Role of Emotions and Conflicting Online Reviews on Consumers’ Purchase Intentions

Abstract

Drawing on dual-process theories, this paper explains how systematic and heuristic information processing of conflicting online reviews can influence the consumer’s purchase decision-making. The study adopts major assumptions of complexity and configural theory in employing fuzzy-set qualitative comparative analysis (fs/QCA) with 680 users of TripAdvisor, to test the complex interrelations among emotions, and systematic and heuristic cues. The results demonstrate that systematic and heuristic processing of online reviews can produce independent impacts on consumer decision-making. Both processing routes can interact with each other due to the online reviews’ contradictory sequence to affect the domination of one route over the other. In the case of the positive-negative sequence, the consumers mainly follow a heuristic processing route, whilst in the opposite sequence, consumers’ concerns regarding the credibility of these reviews leads them to increase the depth of their thoughts (systematic processing) and actively evaluate the argumentation quality and the helpfulness of the online reviews.

Keywords

Online reviews, Heuristic-Systematic processing model, Emotions, Fuzzy set Qualitative Comparative Analysis

1. Introduction

The development of travel review websites has radically changed the hospitality industry. Such platforms are a valuable source of information for consumers and, as a result, are a growing driver of decision-making in booking hotels, restaurants, and attractions. As a critical form of eWOM (electronic Word of Mouth), which refers to ‘any statement made by potential, actual, or former consumers about a product, service, or company, which is available to a multitude of people and institutions via Internet’ (Hennig-Thurau et al., 2004, pp. 215), the online reviews allude to peer-generated evaluations posted on third party websites (e.g. TripAdvisor). Thus, online reviews fall under the eWOM category of asynchronous, ‘one to many’ communications (Bronner and De Hoog, 2011), and constitute one of the most important and influential forms of eWOM because it can directly and significantly explain consumers’ online behaviour (Banerjeeand Chua, 2016). Given the increasing competition in the hospitality industry, the investigation of how consumer-generated reviews affect the consumption decision of tourism services is a key issue.

Dual-process theories provide comprehensive discussions on how individuals process information, establish their validity assessments, and later form decision outcomes (Eagly and Chaiken, 1993). These theories posit that consumers’ process information using two routes: central/systematic processing (i.e. analysis of all relevant pieces of information) and peripheral/heuristic processing (i.e. decision-making by assessing available information as a whole). Dual-process theories recognise consumer information processing and persuasion as a complex combination of systematic and heuristic processing (Petty and Brinol, 2008). The co-occurrence of the two information processing modes means that both types of processing can occur at the same time and affect each other (bias effect). Despite the bias effect having received considerable support in psychological literature, little research has shed light on its examination in the online review literature (Zhang et al., 2014).

While practitioners recognize emotions as being critical to the success of the tourist experience, and researchers agree on the importance of relationships among emotional variables, consumer information processing, and behavioural intentions, there are no conclusive findings with regard to the interplay between emotions and cognition (Bigné et al., 2005; Pappas et al., 2016; Petty and Brinol, 2008). This is probably due to the fact that despite the undoubtedly complex and idiosyncratic nature of the phenomenon, researchers mostly investigate it based on regression techniques, which fail to examine the combined effects of cognitive and affective perceptions on intention to purchase (Pappas et al., 2016).

This knowledge gap motivated the present research, whose conceptual contribution involves the identification and modelling of interesting inter-relationships of emotions with cognitive information cues and behavioural intentions.

Drawing on dual processing theories, this paper adopts the major assumptions of complexity and configurational theories in order to shed new light on the role played by different systematic and heuristic informational cues of online reviews on consumers’ decision-making. The study employs fs/QCA to model the complex causal relationships and detect common patterns among emotions, and heuristic and systematic information cues that can lead to high scores in customer purchase intentions. By successfully combining advantages of both qualitative and quantitative research, the method has increasingly become a significant methodological tool for the analysis of various business topics (e.g. Ordanini et al., 2014; Chatzipanagiotou et al., 2016; Gounaris et al., 2016; Woodside, 2014; Fiss, 2011).

The study contributes significantly to online reviews and eWOM knowledge in the following ways. First, it aims to identify the key configurations of systematic and heuristic information cues for understanding the influence of online reviews in purchase decisions. Second, this study empirically examines if additive or attenuation effects exist when consumers process online reviews between the two processing routes. Third, to the best of our knowledge, this is one of the first studies that expands knowledge on how information cues interact with the sequence of online reviews (the order of positive and negative online reviews) to impact consumers’ purchase intentions.

The study is organised as follows. We first present the theoretical background, followed by developing the conceptual framework and research propositions to explain how consumers process online reviews. Then, we empirically test the model through a fs/QCA with 680 users of TripAdvisor. Finally, we discuss the findings, limitations, and opportunities for future research, summarising the implications for both researchers and practitioners.

2. Literature review

In the dual-process literature, two of the most prevalent models are the elaboration likelihood model (ELM; Petty and Cacciopo, 1986) and the heuristic-systematic model (HSM; Chaiken, 1980). The two models provide similar mechanisms in explaining individuals’ information processing strategies. For instance, the central route in the ELM and the systematic processing in the HSM indicate that individuals use high cognitive effort to elaborate information. By contrast, the peripheral route in the ELM and the heuristic processing in the HSM suggest that individuals adopt heuristic and simple decision rules to quickly form judgments. Both models apply to online reviews, although the HSM is still under-researched.

The HSM (Chaiken, 1980) is a widely-recognised communication model that attempts to explain how people receive and process persuasive messages. Under HSM, there are two models of information processing: heuristic processing and systematic processing (Chaiken, 1980). Heuristic processing uses a few informational cues such as simple decision rules to reach a conclusion by assessing available information. For example, the cue of ‘source credibility’ may trigger the rule ‘credibility implies correctness’, leading a message recipient to favourably assess the validity of a message received from a more credible source. During systematic processing, a message recipient examines all relevant pieces of information carefully for their relevance and importance to the task in order to form the final decisions.

