Generating Electronic Word of Mouth (eWOM) in the Accommodation Sector

Abstract

Electronic word of mouth (eWOM) is a powerful form of online communication that strongly influences consumer purchasing behaviour. However, what remains less clear is the combined influence of situational factors versus personality traits when assessed simultaneously. The aim of this paper is to address this gap by developing an integrative conceptual model to assess the comparative relevance of situational factors and personality traits in driving eWOM generation in the Romanian accommodation sector. To implement the research scope, an empirical, quantitative, questionnaire-based investigation was pursued, data being collected from 291 tourists who had previous experience with online accommodation platforms such as booking.com, Airbnb, Trivago, etc. Based on the proposed conceptual model, data were analysed by means of structural equation modelling via SmartPLS 4.0. The research extends previous knowledge based on the Theory of Reasoned Action (TRA) and the Theory of Cognitive Dissonance (TCD), showing the combined multiple effects of situational factors and personality traits on consumers’ behaviour in generating eWOM in the accommodation sector. The results show that acquisition regret strongly drives eWOM generation intention, this regret being significantly increased by the unpleasantness, unacceptability, and importance of the consumer’s situation. Consumer expressivity predicts eWOM generation and is positively influenced by perceived social support, a relationship newly validated in the literature.

1. Introduction

Electronic word of mouth (eWOM) constitutes a widely known marketing communication phenomenon, whose impact on consumers has been studied exhaustively [1]. Negative or positive eWOM impacts strongly companies’ growth [2]. eWOM is a tool that helps companies, products, and/or brands to both gain customers’ attention and trust and also to disseminate messages and news faster than any other means of communication paid for by the company [3]. Given the large amount of time people spend online, studying eWOM has become increasingly important, as it influences the spontaneous consumption of products [4]. Consumers randomly exposed to eWOM end up spontaneously interacting with a product, which makes them more likely to buy it [5].

With increased access to online information, consumers are able to instantly compare and/or obtain prices, characteristics, quality, and data on peer experiences regardless of the field of activity from which the goods emanate. The power of traditional communication in determining consumption has declined considerably, alternative sources like eWOM becoming more important in the emerging technologies era [6]. eWOM has high credibility among consumers, is free, and has great potential to go viral. This is why it is necessary to investigate the factors that lead consumers to generate eWOM [7], especially in the accommodation sector, through the online environment. With the development of multiple online platforms on which consumers can freely express themselves, it is of utmost importance to recognise the role played by these factors in generating positive eWOM [8] in the Romanian accommodation sector.

Contemporary research has analysed the impact of eWOM on purchase intention, but much less attention has been paid to the determinants that generate eWOM in the field of accommodation. When the antecedents of eWOM were analysed, emphasis was placed on situational factors (consumer satisfaction, brand trust, quality of experience). Even when personal factors were studied, they were investigated separately from situational factors [9].

The quality of the arguments presented in eWOM was also identified as a source of persuasion [10]. eWOM, along with other influencer-type marketing, determines the credibility that consumers attach to brands, products, or companies [11] and has a strong impact on purchase intention [5]. This impact is strongly determined by the credibility of eWOM, numerous factors affecting this credibility, such as those relating to source, platform, message form, brand, and industry [12], while others lead to eWOM generation [13]. The previous literature has mainly focused on two separate streams: on the one hand, it has analysed situational factors (i.e., scarcity [14], consumer satisfaction [15], or quality of the experience [16]; on the other hand, the literature has investigated personal factors and personality traits (i.e., extraversion [17], self-confidence [18] or leadership [19], source attractiveness and/or credibility, and persuasiveness of an eWOM message [10,20,21,22]). What has been less determined is the combined influence of both directions, specifically how situational triggers and stable personality traits jointly predict a consumer’s intention to generate eWOM [23]. A model that simultaneously examines the influence of both situational (relating to the consumption experience) and personal factors (consumers’ characteristics regardless of their consumption experience) in generating eWOM intention in the accommodation sector, has not been yet developed. Therefore, the current paper tests the synergic influence of situational factors and personality traits in eWOM generation.

To cover the gap in the literature, a conceptual model, including both personality traits and situational factors, was built. Constructs such as unpleasantness, unacceptability, and importance are included in the situational factor category mirrored by the Theory of Reasoned Action (TRA-based constructs), with a direct influence on acquisition regret and finally on eWOM intention, while self-esteem, perceived social support, risk prevention, and risk promotion are personality traits mirrored in the Theory of Cognitive Dissonance (TCD-based constructs), impacting expressivity, which finally leads to eWOM generation. Furthermore, the eWOM literature typically analyses online travel agencies, while shared economy research is heavily focused on North American and Western European markets [24]. By focusing on Romania, this study helps to fill the geographical and cultural gap, providing valuable insights from an Eastern European perspective.

This paper aims to determine the synergic effect of situational factors and personality traits on sustaining eWOM intention in the Romanian accommodation sector. The analysis is integrated into the theoretical framework of the Theory of Reasoned Action (TRA) and Theory of Cognitive Dissonance (TCD). To implement the research question, the literature was reviewed, and hypotheses, along with a conceptual model, were developed, reflecting the effects of situational and personal factors in sustaining regret for a given consumption situation, increasing consumer expressivity, and their impact in generating eWOM intentions. The results and implications are multiple, ensuring a better understanding of the eWOM intention, determining the impact of situational factors and personality traits on eWOM generation, and showing their specific role in influencing consumer’s attitudes and behaviours. The research proves that all situational factors play a role in determining eWOM intention and also that they are more important than personality traits.

The novelty of this paper lies in its integrative approach, which simultaneously assesses the combined influence of situational factors and personality traits on eWOM intention in the accommodation sector of an emerging country. The existence of such a conceptual model allows us to understand not only the factors that play a role in generating eWOM but also their comparative relevance. The findings suggest that all relevant situational factors are strongly associated with eWOM intention, and within the model, they demonstrate a stronger relative influence than personality traits.

The paper is structured as follows: The introduction section is followed by the literature review in Section 2, including the theoretical framework, hypotheses, and conceptual model development. Section 3 presents the research methodology, followed by the results in Section 4, the discussion in Section 5, and finally the conclusions in Section 6.

