_F1) on the propensity to write a review (Q4) is 0.292, p < 0.001, which shows that Q3_F1 has a significant positive influence on Q4. Therefore, H5a is supported.

The standardized path coefficient of the hedonic underlying motivations to write a review (Q3_F2) on the propensity to write a review (Q4) is 0.104, p = 0.003, which shows that Q3_F2 has a significant positive influence on Q4. Hence, H5b is supported.

4.2. Mediating Effects

To determine the possible mediating effects on the relationship between Q1 and Q4, we used the mediating effect test method proposed by [54].

First, we fit the regression model explaining Q4 in terms of Q1. Afterwards we fit the path analysis of the mediation model introducing factors of both Q2 and Q3. Results can be found in Supplementary Tables S10–S12.

Figure 3 shows the results of the Path Analysis of the mediation model (first factor of Q2 on the effect of Q1 on Q4). Each path of the model is significant and, as expected, there is a clear decrease (from 0.251 to 0.192) in the estimate of the coefficient of the path Q1 to Q4. Hence, the first factor of Q2 has a mediating effect and H6 is supported. None of these results apply to the second factor and third factors (paths are not significant and the coefficients stay the same), so the second and third factors of Q2 do not have a mediating effect.

Similar results were obtained when studying factors of Q3 (see Figure 4). Only the first factor has a mediating effect: each path of the model is significant, and the estimate of the coefficient decreases from 0.251 to 0.158. Hence, the first factor of Q3 has a mediating effect and H7 is supported. None of these results apply to the second factor, so it does not have a mediating effect.

Figure 4 illustrates a summary of the standardized coefficients and p-values of the mediating effect of Q3 among Q1 and Q4.

4.3. Moderating Effects

To establish the moderating effect of age, we used AMOS to conduct multiple group analysis with an indirect method, using a Chi-square difference test. We observed no significant difference between the models (p = 0.150), suggesting that age has no moderating effect. Therefore, we declare hypothesis H8a as false.

Regarding gender, no significant difference was observed (p = 0.609), suggesting that gender has no moderating effect. Hypothesis H8b was declared false. Results can be found in Supplementary Tables S13 and S14.

5. Discussion

A model was built to explain the propensity to write a review in terms of a set of variables that includes: level of influenceability by opinions, underlying motivations to write and most influential factors at the time of reading reviews. The authors also investigated potential moderating and mediating effects.

The contributions of our research are applicable mainly in the field of consumer behavior regarding online reviews (eWOM), which is critical in the hospitality and tourism industry, being part of a social media Strategy to foster the Earned media of the hotel. The results allow a better understanding of the reasons that make a customer more inclined to write a hotel review, considering the underlying motivations to write and the most influential factor when reading reviews. Results are also relevant for related businesses such as restaurants and entertainment parks to name a few. Results of the hypothesis testing are summarized in

Table 8. Hypothesis testing.

Hypotheses Testing	True/False
Hypothesis 1 (H1). The level of influenceability by reading reviews has a positive effect on the influential factors when reading reviews.	True
Hypothesis 2 (H2). The level of influenceability by reading reviews has a positive effect on the underlying motivations to write a review.	Partially true
Hypothesis 3 (H3). The level of influenceability by reading reviews has a positive effect on the propensity to write a review.	True
Hypothesis 4 (H4). The influenceability by volume of opinions has a positive effect on the underlying motivations to write a review.	True
Hypothesis 4b (H4b). The influenceability by value of opinions has a positive effect on the underlying motivations to write a review.	True
Hypothesis 4c (H4c). The influenceability by quality of opinions has a positive effect on the underlying motivations to write a review.	True
Hypothesis 5a (H5a). The altruistic underlying motivations to write a review has a positive effect on the propensity to write a review.	True
Hypothesis 5b (H5b). The hedonic underlying motivations to write a review has a positive effect on the propensity to write a review.	True
Mediating effects
Hypothesis 6 (H6). The most influential factor from hotel reviews mediates the relationship between the Level of influenceability by opinions at the time of booking a hotel and the Propensity to write a review.	True
Hypothesis 7 (H7). The underlying motivations to write a review mediates the relationship between the Level of influenceability by opinions at the time of booking a hotel and the Propensity to write a review.	True
Moderating effects
Hypothesis 8a (H8a). Age is a moderating factor of the propensity to write a review.	False
Hypothesis 8b (H8b). Gender is a moderating factor of the propensity to write a review.	False

.

The analysis of our model results sheds light on a very critical issue in Digital Marketing and Social Media within the context of hospitality and tourism communication, which is the content creation by customers, what is called the Earned content in the POEM model, a good representation of it being eWOM reviews.