The HSM model does not treat dual processes as simply traded off (which is indicated by the ELM), but suggests that they can occur concurrently and affect each other in complex ways (Eagly and Chaiken, 1993). Their interaction can be explained through the following 3 effects: (i) the additivity effect (which produces independent effects on consumers’ decision making; (ii) the attenuation effect which explains how the systematic mode of persuasion can attenuate the heuristic mode; and, (iii) the bias effect which means that heuristic processing can bias systematic processing by affecting individuals’ expectations or inferences about the validity of arguments (Zhang et al., 2014).

Drawing on the HSM model, the study investigates how different systematic and heuristic informational cues of online reviews interact with and influence consumers´ purchase decision-making.

2.1 Systematic processing route: cognitive processing

Argument quality has been defined as ‘the strength or plausibility of persuasive argumentation’ (Eagly and Chaiken, 1993, pp.35). Argument quality refers to perceptions of strong and convincing arguments rather than weak and unreal ones. When applying the dual-process theory of human information processing in the online reviews context, researchers tend to consider the quality of reviews (argument quality) to manifest central processing (Cheung and Tadani, 2012). Previous research on online reviews has analysed argument quality as a composite construct, failing to discern the importance of different argument quality dimensions on consumers’ information processing (Zhang et al., 2014). We focus on the behavioural outcomes of argument quality through two dimensions: informativeness and persuasiveness.

Informativeness is defined as consumers’ overall perceptions as to whether the online review provides users with complete, consistent, accurate, or adequate information (Sussman, and Siegal, 2003). In this way, persuasivenessrepresents consumers´ perceptions regarding the strength of relevance embedded in online reviews (Zhang, 1996). High-quality online consumer reviews are persuasive because the information is relevant in evaluating the product and contains understandable, reliable, and sufficient reasoning. High quality informative arguments have been found to contribute to favourable decision outcomes (Cheung et al., 2012; Zhang et al., 2014). In the context of online review websites, we therefore expect that if a consumer finds reviews about a particular restaurant to be highly informative and relevant, then the consumer ismorelikely to visit the restaurant.

In addition, perceived helpfulness of a review refers to the extent to which a peer-generated evaluation is perceived by potential consumers as useful and valuable in their decision process of choosing a product/service (Yin et al., 2014). Therefore, helpfulness can be understood as a measure of information diagnosticity. We argue that perceived helpfulness interacts with argument quality. As messages with high argument quality provide more completeinformation, a message with stronger arguments is expected to influence positively how people perceive the usefulness of information (Cheung and Thadani, 2012). Perceived helpfulness has been considered as a cognitive cue because consumers need to read a review carefully to assess if the review is useful for consumer decision-making. According to the Technology Acceptance Model (TAM), perceived usefulness is a fundamental predictor of user adoption (Davis, 1989). Building on the TAM model, this paper posits that if consumers think the information in an online review is helpful, they will have higher purchase intentions.

2.2. Heuristic processing route: affective processing

Online review credibility refers to whether consumers perceive the online recommendation as believable, true, or factual as a whole (Nabi and Hendriks, 2003). In this research, the subject of the credibility assessment refers to the online recommendation or review, and not trusting beliefs about a person or an organisation. Results from literature about the impact of credibility of online information on sales are inconclusive. Some researchers claim that as online information is posted by experienced travellers, it is more credible than information from traditional media (e.g. Fang et al., 2016). On the other hand, some studies posit that the online information can be posted by any individual without any editing or fact-checking processes and, therefore, is less credible than other types of information sources (e.g. Magnini, 2011). This paper supports previous research on online reviews regarding the influence of review credibility on consumer purchase intentions (Cheung and Thadani, 2012).

2.3. The role of emotions

There has been much debate in the previous literature about the definition of emotions. This study adopts the definition of emotion as a mental state of readiness that arises from cognitive appraisals of events or thoughts’ (Bagozzi et al., 1999, p. 184). The choice to include this study’s conceptualisation of emotions instead of other concepts that could be found under the umbrella term of ‘affect’ (e.g. moods) is based on the fact that emotions have very often been associated with a specific, known source, and influence consumers’ tendency to act and behave in a specific way (Lerner and Keltner 2000). In addition, researchers widely recognise the role of emotions in the online environment and stress the importance of investigating the substantial role of emotions in consumers’ processing and interpretation of online reviews (e.g. Yin et al., 2014)

According to Russell’s model (1980), emotions consist of two independent dimensions, that is, pleasure and arousal. The intensity dimension of affective reactions is operationalized as arousal. Arousal is defined as a feeling state varying along a single dimension from drowsiness to frantic excitement. Besides their intensity, affective reactions can be characterized by their polarity (i.e. pleasure versus displeasure). Pleasure refers to the degree to which a person feels good, joyful, or happy. In recent marketing studies, there has been considerable consensus with respect to the bi-dimensional character of emotions (e.g. Bigné et al., 2005), and this reflects the degree to which different individuals incorporate subjective experiences of pleasantness/unpleasantness and activated/deactivated subjective feelings into their emotional experiences.

On the other hand, empathy is the extent to which readers find resonance with the reviewer and think about how they would feel in a situation described in the review (Xia and Bechwati, 2008). An online review of a restaurant could be highly relevant if the reviewer fits the same profile as the reader or describes a situation with which the reader is familiar. Empathy could affect consumer behaviour indirectly by making salient to the consumer the product/service benefits that are being enjoyed by other consumers; or, it could affect consumers via a direct emotional ‘contagion’, as when one finds oneself laughing when reading a funny review or feeling disgusted as a result of a distressing review. From this perspective, the enthusiasm of a reviewer describing the joys of a particular experience at a restaurant could generate similar feelings in the minds of the readers.

