credibility of information on e-commerce websites

25
PROCEEDINGS OF THE 13 th ANNUAL CONFERENCE ON WORLD WIDE WEB APPLICATIONS 14-16 September 2011 Johannesburg South Africa Editors: A. Koch P.A. van Brakel Publisher: Cape Peninsula University of Technology PO Box 652 Cape Town 8000 Proceedings published at http://www.zaw3.co.za ISBN: 978-0-620-51918-2

Upload: uct

Post on 16-Jan-2023

0 views

Category:

Documents


0 download

TRANSCRIPT

PROCEEDINGS OF THE 13

th ANNUAL CONFERENCE

ON WORLD WIDE WEB APPLICATIONS

14-16 September 2011 Johannesburg South Africa

Editors:

A. Koch P.A. van Brakel

Publisher:

Cape Peninsula University of Technology PO Box 652 Cape Town

8000

Proceedings published at http://www.zaw3.co.za

ISBN: 978-0-620-51918-2

2

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

TO WHOM IT MAY CONCERN The full papers were refereed by a double-blind reviewing process according to South Africa‟s Department of Higher Education and Training (DHET) refereeing standards. Before accepting a paper, authors were to include the corrections as stated by the peer-reviewers. Of the 59 full papers received, 41 were accepted for the Proceedings (acceptance rate: 69.5%). Papers were reviewed according to the following criteria:

Relevancy of the paper to Web-based applications Explanation of the research problem & investigative questions Quality of the literature analysis Appropriateness of the research method(s) Adequacy of the evidence (findings) presented in the paper Technical (e.g. language editing; reference style).

The following reviewers took part in the process of evaluating the full papers of the 13th Annual Conference on World Wide Web Applications: Prof S Berman Department of Computer Science University of Cape Town Cape Town Prof RA Botha Department of Business Informatics Nelson Mandela Metropolitan University Port Elizabeth Mr AA Buitendag Department of Business Informatics Tshwane University of Technology Pretoria Prof AJ Bytheway Faculty of Business Cape Peninsula University of Technology Cape Town Dr A Chigona E-Learning Support and Innovation Unit University of the Witwatersrand Johannesburg Prof T du Plessis Department of Information and Knowledge Management University of Johannesburg Johannesburg

3

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

Dr L Harrison Educational Technology, CELT Durban University of Technology Durban Prof M Herselman Meraka Institute, CSIR Pretoria Mr B Kalema Department of Business Informatics Tshwane University of Technology Pretoria Ms F Mohsam Faculty of Business Cape Peninsula University of Technology Cape Town Dr J Mostert Centre for Development Support University of the Free State Bloemfontein Prof L Nagel Department of Education Innovation University of South Africa Pretoria Ms C Muir Department of Strategic Communication University of Johannesburg Johannesburg Mr R Proske University Library Cape Peninsula University of Technology Cape Town Mr F Schwenke Faculty of Informatics and Design Cape Peninsula University of Technology Cape Town Prof A Singh Business School University of KwaZulu-Natal Durban

4

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

Prof JS van der Walt Department of Business Informatics Tshwane University of Technology Pretoria Prof D van Greunen School of ICT Nelson Mandela Metropolitan University Port Elizabeth Dr SC Warden Faculty of Informatics and Design Cape Peninsula University of Technology Cape Town Further enquiries:

Prof PA van Brakel Conference Chair: Annual Conference on WWW Applications Cape Town +27 21 469 1015 (landline) +27 82 966 0789 (mobile)

5

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

Credibility of Information on e-commerce websites

E Cloete Department of Information Systems, University of Cape Town

South Africa [email protected]

M Tanner

Department of Information Systems, University of Cape Town South Africa

[email protected]

M Pollock Department of Information Systems, University of Cape Town

South Africa [email protected]

Abstract The purpose of this research is to better understand how consumers assess the credibility of information on e-commerce websites. Particularly of interest is the case where there are possibly multiple resources supplying information about the product on the same website. A qualitative research methodology was pursued. Eight internet users were interviewed and the transcribed data was analysed using thematic analysis. The researchers also applied corroboration techniques across media, websites, within websites and against prior knowledge to verify claims in the information. The findings show that Internet users judged the credibility of information by cues and judgements on website and information elements and the perceived source of the information. The credibility of the source greatly influences judgements on the information. Sources that are perceived to be trustworthy and objective are judged to be credible and are hence mentioned in this research. The findings produced mainly theoretical implications for the field of online consumer research.

Keywords: e-Commerce, credibility of information, thematic analysis

1. Introduction The online shopping environment presents certain risks to the consumer - financial risk and risk that the product on sale is not as described on the website. In the online medium, where physical first-hand inspection of the product is impossible, the consumer has to rely on intangible textual descriptions and pictures (Jiang & Benbasat, 2007). Through this process of trying to gain an understanding of the product – its features, quality and experience of usage – the customer engages in processing the information. One aspect of processing is assessing whether the information is true, believable or credible (Wathen & Burkell, 2002; Metzger, 2007). One of the consequences of strongly believing the information is that the customer will be strongly persuaded by the message. Whether it is a recommendation to purchase the product or not, it leaves a strong impression on the shopper. In this respect, credibility of information can be a powerful force behind decision-making and trusting beliefs (Rieh, 2002; Pornpitakpan, 2004). In the traditional sense of credibility, it is easy to understand a phrase like “this person is credible”. The implication associated with this phrase is that whatever this person says

6

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

would also be credible (Fogg, 2002; Hong, 2006). In the online realm, such an assumption cannot be made because information can be posted by various sources to one website. Information can vary in credibility within a website (Rieh & Danielson, 2007). Moreover, it is difficult to exactly identify the source of information, since indicators like identity and reputation are not commonly disclosed. (Fritch & Cromwell, 2002; Metzger, 2007). This research will focus on information credibility in the e-commerce domain because there are significant implications as highlighted earlier. Since different bits of information on a single website can originate from different sources (hence result in different levels of credibility), this research will specifically focus on the scenario with multiple sources. However, on the web, sources of information are not necessarily disclosed, so readers merely perceive sources. The purpose of this research is thus to better understand how consumers assess the credibility of information on a website in the case where the consumer perceives more than one source of information. The main research questions are as follows.