Regarding H1, we can state that the higher the level of influenceability of reviews over a customer, the more importance she attaches to the 3 factors of Q2 (i.e., volume of reviews, quality of reviews, valence of reviews), being volume of reviews (F1) the factor with the highest coefficient, meaning that is the one with larger impact. We can therefore conclude that achieving a large volume of reviews should be a critical goal for hotel managers.

The results of the effect of Q1 in Q3, tested in H2, gives clarity about the varied relationships between the level of influenceability when reading a review and the different factors identified as underlying motivations to write a review, resulting that H2 is only partially true, as the Hedonic factor (F2) has a negative coefficient, meaning that the higher the influenceability, the lower the hedonic motivation. The Altruistic factor (F1) has a positive relationship, meaning that the higher the influenceability, the higher the altruistic motivation. We can suggest that customers who are largely influenced by reading hotel reviews will be more likely to have an altruistic motivation. Finally, we observe no significant relationship with the Conflicted factor (F3), which was withdrawn from the figure.

The main finding of this research is the fact that the level of influenceability by reading reviews has a positive effect on the propensity to write a review (H3), hence we manage to link the reading and the writing of reviews using additional explicative variables which improve the simplest model, which has a direct relationship only. We also discovered indirect effects using a path analysis model. Furthermore, we discovered the mediating effects of Q2 (most influential factor from hotel reviews) and Q3 (underlying motivations to write) on the relationship between Q1 (level of influenceability by opinions) and Q4 (propensity to write a review).

The H4 is split in three strains according to each of the identified influential factors of Q2, i.e., volume, quality and valence of reviews. Being all these factors significant in the relationship between Q2 and Q3. Let’s recall that Q3 also has two factors from the EFA, i.e., altruistic and hedonic motivations to write a review. As all the coefficients are positive among the different factors, we suggest that the three influential factors from Q2 have positive effects on the two underlying motivations to write a review of Q3, nevertheless the relationship with the highest coefficient is from Quality of reviews to the Altruistic motivation, meaning that the increase in the influenceability from the Quality of reviews increases the Altruistic underlying motivation to write reviews.

The results in H5a and H5b are in line with previous studies of eWOM by Magno et al. [8] in indicating that altruistic motivation is the most important one when it comes to writing hotel reviews and consequently, managers should focus on creating positive emotions during the hotel experience and asking customers to share their experience in writing to help others, rather than to help the hotel.

The mediating effects studied in H6 and H7 indicate that the independent variable (Q1) and the dependent variable (Q4) are linked by the mediating variables (Q2 and Q3), allowing a better explanations of their relationship, meaning that the most influential factor when reading reviews (what inspires the most) and the underlying motivations to write a review (what the main driver is) are meaningful in the propensity to write reviews. By adding the mediating variables, we help explain the cause-and-effect relationship between the two main variables.

The moderating effects of age (H8a) and gender (H8b) were tested in the relationship between Q1 and Q4, considering the global model. In this research we find that age and gender do not change the effect that the independent variable “Level of influenceability by reading” (Q1) has on the dependent variable “Propensity to write review” (Q4). Hence managers should treat all customers the same in their pursuit of the priceless review, irrespective of their age and gender. These results do not align with some previous studies, such as [36] which stablished a relationship between age and positive eWOM, and [46] who suggest that age and gender are significant factors in eWOM writing, being women more prone to write reviews, according to their study.

It has been already discussed that customer satisfaction does not guarantee customers’ involvement in spreading positive eWOM [13] so the focus should be on customer profiling to identify those customers with the biggest potential to write. This does not mean that managers can neglect the provision of service quality and positive emotional experience that leads to customer satisfaction, being a necessary but not sufficient condition to generate eWOM; previous investigation and customer profiling is required. The right profile for writing reviews are those customers who declare that they have been influenced by reading reviews and in particular those who are more interested in the volume of reviews.

5.1. Contributions to Research and Managerial Implications

The results of this study contribute to the growing body of research in social earned media from a different perspective to previous studies. The effect of eWOM on decision making has been well investigated in the hotel context, and to a lesser extent also the motivations for providing eWOM. However, few studies have attempted to link the reading of hotel reviews to the writing of reviews [30,56]. This is very relevant in the context of eWOM as a key digital marketing strategy. In this regard this study successfully confirmed the significant effect of the sensitivity to reading reviews on the propensity to write them in turn. Hence the authors consider that from a theoretical perspective, this study enriches the literature on eWOM behavior as follows.