Previous research has found empirical evidence supporting the relationship between emotions and several consumer outcomes, such as attitude towards travel review websites (Ruiz et al., 2016) or consumers’ intention to re-patronise (e.g. Hume, 2008).

2.4 Interactions in systematic and heuristic online review processing

The debate over the interplay between emotions and cognition continues to be a popular topic in psychology (Bigné et al., 2008; Chebat and Michon, 2003; Pappas et al., 2016). On the one hand, there is the emotions-lead-to-cognition approach (Zajonc and Markus, 1985). According to this approach, positive affect enables subjects to handle greater informational complexity. The same reasoning can be applied when the consumer has negative feelings. For example, the information in an online review written by a person who has been intoxicated in a restaurant may generate intense feelings of arousal and fear in the reader, and therefore, the need to look for more information about this restaurant. On the other hand, the cognition-leads-to-emotions school of thought (Lazarus, 1991), posits the causal role of cognition as a necessary but not sufficient condition to elicit emotions. Bigné et al. (2008) show how cognitive disconfirmation leads to pleasure and arousal. In the tourism context, for instance, a positive appraisal of the surrounding environment might generate feelings of pleasure associated with visiting a winery. Pappas et al. (2016) present various configurations between cognitive and affective consumer perceptions, showing the conditions under which quality of personalization, benefits, message quality, and emotions lead to increased intention to purchase personalized services.

Recent studies link emotions, credibility, and helpfulness. Yin et al. (2014) highlight the influence of negative emotions elicited by online reviews in the perceived helpfulness of the review. Empathy may interact with perceived helpfulness. The more intense the emotion perceived by the reader, the higher the perceived cognitive effort made by the reviewer and, therefore, the higher the perceived helpfulness of the online reviews. Empathy also can be linked with credibility and argument quality. The rationale is that, when reading an online review, if a reader senses resonance with the reviewer (similarity), he/she may perceive the review as more credible and trustworthy (Bickart and Schindler, 2001) and, hence, more persuasive.

Previous research also supports the interaction between credibility and argument quality. On the one hand, a large number of studies have found that more credible sources result in stronger persuasion than less credible ones (e.g. Bickart and Schindler; Wathen and Burkell, 2002). Wathen and Burkell (2002) pointed out that a key early stage in the information persuasion process is the receiver’s judgment of the information credibility; it determines how much weight individuals give the information within a specific review. On the other hand, according to ELM (Petty and Cacioppo, 1986), central cues, informativeness, and persuasiveness determine one’s attitude towards a message. Therefore, readers judge the credibility of online recommendations based on the argument strength of the online review message (Cheung et al., 2009; Wathen and Burkell, 2002).

Information helpfulness, online review credibility, and online review adoption are also theoretically explained by the information adoption model (Cheung and Thadani, 2012; Sussman and Siegal, 2003) which was adapted from the ELM (Petty and Cacioppo, 1986). The processing of received advice (online review) leads to the judgment of the usefulness of this advice (helpfulness), and affects the degree to which the information is adopted (Sussman and Siegal, 2003). Following previous research (Cheung and Thadani, 2012; Maslowska et al., 2017), we posit that if the readers of an online comment perceive it as helpful and credible, they will have more confidence to use online review comments for making decisions (whether to visit the restaurant or not).

The theoretical reasoning and empirical findings outlined above suggest an interaction effect between consumer emotions and information cues. Therefore, we posit that the interaction effect between the systematic and heuristic information cues influence the consumer’s decision-making.

2.5 Research propositions of the study

Based on the above discussion, we summarise in Figure 1 the study’s main research propositions. Based on the rationale of dual-processing theories (Chaiken, 1980; Petty and Cacioppo, 1986), the model does not treat information processes (systematic and heuristic) simply as trade-offs, but suggests that they can occur concurrently and affect each other in complex ways. Emotions are integrated as both systematic and heuristic cues (Bigné et al., 2008; Lazarus, 1991).

Insert figure 1

Figure 1 shows the interrelationships among the study’s main conditions demonstrating simultaneously the configural nature within the routes. Based on the above discussion, we assume that:

Proposition 1. Sufficient configurations of the elements constituting the systematic processing route lead to high scores in consumers’ intention to visit the restaurant.

Proposition 2. Sufficient configurations of the elements constituting the heuristic processing route lead to high scores in consumers’ intention to visit the restaurant.

Proposition 3. Sufficient configurations of the elements constituting both systematic and heuristic routes lead to high scores in consumers’ intention to visit the restaurant.

The conceptual framework also integrates the interactions of the sequence of reviews with the consumers’ systematic and heuristic processing routes. Previous research has demonstrated the influence of reviews on product sales (Fang et al., 2016) and consumer decision-making (Huang and Korfiatis, 2015); however, this impact differs depending on the review valence and sequence (Kim and Lee, 2015).

Two-sided reviews have been shown to increase the probability of a purchase being made (Huang and Korfiatis, 2015), but previous research has reported inconclusive results in regard to the specific effects of two-sided review effects on product choice. Chevalier and Mayzlin (2006) found evidence of confirmatory bias that drives consumers to look for affirmative evidence supporting an already made product choice. On the other hand, Cui et al. (2012) suggest that when consumers are neutral, negative reviews tend to be more salient than positive reviews (negative bias).