2. Theoretical Framework, Hypotheses, and Conceptual Model Development

2.1. The Theory of Reasoned Action (TRA) and of Cognitive Dissonance (TCD)

The Theory of Reasoned Action (TRA) aims at understanding the attitude–intention–behaviour relationship of individuals; it postulates that a person’s intention to perform a certain behaviour is determined by their attitude towards the targeted behaviour and also by their expectations regarding the performance of the behaviour [25]. The TRA was later revised and expanded to overcome any discrepancies in the attitude–behaviour relationship with the Theory of Planned Behaviour, which also includes perceived behavioural control [25]. The TRA originally came from the expectancy-value theory, which highlighted the achievement motivation of individuals in education [26].

Attitudes and subjective behavioural norms explained by the TRA are often associated with eWOM [27,28,29] because they are major elements in influencing the intentions of a person to engage in a behaviour such as sharing or acting upon eWOM. The positive or negative feeling a consumer has towards eWOM influences whether or not to adopt eWOM in their life (attitude) [30]. The extent to which an individual feels that using eWOM brings value to their life is known as subjective norm. The TRA suggest that both attitude and subjective norms directly shape a person’s intention to perform a behaviour [31]. Consumer attitude and subjective norms have a direct influence on the eWOM–purchase intention relationship. Both determinants are significant in predicting whether a consumer will buy a product after encountering eWOM or not [32,33].

Cognitive Dissonance Theory (TCD) posits that a person’s actions are determined by the psychological need to maintain consistency between their attitudes and perceptions of reality and ongoing behaviours, referring to a situation involving conflicting attitudes, beliefs, and behaviours [34]. This need drives the individual to engage in any effort, following the path of least resistance to lessen the dissonance between these states. A person who experiences internal inconsistency tends to become psychologically uncomfortable and is motivated to reduce the cognitive dissonance. Attitudes serve as action tendencies and those efforts to reduce dissonance enable the individual to engage in decisive action [35].

Cognitive dissonance is the mental discomfort experienced when a person has two conflicting beliefs: to reduce this discomfort they need to change their thoughts or actions to justify their behaviour. Cognitive dissonance appears when there is a conflict between what individuals believe and how they act in a certain situation. This inner conflict creates tension, which forces people to change their behaviour or to adjust their beliefs [36]. The literature explains various phenomena and processes by means of cognitive dissonance, applied, for example, to the behaviour of investors in financial markets [37], consistency in leadership [38], increasing customer satisfaction, and encouraging loyalty under the influence of trust and cognitive dissonance [39].

Reducing cognitive dissonance can also be achieved by seeking social support. The dissonance between beliefs that contradicts objective and subjective reality, following a certain event that disproves those beliefs can be diminished by obtaining confirmation from third parties about the validity of the disproved beliefs. For instance, people convinced that an apocalyptic event is to occur on a fixed date, after that time has passed, have radically increased their proselytising behaviours, trying to convince their peers to join them, even though previously before the apocalyptic event, their interest in proselytising was relatively low [34]. Cognitive dissonance does not manifest itself in irrelevant cognitions, the lack of relevance being a classic mechanism for reducing cognitive dissonance [40]. Conversely, dissonance manifests itself when an important decision is at stake for the consumer. This social support-seeking mechanism is exactly what drives consumers to generate eWOM. When they have a consumption experience that is dissonant with their expectations (positive or negative), they seek to validate it by supporting others, seeking to gain a reaction from them. Under conditions of high cognitive dissonance, consumers are more likely to spread negative eWOM, but there is a higher probability that they falsify information to exaggerate negativity and lie through eWOM [21]. Users of social media platforms often experience discomfort, as they are sometimes exposed to information that contradicts their own knowledge or previous experience, hence the appearance of cognitive dissonance [41]. To influence online purchase intention and to improve consumers’ digital engagement strategies [42], companies should focus on reducing cognitive dissonance. If left unmanaged, dissonance can transform into regret or a feeling of injustice, leading to negative online review behaviours (negative eWOM) [43]. Thus, the strategic use of visual eWOM or social support is crucial to reduce this dissonance and prevent negative eWOM.

Cognitive dissonance is often to be seen as influencing individuals’ behaviour, having the power to change their attitude and ultimately their behaviour [44]. Human attitudes and beliefs lead to specific behaviour (supported by the TRA), while conversely, behaviour leads to attitude changes (supported by TCD). The TRA is a predictive theory of human behaviour because the main scope of the model is to explain and predict behaviour by measuring the intention of an individual to engage in a certain action [25]. The TCD is a determinant of changes because it explains how individuals are motivated to change their attitudes and/or behaviours to reduce the psychological discomfort caused by holding conflicting attitudes and/or behaviours. In the case of eWOM, from the perspective of the TRA, if a consumer believes that eWOM represents an important communication tool (attitude) and is able to transmit important information to other consumers (subjective norm), they will have strong intentions to practise eWOM [45]. If a consumer who believes that eWOM is good (attitude) does not transmit information through eWOM (behaviour), they will experience dissonance (the Cognitive Dissonance Theory).

For both the TRA and TCD, the association between cognitions that induce dissonance and eWOM has been previously investigated in the literature. By generating eWOM, healthy-oriented consumers tend to reduce their dissonance [46]. Being confronted with a negative consumption experience could cause a shift in consumers’ attitudes regarding the respective behaviour [47]. The TRA predicts that when exposed to a negative experience, consumers end up regretting their purchase decision, while the TCD would predict that once the consumers start to regret their purchase experience, they will seek ways to externalise that regret, including public displays of regret through negative eWOM. Both theories can also explain variation in intention according to personal factors.

Considering that both the TRA and TCD have, as a pivotal element, consumer’s attitude, the variation according to personal factors is explained because the attitude of an individual is strongly influenced by various personality traits. The TRA has been criticised for not taking sufficient account of unconscious factors [48]. Therefore, new research based on the TRA and TCD has investigated, by means of implicit association tests, the role of cognitive dissonance; when an internal psychological conflict regarding a consumption decision exists, cognitive dissonance is generated [49,50].