1. What factors influence the online consumer perception of a source of information and the credibility of that source?

2. How does the user rationalize the credibility of product information where multiple information sources are perceived?

This report commences with a thorough literature review, which aims to ground this research in the existing literature. Thereafter, the methodology used to answer the research questions is detailed. The following section then covers the analysis of the collected data. A discussion and various implications in the context of the research are then detailed, followed by the conclusion.

2. Literature review

In e-commerce research, there has been much interest in understanding the concept of trust. Trust has been found to the top antecedent to e-commerce websites acceptance and usage (Koufaris, 2002; Krauter & Kaluscha, 2003; Pavlou, 2003). Sometimes, trust and credibility can be thought as being the same, but the two concepts are different. Trust in e-commerce denotes “a positive belief about the perceived reliability of, dependability of, and confidence in a person, object or process” (Tseng & Fogg, 1999, p.41). Credibility, on the other hand, is synonymous with “believability”. Credible people are believable people, and credible information is believable information (Fogg, 2002). 2.1. Credibility in the context of e-commerce research

Within e-commerce research, credibility has received attention from multiple fields - communication, human computer interaction (HCI), consumer behavior, information processing, and marketing. Owing to the richness of literature from various disciplines, Rieh and Danielson (2007) did an extensive literature review in order to develop a multidisciplinary framework for credibility. They found three levels of analysis at which credibility can be assessed – internet as medium, websites and web information, as summarized in Table 1.

7

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

Table 1 - Levels of analysis for credibility

2.2. Credibility of web information In describing information credibility, some key concepts, which have been studied in their own rights, are entangled in its definition. These refer to cognitive authority, quality, accuracy and truthfulness (Liu, 2004; Rieh & Danielson, 2007). Even concepts like accuracy and truthfulness are dimensions of so-called information quality. Once reduced, authority and quality are the two key concepts that are associated with credibility. A recent model of information credibility judgement behaviour named the unified framework for credibility assessment has been proposed by Hilligoss and Rieh (2007). This model proposes that users make a series of credibility judgements that result in the acceptance or rejection of information; consequently web information is judged on three distinct levels: construct level, heuristic level and interaction level. All the approaches introduced so far, have determine the credibility of information by looking at information, sources, users, context and understanding the various stages or levels of judgment. Tseng and Fogg (1999) approached this in a different way. They expressed four different types of credibility:

Presumed credibility results from the user‟s general assumptions on the object.

Reputed credibility results from the user‟s interpretation of third-party accounts of the object.

Surface credibility results from the user‟s interpretation of the superficial features of the object.

Experienced credibility results from the user‟s own experience with the object (Tseng & Fogg, 1999).

The “object” in the above definitions refers to the object of assessment. In the context of this research, it would be replaceable with “information”. 2.3. Information Characteristics that affect Information Credibility According to Rieh (2002), the users would assess the information object characteristics to judge its quality and authority, as described in Table 2.

8

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

Table 2 – Assessment of Information Object Characteristics

Considering that the above factors influence information quality judgements and not credibility per se, they are still worth considering because the two concepts are very similar and their meanings and dimensions are entwined with each other (Rieh & Danielson, 2007). Quality, in Rieh‟s (2002) study consisted of five facets: accuracy, goodness, importance, currency and usefulness. 2.4. Source characteristics that affect information credibility From the understanding that “credible sources are seen as likely to produce credible messages and credible messages are seen as likely to have originated from credible sources” (Rieh & Danielson, 2007, p. 310), it is expected that perceptions of the source and its characteristics would influence the credibility of information. In short, by determining the factors that makes a source credible, one can determine whether information is credible. “Source” refers to the author of information in the context of this research, since the object under credibility judgement is information. This kind of definition is not always apparent. In the unifying framework of credibility assessment, Hilligoss and Rieh (2007) do not distinctly express what “source” means when they discuss notions of source credibility and source characteristics. Hilligoss and Rieh (2007) however explain how peripheral and heuristic source cues impact the credibility assessment of information. Source cues are qualities like reputation, affiliation and author credentials. Here, sources refer to the author of the information. In the same explanation of source cues, Hilligoss and Rieh (2007) mention that past experience with using a website impacts the credibility judgment of information on that website. Here, the source now refers to the website. Despite this incongruence, the following source characteristics can be extracted, assuming that the source refers to the author specifically:

Reputation

Affiliation,

Educational background

Other factors, which are not specially characteristics, can also affect the credibility of information. These factors are the familiarity of the source and whether the source is the primary or secondary source. These factors allow the user to make judgements based on heuristics. If the user is familiar with the source, for example, the user knows the author personally, then the information is likely to be judged as more credible than if the same information were from an unknown source. Additionally, primary sources are more trusted since they produce the original information. This makes information written by primary sources to be more credible (Hilligoss & Rieh, 2007).

9

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

Following is a list of information source characteristics that impact the perception of cognitive authority. By definition, cognitive authorities are people or objects that are trustworthy, reliable, official, scholarly, credible and authoritative (Rieh, 2002). Hence the characteristics of cognitive authorities also cover the characteristics of credible sources.

URL domain – Domains like .edu and .gov influence higher authority perceptions as compared to .com domains. For example: the mention of URL domain reiterates the confusion around who the source is. It is palpable that URL domain as a feature refers to a website as a source, whereas, a feature like author credentials refers to the author of the information. Still, the URL domain, as a website feature could influence the credibility judgement of information.

Type – Sources identified as organizations were seen to have more authority than individuals.

Reputation – Well-known sources that have good reputations are the most trusted.

One vs. collective – Collective sources such as forums have different authority level to individual sources.