According to two-sided literature, the sequence of online consumer reviews (heuristic cue) impacts how consumers process online review information (Huang and Korfiatis, 2015). Justice literature suggests that evaluations are influenced more by what comes first than by what is received subsequently (Van den Bos et al., 1997). Sparks and Browning (2011) demonstrated that information (hotel online reviews) received early, especially if negatively worded, is likely to be more influential on consumer evaluations. We posit that when positive reviews are presented proximally, consumers may interpret that the company is making improvements and receiving favourable evaluations, and positive reviews can be more salient than negative ones. Conversely, when the negative reviews are more proximate, it may imply that the company’s recent performance is poor. Thus, we assume that;

Proposition 4: The online review sequence modifies how different configurations of systematic and heuristic processing route elements contribute toward producing high scores in consumers’ intention to visit the restaurant.

3. Methodology

3.1 Data Collection

Data was collected in January 2016. Participants responded to an online survey. The respondents were Spanish Internet users of TripAdvisor, aged 18 or older and who had used TripAdvisor when choosing a restaurant at least once in the last year. A total of 680 individuals who participated during the study were geographically dispersed in Spain. The study focused on heavy users, considering those who looked up information about restaurants at least five times or more in the last month. Heavy users are the most attractive segment for review website platforms. Therefore, understanding how they perceive online reviews and encourage them to provide active recommendations to other members is important both from retention as well as acquisition points of view. Participants were instructed to imagine a situation where they were going out for dinner to an Italian restaurant with friends; they were told to read a total of 10 reviews about an Italian restaurant in the same order they were displayed and answer the questions that followed. Both positive and negative reviews were included. Each review consisted of a satisfaction rating (out of five stars) and a comment that matched the rating. The reviews were modified from the real reviews posted on TripAdvisor and were validated from our pretexts in terms of their realism of content, and appropriateness of length, and readability (control variables). None of the reviews included pictorial content or any information about the reviewer. We also omitted the name of the restaurant to avoid bias due to familiarity. The total sample comprises 58% females and 42% males, aged 35-54. The respondents looked up information about restaurants in TripAdvisor at least five times or more in the last month, and 28.4% of the sample more than six times in the same period. As regards measuring the variables, we use 7-point Likert scales (see Table 1), as they are most appropriate for electronic distribution of questionnaires (Finstad, 2010).

Insert Table 1

3.2 Fs/QCA

Fs/QCA addresses the inadequacy of regression-based techniques for dealing with the complex causality surrounding consumers’ online behaviour (Pappas et al., 2016) and thus, it is uniquely fitted to the nature of the study’s research propositions (Ragin, 2008).

As a set-theoretic method, fs/CQA allows the classification of each antecedent conditions and the outcome of interest in the form of sets. It represents each of the cases as a complex entity based on the degree of membership of each case in the relevant sets. The alteration of study’s raw variable scores into set memberships alludes to the ‘data calibration’ phase. Following the most widely used method of calibration, the direct method proposed by Ragin (2008), the study uses three qualitative anchors (1.0 = full membership, 0 = full non-membership, and 0.5 = point of indifference) to assess varying degrees of the study’s variables membership. For example, and as Table 2 shows, in order to locate the qualitative anchors and calibrate consumers’ intention to visit, we set cases in the highest quintile equal (517) to 0.95 membership; cases in the middle quintile (357) at 0.50; and calibrated scores for the lowest quintile (181) at 0.05. The same procedure governs the calibration of all the study’s concepts.

Insert Table 2

Following this holistic approach in terms of case representation, fs/QCA detects not only the qualitative composition of the cases but also common causal patterns in a systematic cross-case analysis (Rihoux and Marx 2013). The latter revolves around the identification of necessary and sufficient conditions. Necessary conditions are conditions that are required for an outcome to occur and thus it constitutes a super-set of the given outcome. The conditions or configurations of conditions that are always produce the outcome refer as sufficient conditions/configurations indicating their role as a sub-set of the outcome (for detailed discussion see Ragin 2008; Goertz and Starr 2003).

Two set-theoretic indices, namely: consistency and coverage are in use to estimate and interpret the results (Ragin, 2008; Wagemann and Schneider, 2010). Consistency expresses the degree of the subset relationship between the antecedent conditions (or a combination of them) and the outcome of interest, and signals the significance of each of the derived configurations and the entire solution. The coverage gauges the empirical relevance by demonstrating the degree to which an antecedent condition or configuration of antecedent conditions can explain instances of the outcome of interest (Ragin, 2008). Following Ragin (2008) suggestions for consistency above 0.75 the study sets .80 as the minimum threshold for the examination of the derived solutions’ consistency and two cases as frequency cut-off point of observations (Fiss, 2011). The study also incorporates, sensitivity checks based on different levels of frequency (one, three, and six cases) and consistency thresholds (range from 0.81 to 0.90) (Skaaning, 2011). The results did not challenge the robustness of the aforementioned choice (the results are available upon request).

Furthermore, the study detects core-periphery models (Fiss, 2011) permitting the identification of the conditions, which demonstrate a causally intense connection to the outcome of interest (core causes), as well as the conditions with a weaker or less important relationship with the outcome (periphery causes). These core-periphery distinctions lead to a better interpretation of the derived solutions whilst further explaining the role of each of the antecedent conditions on the outcome of interest.

4. Results

The first stage of data analysis includes the descriptive statistics, the identification of the correlation coefficients among the study’s conditions as well as the quintile analysis between the causal conditions and the outcome. Table 3 demonstrates these results and confirms that no symmetrical relations occur (all coefficients are below the .80 threshold) (Woodside, 2014).

Insert Table 3

The results of the quintile analysis indicate a positive effect between antecedent conditions and the consumers’ intention to visit (phi coefficients between .440 and .699). However, both positive and negative contrarian cases occur, which further confirm the asymmetric relationship among the study’s conditions and consumers’ intention to visit (i.e. Woodside, 2014).