2.2. Hypotheses and Conceptual Model Development

eWOM, unlike other forms of mediated communication, represents a person-to-person form of communication that can strongly influence consumer buying behaviour [51]. Recently, more and more attention has been paid to eWOM. As the nature, quality, and structure of the arguments included in eWOM affects purchase intentions, social media platforms play a crucial role in developing the credibility of eWOM [20] in all fields of activity. Brand image is often a mediator of eWOM effects on purchase intentions [52]. Moreover, in crisis situations, when an organisation is confronted with a public scandal or suffers some controversy, eWOM has become one of the main pillars of successfully combatting image loss, mainly due to consumers’ trust in it [51]. Overall, the relationship between eWOM and company performance has been well studied in the literature; positive consumer experiences are reflected in eWOM, which in turn facilitates loyalty towards the company. Negative consumer experiences often lead to negative eWOM which, in turn, decreases consumer loyalty [53].

eWOM is influenced by situational factors and personality traits. Situational factors are those concerning consumption situations and evaluation by consumers, including regret, unpleasantness, unacceptability, and importance [41]. The present research assumes that situational factors such as unpleasantness, unacceptability, and importance influence consumers’ acquisition regrets, which have an impact on eWOM intention. Regret often plays a role in generating eWOM [22,54], being based on negative emotions caused by unpleasant circumstances or by situations beyond a consumer’s control, sometimes involving self-blame. Consumers’ regret has a significant effect on negative eWOM [55]. Regret relating to external factors generates anger, which in turn leads to negative eWOM [56].

Regret associated with internal factors and attributed to one’s own mistakes generates sadness, which, in turn, does not favour negative eWOM [57]. Therefore, regret appears to be a crucial psychological consequence of negative situational factors and, in turn, becomes a primary driver for generating eWOM [54]. This context can be linked to the Theory of Reasoned Action as an individual’s voluntary behaviour (writing positive/negative reviews) that generates that individual’s action (eWOM). Based on these arguments, the following can be stated:

Hypothesis 1 (H₁). 

Acquisition regret will positively influence consumer’s eWOM intentions.

Unpleasantness reflects how unsatisfying the consumption experience is for the consumer [58], a negative consumption experience decreasing repurchase intention [59]. Unpleasantness is a TRA-based situational factor associated with eWOM. Consumer satisfaction with the consumption experience represents an antecedent of both positive and negative eWOM. Analysing the evaluation of various attributes relating to food services (taste, atmosphere, authenticity), it was discovered that negative consumption experiences lead to negative eWOM and that positive consumption experiences lead to positive eWOM [58]. When consumers feel unsatisfied and regret their purchases, they will write negative online reviews [22]. This trend may also apply to the accommodation sector. The following hypothesis is defined:

Hypothesis 1A (H_1A). 

Unpleasantness will positively influence acquisition regret.

Unacceptability means how mismatched the consumption experience is with consumer expectations [60]. Unacceptability is also a TRA-based situational factor associated with eWOM. The effect of acceptability on eWOM has not been analysed to date, as an unpleasant experience is automatically unacceptable, and a pleasant experience is automatically acceptable. It is possible for an experience to be perceived as unpleasant and yet, for various reasons, the consumer finds it acceptable (offering of compensation, expecting the experience to be unpleasant, etc.). Disconfirmation is a phenomenon where a consumer reads reviews that are contrary to their experience and therefore contrary to their expectations [61]. Disconfirmation was also found to have an effect on eWOM, because of the consumers who read long and more detailed reviews, contrary to their expectations [15]. Unacceptability, or failure to meet expectations would, in turn, generate eWOM [60] in the accommodation sector. The next hypothesis is stated:

Hypothesis 1B (H_1B). 

Unacceptability will positively influence acquisition regret.

eWOM is an important communication tool in the accommodation sector online environment. The quantity and quality of eWOM has a positive influence on the purchase intention of accommodation services, while the credibility of eWOM has no significant impact on purchase intention [62]. When eWOM is considered as being credible and useful, it positively affects consumer attitudes and purchase intention [63]. The importance of eWOM in the online environment is growing because the Internet plays a major role in most people’s lives [64]. eWOM exists in most online media, supporting users in generating their own content [21]. Crucially, it influences the buying behaviour of individuals, importance being a situational factor reflecting how relevant the consumption experience is for people. eWOM affects the whole consumer behaviour [65] and helps individuals to make wise decisions in order to avoid the risk of having regrets after purchasing, i.e., accommodation services. Therefore, the following is presumed:

Hypothesis 1C (H_1C). 

Importance will positively influence acquisition regret.

Situational factors are always context dependent, while personality traits are those that remain stable to consumers regardless of the consumption situation [66]. Among these, one can include expressivity, self-esteem, regulatory focus and perceived social support [67]. These variables are considered personality traits, of which the construct ‘expressivity’ deserves special attention because it acts as mediator between personality traits and eWOM intention, being a TCD-based variable [68]. Personality, expressiveness, and individual ability to become close to other people, combined with cultural factors, are all relevant in generating eWOM towards a product, brand, company, or consumption situation [69]. This relevance is also in relation to neuro-management decision-making and cognitive algorithmic processes in the technological adoption of mobile commerce apps [70], algorithmic predictive modelling, and customer behaviour analytics in the multisensory extended reality metaverse [71], together with consumers’ decision-making processes on social commerce platforms in terms of online trust, perceived risk, and purchase intentions [72]. Through interactions with other people, an individual tends to manifest and express their own identity [73]. Therefore, it is far easier for consumers to initiate eWOM as a form of identity expression. Furthermore, the expressiveness of an individual in the future dissemination of positive and negative information as a result of increased/decreased satisfaction in the acquisition/consumption of certain goods/services is a determining factor of the intention to inform as many individuals as possible of their acquisition experiences [74].

eWOM is more easily generated when consumers identify themselves with a certain situation, when they trust the brand or company, and when they are satisfied by what they have obtained [75]. Basically, once consumers try a certain product or brand or interact with an advertisement that is congruent with their identity, they will want to spread the word about that product, brand, company, or message. In such a moment, self-esteem is strongly activated, and the desire to share with other people the positive (or maybe negative) acquisition/consumption experience is extreme [76]. The more expressive a person is, the more they will want to show their identity and will be more inclined to generate eWOM [6]. Self-respect plays an important role in how a person acts in life in any given situation [77]. Therefore, it can be assumed that there is a close connection between self-esteem and expressivity. That is why the negative or positive quality of a consumer’s self-image has an impact on building eWOM [78] in the accommodation sector too. Therefore, the following hypotheses are stated:

Hypothesis 2 (H₂). 

Expressivity will positively influence eWOM intention.

Hypothesis 2A (H_2A). 

Self-esteem will positively influence expressivity.

