Persuasion Processes in Consumer Intent to Read Online Product Reviews: A Study Based on the Elaboration Likelihood Model

4026 | P a g e M a r c h 2 8 , 2 0 1 4 Persuasion Processes in Consumer Intent to Read Online Product Reviews: A Study Based on the Elaboration Likelihood Model Meng-Lin Shih, *Shu-Hui Chuang Department of Information Technology, Ching Kuo Institute of Management and Health, Fu Hsin Rd., Keelung 20301, Taiwan mlshih@mis.ccu.edu.tw *Department of Business Administration, Asia University, Liufeng Rd., Wufeng, Taichung, Taiwan 41354 joyce@asia.edu.tw ABSTRACT


INTRODUCTION
In the past two decades, the mode of shopping for merchandise has shifted dramatically from buying during a physical visit to a retail store to reviewing and purchasing merchandise through an online retail outlet in the worldwide web. Shopping at online retail outlets is particularly strong during holiday seasons, and contributes substantially to overall retail sales at these times, demonstrating the economic importance of online shopping. In contrast to shopping via physical visits to retail stores, however, shopping at online retail outlets cannot employ the five human neural senses in selecting merchandise for purchase, obligating the consumer to instead base their purchasing decisions on physically limited representations of a product such as photographs and text descriptions presented at a website. Given this situation, online product reviews appear to be exceptionally important to consumers, who seem likely to regard these reviews, which convey the reactions of other consumers of a product, as evidence of the intrinsic nature and value of the product (Burnkrant & Cousineau, 1975 (Chatterjee, 2001) Unlike such oral evaluations, however, online product reviews express the views of persons unknown to the viewer, and therefore carry less credibility for the viewer than do direct consumer opinions from family members or friends (Ratchford et al., 2001;Xie et al., 2011). Despite this, there is compelling evidence that many consumers have the intent to read online reviews of a product before shopping online for that product. Exemplifying this is that half of consumers who visit online shopping malls consider it important to read product reviews in making their buying decisions (Piller, 1991). In a survey of 1000 individuals shopping online for components of a vacation, 53% stated that apart from price, other persons' reviews were the most important factor in their choice of vacation destinations and accommodations (Walsh, 2007). The reason for the willingness of so many consumers to rely on others' online reviews must be that most consumers have already been persuaded by such reviews and must have thus formed the intent to refer to them when shopping online. Unfortunately, neither the factors responsible for the formation of this intent nor those that persuade online consumers to buy a product have been defined.
The study reported here was conducted to examine the processes responsible for consumers' intent to read online product reviews. It used the elaboration likelihood model (ELM) (Eagly & Chaiken, 1993;Petty & Cacioppo, 1986) of persuasion to identify these processes, to discern which of these process is most effective in persuading consumers to seek online reviews, and to identify any factors that modify these processes. Although the study is still new and continuing, it may help the understanding of consumers' motivation to read online product reviews.

THEORY AND HYPOTHESES
The social psychology literature has described the use of the ELM in determining the role of persuasive processes in shaping a targeted human behaviour (Petty & Cacioppo, 1986;Petty et al., 1981;SanJosé-Cabezudo et al., 2009). The ELM is founded on the concept that two routes of persuasion, a central route and a peripheral route, which differ in the degree of reflective information processing they involve, govern the formation of individual perceptions. The central route requires an individual to think carefully about issue-related arguments in an informational message, and to scrutinize the relative merits and relevance of those arguments before establishing a perception that contributes to a target behaviour. The peripheral route of persuasion is more likely to be used in processing simple cues that induce the formation of a perception without ISSN 2277-3061 4028 | P a g e M a r c h 2 8 , 2 0 1 4 requiring reflection on the rational merits of the information being presented. This two-route concept has been applied in the context of usage of information technology (IT) (Bhattacherjee & Sanford, 2006), where it has shown that both the content and sources of information about a particular IT are equally effective in shaping users' attitudes toward using that IT. This means that an individual user may be persuaded to accept an IT by experts' opinions about it (peripheral cues), although such acceptance may sometimes rely on the quality of the experts' arguments in terms of their rationale or the evidence they provide (argument quality). On the premise that online consumers are also and simultaneously users of IT (Koufaris, 2002), the ELM appears to be as readily applicable to investigating their online information-seeking and purchasing behaviour toward merchandise presented via IT as toward IT itself.