The second stage of the data analysis includes the investigation of the study’s research propositions. Specifically, Table 4 summarizes the results of P1, P2, which examine the reasoning of systematic and heuristic processing routes and how each route can individually sufficiently explain high scores in consumers’ intention to visit. Specifically, Table 4 (panel A) depicts the results of the systematic processing route leading to high scores in consumers’ intentions to visit the specific restaurant with high overall consistency (= .85) and coverage (=.51). Two sufficient solutions explain the different pathways that consumers follow in order to make this decision. Consumers’ emotions, in terms of high pleasure and degree of online review stimulation, constitute core causes which, in combination with either online reviews helpfulness (solution 1), or informativeness (solution 2), lead to high scores in consumers’ intentions, providing support to P1.

Insert Table 4

In addition, Table 4 (panel B) summarizes the results in regard to the heuristic processing route (P2). One solution can sufficiently explain consumers’ intention to visit the specific restaurant. The configuration of the consumers’ high degree of pleasure, arousal, and empathy with the online reviews sufficiently lead to high scores in consumers’ intention to visit the specific restaurant, providing support to P2.

Interestingly enough, both systematic and heuristic processing routes incorporate a great degree of emotion which play a core role in the functioning of both these routes.

In addition, the interaction of the two processing routes is examined in P3. The results detect two solutions that sufficiently predict high scores in consumers’ intentions with high overall consistency (.85) and (.50) and coverage, providing support to P3.

Insert Table 5

As Table 5 shows, consumers’ emotions in terms of pleasure and arousal constitute core casual conditions in both solutions. Specifically, solution 1 indicates that online reviews that consumers find pleasurable and stimulating in combination with the online reviews’ helpfulness lead consumers to choose specific restaurants. Solution 2 further demonstrates that online reviews that consumers find informative (in terms of relevance, completeness, and timeliness) coupled with high scores of consumers’ emotions (pleasure and arousal) facilitate their decision to visit the specific restaurant. The components of the heuristic processing route (empathy and credibility) are incorporated in the second solution, although in a peripheral role. Note that both solutions highlight the core role of online reviews’ informativeness and helpfulness, highlighting that mostly consumers follow a systematic processing route, with emotions also playing a critical, core role in interpreting online reviews and deciding positively on the specific restaurant.

The identification of the role that the online review sequence has on consumers’ intentions to visit the specific restaurant is presented in Table 6. Specifically, as Table 6 demonstrates, the online review sequences (positive-negative or negative-positive) can modify the way consumers decide to assess these reviews.

Insert Table 6

Table 6 (panel A) demonstrates that when online reviews’ contradictory sequences start with positive comments, consumers’ emotions of pleasure and arousal can alone predict high scores in their intentions to visit the specific restaurant (solution 1, with high overall consistency .87 and .61). Solution 2 detects that the elicitation of consumers’ positive emotions of pleasure in combination with a high degree of credibility and empathy can make consumers willing to visit the specific restaurant. In other words, the majority of consumers follow a heuristic processing route with their emotions to constitute core causal conditions. The minority of consumers follow solution 3 which includes a combination of systematic and heuristic processing components. Solution 3 suggests that when consumers do not find the online reviews stimulating and cannot be empathetic towards them, both systematic (perceived informativeness, persuasiveness, usefulness) and heuristic components (perceived credibility) are combined in order for consumers to make their decisions.

In the case of negative-positive sequence, Table 6 (panel B) summarizes two solutions. Both solutions highlight that when online reviews start with negative comments, consumers tend to follow a systematic processing route. Specifically, solution 1 indicates that consumers’ emotions of pleasure and arousal constitute core causes, which, in combination with either online reviews’ helpfulness (solution 1) or perceived argumentation quality (solution 2), lead to high scores in consumers’ intentions to visit. The heuristic processing route components of credibility (in its absence) and empathy do contribute but in a peripheral role.

Results show that both routes can interact with each other, with the online reviews’ contradictory sequence affecting the domination of one route over the other. In the case of the positive-negative sequence, the consumers mainly follow a heuristic processing route, which focuses either on high scores in consumers’ affective states of pleasure and arousal or on consumers’ empathy and online reviews’ credibility. Therefore, consumers’ positive emotions tend to favour the heuristic processing route in which consumers do not actively involve themselves in the examination of online reviews’ argumentation quality. They mainly focus on their feelings and their general perception regarding the online reviews’ credibility. In contrast, and when consumers face the opposite online reviews’ contradictory sequence (negative-positive), the results demonstrate that consumers follow a systematic processing route and actively evaluate the argumentation quality presented in online reviews and the helpfulness of these online reviews.

5. Conclusions

5.1. Theoretical implications

The study extends our knowledge regarding dual-process theories and significantly contributes to the identification of the role of emotions in consumers’ online decision making. Deviating from general correlational associations, the study employs fs/QCA to demonstrate the combinatorial nature of the causal relationships among emotions, and systematic and heuristic cues. The derived solutions demonstrate that consumers’ emotions play a potent, influential, and pervasive role in both systematic and heuristic processing routes and they are effectively integrated with either systematic or heuristic cues in order to further inform consumers’ decision making.

Regarding the operationalisation of HSM, the results reveal that both systematic and heuristic processing routes can have an independent impact on consumers’ intention to visit providing further support to the additivity effect (Zhang et al., 2014). However, the examination of the interplay between the two processing routes reveals that the dominant role of the systematic processing route in an online review context, (with the core causal role of online review helpfulness and informativeness), is the attenuation of the heuristic route through the peripheral role of online review, perceived credibility, and consumers’ empathy.

Interestingly enough, the contradictory sequence of online reviews further explains consumers’ online behaviour. The online reviews’ incongruence plays an adaptive role, and serves as a signal enabling consumers to detect whether a situation needs more cognition or not. In the positive-negative sequence, consumers base their decision making in either, strong positive emotions, or they alter their concerns and base it on the increased online credibility following a heuristic processing route, which eventually biases systematic processing. In the form of negative-positive sequences, consumers are keen to improve their depth of thought, recognising that the situation demands further attention. They rely on the systematic processing route and especially on the online reviews’ helpfulness or the combination of online reviews’ informativeness and persuasiveness in order to increase their judgmental confidence. The results reveal the attenuation of heuristic processing. In both the above cases, consumers’ emotions constitute an integral part of the process in which consumers assess online reviews and incorporate them in their decisions.