Author credentials

2.5. User Characteristics that affect information credibility

It must be noted that credibility is always a perceived value. Simply, two people cannot find the same object to have exactly the same level of credibility. Their individual user characteristics such as propensity for trusting behaviour, personality and past experiences can influence the credibility judgement (Hong, 2006; Metzger, 2007; Rieh & Danielson, 2007). One of the major influences is pre-existing knowledge of the subject matter (Wathen & Burkell, 2002; Hilligoss & Rieh, 2007). In fact, this is “the primary method by which people interact with content from a credibility assessment perspective” (Hilligoss & Rieh, 2007, p. 1478). The user compares the information to his/her pre-existing knowledge and actions are based on how well it matches or not. This is also coherent with what Tseng and Fogg (1999) call “experienced credibility”. However, at times when the user lacks pre-existing knowledge, acts to verify data become important. An example strategy is to check that there are multiple sources that say the same thing (Hilligoss & Rieh, 2007). Another type of pre-existing knowledge, which is mentioned separately in literature, is the level of internet experience. The amount of experience the user has with using the internet can affect their critical approach to information on the net. Non-experts are more easily persuaded and are likely to believe the information more readily. Instead, more experienced users would be more skeptical of the information and are likely to make incredulity errors (Fogg & Tseng, 1999). As users gain experience through the natural course of using the internet, they gain more skills to critically evaluate information. The fact is that credibility perceptions change as their experience grows (Freeman & Spyridakis, 2004; Metzger, 2007). 2.6. Situations and context and its influence on information credibility

According to the Elaboration Likelihood Model (ELM) by Petty and Cacciopo (1986), which has been incorporated into the credibility models by Wathen and Burkell (2002) and Metzger (2007), the way a user assesses credibility of information is either through the central or peripheral route. User involvement is identified as the factor that determines which route is taken. Involvement is defined as the motivation to process information on a topic of interest.

10

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

A highly involved user will take the central route and critically evaluate the information. This requires more cognitive effort. Low user involvement means that the user will take the peripheral route and assess the information quickly and without much interest in the consequences of the assessment (Wathen & Burkell, 2002; Metzger, 2007). The kind of information under scrutiny is another contextual factor that comes into play. For instance, users found that trustworthiness and reliability of information were more critical when seeking health information. Likewise, for health information, users wanted the basic information, without complex jargon, whereas, for subjects on travel and computer, they wanted detailed information (Rieh, 2002). Concerning information quality, some of the facets are more important than others depending on the context – what kind of information is the user looking for and how is the information going to be used? For example, accuracy is very important for health information, while currency is more important for news information. At the same time, currency is not necessarily important if the user is looking for old news (Eysenbach, Powell, Kuss & Sa, 2002; Rieh, 2002; Warnick, 2004; Metzger, 2007). 2.7. Type of information about products

Shopping in the online medium is limiting in that the product cannot be physically handled and judged by tangible means. Instead, the shopper has to rely on textual and graphical information and maybe virtual models. The product characteristics, such as look and feel, that would have been validated in-store, now gets transferred into textual information describing the experience attributes of the product (Graefe, 2003). The experience attributes of a product are, strictly speaking, the qualities associated after purchasing it and using it. By contrast, the search attributes of a product refer to qualities discerned before purchasing it (Jiang & Benbasat, 2007). 2.8. Sources of product information

There has been recent interest in the online word-of-mouth phenomenon. The output of this phenomenon has seen many e-commerce websites now with the features to allow website visitors (who could be consumers) to post comments. These consumer-created reviews have been found to be useful and understandable because the wording is familiar and users can relate to the source (Chen & Xie, 2004). This is an example where the content might not be typically described as good-quality content but is deemed credible because of the source. On e-commerce websites, users do read textual reviews and do not solely rely on rating scales (Chen, Wu & Yoon, 2004). Similarly, in a study on the effect of online reviews on cinema ticket sales, ratings do not affect the purchasing decision. Rather, sales were significantly affected by the quantity of online posting. This relationship is classified as the “awareness effect” of the reviews – i.e. the greater the number of reviews, the more awareness there is amongst the consumers, which then leads to larger sales figures. This scenario basically describes the word-of-mouth phenomenon in the online context (Duan, Gu & Whinston, 2008). Research on online consumer behaviour has studied specific players that are responsible for product information on e-tailer websites – sellers, other consumers and experts or well-known people (Senecal & Nantel, 2004; Park et al., 2007). Each of these players can post product information and each of them can influence the buyer. The availability from these multiple sources together on one website improves the shopping experience and increases trust in the website (Koufaris, 2002). From the perspective of credibility, it also improves the overall perceived fairness of the information leading to higher trust in the website (Fogg, 2002).

11

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

3. Research methodology The research was interpretive, and exploratory in nature. Qualitative data was collected through the think-aloud method and semi-structured interviews. A purposive sampling was used to obtain a sample of eight participants (P001, P002, P008). These participants were selected so that their age, race, education level and internet experience were varied. This variety was sought to avoid the risk of specific narrow data that might result from a set of similar people. 3.1. Data Collection Since the data collection method involves the researcher (to observe the participant in the process of reading a website), the locality was limited to Cape Town, South Africa. It was not feasible in the short period of data collection time to obtain participants from outside the greater Cape Town region.

Before the interview, participants were asked to name one product which they are interested in purchasing on the internet, or purchasing offline but want information from the internet. The participants were asked to be as specific as possible - provide the brand name and if applicable the model number. The researcher then selected a website for each product. This is the website that the participant would visit and actively learn about the product from. Furthermore, the website was only selected if the researcher deemed that it did host multiple information sources and where there was no explicit authorship. These criteria enforced the context and scope of this research. In the first part of the interview sessions, the participants were directed to their pre-selected website, to learn about their product of interest. The participants were told to take as much time as needed to achieve this goal. A by-product of this practical experience was that they were prepared for in depth questions that followed in the interview as the recent activity was fresh in their minds. The participants were asked to talk out loud their thoughts on the believability of the information as they process the information. The command that was given to the participant was: “Please use this website to learn about your product of interest. As you are going through this process, talk out loud about what you are thinking and comment on whether you believe the information you have read and why.” The second part was the interview. It followed a semi-structured approach. The researcher also used any interesting comments from the think-aloud session to support the questions or ask more as necessary. 3.2. Data Analysis The researcher transcribed all the voice data, together with researcher notes on participant observations. The transcriptions were coded by incorporating the thematic analysis method as described by Owen (1984). Thematic analysis is a suitable choice of analysis methods as it has been used extensively in qualitative research to analyse interview data (Huberman, 1994).

4. Findings

4.1. Perception of source and credibility of source In the process of determining the source of information, interviewees looked upon a combination of information, user and website characteristics. Each of these characteristics

12

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

Information

characteristics

impacts the perception and the credibility of the source. The general framework is proposed in Figure 1.

