Trust and distrust on the web

FIND A SOLUTION AT Academic Writers Bay

Trust and distrust on the web: User experiences and website
characteristics
Mirjam Seckler ⇑, Silvia Heinz, Seamus Forde, Alexandre N. Tuch, Klaus Opwis
University of Basel, Department of Psychology, Center for Cognitive Psychology & Methodology, Missionsstr. 62a, 4055 Basel, Switzerland
a r t i c l e i n f o
Article history:
Available online 17 December 2014
Keywords:
Trust
Distrust
Experience
Critical incidents
Design
Websites
a b s t r a c t
The aim of this research is to study the content of trustful and distrustful user experiences on the web to
identify website characteristics that enhance trust or cause distrust. We collected users’ reports about
critical incidents and quantitative questionnaire data by means of an online survey. Results from
N = 221 participants suggest that distrust is mostly an effect of graphical (e.g., complex layout) and structural (e.g., pop-ups) design issues of a website, whereas trust is based on social factors such as reviews or
recommendations by friends. The content of a website affects both trust and distrust: privacy issues had
an effect on distrust and security signs enhanced trust. Furthermore, we show how trustful and distrustful user experiences differ in terms of perceived honesty, competence, and benevolence. High honesty
and competence characterize a trustful experience, whereas a distrustful experience is associated with
missing honesty and missing benevolence. We discuss how different website characteristics help to
enhance trust or to prevent distrust and how this impacts the allocation of design resources.
2014 Elsevier Ltd. All rights reserved.
1. Introduction
Designing for trust in technology-mediated interaction is an
increasing concern in human–computer interaction
(Riegelsberger, Sasse, & McCarthy, 2005). As the online environment features many possibilities for fraud such as identity theft,
credit-card fraud and unfulfilled product promises, users are eager
to find out whether a particular website is trustworthy or not. In ecommerce, trust was found to be one of the main factors for customers buying a product or in the event of distrust, aborting the
shopping process (Jarvenpaa, Tractinsky, & Saarinen, 1999;
Schlosser, White, & Lloyd, 2006). For information websites, judgments about their quality are based on trust in the website
(Wathen & Burkell, 2002). Moreover, users’ trust is a predictor
for the usage of social network sites (Sledgianowski & Kulviwat,
2009) and leads to a higher intention to send and receive information in virtual communities (Ridings, Gefen, & Arinze, 2002).
In the last 15 years, a considerable amount of research has
investigated how to increase trust in the online context (see
Beldad, De Jong, & Steehouder, 2010). However, comparatively
little research has investigated how to prevent distrust. Recent
studies suggest that trust and distrust are two distinct constructs
and differ qualitatively from each other (e.g., Ou & Sia, 2010).
Nonetheless, only a few studies about website characteristics have
integrated both trust and distrust in the same empirical research
(Andrade, Lopes, & Novais, 2012; Chang & Fang, 2013; Cho, 2006;
McKnight & Choudhury, 2006; Ou & Sia, 2010). As Chang and
Fang (2013) noted, there is a need for studies that examine
whether trust and distrust have different antecedents. It is not
clear what web users watch out for when they decide whether a
website is trustful or distrustful. Moreover, determining whether
trust and distrust are distinct constructs has significant implications for website design and management because different website characteristics may need to be managed in order to enhance
trust and to reduce distrust (Ou & Sia, 2010).
To address this gap, the present study aims to simultaneously
investigate web trust and distrust by means of the critical incidents
technique (Flanagan, 1954) and subjective questionnaire data. We
analyze the content of 221 incident reports on trust and distrust
obtained from an online study about users’ past web experiences.
This method enabled us to gain insight into how and why people
trust or distrust a website and to gather information about specific
website characteristics related to trust and/or distrust. The present
research aims to provide new perspectives explaining how the formation of web trust and distrust is significant. We show that web
trust and distrust are affected by different antecedents and that
trustful and distrustful user experiences differ in terms of
http://dx.doi.org/10.1016/j.chb.2014.11.064
0747-5632/ 2014 Elsevier Ltd. All rights reserved.
⇑ Corresponding author at: University of Basel, Department of Psychology, Center
for Cognitive Psychology and Methodology, Missionsstrasse 62a, CH 4055 Basel,
Switzerland. Tel.: +41 (0) 61 267 06 17; fax: +41 (0) 61 267 06 32.
E-mail address: mirjam.seckler@unibas.ch (M. Seckler).
Computers in Human Behavior 45 (2015) 39–50
Contents lists available at ScienceDirect
Computers in Human Behavior
journal homepage: www.elsevier.com/locate/comphumbeh
perceived honesty, competence, and benevolence. Furthermore, we
highlight important implications for web designers and managers
on how to enhance users’ trust or to prevent distrust by optimizing
specific website characteristics.