5.2. Managerial Implications

The findings provide practical implications for managers and online review platform service providers. Acknowledging the importance of the online review content, they should pay extra attention to encouraging consumers to write realistic, informative, and timely comments, in order to increase the online review argumentation quality and perceived helpfulness. Managers and online review platform service providers should advise consumers to provide a sufficient, vivid description of all the different components of their experience such as atmosphere, location, customer service for improving the holistic evaluation of online reviews. To ensure the online review content, they should encourage users to volunteer any online reviews that are extremely non-informative or have fake content.

Furthermore, new contextual classifications of online reviews in terms of their completeness; informativeness, credibility, etc. based on multiple elements that the users can provide (e.g. text, pictures) could lead to more accurate and useful online review classifications for consumers’ decision making.

5.3 Limitations and directions for future research

The study refers to specific national context, industry, and time period. The replication of the study in different contexts and the longitudinal assessment of the model would be an essential future research priority. In addition, the model could be extended to include the impact of other heuristic cues such as likeability, attractiveness, and expertise of the communicator. Furthermore, an emerging area of interest in the literature is visual online review. The investigation of the role of visual content in online review adoption and behavioural intentions would be extremely interesting. Consumers may present different perceptions of argument quality according to their experience. Thus, comparisons between experienced and novice users of TripAdvisor would provide fruitful insights.

References

Bagozzi, R., Gopinath, M., and Nyer, P. (1999). The role of emotions in marketing. Journal of the Academy of Marketing Science, 27(2), 184-206.

Banerjee, S. and Chua, A.Y., 2016. In search of patterns among travellers’ hotel ratings in TripAdvisor. Tourism Management, 53, pp.125-131.

Bickart, B., and Schindler, R. M. (2001). Internet forums as influential sources of consumer information. Journal of Interactive Marketing, 15(3), 31-40.

Bigné, E., Andreu, L., and Gnoth, J. (2005). The theme park experience: An analysis of pleasure, arousal and satisfaction. Tourism management, 26(6), 833-844.

Bigné, E., Mattila, A., and Andreu, L. (2008). The impact of experiential consumption cognitions and emotions on behavioral intentions. Journal of Services Marketing, 22(4), 303 – 315.

Bronner, F. and De Hoog, R. (2011). Vacationers and eWOM: Who posts, and why, where, and what? Journal of Travel Research, 50(1), 15-26.

Chaiken, S. (1980). Heuristic versus systematic information processing and the use of source versus message cues in persuasion. Journal of Personality and Social Psychology, 39(3), 752-766.

Chatzipanagiotou, K. and Veloutsou, C. and Christodoulides, G. (2016). Decoding the complexity of the consumer-based brand equity process. Journal of Business Research, 69 (11), 5479-5486.

Chebat, J.C. and Michon, R., (2003). Impact of ambient odors on mall shoppers’ emotions, cognition, and spending: A test of competitive causal theories. Journal of Business Research, 56(7), 529-539.

Cheung, C. M., and Thadani, D. (2012). The impact of electronic word-of-mouth communication: A literature analysis and integrative model. Decision Support Systems, 54(1), 461-470.

Cheung, C., Sia, C., and Kuan, K. (2012). Is this review believable? A study of factors affecting the credibility of online consumer reviews from an ELM perspective. Journal of the Association for Information Systems, 13(8), 618-635.

Cheung, M., Luo, C., Sia, C., and Chen, H. (2009). Credibility of electronic word-of-mouth: Informational and normative determinants of online consumer recommendations. International Journal of Electronic Commerce, 13(4), 9-38.

Chevalier, J.A. and Mayzlin, D. (2006). The effect of word of mouth on sales: online book review. Journal of Marketing Research, 43(3), 345-354.

Cui, G., Lui, H., and Guo, X. (2012). The effect of online consumer reviews on new product sales. International Journal of Electronic Commerce, 17(1), 39-58.

Davis, F. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology, MIS Quarterly, 319-340.

Eagly, A., and Chaiken, S. (1993). The psychology of attitudes. Hartcourt Brace: Jovanovich College Publishers:

Fang, B., Ye, Q., Kucukusta, D., and Law, R. (2016). Analysis of the perceived value of online tourism reviews: influence of readability and reviewer characteristics. Tourism Management, 52, 498-506.

Finstad, K. (2010). Response interpolation and scale sensitivity: evidence against 5-point scales. Journal of Usability Studies, 5(3), 104-110.

Fiss, P. C. (2011). Building better casual theories: A fuzzy set approach to typologies in organizational research. Academy of Management Journal, 54(2), 393.

Gounaris, S., Chatzipanagiotou, K., Boukis, A., & Perks, H. (2016). Unfolding the recipes for conflict resolution during the new service development effort. Journal of Business Research, 69(10), 4042-4055.

Hennig-Thurau, T., Gwinner, K., Walsh, G., and Gremler, D. (2004). Electronic Word-of- Mouth via Consumer-Opinion Platforms: What Motivates Consumers to Articulate Themselves on the Internet? Journal of Interactive Marketing, 18 (1), 38-52.

Huang, G., and Korfiatis, N. (2015). Trying before buying: The moderating role of online reviews in trial attitude formation toward mobile applications. International Journal of Electronic Commerce, 19(4), 77-111.

Hume, M. (2008). Developing a conceptual model for repurchase intention in the performing arts: the role of emotion, core service and service delivery. International Journal on Arts Management, 10 (2), 40-55.