Figure 1 Framework on the factors influencing perception of a source and credibility of a source

As depicted above, the themes presented by the interviewees are:

Website characteristics affect the perception of source and its credibility

Information characteristics affect the perception of source and its credibility

User characteristics affect the perception of source and its credibility

Source credibility and perception of source is closely related

Theme: Website characteristics affect the perception of source and its credibility

There are certain physical features of the website which the interviewees pointed out as clues to identifying the source of the information. All interviewees who recognized the format of the information to be forum-like (where comments can be posted) either classified the sources to be the general public or unknown individuals. P003, who reviewed a CNET webpage, commented on how the layout of the webpage implied that the source is an online magazine.

“Because it‟s very article like. And I know of a typical magazine that review games, like a gaming magazine look exactly like this. They give it a score. The pictures, the text and score as well. It‟s very similar to the gaming magazines and same structure.” (P003)

The overall offering of the website can affect how competent it is perceived to be in the subject matter. If the website is truly an expert in the subject of its content, then it would need to be a reference point for the users who are very knowledgeable on the subject matter. P003 recognised CNET as an online magazine and found them to be credible, but not as credible as a „proper‟ gaming website. He found that the information presented by CNET is true, but that most of that information was gathered from another central source – a higher authority. CNET was seen as the middleman relaying information.

Theme: Information characteristics affect the perception of source and its credibility

The content of the website as interpreted by the reader provides a means to perceive the source of the content. These information characteristics are discussed in their sub-themes.

Writing style and grammar

Website

characteristics

User

characteristics

Perception of source and its

credibility

13

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

When information is interpreted to have a strong motivational or promotional writing style, it is believed to originate from a source that is trying to encourage a sale. Interviewees have classified such sources as for-profit companies or sales persons.

“And then the product description is motivational, it‟s sort of written in a way that makes you want to buy the product. That is obviously written by the company because they‟re trying to sell the product.” (P001)

Motivational writing can also be described as the opposite of factual emotionless writing. However, factual content like a table containing the technical specifications is seen to come from the manufacturer, which is also a for-profit company. In the context of existing literature, this is explainable in terms of product search and experience attributes. Product search attributes are likely viewed to originate from the manufacturer of the product. However, descriptions on product experience attributes are judged on multiple bases in order to identify its source. The style of writing is one of those bases. When the information is not motivational, but well-structured, well-reasoned and non-emotional, the interviewees tended to think that the source is an expert in the field or product. However, P007 warns not to get fooled into thinking that the source of information is an expert by judging on the writing style.

Detail and thoroughness of coverage The breadth and depth of coverage of the subject area mainly affects the impression of expertise. There is a unanimous view amongst the interviewees that content which offers high detail and thoroughness originates from an expert source. This expert source could be a for-profit company or an independent party.

“He covered everything, there‟s photos, in terms of the content and they way it‟s laid out. It‟s quite easy to read. It covers all the important issues about the phone. It gives the impression of an expert cellphone reviewer or cellphone user.” (P004)

Content features

The most widely mentioned content feature refers to profile information on the author. When asked what particular features could improve the credibility of the source, P002, P005 and P007 vehemently answered that explicit authorship, or at least credentials and even a screen name would be better than having no profile information. Profile information helps them relate to the source in a more direct manner than judging through the work of the author. P007 said that explicit authorship is also a mechanism of showing the readers that the author is proud of the work and willing to risk his/her reputation. This improves the credibility of the source as well as the information. This is a similar concept to website features like Web 2.0 tools that improve credibility judgments.

Information layout From observation, the labeling of information sections almost subconsciously implies certain sources. It was even difficult for the interviewees to answer a simple question like “Who do you think wrote this piece of information?” because it seemed so obvious to them. Sections labeled as “Comments” or “Customer reviews” or “Reader reviews” are automatically regarded to originate from common internet using public. This perception is a strong conjecture and requires no deliberation.

Theme: User characteristics affect the perception of source and its credibility

This is perhaps the most obvious theme because everything about credibility and about the interviews is entirely based on how each participant interprets the websites. The major

14

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

domains that have influenced perceptions of source and credibility judgements are bases on prior knowledge and internet experience.

User prior knowledge, experience and assumptions

This sub-theme encompasses the effects of the way the user (or the interviewees particularly) interprets the information. Such an interpretation could be rational or irrational and based on prejudices. Interviewees also approach the website with preconceptions about the information or about a particular source. The concept is that the interviewees involved prior knowledge in assessing the credibility of sources and perception thereof. Sometimes, general assumptions are made that “the people that have used them obviously have more knowledge on what they can do.” (P001). These assumptions help perceive the source and

assess the credibility of the source. P002 perceived the source to be unknown individuals of the general public based on the assumption that they are the users of the product. Product usage is an assumed trait of the general consuming public. Previous experiences with a source influence the credibility of the source even though it might appear on a different website. In this case, the source is the general public as a collective. Even though the website under assessment provided completely new individuals whom the participant has never been exposed to, the collective source was altogether rejected as a result of past experiences.

User internet experience

The less experienced internet users like P002 and P008 commented on the use of jargon as a trait of an expert source. When they came across the jargon that they did not understand, they conceded the information and the source to be credible. On the contrary, P007 with vast experience with using the internet and himself working in the online marketing industry is skeptical to trust the source. Generally, all the participants were aware that information found on the internet is not necessarily edited and that it might not be true. Some of the interviewees were not even comfortable to use the internet as a source of information. P005 commented that she would rather see a real-life practitioner about the Flugo medicine. She admitted that she had never really gone onto the internet to learn about a product like this. Again, this could be due to her lack of experience with using the internet and the fear around trusting authorless information. She did mention that author credentials and reputation was important to her. Certainly, this lack of confidence with using the internet also impacts the credibility perceptions of information on the internet.

Theme: Source credibility and perception of source is the same concept

So far, source perception has been interpreted as the act of naming the source – calling the source “expert” or “for-profit company” or “general public”. This naming is useful because it is an easier way of differentiating the bits of information with possibly different levels of credibility. However, source credibility, as a concept, can exist without having to name exactly who or what this source is. The source is simply the source of information. However, users find it useful to name sources as this conjures up typical credibility profile of such a named source. Earlier, it was explained that a motivational writing style causes the reader to perceive the source and the source‟s motives to sell the product. The perception of source and motives (or lack of motives) probably happens at the same time. These perceptions influence the credibility of the source through the presumed credibility profile of the named source. P001 recognised a for-profit company as one of the sources. She judged that the company‟s motive is to sell the product hence she became wary that the claims could be overstated. She also perceived the general public to be the authors of the “Customer reviews” section.