Effects of perceived review quality and consistency on intent to read online reviews
The postulate for the study was that consumers' intent to read online reviews of a product is founded in the central and peripheral routes of persuasion. Typically, research based on the ELM operationalizes "argument quality," or the persuasive strength of arguments embedded in an informational message, as the key element or construct in persuasion through the central route (Lee & Xia, 2011). However, the present study relies on "perceived review quality," or the quality of the contents of a review from the perspective of their informational characteristics, as the key element or variable in persuasion through the central route. This is based on the demonstrably persuasive effect of this variable on consumers' intent to buy a certain product found in other studies of central-route processing based on the ELM (Park et al., 2007). Furthermore, in contrast to classical ELM-based research, which uses "source credibility," or the extent to which users perceive an information source as believable, competent, and trustworthy, as a peripheral cue in persuasion (Petty et al., 1981), the present study utilizes "perceived review consistency," or the extent to which a review is consistent with the consumer's prior knowledge about a product (Zhang & Watts, 2004), as a peripheral cue. A similar construct, called "information consistency," has also been considered a peripheral cue and has demonstrated its persuasive power among users of online communications (Zhang & Watts, 2004). In short, the perceived quality and consistency of an online product review, both of which are pivotal in shaping a consumer's intent to read other online reviews, are used as the respective centraland peripheral-route constructs in the present study. Accordingly, it was hypothesized that: H1: Perceived review quality positively affects consumers' intent to read online product reviews.
H2: Perceived review consistency positively affects consumers' intent to read online product reviews.

Moderators of persuasion processes
Individuals choose one of the two routes of persuasion set forth in the ELM on the basis of their ability and motivation to process arguments, which in the ELM fall into the category known as likelihood of elaboration (Petty & Cacioppo, 1986).
Individuals with a high likelihood of elaboration are energized, ready, and willing to engage in careful scrutiny or thoughtful processing of an informational message, and therefore tend to be more persuaded by the quality of an argument than by peripheral cues. Contrastingly, individuals with a low likelihood of elaboration lack the motivation or ability to deliberate thoughtfully, and are more likely to be persuaded by peripheral cues than by the quality of an argument. In short, individual motivation and ability are presumed to moderate the effects of argument quality and peripheral cues on a target behaviour information presented in a product review, and operationalized the ability aspect of the likelihood of elaboration as consumer expertise, defined as the ability to read an online product review.
Consumers deeply involved in the process of reading an online product review engage in processing the information it contains through the central route, by focusing on content cues such as the perceived quality of the review. Conversely, consumers with a low level of involvement in the review will be less likely to engage in elaboration and more likely to be affected by cues processed via the peripheral route of persuasion, such as the perceived consistency of the review.
Therefore, two further hypotheses for the present study were that: H3: The greater the individual involvement in the topic of an online review, the more the perceived quality of the review will affect the individual's intent to read online product reviews.

H4:
The greater the individual involvement in the topic of an online review, the less the perceived consistency of the review will affect the intent to read online product reviews.
Beyond this, the greater the individual level of expertise in the topic of an online product review, the more important will be its argument-relevant informational content in persuading the reader to use online reviews, and the greater will be the ability of the individual to understand that content. This in turn increases the likelihood of central-route processing and reliance on cues such as perceived review quality, and reduces the likelihood of reliance on peripheral cues such as review consistency (Ratneshwar & Chaiken, 1991). Therefore:

H5:
The greater the individual expertise in the topic of an online review, the more will perceived review quality affect the individual's intent to read online product reviews.

H6:
The greater the individual expertise in the topic of an online review, the less will the perceived review consistency affect the individual's intent to read online product reviews.