Kim, E., and Lee, C. (2015). How do consumers process online hotel reviews? The effects of eWOM consensus and sequence. Journal of Hospitality and Tourism Technology, 6(2),113-126.

Lazarus, R.S. (1991), Emotion and Adaptation. New York, NY: Oxford University Press.

Lerner, J.S. and Keltner, D., (2000). Beyond valence: Toward a model of emotion-specific influences on judgement and choice. Cognition & Emotion, 14(4), 473-493.

Magnini V. (2011). The implications of company sponsored messages disguised as word-of-mouth. Journal of Services Marketing, 25(4), 243–251.

Maslowska, E., Malthouse, E.C. and Bernritter, S.F., (2017). Too good to be true: the role of online reviews’ features in probability to buy. International Journal of Advertising, 36(1), 142-163.

McCullough, M., Worthington Jr, E.., and Rachal, K. (1997). Interpersonal forgiving in close relationships. Journal of Personality and Social Psychology, 73(2), 321-336.

Nabi, R., and Hendriks, A. (2003). The persuasive effect of host and audience reaction shots in television talk shows, Journal of Communication, 53, 527-543.

Ordanini, A., Parasuraman, A., & Rubera, G. (2014). When the recipe is more important than the ingredients: A qualitative comparative analysis (QCA) of service innovation configurations. Journal of Service Research, 17(2), 134-149.

Pappas, I. O., Kourouthanassis, P. E., Giannakos, M. N., & Chrissikopoulos, V. (2016). Explaining online shopping behavior with fsQCA: The role of cognitive and affective perceptions. Journal of Business Research, 69(2), 794-803.

Petty, R. E., and Brinol, P. (2008). Persuasion: From single to multiple to metacognitive processes. Perspectives on Psychological Science, 3(2), 137-147.

Petty, R., and Cacioppo, J. (1986). Communication and Persuasion: Central and Peripheral Routes to Attitude Change. New York: Springer-Verlag.

Ragin, C. C. (2008). Redesigning social inquiry: Fuzzy sets and beyond. Chicago: Chicago University Press.

Rihoux, B., and Marx, A. (2013). QCA, 25 Years after “The Comparative Method” Mapping, Challenges, and Innovations—Mini-Symposium. Political Research Quarterly, 66(1), 167-235.

Ruiz, C., Tronch, J., and Sanz, S. (2016). The role of emotions and social influences on consumer loyalty towards online travel communities. Journal of Service Theory and Practice, 26(5), 534-558.

Russell, J.A. (1980). A circumplex model of affect. Journal of Personality and Social Psychology, 39(6),1161-78.

Skaaning, S. E. (2011). Assessing the robustness of crisp-set and fuzzy-set QCA results. Sociological Methods & Research, 40(2), 391-408.

Sparks, B.A. and Browning, V. (2011). The impact of online reviews on hotel booking intentions and perception of trust. Tourism Management, 32(6),1310-1323.

Sussman, S., and Siegal, W. (2003). Informational influence in organizations: an integrated approach to knowledge adoption. Information Systems Research, 14, 47–65.

Van den Bos, K., Lind, E., Vermunt, R., and Wilke, H. (1997). How do I judge my outcome when I do not know the outcome of others? The psychology of the fair process effect”. Journal of Personality and social Psychology, 72(5),1034-1046.

Wagemann, C., and Schneider, C. Q. (2010). Qualitative comparative analysis (QCA) and fuzzy-sets: Agenda for a research approach and a data analysis technique. Comparative Sociology, 9(3), 376-396.

Wathen, N., and Burkell, J. (2002). Believe it or not: Factors influencing credibility on the Web. Journal of the American Society for Information Science and Technology, 53(2), 134-144.

Woodside, A. (2014). Embrace perform model: Complexity theory, contrarian case analysis, and multiple realities. Journal of Business Research, 67(12), 2495-2503.

Xia, L., and Bechwati, N. (2008). Word of mouse: the role of cognitive personalization in online consumer reviews. Journal of Interactive Advertising, 9(1), 3-13.

Yin, D., Bond, S., and Zhang, H. (2014). Anxious or angry? Effects of discrete emotions on the perceived helpfulness of online reviews, MIS Quarterly, 38(2), 539-560.

Zajonc, R.B. and Markus, H. (1985). Must all affect be mediated by cognition? Journal of Consumer Research, 12(3), 363-364.

Zhang, K., Zhao, S., Cheung, C., and Lee, M. (2014). Examining the influence of online reviews on consumers’ decision-making: A heuristic–systematic model. Decision Support Systems, 67, 78-89.

Zhang, Y. (1996). Responses to humorous advertising: the moderating effect of need for cognition. Journal of Advertising, 25 (1), 15-32.

Figure 1. Study’s conceptual framework- Interrelationships among the study’s conditions. Sufficient algorithms predicting high scores in consumers’ intentions to visit.

Table 1. Study’s measures

Variable	Items	Source
Online review Credibility	I think these reviews are factual I think these reviews are accurate I think these reviews are credible	Cheung et al., (2012)
Online review informativeness	These reviews provide relevant information about the restaurant These reviews provide complete information about the restaurant These reviews provide timely information about the restaurant	Zang et al., (2014)
Online review persuasiveness	The arguments of these reviews are convincing The arguments of these reviews are persuasive The arguments of these reviews are good The arguments of these reviews are strong
Online review helpfulness	Using the scales below, how would you describe the above consumer reviews?   – not at all helpful/very helpful – not at all useful/very useful – not at all informative/very informative	Yin et al., (2014)
Empathy	-While reading this review, to what extent did you feel like you were experiencing the same emotions as the reviewer?   – While reading this review, to what extent did you feel concerned for the reviewer? – While reading this review, to what extent did you feel moved by the review?	McCullough, et al. (1997)
Emotions	PLEA1 Angry-content PLEA2 Unhappy-happy PLEA3 Displeased-pleased PLEA4 Sad-joyful PLEA5 Disappointed-delighted PLEA6 Bored-entertained	Russell (1980)
	AROU 1 Depressed-cheerful AROU2 Calm-enthusiastic AROU3 Passive-active AROU4 Indifferent-surprised AROU5 Quiet-anxious AROU6 Relaxed-nervous