Perceived social support, a TCD-based construct, represents the consumer’s perception of the aid and protection offered by peers in various circumstances and also by means of eWOM. Perceived social support constitutes a relevant factor in trust development towards others [79] and also plays a role in generating eWOM towards a company, its brands or products because it increases customer engagement and loyalty, leading to increased intention to share positive experiences regarding the company and its products. Social capital is an important variable in eWOM generation [75]. Perceived social support improves psychological well-being while moderating the adverse effects of social comparison on self-esteem [80]. Marketing communication practitioners have underlined the strength of social media marketing strategies in spreading eWOM in the digital environment [63]. Perceived social support is often the starting point for people to communicate [81], while emotions are always present in the social media architecture of communication [82]. If perceived social support facilitates a person’s overall openness to communication, it might also facilitate their openness towards eWOM and the different emotions generated by the communication channel. When people perceive a high level of social support, they tend to openly express their emotions. Conversely, individuals suppress their emotions when they perceive a low level of social support [81]. Therefore, the following is presumed:

Hypothesis 2B (H_2B). 

Perceived social support will positively influence expressivity.

Regulatory focus represents the consumer’s tendency to prevent or accept risk for a potential gain. In psychology, the Regulatory Focus Theory is defined as a framework that analyses the way individuals’ goal-oriented behaviour differs based on individual motivation, distinguishing between promotion focus (approach-oriented and based on positive outcomes) and prevention focus (avoidance-oriented and based on safety while avoiding negative outcomes) [83]. Regulatory focus uses two different systems oriented toward motivation—promotional and preventive—in order to determine how individuals approach positive goals and avoid negative goals [84]. Consumers who develop a regulatory focus oriented towards prevention tend to analyse their own communication efforts more carefully [85], while those who have a promotion-oriented regulatory focus tend to act more impulsively in certain contexts [86]. Clients who exert a regulatory focus oriented towards promotion are more likely to generate eWOM compared with those who highlight a regulatory focus oriented towards prevention [87]. Regulatory focus is a TCD-based personality trait, which has an impact on an individual’s well-being, but in the literature, there are few empirical investigations in this direction [88]. As a result, there is also a lack of studies that validate the relationship between promotion-oriented focus and expressiveness. A strong regulatory focus suggests a weak ability to react in a positive manner to novel and creative ideas, while individuals with a strong promotion focus tend to perceive more novelty and creativity [89]. The following can thus be postulated:

Hypothesis 2C (H_2C). 

Promotion-oriented regulatory focus will positively influence expressivity.

Hypothesis 2D (H_2D). 

Prevention-oriented regulatory focus will negatively influence expressivity.

Based on these arguments, the conceptual model from Figure 1 is proposed. This shows the connection between the studied constructs, grouped in constructs reflecting situational factors and constructs reflecting personality traits.

3. Research Methodology

3.1. Research Context and Design

The aim of this research was to identify the factors that generate eWOM; subsequent research objectives were to identify relevant (1) situational factors, (2) personality traits, and (3) their composite interactions in generating eWOM in the accommodation sector. Therefore, a quantitative-based survey research design was implemented. This choice relied upon the fact that the composite influence of situational factors and personality traits can easily be analysed, the chosen constructs depicting a higher construct validity. The authors relied for all concepts on scales that had already been validated by the literature and that could easily be adapted to serve the given research scope. Respondents had to assess the different statements on a seven-point Likert scale, ranging from total disagreement (1) to total agreement (7)—see

Table 1. Scales and measurement (items).

Item	Measure	Loading > 0.7	VIF < 3.3
Self-esteem (SE) adapted after [78].
SE1	Overall, I am satisfied with myself.	0.800	2.250
SE2	I have a positive attitude about myself.	0.833	2.524
SE3	I think I have enough positive qualities.	0.853	2.656
SE4	I am capable of doing things at least as well as others.	0.864	2.848
SE5	I feel I am a valuable person.	0.858	2.902
SE6	I consider myself to be as valuable a person as anyone else.	0.861	2.879
Perceived Social Support (PSS) adapted after [79,80,81]
PSS1	My friends are really trying to help me.	0.880	1.153
PSS2	There is a special person in my life who cares about my emotions.	0.812	1.550
PSS3	My family is willing to help me make decisions.	0.713	1.339
Expressivity (EX) adapted after [6,75].
EX1	I have friends with whom I can share sorrows and joys.	0.929	3.128
EX2	I can talk about my problems with my friends.	0.914	3.019
EX3	I can rely on my friends when needed.	0.933	3.236
Promotion-oriented regulatory focus (PORF) adapted after [83,87].
PORF1	When I was little, we did not ‘cross the line’, and we did not do things that my parents would not tolerate.	0.873	2.126
PORF2	I often achieved things that motivated me to work harder.	0.826	1.678
PORF3	When I was little, I followed the rules set by my parents.	0.902	2.446
Prevention-oriented regulatory focus (PREV) adapted after [83,87].
PREV1	I often do well with the activities I try.	0.837	1.557
PREV2	When it comes to accomplishing things that are important to me, I notice that I do not perform as well as I would like.	0.939	1.557
Unpleasantness (UPN) adapted after [22,58,59].
UPN1	I am unhappy with the room I received.	0.838	2.391
UPN2	My experience with this accommodation unit is a negative one.	0.901	3.130
UPN3	The ratio of quality–price is not what I expected.	0.894	3.217
UPN4	I want to stay somewhere else.	0.815	1.958
UPN5	This situation is very unpleasant for me.	0.837	2.250
Acquisition regret (AR) adapted after [22,54,55,56].
AR1	I want my money back.	0.850	2.393
AR2	I regret making this purchase.	0.860	2.254
AR3	I will do everything I can to change this situation.	0.844	2.142
AR4	I regret that I trusted the accommodation unit.	0.777	1.665
Electronic Word of Mouth (eWOM) adapted after [90]. I will…
eWOM1	…post on my social media page how unhappy I am about this situation.	0.707	1.917
eWOM2	…search the hotel’s social media page to leave a negative review regarding this situation.	0.853	3.022
eWOM3	…look for social media groups (accommodation/tourism) to warn other consumers not to stay here.	0.784	3.134
eWOM4	…pass on my displeasure about this accommodation to others.	0.722	1.494
eWOM5	…look for travel review apps/sites to leave a negative review for the hotel.	0.818	2.915
eWOM6	…leave negative reviews on the booking platform where I rented the location.	0.812	2.850
Unacceptability (UAC) adapted after [60].
UAC1	This situation is completely unacceptable.	0.827	2.263
UAC2	Accommodation establishments should be prohibited from having such practices.	0.883	3.151
UAC3	I have been deceived.	0.922	2.446
UAC4	I will complain to the owner about this situation.	0.898	2.394
UAC5	It is my right to have my situation remedied.	0.911	2.787
Importance (IMP) adapted after [21,41].
IMP1	It is especially important for me to be accommodated in an upgraded room.	0.785	2.286
IMP2	My values do not allow me to accept this situation.	0.860	2.946
IMP3	This situation causes me great discomfort.	0.910	2.981
IMP4	This situation is landing me in a state of stress.	0.873	2.900
IMP5	I am angry about this situation.	0.847	3.004

Note: VIF = Variance inflation factor.