Questionnaire development
The questionnaire used in the study was developed to quantify the following five constructs of interest in the study: (1) intent to read online reviews, (2) perceived review quality, (3) perceived review consistency, (4) consumer involvement in the topics of online reviews, and (6) consumer expertise in the topics of online reviews. All five constructs were quantified on the basis of items validated in prior research, which were reworded to relate specifically to the context of reading online product reviews (see Appendix A). The construct for the intent to read reviews was adapted from Ajzen and Fishbein (1980) and Taylor and Todd (1995), and was modified to suit the context of online product reviews. The construct designated as perceived review quality, derived from a five-item measurement drawn from Park et al. (2007). focuses on the relevance, objectivity, understandability, and adequacy of explanation on online product reviews. Perceived review consistency was assessed with a three-item instrument adapted from an information-consistency scale developed by Zhang and Watts (2004) on the basis of a study by Vandenbosch and Higgins (1996). The quantification of consumer involvement and expertise was based on the original scales for these two measures that were developed and validated by Stamm and Dube (1994).
Before conducting the main survey, we conducted a pretest and a pilot test to validate the questionnaire used as the survey

Data collection
The research model for the study was tested with data from reviewers on the Mobile01 (www.mobile01.com) website. This The survey yielded 265 responses, which were screened for usability and reliability, and of which 16 were eliminated as having come from respondents who were obviously unconcerned about their responses (e.g., those giving the same rating for all items in the questionnaire), leaving 249 questionnaires that were considered suitable for data analysis and from which the pretest data were derived. Of the survey respondents, approximately, 54.8% were male; 59.1% were between 20 and 39 years of age; 93.5% were educated to the level of a college degree or above; and 50% used the Internet for more than 18 hours each week. Table 1 presents the descriptive statistics for the main constructs used in the study, including their means and standard deviations. Because the responses to the survey questionnaire were derived through a nonrandom method of sampling, a test of homogeneity on the survey respondents' demographic variables. All items in the five study constructs were tested against demographic control variables (gender, age, level of education, and Internet usage) through analysis of variance (ANOVA), as suggested by Cho (2006). The mean scores of all of the items tested were nonsignificantly (P > 0.05) represented by the demographic control variables.

RESULTS
The study data were analyzed with the partial least squares (PLS) method, a technique of structural equation modeling that has become widely accepted because of its accuracy and utility. The PLS method also puts minimal restrictions on sample size and residual distribution. The data analysis was done in two stages. In the first stage, all of the measurement scales used in the study were tested for reliability and construct validity using confirmatory factor analysis (CFA). In the second stage of the analysis, the structural model for the study was analyzed. The two-stage model was selected in preference to an ISSN 2277-3061 4031 | P a g e M a r c h 2 8 , 2 0 1 4 alternative one-stage approach in which measurement scales and structural models are examined simultaneously, because the two-stage approach provides a more complete and robust test of measurement validity by examining potential associations between constructs that may not otherwise be theoretically specified in the one-stage structural model.    Table 3 shows the discriminant validity of the measurements used in the study. For good discriminant validity, the square root of the AVE of a construct should exceed that of the construct's correlations with the other constructs in a structural model (Fornell & Larcker, 1987). The data in the study indicated that the shared variance among variables was smaller than the variances extracted for the constructs, which were the values for the diagonals connecting the constructs, or independent variables in the study model, with the main outcome measure, or dependent variable, of intent to read online reviews. This shows that the constructs were empirically distinct.

Testing of hypotheses
We used PLS with the bootstrapping resampling procedure to test our model. The examination of a structural equation The product indicator approach proposed by Chin et al. (2003) was used to measure the model of moderating effects on the study constructs. This model tested the extent to which consumer involvement and consumer expertise, representing elaboration motivation and ability, respectively, moderated the main effects comprised in hypotheses 1 and 2. An examination of individual routes in the moderating-effects model revealed that consumer involvement had a significantly positive moderating effect on the association between perceived review quality and the intent to read online product reviews (β = 0.169, P < 0.01), as posited in hypothesis 3. Conversely, consumer involvement negatively influenced the effect of perceived review consistency on the intent to read product reviews (β = -0.096, P < 0.05), as posited in hypothesis 4. Consumer expertise positively moderated the effect of perceived review quality on the intent to read online product reviews (β = 0.156, P < 0.01), in accord with hypothesis 5, and negatively moderated the effects of perceived review consistency on the intent to read such reviews (β = -0.111, P < 0.05), in accord with hypothesis 6. Hypotheses 3 and 4 repeat this theoretical argument for the moderators of involvement, since involvement also increases the likelihood of elaboration. Hypotheses 5 and 6 posit that expertise will moderate the effects of perceived review quality and perceived review consistency on the intent to read online product reviews. Our model suggests that for individuals with high levels of expertise, the quality of online product reviews will strongly affect the intent to read such reviews, whereas for those with lower levels of expertise, cuses relating to consistency will have a stronger effect.