2. Related work
2.1. Trust in an online context
Trust is an essential factor in many kinds of human interactions, allowing people to act under uncertainty and with the risk
of negative consequences (Flavián, Guinalíu, & Gurrea, 2006). It
also plays a crucial role in human–computer interaction due to
the high complexity and anonymity associated with e-commerce,
e-banking or information search (Wang & Emurian, 2005). Presently, however, researchers have difficulty in operationalizing
what exactly trust is and there exist multiple definitions in the
literature. This is likely because trust is an abstract concept and
is often used interchangeably with related concepts such as credibility, reliability, or confidence. Thus, to define the term and to
delineate the distinction between trust and its related concepts
have proven challenging for researchers (e.g., Wang & Emurian,
2005). Moreover, although trust has been widely studied in many
disciplines, but each discipline has its own understanding of the
concept and different ways to operationalize it. In their review
about trust in the context of the online environment, Wang
and Emurian (2005) highlighted two characteristics that most
definitions have in common. First, there must exist two specific
parties in any trusting relationship: a trusting party (trustor)
and a party to be trusted (trustee). In online trust, the trustor
is typically a user who is browsing a website, and the trustee
is the website, or more specifically, the merchant that the
website represents. Second, trust involves vulnerability. Trust is
only needed, and actually flourishes, in an environment that is
uncertain and risky. Users are often uncertain about the current
risks and their full consequences when transacting or visiting
online websites.
As suggested in the literature, trust is a multidimensional construct (Chen & Dhillon, 2003), consisting of three different facets:
benevolence, honesty, and competence (e.g., Casaló & Cisneros,
2008; Casaló, Flavián, & Guinalíu, 2007; Chen & Dhillon, 2003;
Flavián et al., 2006). Benevolence is related to the user’s belief that
the other party is interested in his welfare, motivated by a search
for a mutually beneficial relationship and without intention of
opportunistic behavior (Flavián et al., 2006); namely, that a website is concerned with the present and future interests, desires
and needs of its users and gives useful advice and recommendations. Honesty is the belief that the other party will keep his or
her word, fulfill promises, and be sincere (Doney & Cannon,
1997). For websites, this means that there are no false statements
and the information on the site is sincere and honest. In turn, competence means that the website has the resources (whether technical, financial, or human) and capabilities needed for the successful
completion of the transaction and the continuance of the relationship (Casaló & Cisneros, 2008).
In recent years, a lot of research has been conducted into the
importance of trust in an online context. In e-commerce, trust
has been shown to have an important positive influence on the
intention to buy a product (Bart, Shankar, Sultan, & Urban, 2005;
Jarvenpaa et al., 1999; McKnight, Choudhury, & Kacmar, 2002;
Schlosser et al., 2006). On social networks, users are more likely
to contact friends and to connect with other users if they trust
the website (Almadhoun, Dominic, & Woon, 2011). Additionally,
people’s intentions to share more of their personal information
increases if they trust a website (Bart et al., 2005; McKnight
et al., 2002).
2.2. Trust and distrust as distinct constructs
Although the extant research on trust has revealed how trust
can be built and maintained, the topic of distrust has been relatively neglected. For a long time, researchers viewed trust and distrust as extreme values along the same dimension (Schoorman,
Mayer, & Davis, 2007). However, in more recent research it is
argued that trust and distrust are not opposite ends on the same
conceptual spectrum but actually two distinct constructs that
coexist (for an overview see Chang & Fang, 2013). Distrust is
defined as unwillingness to become vulnerable to the trustee based
on the belief that the trustee will behave in a harmful, neglectful,
or incompetent manner (e.g., Benamati, Serva, & Fuller, 2010). As
antecedent of this unwillingness, users’ generally have negative
expectations regarding a website’s conduct, characterized as suspicion, wariness and fear of transactions (e.g., Lewicki, McAllister, &
Bies, 1998).
The deliberation of trust and distrust can be traced back to
ambivalence theories on examining positive-valent and negativevalent attitudinal reactions (Ou & Sia, 2010). Two main arguments
have been used to defend this approach (Andrade et al., 2012): (a)
distrust may co-exist with high trust at the same time (e.g.,
McKnight & Choudhury, 2006) and (b) high trust does not necessarily mean low distrust, and the absence of trust is not enough
to necessarily create distrust (Lewicki et al., 1998). Furthermore,
evidence from neuroscience theories and functional brain-imaging
studies have shown that trust and distrust are connected to different cortical regions. Whereas distrust is associated with the amygdala and the right insular cortex, trust is linked to the caudate
nucleus and the medial prefrontal cortex (Dimoka, Pavlou, &
Davis, 2007).
However, Schoorman et al. (2007) raised concerns about the
deliberation of trust and distrust as distinct constructs, arguing
that most studies do not account for different attribution factors.
The authors concluded that it is possible to experience distrust
and high trust at the same time due to attribution factors such as
trusting a colleague to do a good job collaborating on a research
project but not trusting him/her to do a good job teaching your
class in your absence.