.

This seven-point Likert scale was chosen, as the psychometric literature shows the advantage of increasing the points measuring opinion or attitudes [91]. For each of the considered constructs, several statements/items were chosen (Table 1). The questionnaire also contained some socio-demographic questions. Before implementing the main research, a pilot study on 30 respondents was conducted to properly assess the consistency of the statements and to pinpoint any ambivalent meanings. Small corrections were made, and the final questionnaire was then distributed.

The research context, Romania, was chosen, as the country ranks high globally in Internet speed, being ranked in the top 15 in 2025 for fixed broadband and in the top 25 for mobile Internet. Its strong performance is attributed to a competitive market and early adoption of fibre optic technology [92]. Romanians demonstrate high digital engagement and are among the most active social media users in Europe, with 68.6% of the total population being active users [93]. This high level of digital interaction suggests a population that is comfortable with, and actively participates in, creating and consuming user-generated content (UGC). As eWOM is a primary form of UGC (e.g., reviews, ratings), this population represents an ideal cohort for studying its impact. Platform adoption is currently still quite low in Romania but “rapidly increasing” [94], making this emerging and dynamic market in transition a compelling case study. This context allows for the investigation of the factors influencing how and why consumers are shifting from traditional booking methods to platform-based ones. It also provides a unique window to observe the specific role eWOM plays during the critical adoption phase—insights that might be missed in a mature, saturated market.

To implement the research scope, the respondents had to think about a negative consumption experience they had encountered concerning accommodation during the last three years. They were asked to assess the statements in the online questionnaire, thinking of an accommodation unit that did not meet the promoted and expected quality standards. The experiment was formulated as follows: ‘You booked an apartment through a booking service (Booking.com, AirBNB, Trivago, etc.) for a weekend stay. When you check in, you notice that contrary to the advertised pictures, the actual room you are staying in has outdated furniture, some defects on the walls and the TV does not work’. In this negative consumption situation, several of the consumers’ normal expectations are not met. Once exposed to this experimental situation, the respondents had to complete the online survey, which consisted of statements regarding situational factors and personality traits and also regret and their expressivity towards the situation.

3.2. Sampling and Data Collection

The online survey-based research was implemented in late 2022 via Google Forms. From more than 500 persons invited to take part, only 291 respondents were retained. Only respondents who had booked accommodation via a booking service like AirBNB, Booking.com, etc., during the last three years were included in the final sample, data with missing information being deleted. Only correctly completed questionnaires were considered for further processing. The final sample consisted of 195 women (67.01%) and 96 men (32.99%). Most respondents were young, 107 (36.77%) of them being between 15 and 20 years, 102 respondents were aged between 21 and 30 years (35.05%), 35 respondents aged between 31 and 40 years (12.03%), 27 respondents aged between 41 and 50 years (9.28%), and 20 being older than 50 years (6.87%).

Thirty-two respondents came from rural areas (11%), thirty-five from small urban areas with fewer than 10,000 inhabitants (12.03%), and two hundred twenty-four from large urban areas (76.98%). Forty-eight respondents were unemployed or still studying (16.49%); one hundred one were qualified with a bachelor’s degree (34.71%); sixty-eight were qualified with a master’s degree (23.37%); eighteen were employed in an executive, decision-making position (6.19%); seventeen were self-employed (5.84%); and thirty-four were entrepreneurs (11.68%).

4. Results

4.1. The Evaluation of the Measurement Model

The authors analysed the conceptual model shown in Figure 1 using structural equation modelling (SEM) in SmartPLS to investigate the developed hypotheses. The operationalisation of the constructs is presented in Table 1, which also contains the exact measures according to references. Data were checked for reliability and validity, all items having a loading above 0.7 as recommended by the literature [95,96]. The measurements (items) were also checked for collinearity, the variance inflation factors (VIF) being under the threshold of 3.3 [95,96]; thus, there is no multicollinearity. The items measure exactly what they are supposed to.

The reflective constructs from Table 1 were further analysed by computing reliability and validity indicators, such as Cronbach Alpha, the average variance extracted (AVE), and the composite reliability (CR). The results (

Table 2. Constructs and items: Cronbach Alpha, average variance extracted (AVE), and composite reliability (CR).

Construct	Cronbach Alpha > 0.7	AVE > 0.5	CR > 0.7
Self-esteem (SE)	0.92	0.71	0.94
Perceived Social Support (PSS)	0.73	0.65	0.85
Expressivity (EX)	0.92	0.86	0.95
Promotion-oriented regulatory focus (PORF)	0.84	0.75	0.90
Prevention-oriented regulatory focus (PREV)	0.75	0.79	0.88
Unpleasantness (UPN)	0.91	0.74	0.93
Acquisition regret (AR)	0.85	0.70	0.90
eWOM	0.88	0.62	0.92
Unacceptability (UAC)	0.93	0.79	0.95
Importance (IMP)	0.91	0.73	0.93

) show that all constructs are above the recommended thresholds [95,96].

In the next step, the discriminant validity was assessed for each construct by means of the Fornell–Larcker criterion (

Table 3. Discriminant validity analyses (Fornell–Larcker criterion).

Construct	EX	IMP	UAC	UPN	PREV	eWOM	AR	PORF	SE	PSS
EX	0.926
IMP	0.244	0.856
UAC	0.323	0.632	0.889
UPN	0.337	0.542	0.823	0.858
PREV	0.276	0.160	0.268	0.304	0.890
eWOM	0.379	0.597	0.554	0.415	0.121	0.784
AR	0.306	0.746	0.878	0.852	0.287	0.572	0.834
PORF	0.564	0.315	0.483	0.489	0.303	0.397	0.443	0.867
SE	0.467	0.319	0.416	0.376	0.186	0.410	0.388	0.743	0.845
PSS	0.866	0.327	0.396	0.394	0.295	0.411	0.374	0.640	0.571	0.805

Note: EX: Expressivity; IMP: importance; UAC: unacceptability; UPN: unpleasantness; PREV: prevention-oriented regulatory focus; eWOM: e-word of mouth; AR: acquisition regret; PORF: promotion-oriented regulatory focus; SE: self-esteem; PSS: perceived social support.