DISCUSSION
At the beginning of this report, we described three major points relating to the understanding of consumer intent to reading online product reviews. The first was that a consumer's perceptions from reading reviews at a retail website determine whether the consumer maintains an intention to read reviews for each purchase made at the website. A consumer who perceives that most online product reviews at a website are of high quality and consistent with the consumer's prior knowledge about the reviewed products will be more favorably disposed to reading other reviews posted at the website. This is consistent with the postulate in the ELM that central-route (e.g., perceived review quality) and peripheral-route (e.g., involving perceived review quality, and negatively moderated the peripheral-route process involving perceived review consistency, as theoretically expected, consumer expertise had a positively moderating effect on central-route processes and a negatively moderating effect on peripheral-route processes. The analysis of the study data resolved this potential anomaly by revealing that while perceived review consistency was a salient peripheral cue for consumers with low elaboration, it was also a review-relevant quality for consumers with high elaboration. Combining these two groups of consumers into a single group and testing the resulting collective moderating influence on the effect of perceived review consistency masked the differential nature of the effects of the two elaboration states. The third point raised in the opening discussion was that the present study is as important for electronic retailers as other, prior research in this area. However, unlike prior research, which was primarily based on study of the ways in which positive and negative online reviews respectively helped or hindered online product purchases, the present study examined the factors that motivate consumers to read online product reviews, with the goal of identifying consumers' views of such product reviews themselves. The hypothesis behind this was that before shopping online, consumers will read reviews that are logical, persuasive, and provide sufficient rationale for buying a product on the basis of specific facts about the product. Conversely, consumers will not read online reviews that are abusively negative, emotional, or deficient in the quality of the reasoning used to support their statements, in the form of reasons based on facts about a product. The basis for the study was that such evidence can help online vendors to decide whether or not to add an online discussion forum to their websites, and how to properly manage the product reviews for such a forum if they decide to do this. Even though some online vendors may hesitate to adopt the strategy of providing a venue at which consumers can voice their opinions, from fear of product defamation through such a venue, they ought nevertheless to take into account the product reviews provided by third parties (e.g., at eopinion.com, or www.mobile01.com), because reviews of high quality that are consonant with consumers' prior experience with products discussed at these online for a strongly influence consumers' purchasing decisions.

CONCLUSIONS
The findings in the present study show that both the central and peripheral routes of persuasion are viable pathways through which to persuade consumers to read online reviews. In the central route, consumers engage in reflective processing of issue-related arguments in an informational message, whereas in the peripheral route they simply attend to cues about the consistency of an online product review. These mechanisms of behavioural influence shape consumers' intent to read online product reviews. The results of the present study confirm that both of these routes of influence are moderated by consumers' involvement and expertise in processing issue-relevant arguments. Consumers with higher elaboration involvement and expertise tend to be more influenced by the central route of persuasion, whereas those with lower involvement and expertise are more influenced by the peripheral route.
Although this study offers important insights into an unexplored topic, it is not without limitations. The study surveyed the views or perceptions that consumers kept in mind after reading product reviews at a functioning website M a r c h 2 8 , 2 0 1 4 (www.mobile01.com), and examined these consumers' willingness to refer to related product reviews at that same website if buying a product online. The study data showed that the website used in the study is not the online retail outlet most frequently used by our sample population, but does permit online shopping as well as providing a forum for consumer discussion. In Taiwan, the most commonly utilized consumer websites include Yahoo.com, PChome.com, and ETMall.com.
A more influential study, of greater significance, would use product reviews at these online outlets for a model of the effect of online product reviews on consumer behaviour. Although this deficiency may have hurt the internal validity of our findings, the field-based data collection done in the present study improves the external validity or generalizability of the study findings as compared to that of studies conducted in laboratory settings, as have been most of the studies so far conducted of consumers' online retail behaviour. We urge that future studies use a controlled experimental design to verify the internal validity of the results of the present study.