Firstly, this study complements the concept of personal conditions developed by Ismagilova et al. [30] where factors affecting eWOM providing behavior were divided into four groups: personal conditions, social conditions, perceptual conditions and consumption-based conditions. Within the personal conditions, the concept of influenceability by reading reviews used in this work comprises two factors considered in previous studies, opinion seeking and information usefulness. Opinion seeking is a behavior focussed on looking for eWOM communications [57]. According to Davis [58], information usefulness refers to how much information helps consumers make purchasing decisions. Previous research conducted by [59,60] demonstrated that perceived usefulness of information can have a significant impact on consumer behavior. Bobkowski [59] used information utility theory to show that perceived usefulness of information has a positive effect on the intention to share news. Therefore, the present study expanded on the previous framework by offering a comprehensive approach to investigate and categorize factors that are associated with eWOM behavior in different settings.

Furthermore, the present study integrated prior research and resolved conflicting findings, resulting in cumulative insights into eWOM behavior. Previous studies, such [7,57,61] examined factors that influence eWOM behavior but have yielded varying results. However, this study’s findings indicate that factors commonly investigated in prior research, which some previous studies did not find to be significant, such as altruism are indeed significantly related to eWOM providing behavior. On the other hand, this study found that mediating factors such as age and gender, which were considered relevant in other studies, do not have a significant relationship with eWOM providing behavior.

The results of this work offer researchers the chance to observe where ideas intersect and differ, creating opportunities for exploring and investigating new research questions in a broader context. This, in turn, facilitates the advancement of knowledge regarding the dissemination of online information.

With regards to managerial implications, there is a high reward in achieving social earned media through eWOM. When receiving customers in their hotels, managers may be able to identify in most cases the source of the booking if the reservation come from one of the OTAs where the hotel is listed, and in the case of a direct booking, they may find out the profile of their customer by asking simple questions like the online website where they first saw their hotel and if they perused reviews of the hotel, and most importantly to ask if those reviews which they read were of any help in their decision when making the reservation. Those customers who declare higher influenceability of their decision by reading eWOM reviews will be the most likely to write reviews of the hotel. Managers should encourage those customers suggesting the value of their future reviews to help other customers, rather than to help the hotel, fostering their altruistic underlying motivation to write, regardless if their reviews are positive or negative. This is in line with results from [8]. Within the field of emerging manipulation strategies, one noteworthy approach that has garnered increasing attention in the context of hospitality businesses involves the solicitation of reviews from guests [11]. Prominent review platforms like TripAdvisor have acknowledged the legitimacy of review solicitation, provided certain ethical guidelines are adhered to. These guidelines emphasize that solicitation should not be discriminatory, meaning that it should not exclusively target satisfied guests. Furthermore, ethical standards mandate that the solicitation process should not influence the sentiment or valence of the reviews. In other words, guests should not be explicitly prompted to write positive reviews (TripAdvisor, 2023), as it is considered fraud to “Selectively soliciting reviews (by email, surveys or any other means) only from guests who have had a positive experience”.

This work also confirmed that the volume of reviews seems to be the most relevant factor when deciding to book a hotel, pointing at the importance for hotel managers of achieving a large number of reviews. This result is aligned with [34]. We also suggest that the altruistic underlying motivation to write reviews is more prevalent among those customers who are influenced by the quality of reviews at the time of booking their accommodation.

5.2. Limitations and Opportunities for Further Research

Our research has limitations that can be addressed in future research. First, our sample distribution is uneven. Since there is a bias in Education and Level of Income, the moderating effect of age and gender may be disguised and can be an object of study in a more even sample to investigate if age and gender are moderating factors or not. In this context there is a chance that higher income and higher education has a homogeneity effect in customer behavior regardless of age and gender, which may not exist in the general population.

Another study limitation is the non-probability sampling due to the gathering of survey responses with a snowball approach, where initial contacts were asked to redistribute their questionnaire to their own contacts. This sampling strategy limits the generalization of results.

At the same time, it will be valuable to know on which platforms the reviews are posted. Whether it was simply sharing a photo on Social Networks (e.g., Instagram or Facebook) with a comment or in a more conventional hotel review platform (e.g., TripAdvisor, Expedia or Booking). These are very different channels that would require individualized treatment.

Future work will include the development of some new lines of research, including the development of field work in a different country, like the UK or USA, in order to compare results.

6. Conclusions

This research introduced the relationship between reading and writing hotel reviews in the context of eWOM as part of the social media communication strategy of hotels. Although some of the hypotheses may seem evident, this research has confirmed some of them only partially and some moderating effects were rejected. Authors have demonstrated the importance of some mediating variables and identified the most important factors driving the propensity to write reviews. They also have shown that propensity to write reviews is not related to sociodemographic factors but to attitudinal factors such as the most influential factor when reading and underlying motivations to write.

We expect the importance of this topic to increase for marketers in hospitality and tourism as social media continues to gain favor among consumers and hotel eWOM drives decision making at the time of booking an accommodation.