). All values are within the recommended thresholds [96], indicating that the constructs highlight discriminant validity.

The conceptual model (Figure 1) was further analysed by means of structural equations using SmartPLS 4.0 [97]. Therefore, the measurement model was assessed, followed by the structural model and the relations between the concepts. The bootstrap procedure was used to test the hypotheses and relationships between constructs, several hypotheses being supported with a significant and positive relationship based on t-statistics.

4.2. The Evaluation of the Structural Model

When analysing the collinearity of the constructs, it was found by means of the VIF that the highest value of the inner model is 3.257 < 3.3 (UPN → AR), so there are no multicollinearity issues. With a value of 0.064 < 0.08 (saturated model) and of 0.070 < 0.08 (estimated model), the square root mean residual (SRMR) pinpoints an acceptable goodness of fit. Self-esteem, perceived social support, prevention-oriented regulatory focus, and promotion-oriented regulatory focus explain 75.4% of the variation of expressivity (R² = 0.754). Unpleasantness, unacceptability, and importance explain 87.8% of the variation of regret (R² = 0.878). The acquisition regret and expressivity explain 37.3% of the variation in generating eWOM. All these values define a high predictive power of the structural model (Figure 2).

The results of the hypotheses testing are shown in

Table 4. The direct path coefficients of the structural equation model.

Paths	Path Coefficients	Standard Deviation	T-Value	p-Value	Hypotheses
AR → WOM	0.503	0.053	9.562	0.000 ***	H1-Confirmed
UPN → AR	0.379	0.040	9.583	0.000 ***	H1A-Confirmed
UAC → AR	0.375	0.041	9.103	0.000 ***	H1B-Confirmed
IMP → AR	0.304	0.026	11.868	0.000 ***	H1C-Confirmed
EX → eWOM	0.225	0.051	4.406	0.000 ***	H2-Confirmed
SE → EX	−0.075	0.053	1.400	0.162 n.s.	H2A-Rejected
PSS → EX	0.865	0.030	28.766	0.000 ***	H2B-Confirmed
PORF → EX	0.062	0.055	1.115	0.265 n.s.	H2C-Rejected
PREV → EX	0.017	0.030	0.548	0.584 n.s.	H2D-Rejected

Note: *** p < 0.001; n.s.: not significant; EX: expressivity; IMP: importance; UAC: unacceptability; UPN: unpleasantness; PREV: prevention-oriented regulatory focus; eWOM: electronic word of mouth; AR: acquisition regret; PORF: promotion-oriented regulatory focus; SE: self-esteem; PSS: perceived social support.

.

Hypothesis H₁ postulates that acquisition regret positively influences eWOM intention. The findings confirm previous research [22], which considers regret as being a relevant mediating factor between situational factors and eWOM [54]. Regret generates negative eWOM [55], but under the influence of external factors, it will influence consumers’ anger, impacting the negative eWOM [56]. The results (β = 0.503; T-value = 9.562; p < 0.000) pinpoint a strong, positive, and significant relationship between the two constructs, confirming hypothesis H₁. Hypothesis H_1A asserted that unpleasantness of the consumer’s situation will positively influence acquisition regret and that a high level of unpleasantness will generate a high level of acquisition regret. This relation was confirmed previously by the literature [58,59], which discovered a connection between a negative consumption experience and purchase intention, respectively, on a negative eWOM. The relationship between consumer satisfaction and the intention to write negative reviews was previously validated [22]. As the results (β = 0.379; T-value = 9.583; p < 0.000) indicate a strong, positive, and significant relationship between these constructs, H_1A is confirmed.

Hypothesis H_1B presumed that the unacceptability of the consumer’s situation will positively influence acquisition regret and that a higher level of unacceptability will generate higher acquisition regret. The effect of acceptability on eWOM has not been analysed to date, but the literature [60] defines unacceptability as how mismatched with consumer expectations the consumption experience is. A similar concept was further investigated in the literature [60]: disconfirmation, which has a negative effect on eWOM, when consumers read long and more detailed reviews contrary to their expectations. The results (β = 0.375; T-value = 9.103; p < 0.000) indicate that the relationship between the two constructs is strong, positive, and significant, confirming, for the first time in the literature, hypothesis H_1B.

Hypothesis H_1C assumed that the importance of the consumer’s situation will positively influence acquisition regret and that a higher level of importance will generate a higher level of acquisition regret. Importance is a crucial situational factor that can influence consumers’ purchasing behaviour [98]. The results (β = 0.304; T-value = 11.868; p < 0.000) indicate a strong positive and significant relationship; thus, H_1C can be confirmed, this hypothesis being validated for the first time in the specialised literature.

Hypothesis H₂ posited that consumers’ expressivity will positively influence WOM intention and that with a higher level of consumer expressivity, WOM intention will also be higher. Expressivity was not studied in the literature, only consumer identity (identity expression/expressiveness), demonstrating that the more expressive a person is, the more they will want to express their own identity, generating eWOM [6,75]. The results (β = 0.225; T-value = 4.406; p < 0.000) indicate that the relationship between these two constructs is positive and significant, confirming hypothesis H₂ for the first time in the literature.

Hypothesis H_2A assumed that consumer self-esteem will positively influence consumer expressivity and that with a higher level of self-esteem, expressivity will be higher. The literature [78] validated the relationship between consumers’ self-image and eWOM but not between self-esteem and consumers’ expressivity. The results (β = −0.075; T-value = 1.4; p = 0.162) indicate that the relationship between these two constructs is weak, negative, and insignificant, failing to confirm H_2A. Thus, H_2A was rejected.

It was hypothesised (H_2B) that consumer perceived social support will positively influence consumer expressivity and that with a higher level of perceived social support, expressivity will be higher. Perceived social support is a TCD-based construct that improves psychological well-being while moderating the adverse effects of social comparison on self-esteem [80]. The relationship between perceived social support and consumer expressivity has not been analysed in a direct way in the literature, even if perceived social support proved to be a developing factor of trust in the market [79] and an opening determinant to communication [81]. The results (β = 0.865; T-value = 28.766; p < 0.000) indicate that the relationship between these two constructs is moderate, positive, and significant, confirming, for the first time in the literature, hypothesis H_2B.