15

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

She judged that the general public is more reliable and trustworthy than the company because they have no incentive to stretch the truth. Consequently, the perception of a source can impact the credibility of the source through sensing ulterior motives (or lack of motives) that a source might have. 4.2. Credibility of information Participants behaved differently in assessing the information because of contextual differences. These behaviors, together with perceptions of certain qualities of information

and source credibility, eventually impact the credibility judgment of the information.

Information processing behaviour determines credibility

Flow of content and generalisation of message

When the participants start reading the content, they follow the flow of information as laid out on the webpage. If the information flows top down, this is how it will be read. If there is a link to the second page, then the participant might click on it, depending on whether the content holds enough attention to keep reading. This flow of information determines whether the user notices the various bits of information. From participant observation, the following points were identified. When the content is split up over multiple pages, prompting the user to press a link to the next page, all participants who viewed websites with this layout, namely P002, P003, P006, P007 and P008, were slightly hesitant to carry on reading onto the next page. Once the interviewer assured them that they can browse as they wished, they carried on. However, P006 and P008 did not read all the content pages. They became disinterested in the lengthy product description. Both their websites covered over 7 web pages of content. P008 also found the content difficult to understand as there was some jargon and she was not interested in the technical details of the product. The end effect was that the impression of the initial set of content is generalised over the entire content. The participants judged the credibility of information based on that generalization. To a lesser extent, it was noticed that P006 and P007 started to take less and less time to read as they progressed further along. This action of skimming indicates that the participant had already built up an opinion of the information and the skimming now simply tests this opinion. The participants‟ disinterest in the details means that they are interested in the general message. This translates to information credibility being judged on the general idea instead of the specific details. An interesting observation made with P002 and P006 is that they did not read the content consecutively from page to page. They jumped from page to page randomly based on the labeling of the links. The participants were able to navigate to the sections that interested them and in the order that they wished to read. The impact of this feature is that only the relevant information is judged.

Relevance judgement

Opinionated text can be interpreted to be “marketing talk” especially if the source is perceived to profit from the sale of the product. This can be otherwise explained that the reader recognizes the context of the information and its source and judges whether to accept or reject information based thereon. Participants P002 and P007 encountered some pieces of information, which they deemed were not applicable or not relevant to their own situation. However, they still found the information to be credible.

16

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

“The area that I‟m staying in, it doesn‟t have rough surfaces, so I guess my ride will be smooth. As far as room is concerned, they talk about “cramped for space”. I don‟t think that‟s a big issue because I don‟t have kids. Single at the moment, so it‟s only me driving the car, so that shouldn‟t be a problem.” (P002) “[expert] I wouldn‟t trust him. But for example, if he says: “It is easy to use”. I would take his word for it. But I can only know once I‟ve used it myself. I still wouldn‟t really consider what he says towards making my decision. So whatever he says doesn‟t necessarily apply to me, but I can accept if it applies to him. I would actually trust the other guys [general public] more.” (P007)

Application of effort and interest

Before the reader decides to engage with the text, a decision is first made on how to engage with it – whether to read it thoroughly, or skim through it, or disregard it. This decision is based on certain combinations of user and information characteristics. P003, P004 and P007 who already had intimate knowledge of their respective products, skimmed through the text, looking for keywords and matching the gist of the content with their prior knowledge. The credibility judgement of the information is hence based on the comparision between prior knowledge and keywords and phrases found in text. The implication of discontinuous information is discussed in the next part titled “Corroboration and dealing with contradictory information”. On the contrary, those with limited knowledge of the product, such as P002 and P005, started reading word for word and attempted to absorb the information. P007 mentioned another situation where he would rather browse through text rather than read carefully. If there was a lot of content to sift through and he wasn‟t sure whether it was worth reading, he would search for keywords or phrases that raise interest and relevance.

“A lot of the times, information is not really useful. I just get sidetracked, because some people just put a link to another website where you can get more prices and you keep going there and lose yourself. In this situation, I would browse through looking for somebody who has actually used the Vespa, instead of someone who just talks about the Vespa 25 years ago etc.” (P007)

This behaviour of browsing builds weak sentiment around the credibility of information because the reader bases it on partially read text, most of which bears no interest. The evidence lies in the hesitation and confusion when asked whether the information was found to be trustworthy or believable. This kind of question is not relevant to the participant who has not built up a judgement from reading the raw text. Information credibility becomes irrelevant when the reader does not see or read the content. P007 also found that message features like the date of publishing and geographical location are relevance factors in determining whether to read or not. He did not want to accept the information fully because the text was not very recent or the source was based in a different country.

“Because I end up judging how important the information is based on the information around the specific article… This was written in April 2008… This other guy here, he‟s from the states. I don‟t really interest myself with that.” (P007)

17

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

The level of enthusiasm perceived to be associated with the webpage can indicate usefulness of information. P004 recognises high number of comments that the general public publishes as an indicator of enthusiastic readers who find the main article interesting enough to comment. Similary, P007 finds that enthusiastic readers would publish their full profiles with pictures and the comments made by such enthusiastic public is worth reading.

“Also, I will check the number of comments – if it shows 2 comments, then I don‟t bother to read it. It means it‟s not interesting. If it shows 70 or 100, then I will bother to read it, just as a matter of interest.” (P004) “Because I end up judging how important the information is based on the information around the specific article. If he was really enthusiastic, he would have a picture.” (P007)

In brief, the application of effort into reading directly affects the importance or relevance of information credibility in decision-making. The application of effort is determined by:

how much the user already knows about the product

how lengthy the text is

geographical location of author and publishing date of the content

enthusiasm perceived from fellow general public

These above four points converge on the concept that it is “just a matter of interest” (P004). The user‟s perception of relevance and interest affects the amount of effort that goes into reading, which then impacts the credibility of the read information.