Table 3. Correlations among the study’s concepts

	Mean	St.dev	1	2	3	4	5	6	7	8
INFORMATIV.	4.96	1.08	1
PERSUASIVENESS	4.78	1.11	.757**	1
HELPFULNESS	4.98	1.28	.643**	.610**	1
CREDIBILITY	5.02	1.15	.737**	.756**	.534**	1
EMPATHY	4.66	1.33	.638**	.624**	.502**	.567**	1
PLEASURE	4.29	1.10	.372**	.371**	.354**	.305**	.360**	1		.
AROUSAL	4.12	.87	.316**	.341**	.270**	.237**	.368**	.564**	1
IN. TO_VISIT	4.16	1.68	.359**	.292**	.279**	.237**	.263**	.587**	.387**	1
**Correlation is significant at the 0.01 level.

Table 2. Original and Calibrated Consumers’ intention scale and frequency of cases by scores

Original values after quintile analysis	Using 5 scores   Calibrated	Using Fuzzy Scores	Frequency	Percent (%)	Cumulative   Percent (%)
24.000	.05	.00	47	6.9	6.9
50.500	.05	.01	6	.9	7.8
59.500	.05	.01	12	1.8	9.6
69.500	.05	.01	8	1.2	10.7
100.000	.05	.01	53	7.8	18.5
131.000	.05	.02	9	1.3	19.9
140.000	.05	.02	9	1.3	21.2
150.000	.05	.03	11	1.6	22.8
181.000	.05	.05	51	7.5	30.3
213.000	.15	.08	13	1.9	32.2
228.500	.15	.10	18	2.6	34.9
243.500	.15	.13	12	1.8	36.6
281.500	.15	.22	64	9.4	46.0
326.000	.15	.37	25	3.7	49.7
357.000	.50	.50	37	5.4	55.1
370.500	.67	.65	30	4.4	59.6
390.000	.67	.83	69	10.1	69.7
490.500	.67	.92	32	4.7	74.4
517.500	.95	.95	22	3.2	77.6
540.000	.95	.97	23	3.4	81.0
579.000	.95	.98	55	8.1	89.1
614.000	.95	.99	15	2.2	91.3
630.500	.95	.99	18	2.6	94.0
643.500	.95	1.00	8	1.2	95.1
664.000	.95	1.00	33	4.9	100.0
Total			680	100.0
Mean =340 (st.dev. =196.02); Median= 357; Mode= 390 Cut points: 181=0.05; 357=0.50; 517=0.95 Note: Using the above cut points the fs/QCA software set scores for the second and the fourth quintile

Table 4. Core–periphery models of Systematic and Heuristic processing routes predicting high scores consumers’ intention to visit (P1 & P2)

Systematic processing route   Consumers’ Intention to visit   (1) (2) Perceived Informativeness    Perceived Persuasiveness   Perceived Helpfulness    Pleasure   Arousal   Raw Coverage .48 .41 Unique Coverage .10 .03 Consistency .86 .87 Overall Consistency .85 Overall Coverage .51

Heuristic processing route   Consumers’ Intention to visit   (1) Perceived Credibility   Empathy  Pleasure  Arousal  Raw Coverage .43 Unique Coverage .43 Consistency .87 Overall Consistency .87 Overall Coverage .43

Note: The black circles indicate the presence of a condition, and circles with “x” indicate its absence. The large circles indicate core conditions; the small circles indicate peripheral conditions. Blank spaces in a pathway indicate “don’t care”. The Analysis of Necessary Conditions (NC) do not confirm the existence of any N.C

Table 5. Core–periphery models from the interaction of Systematic and Heuristic processing routes predicting high scores consumers’ intention to visit (P3)

	Consumers’ Intention to visit
	(1)	(2)
Perceived Informativeness		
Perceived Persuasiveness		
Perceived Helpfulness	
Perceived Credibility		
Empathy		
Pleasure		
Arousal		
Raw Coverage	.48	.33
Unique Coverage	.17	.02
Consistency	.86	.89
Overall Consistency	.85
Overall Coverage	.50

Note: The black circles indicate the presence of a condition, and circles with “x” indicate its absence. The large circles indicate core conditions; the small circles indicate peripheral conditions. Blank spaces in a pathway indicate “don’t care”. The Analysis of Necessary Conditions (NC) do not confirm the existence of any N.C

Table 6 – Core–periphery models from the interaction of Systematic and Heuristic processing routes predicting high scores consumers’ intention to visit for different online reviews’ sequence (P4)

Sequence I: Positive –Negative online reviews comments   Consumers’ Intention to visit (1) (2) (3) Perceived Informativeness      Perceived Persuasiveness      Perceived Helpfulness      Perceived Credibility     Empathy     Pleasure     Arousal      Raw Coverage .61 .45 .11 Unique Coverage .22 .04 .03 Consistency .87 .91 .84 Overall Consistency .85 Overall Coverage .72

Sequence II: Negative- Positive online reviews comments   Consumers’ Intention to visit   (1) (2) Perceived Informativeness    Perceived Persuasiveness    Perceived Helpfulness    Perceived Credibility    Empathy    Pleasure   Arousal   Raw Coverage .49 .39 Unique Coverage .12 .02 Consistency .90 .90 Overall Consistency .89 Overall Coverage .52

Note: The Analysis of Necessary Conditions (N.C) do not confirm the existence of any N.C