Hypothesis H_2C asserted that promotion-oriented consumer regulatory focus will positively influence consumer expressivity and that the level of promotion-oriented regulatory focus favours a higher increase in expressivity. People who exert a regulatory focus oriented towards promotion are more likely to generate eWOM compared with others [87]. Regulatory focus is a TCD-based personality trait impacting individual well-being, an under-researched construct in the literature [88]. Individuals with a strong promotion focus tend to perceive more novelty and creativity [89] while acting more impulsively in certain contexts [86]. The results (β = 0.062; T-value = 1.115; p = 0.265) indicate that the relationship between these two constructs is weak, positive, and insignificant, failing to confirm H_2C, which is hence rejected.

Hypothesis H_2D postulates that prevention-oriented consumer regulatory focus will negatively influence consumer expressivity and that with a higher level of prevention-oriented regulatory focus, expressivity will be higher. In the literature, the presumption is that people who have a regulatory focus oriented towards promotion are more likely to generate eWOM compared with those who have a regulatory focus oriented towards prevention [87]. Regulatory focus has received little attention in the literature [88]. People with strong regulatory focus show a weak ability to react in a positive manner to novel and creative ideas [89]. The results (β = 0.017; T-value = 0.548 and p = 0.584) indicate that the relationship between these two constructs is weak, positive, and insignificant, failing to confirm H_2D, which is not supported by the data.

5. Discussion

eWOM represents an advertising form rapidly and strongly influenced by digitalisation, consumers supporting its early spread. All factors identified exert both an individual and a synergic effect on eWOM in the accommodation sector, extending previous knowledge based on both the TRA and TCD. While the influences of situational factors such as unpleasantness, unacceptability, and importance (H_1A, H_1B, H_1C) on eWOM intention were fully accepted, the effects of personality traits (expressivity, perceived social support) could only be partially accepted (H₂, H_2B). The three hypotheses indicating a possible influence between two personality traits were rejected. The first of these was the relationship between consumer self-esteem and expressivity (H_2A), the second between promotion-oriented consumer regulatory focus and consumer expressivity (H_2C), and the third between prevention-oriented consumer regulatory focus and consumer expressivity (H_2D). The rejection of possible links indicates that the constructs do not develop interdependent relationships. The only TCD-based constructs that could influence each other were the mediator variable expressivity, which influences eWOM intention, and perceived social support, which influences expressivity.

Conversely, the only personality-related trait that demonstrated a powerful effect was perceived social support (H_2B), an inherently collectivist construct that measures the perceived strength of one’s in-group (e.g., family, friends) [99]. This finding suggests that the decision to express oneself (via expressivity) is less an act of individual assertion (driven by self-esteem) and more a behaviour contingent on the perceived backing of one’s social network. This cultural dimension offers a substantive explanation and highlights a valuable avenue for future cross-cultural research.

Acquisition regret concerning accommodation services positively influenced eWOM intention, a hypothesis tested and confirmed in the literature [22] and accepted as valid for Romania, an emerging country. In the proposed conceptual model, regret influenced eWOM intention [100]. The same was valid for the relationship between unpleasantness, unacceptability, and importance, which all positively influenced acquisition regret, which further impacted eWOM. The personality traits perceived social support and consumer expressivity positively influenced eWOM intention.

Hypothesis H_1B, which presumed that the unacceptability of a consumer’s situation positively influenced acquisition regret, and H_1C, assuming that importance positively influenced acquisition regret, were both validated for the first time in the literature. Furthermore, H₂, which posited that consumer expressivity positively influenced eWOM intention, and hypothesis H_2B, confirming that consumers perceived social support positively influenced consumer expressivity, were also validated. The findings suggest that the relevant situational factors were strongly associated with consumers’ eWOM intention in the accommodation sector, being more important than personality traits. Contrary to expectations, only expressivity mediated the effect of perceived social support on eWOM intention [101]. The theoretical model allows for a more complex understanding of eWOM intention.

Contrary to the literature [102], self-esteem, perceived social support, and regulatory focus, mediated through respondents’ expressivity, self-esteem, and regulatory focus, did not have a significant role on eWOM generation. This suggests that although taken individually, these factors may contribute to the intention to generate eWOM. Although surprising, the result is explainable, since perceived social support indicates how much the respondents felt that their family, friends, and relatives were willing to be with them, providing them with resources and psychological support [60]. Since this construct directly measured respondents’ relationship with peers, it is expected that this factor would be more relevant in causing them to seek support in the case of a negative consumption experience and to generate eWOM among their peers [103]. These findings show that the TRA was more relevant compared with the TCD in analysing consumers’ behaviour regarding the accommodation sector in Romania.

A notable finding was the rejection of hypotheses H_2A, H_2C, and H_2D, which linked self-esteem and regulatory focus to expressivity. Although the literature initially cited suggested that these traits should be relevant [78,98], the interpretation requires a more substantive theoretical explanation. The failure of hypotheses H_2C and H_2D (regulatory focus) may be explained by the contextual mismatch of the theory. Regulatory focus theory centres on how individuals pursue goals, either by seeking gains (promotion) or by avoiding losses (prevention) [104]. This framework was highly relevant before a purchase, during the decision-making phase. However, the experimental scenario placed the consumer in a negative post-purchase situation, where the ‘loss’ had already occurred [58]. The behaviour (eWOM intention) was no longer governed by goal pursuit but rather by cognitive dissonance reduction (TCD) or the need for emotional venting. As the psychological goal shifts from “goal-pursuit” to “dissonance-reduction,” it is logical that the personality traits governing the original goal (regulatory focus) would lose their predictive power.

Cultural Implications of the Research

The research results show that when a consumer of accommodation services is exposed to unfavourable, even negative situations, they will change their attitude towards the purchase and consumption process, tending at the same time to externalise their disappointment. This behaviour is specific to Romanians (and probably to other nationalities as well) and may stem from cultural differences between the world nations. Romanians are an open, sociable, communicative, hospitable people who have not given up on Christian values [105], which is why they are willing to share their experiences, whether good or bad, of accommodation and, implicitly, of tourism and other areas, with the intention to help their peers to make decisions in their own lives [106]. As shown by the research results, Romanians are an expressive people (expressivity), who freely share their opinions on one subject or another [107] and who place a high value on family, friends, and communication with peers (eWOM intention). That is why eWOM is so extensively practised in Romanian culture. It is not only about communication between people but also about self-respect (self-esteem), an important feature in Romanian culture [108].