Corroboration and dealing with contradictory information

When asked whether certain information was believable, half the answers were hesitant and unsure. The participants proceeded to justify for their hesitancy by acknowledging that they have no way to check whether the information was correct unless they compared it to other sources. Most of the participants either wanted explicit authorship profiles or corroboration of information in order to surely say that they believe the information. Participants valued different perspectives which different sources provide. They valued hearing both the good and the bad about the product; it made them feel more comfortable to make purchase decisions knowing all the aspects of the product. When they came across certain claims (i.e. information was not factual) which they were unsure about, they sought to corroborate in different ways. The following methods of corroboration were mentioned:

Validation across media Sometimes, asking a friend or a live human expert is preferable. For P002, whatever he has read on the internet will be double checked with his mechanic.

“I generally believe this statement… What I know about engines, I‟m at the mercy of a mechanic. I do have a person that I know that does mechanical work, so I‟ll probably use that person to steer me in the right direction. Well if this is true what they say here, then there is a little problem with the cam shaft. If I do take a closer look, I will ask the mechanical person to have a look at that specific part of the car – seeing that this statement that they‟ve put here might be true.” (P002)

Validation across multiple websites

18

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

P006 came across some negative comments about the Nokia E75. Although he finds these comments to be credible, he wanted to read other websites to validate them. Similarly, the experienced internet shopper, P004, admits that he gets influenced by whatever he reads, so he has “to read a lot more than just one website..”, (P004) to balance out the influence.

When there are contradictory claims or information that does not match well, P007 quoted that he would try and find support on other websites.

Validation across multiple sources on one website/webpage All the participants who had identified multiple information sources on their respective websites mentioned that they valued information from the different perspectives. This type of validation is strongly linked to source credibility as the motives of each source are considered. P001 recognised sources with two different set of motives. Sources with the similar motives would have similar credibility ratings, in which case, they are effectively one source and would not count towards efforts for corroboration.

Validation within source This type of validation is to ensure that the source is delivering a consistent message. Consistency of message indicates the source‟s ability to communicate with reason and logic. Interestingly P002 tested consistency of message by validating textual information against pictorial information.

“I‟m looking at the part, where it says „can wallow through corners‟. What he‟s said and what the picture has shown, it fits in. It sounds credible.” (P002)

A source such as the general public, because it is made up on many different individuals, is judged to have a consistent message if there is consensus among most of the comments made by these different individuals.

“If a lot of people have a similar experience, and it helps you see a pattern if it‟s really good or not, because you have an input of so many people. The larger the sample size, the more accurate the result.” (P007)

Here, P007 revealed the concept of strength of validation. If there is some sort of consensus, the quantity of opinions will strengthen the validation.

Validation against own user knowledge When reading any bit of information, participants naturally use their own prior knowledge (which includes prejudices and preconceptions) to process the information and understand it. This process will identify bits of information that correlate with their prior knowledge. This is a form of validation that determines whether the information is correct or not. P002 read a statement that the Nissan Almera engine is good and reliable. This confirmed his own preconception about Nissan (whether it originated from hearsay or not) and he deemed the website information to be credible. P008 quotes in distress upon reading unfamiliar text: “They‟ve made it so complicated that that‟s why I trust them. They sound like they know what they‟re talking about so that must be true”. She later on went to say that she would need to

validate this unknown information with many other sources before she can say it is correct.

“I do trust the information. I have no reason not to trust it. I don‟t question the information they put forward. Without comparing it, I wouldn‟t even know where to start. If it tells me this Nokia has a USB and exterior Bluetooth… if I haven‟t read about the E75 elsewhere, I would believe this as is.” (P006)

19

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

Based on this quote, this sub-theme can be summarised that until there is evidence or support about a claim, the reader will not cannot easily perceive its correctness. Correctness of information is very tightly knit to its credibility. In other words, the reader will believe the information if it is thought to be correct. All the various methods of validation are methods that the participants quoted they would use in order to validate the correctness of information.

Qualities that affect information credibility

There are certain qualities or descriptions about the information that contribute towards its credibility.

Correctness Participants could not distinguish between the correctness and accuracy dimensions. However, information that was perceived to be correct and/or accurate directly led to credibility. Correctness is also linked to how factual it seems. Product search attributes, which are mainly shown in a specification table are identified to be factual, hence correct.

“Correctness, I only see it in terms of hard facts, for example this is a 125cc motorcycle. Even if I don‟t know anything about it.” (P007)

However, product experience attributes by nature is not factual and participants had a difficult time determining if the information is correct or not. Owing to this, they were wary to trust opinions completely.

Currency

Some participants were concerned with accepting information only if it was published recently. P007 comments: “The dates of when this review was done and where it was done from makes a difference.” The implication is that out-of-date information is not relevant,

hence not credible for the purposes of the participant‟s information needs.

Provision of evidence

Information that is adequately reasoned and logically explained is certainly more believable than vague statements about the product (P001, P004, P008). Besides writing style, pictorial support to textual information also improves the credibility of the information. The participant effectively wanted to use this graphical information to verify claims in the text. P001 quotes: “To believe it more, I think they could put pictures on it so that you can see what the product does. For e.g. they say it creates curls, so perhaps put a picture showing what it does.”

Objectivity Although participants validate contradictory information that they might find, they still mentioned that they value different opinions, and over above that, they value reading about the pros and cons of the product. P005 mentioned that it is only realistic and truthful that products have both a good and bad side to them.

“I think it would be great if someone said something bad about the product. Because it would be truthful because it might work for someone but not someone else.” (P005)

This provides some insight as to why participants tended to believe the general public more than other sources. The general public as a source are more likely to tell the truth and not be afraid to mention negative aspects of the product – there is no reason for them to be inhibited.

20

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

Goodness The provision of a signal of goodness of information, such as tools that allow viewers to distribute the information, can improve credibility. Web 2.0 tools like bookmarking, tagging and sharing with friends are seen as signals that imply good quality information in the general sense.