The limited effect of certain personality traits may be explained by cultural moderators. The research was conducted in Romania, an emerging economy characterised by relatively strong collectivist values [109,110]. In such a cultural context, individualistic traits like self-esteem (H_2A) and personal goal orientation (Regulatory Focus, H_2C, H_2D), which are often validated in Western-centric literature [111,112], may exert less influence on social behaviours such as eWOM. Conversely, the only personality-related trait that mattered was perceived social support, an inherently social construct. This suggests that in this cultural setting, the decision to express oneself is less an act of individual assertion and more a behaviour contingent on the perceived backing of one’s social network [113].

6. Conclusions, Policy Implications, and Research Limitations

From a theoretical perspective, the paper extends the TRA and TCD by developing a novel conceptual model, which clarifies the mechanism through which situational factors operate, specifically capturing the indirect effect of unacceptability (H_1B) and importance (H_1C) on eWOM generation intention. It is proven that expressivity influences eWOM intention, while perceived social support generates, through consumer expressivity, the eWOM intention. The influence of expressivity on eWOM intention was accepted and validated as well as the mediation effect of acquisition regret on the relationship between unpleasantness, unacceptability, importance, and eWOM intention. While robust, both theories have known limitations. The TRA has been criticised [114] for not sufficiently accounting for unconscious factors or behaviours not under full volitional control, which led to its extension into the Theory of Planned Behaviour. The TCD, similarly, has faced criticism for the difficulty in directly measuring the internal psychological state of ‘dissonance’ and for its broadness [35]. The model attempts to address these criticisms by using concrete, measurable antecedents (e.g., situational factors) as proxies for dissonance triggers and stable personality traits as antecedents of attitudes.

From a managerial perspective, understanding how eWOM intention is generated among Romanians could allow for marketing communication managers from the accommodation sector to model the relevant parameters for professional success in communication (regardless of the industry) and help to better predict the consequences of certain marketing decisions as well as estimating their potential impact. If personality traits had been identified as more relevant, marketers should have focused more on their consumer segmentation efforts to try to exclude, or at least isolate, those who, due to their personality, have a higher potential to generate negative eWOM. Since situational factors were more relevant, marketing communication practitioners should focus more closely on the marketing mix and their promise to blunt the generation of negative eWOM. Given that perceived social support plays a relevant role in the accommodation sector, it would be useful for professionals to approach this segment differently if they can identify it, as it is more likely to generate negative eWOM. The research findings offer more targeted managerial guidance. The model shows that a guest’s perceived social support (the feeling that they have a strong network) and their general propensity to be expressive are significant predictors of their intention to write online reviews. This provides a clear directive for hospitality managers: instead of focusing on abstract psychographics, staff should be trained to identify behavioural cues that signal a guest is highly expressive and feels well-supported by their social network. These customers represent a higher risk for negative eWOM (or a higher opportunity for positive eWOM).

Regarding policy recommendations, the way in which eWOM is generated is the basis for developing more complex models to provide a deeper understanding of eWOM. Once the importance of personality and situational factors are established, a more complex model could analyse the synergic role of temporality, brand, and other factors, culminating in the breakdown of these factors in the final purchase intention. Considering how consumers differ, it is necessary to start from the factors that generate consumer attitudes to better understand their purchasing process. For practitioners, the findings suggest investing in immediate service recovery rather than complex personality-based segmentation. Since the negative situation (via regret) was a much stronger driver of eWOM than stable traits, managers should empower staff to identify and resolve guest complaints on the spot to neutralise the dissatisfaction before it becomes a negative review. Furthermore, staff can be trained to identify high-risk guests—such as those travelling in close-knit groups or those visibly active on social media during their stay—as these individuals are likely to possess the social network and expressiveness that predicts eWOM generation. A proactive satisfaction check with these specific guests can be a highly effective intervention.

The research has several limitations. Among these, the socio-demographics of the sample might constitute a limitation, as the respondents were either young people aged between 15 and 35 or adults between 36 and 60 years. Respondents from lower social classes were underrepresented in terms of occupation. A total of 51.1% of the respondents were students, 17.8% were entrepreneurs or freelancers, and 30.1% were employed in different positions. Only 1% were unskilled workers, which may indicate that the results are not generalisable to the lower social class. Nevertheless, as the consumption situation was familiar regardless of respondents’ social class or occupation and there was no reason to suggest that unskilled people would consume accommodation experiences differently compared to skilled people, this does not constitute a significant problem. Occupation does not necessarily indicate social class, as in the given country, Romania, unqualified employees might earn more than their peers with a higher educational background. Another limitation concerns the fact that data were only collected via an online survey, which might explain the prevalence of younger people (93.1% of respondents being under 50 years). However, considering that the research mainly analysed how eWOM is generated, this limitation is not particularly relevant.

It is plausible that individuals willing to participate in such a survey were also more digitally savvy and more attitudinally inclined to engage in eWOM. This potential bias means the sample might over-represent individuals with a higher baseline propensity to generate eWOM, which could, in turn, inflate the mean scores or the strength of the relationships observed in the model. Future research should, therefore, attempt to replicate these findings using mixed-method or offline data collection techniques to verify the generalisability of the model. It is also plausible that individuals from different socio-economic backgrounds may have different levels of expectations (which would directly influence the perception of unacceptability—H_1B); varying levels of comfort with, or access to, digital review platforms; and different preferred channels for expressing dissatisfaction (e.g., a direct complaint at the front desk versus generating eWOM). Therefore, future research should test this model on more socio-economically diverse samples to explore these potential differences.

Although there are clearly differences between people who have ready access to the Internet and those who do not or do not use it, these differences would not have changed the results because such respondents would not rely on eWOM intentions when considering the given consumption situation. Besides personality and situation factors, future research could include additional factors, such as those relating to a company’s reaction to a negative consumption situation, the brand of the consumed experience, etc. Future research could also consider how eWOM intentions might evolve over time (before, during, and after a consumption situation, even one week or one month later). The felt experience would depend on the memorability of the consumption situation, which might exert a different effect on eWOM intention.