“For example, you can bookmark all of this, email a friend things like that. That goes to show you how credible it is in a way. If it was not really credible, you wouldn‟t get these features. If you don‟t want too many people to find out about it because of plagiarism for example, then you won‟t have these functions.” (P007)

Usefulness

The perceived usefulness of information affects its credibility through relevance judgements. A piece of information that is not useful is not read; hence the credibility becomes unimportant, or not believable in the eyes of the user and in that specific context. P003 explained that he did not read the general public commentary, “because it‟s generally other people‟s opinions… these people just put their comments and opinions on the article, so not interested in what they have to say.” (P003)

Depth and breadth Participants were unanimous about the benefits of displaying detailed information. P003 said that a summary of the information is useful. However when he was asked whether it would be credible if only the summary was there, he quickly retracted and said that there would have to be a link to the detailed information behind the summary. Depth can be achieved by displaying pictures as well. In fact, pictures from multiple angles were desired. Going one step further, P002 mentioned that a video would be more convincing than pictures. Similarly breadth of coverage is important as well. The main product information on a website is especially expected to cover all the aspects of the product.

Credibility hierarchy

Earlier, it was recognized that user characteristics like internet experience and prior knowledge affect how users perceive sources of information and determine source credibility. All of these factors also information credibility. In other words, whatever affects source credibility has the same effect on information credibility. However, these two concepts are not the same. (This is discussed further in the next section.) By and large, whenever sources were found to be believable, participants believed the information to be correct and true. In a relative context, this means that relatively more credible sources produce relatively more credible information. On another scale, P004 had used the assessed website, Gizmodo.com before and owing to his positive pre-conceptions about the website, he found everything about the website, the information sources and the information on it to be trustworthy.

“First thing, I think this website, Gizmodo. I‟ve been there before and I think it‟s reputable. I‟ve been there before, so it‟s quite trustworthy.” (P004)

There seems to be a hierarchy whereby the credibility of the media influences the credibility of the website, which in turn influences the credibility of the information on the website. These influences do not completely define the credibility of the resource – they are merely factors among many others.

21

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

Source credibility affects information credibility

These two concepts seem to be two sides of the same coin. Information from a credible source is mostly credible. Although the final evaluation on the information does not solely rely on the credibility of the source, it certainly has an effect. The persuasive power of a credible person, or a cognitive authority is evident in the quote: “ If the product is endorsed by a credible person, it [the information] would be more credible” (P005). The more experienced user, P004 was doubtful that all the information on the website is correct although the source might be highly credible. “Even expert writers are biased and

can make mistakes” (P004). At the same time, non-expert sources like the general public

produce highly credible information (P001, P002, P006, P007). This challenges the definition of source credibility. The trustworthiness dimension seems to be far more important than the expertise dimension. Furthermore, these participants found the general public source to be more credible because they do not have ulterior motives that might affect the information, and hence they are seen to be more reliable and objective (P001, P002, P006, P007). This suggests objectivity as a desirable dimension of source credibility. The perception of a source brings with it certain associated preconceptions. Without having fully read what the general public had written, most participants were already sure to claim that they would believe them over other sources of information.

4.3. Summary of data analysis

The model presented below in 1 was produced from the data analysis results on both research questions. As per research question 1, there are website, information and user characteristics that affect the credibility of the source (after first identifying the source). Then, these same characteristics also help judge the various qualities of information. At the same time, all of these judgements of quality and source credibility impact how the user behaves to process the information. In the end, all of this affects how the user perceives the credibility

of a piece of information.

22

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

Qualities of information

Objectivity

Correctness

Currency

Depth and breadth

Usefulness

Goodness

Provision of evidence

Source credibility

Trustworthiness

Objectivity

Website and information features

Layout

Presentation and flow

Website functionality

Writing style

Level of detail

Content features

Graphics

User characteristics and context

Past knowledge of subject

Internet experience

General assumptions

Information processing behaviour

Flow

Generalization of message

Relevance judgement

Application of effort

Corroboration

Dealing with contradictions

Information credibility

Source characteristics

Credentials

Authorship

Reputation

Figure 2: Model of information credibility after the data analysis

5. Conclusion This study addressed the lack of theory on information credibility in general. This was achieved by studying e-commerce websites as it could have significant implications on consumer buying decisions. This is an exploratory study since there is not much support from existing literature regarding specifically the credibility of web information. The purpose was to understand how internet users rationalise the credibility of information on e-commerce websites. First, it was necessary to understand how users might perceive the source of information as most e-commerce websites do not reveal authorship and credential detail. Secondly, there can be multiple sources of information on a website. Hence the purpose in context is to understand how users rationalise the credibility of information on authorless e-commerce websites, where multiple information sources persist. Web information does not exist by itself – it exists in the context of the website that hosts it, and the website exists within the internet infrastructure. Credibility relationships between these levels of assessment also mimic the physical hierarchy. Only through the positive assessment of the internet as medium, will websites be even relevant for assessment. Similarly, only if the website itself is credible, will the information in it be relevant for assessment. As such, the credibility of the object/s higher in the hierarchy will always affect

23

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

the credibility of objects lower in the hierarchy. This kind of distinction of the three levels of analysis greatly clarifies the contexts in which past research can be analysed, and on which future research can be based. This theoretical contribution to the field of credibility would also be useful for future depictions of end-to-end credibility evaluation behaviour. In the e-commerce domain, assessment of source credibility is not heavily dependent on perceptions of expertise. What‟s more important is trustworthiness and objectivity. Sources who are perceived to have ulterior motives to sell the product, were judged not be objective in their delivery of messages. This made them relatively less credible than sources like other consumers who have no incentive to provide untrue information. Apart from the obvious contribution of redefining the credibility construct, this research also confirms that dimensions are not fixed. They can change for different objects of assessments, as well as different types of information. Practically, website owners can manage content to ensure that the writing style is not motivational, well-justified and detailed. What's more, the disclosure of authorship details and credentials go a long way to improve perceptions of the source. In the e-commerce context, believable sources produce believable information, which can eventually persuade buyers to follow recommendations. There was focus on understanding the dynamics behind credibility assessments when multiple sources are perceived. The findings are encapsulated in a set of behaviours that internet users adopted. Based on perceptions of website and information elements, along with the interpretation of these elements and judgements of source credibility, users behaved in certain ways that impacted on the credibility judgement of the information. The key contribution is understanding how users dealt with contradictory information found on one website. The users appreciated debate where arguments were supported by evidence. However, when websites contain gross contradictions without support for opinions, users reject all the information. When users felt unsure of the correctness of information, they used various techniques to verify: compare across media, across websites, within the website against different sources, within the source, and against own knowledge. This theoretical contribution is a milestone in understanding online consumer behaviour through the credibility lens. 5.1. Recommendations for future research This study has mostly contributed to advancement in theory – particularly theory in online consumer behaviour because it is studied through the field of credibility and not the popular field of trust. It would be interesting to see how the results of this research align with trust research and if a higher contribution can be made from the merging of the two. In terms of HCI, there have been many studies that have yielded practical implications (Fogg et. al, 2003; Tombros et al., 2003). These have been based on credibility judgements of websites and heavily based on aesthetic features. This study has produced implications based more on the content features and information qualities like objectivity and provision of evidence to name a few. Within HCI, the implications refer to content management and information management on websites. However, in order to implement practical solutions, there needs to be more research done on exactly what qualities are important for what type of information and how these qualities can be inspired. This research has only given a list of blanket qualities of information.

Acknowledgements The authors also acknowledge and appreciate the dedicated work done by Cho Walton in 2009, a post-graduate student at the University of Cape Town that have been involved with this research project (Walton, 2009).

24

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

6. References

Chen, P.Y., Wu, S.Y., & Yoon, J. (2004). The impact of online recommendations and consumer feedback on sales. Proceedings of the International Conference on Information Systems, 711-724.

Chen, Y., & Xie, J. (2004). Third party product review and firm marketing strategy. Marketing Science, 24(2), 218-240.

Duan, W., Gu, B., & Whinston, A.B. (2008). Do online reviews matter? An empirical investigation of panel data. Decision Support Systems, 45, 1007-1016.

Eysenbach, G., Powell, J., Kuss, O., & Sa, E.R. (2002). Empirical studies assessing the quality of health information for consumers on the World Wide Web: A systematic review. Journal of the American Medical Association, 287, 2691-2700.

Fogg, B.J. (2002). Stanford guidelines for web credibility. Stanford Persuasive Technology Lab, Stanford University. Retrieved April 12, 2009, from http://credibility.stanford.edu/guidelines/index.html.

Fogg, B.J., Soohoo, C., Danielson, D.R., Marable, L.,Stanford, J., & Tauber, E.R. (2003). How do users evaluate the credibility of web sites? A study with over 2,500 participants. Proceedings of the 2003 Conference on Designing for User Experiences (DUX‟03).

Freeman, K.S., & Spyridakis, J.H. (2004). An examination of factors that affect the credibility of online health information. Technical Communication, 51, 239-263.

Fritch, J.W., & Cromwell, R.L. (2002). Delving deeper into evaluation: Exploring cognitive authority on the Internet. Reference Services Review, 30(3), 242-254.

Graefe, G. (2003). Incredible information on the Internet: Biased information provision and lack of credibility as a cause of insufficient information quality. Proceedings of the 8th International Conference on Information Quality, 133-146.

Hilligoss, B., & Rieh, S.Y. (2007). Developing a unifying framework of credibility assessment: Construct, heuristics, and interaction in context. Information Processing and Management, 44(4), 1467-1484.

Hong, T. (2006). The influence of structural and message features on website credibility. Journal of the American Society for Information Science and Technology, 57(1), 114-127.

Huberman, M.M. (1994). Qualitative data analysis: An expanded sourcebook. London:

Beverly Hills.

Jiang, Z., & Benbasat, I. (2007). The effects of presentation formats and task complexity on online consumers‟ product understanding. MIS Quarterly, 31(3), 475-500.

Koufaris, M. (2002). Applying the Technology Acceptance Model and Flow theory to online consumer behaviour. Information Systems Research, 13(2), 205-223.

Krauter, S.G., Kaluscha, E.A. (2003). Empirical research in on-line trust: A review and critical assessment. International Journal of Human Computer Studies, 58, 783-812.

Liu, Z. (2004). Perceptions of credibility of scholarly information on the web. Information Processing & Management, 40, 1027-1038.

Metzger, M.J. (2007). Making sense of credibility on the web: Models for evaluating online information and recommendations for future research. Journal of the American Society for Information Science and Technology, 58(13), 2078-2091.

25

Proceedings of the 13th Annual Conference on World Wide Web Applications

Johannesburg, 14-16 September 2011 (http://www.zaw3.co.za) ISBN: 978-0-620-51918-2

Owen, W.F. (1984). Interpretive themes in relational communication. Quarterly Journal of Speech, 70, 274-287.

Park, D., Lee, J., & Han, I. (2007). The effect of on-line consumer reviews on consumer purchasing intention: The moderating role of Involvement. International Journal of Electronic Commerce, 11(4), 125-148.

Pavlou, P.A. (2003). Consumer acceptance of electronic commerce: Integrating trust and risk with the Technology Acceptance Model. International Journal of Electronic Commerce, 7(3), 69-103.

Petty, R.E., & Cacioppo, J.T. (1986). The elaboration likelihood model of persuasion. Advances in Experimental Social Psychology, 19, 123-205.

Pornpitakpan, C. (2004). The persuasiveness of source credibility: A critical review of five decades‟ evidence. Journal of Applied Social Psychology, 34(2), 243-281.

Rieh, S.Y. (2002). Judgment of information quality and cognitive authority in the web. Journal of the American Society of Information Science and Technology, 53(2), 145-161.

Rieh, S.Y., & Danielson, D.R. (2007). Credibility: A multidisciplinary framework. Annual Review of Information Science and Technology, 41, 307-364.

Senecal, S., & Nantel, J. (2004). The influence of product recommendations on consumers' online choices. Journal of Retailing, 80, 159-169.

Tombros, A., Ruthven, I., & Jose, J.M. (2005). How users assess web pages for information seeking. Journal of the American Society for Information Science and Technology, 56, 327-344.

Tseng, H., & Fogg, B.J. (1999). The elements of computer credibility. Proceedings of the SIGCHI conference on Human Factors in Computing Systems, Pittsburgh, Pennsylvania,

80-87.

Walton.C. (2009). Credibility of Information on e-commerce websites. Honours Dissertation, Department of Information Systems, UCT, 2007.

Warnick, B. (2004). Online ethos: source credibility in an “authorless” environment. American Behavioral Scientist, 48(2), 256-265.

Wathen, C.N., & 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.