To sum up, little is known as to how trust is formed differently
in contrast to distrust and to what extent distrust affects behavioral outcomes differently compared with lack of trust (Cho,
2006; Ou & Sia, 2010). However, determining whether trust and
distrust are actually two distinct constructs has significant implications for website design and management (Ou & Sia, 2010).
2.3. Facets of trust and distrust
Several authors found that trust and distrust are built up of the
same three facets, which are – as discussed above – benevolence,
honesty and competence (e.g., Casaló et al., 2007; Cho, 2006). There
is little research, however, that has investigated potential differences between the three facets for distrust and trust experiences
in the web design context. Cho (2006) conducted a study about
business-to-consumer Internet exchange relationships. She identified the benevolence and competence of e-vendors as the two key
antecedents of trust and distrust. The results of Cho’s study (2006)
showed that trust is primarily driven by benevolence whereas distrust is based on a lack of competence.
2.4. Website characteristics
The characteristics of a website are important determinants for
web trust (Shankar, Urban, & Sultan, 2002). McKnight et al. (2002)
suggest that as a first step, users explore a website before
being ready to do transactions. At this initial stage, website
40 M. Seckler et al. / Computers in Human Behavior 45 (2015) 39–50
characteristics such as reviews or content quality play a crucial
role. But it is still unclear which further website characteristics
are relevant for the formation of trust or distrust. Without attempting to identify these characteristics, it is difficult to derive effective
and reliable design principles or implications for enhancing users’
trust or lowering users’ distrust in websites (Wang & Emurian,
2005). A fair amount of research has therefore been carried out
on the influence that website characteristics have on trust (e.g.,
Bart et al., 2005; Ou & Sia, 2010; Wang & Emurian, 2005). For
example, brand strength, third-party statements and user friendliness were found to affect web trust (Shankar et al., 2002). But
effects on distrust have long been overlooked.
The few studies about the effect of website characteristics on
trust and distrust provide a more differentiated view. Regarding
trust, Chang and Fang (2013) and Andrade et al. (2012) showed
that correct and helpful website information as well as informative
customer service (e.g., the possibility to contact a vendor via
e-mail) lead to higher trust but have no effect on distrust. This is
in line with Ou and Sia (2010), who confirm that customer service
leads to higher trust but does not influence distrust. Regarding distrust, Ou and Sia (2010) discovered that the quality of the content,
the technical functionality, and similarity to other websites help to
eliminate users’ distrust. Other characteristics were found to influence trust and distrust, or neither of them. Andrade et al. (2012)
found an effect of unpleasant design on trust and on distrust.
Elements of social proof such as ratings (e.g., ratings on travel
websites) influenced neither trust nor distrust (Ou & Sia, 2010).
For other website characteristics, there exist several discrepancies between different studies. According to Chang and Fang
(2013), distrust is lowered if the website is linked to a positive
image/brand, although there is no effect on trust. Andrade et al.
(2012), however, found the opposite effect (i.e., the brand and
the logo of a bank lead to higher trust but have no effect on distrust). Similar discrepancy is found for the easiness to browse the
website and privacy policy and security indications (e.g., security
symbols by third parties). According to Chang and Fang (2013),
these characteristics have no effect on trust or distrust. Ou and
Sia (2010), in contrast, found that these three factors contributed
to both trust and distrust. And finally, Andrade et al. (2012) concluded that only privacy statements and security signs (such as
lock symbols) influence both trust and distrust, but ease of browsing the site only has an effect on trust (but not on distrust).
The above-described studies have two main shortcomings. First,
they focused only on website characteristics that were defined in
advance and inquired by means of predefined scales. This may lead
to a limited conceptualization of trust and distrust as it is not clear
whether users care for the predefined characteristics and whether
content validity is given for trust as well as for distrust. Second,
previous studies only evaluated website characteristics on a limited number of different websites (mostly one or two; e.g., Ou &
Sia, 2010) and website types (mostly online shops or online banking sites; Andrade et al., 2012; Benamati, Serva, & Fuller, 2006,
2010; Chang & Fang, 2013; Cho, 2006; McKnight & Choudhury,
2006). In our study, we apply a more holistic approach in the sense
that we allow participants to describe personal experiences that
led to trust or distrust without restricting them to predefined evaluation criteria or to a specific type of website. By doing so, we
expect to get a more detailed and comprehensive description of
which website characteristics are important for users and have
an influence on trust or distrust.
2.5. Dimensions for website characteristics
Based on a literature review, Wang and Emurian (2005) found
four dimensions which incorporated the existing website
characteristics that induce trust: (1) The graphic design refers to
the websites’ graphical elements that trigger the users’ first
impressions. This comprises the overall visual design of a site,
including layout, typography, font size, and color schemes used
on the page as well as photo quality. (2) The structure design refers
to accessibility by users to the information displayed on the website and how the website is generally organized. Usability in general and help such as prompts, guides, tutorials, and instructions
in particular contribute to a good structure design, whereas broken
links, ads and inconsistencies lead to a bad structure design. (3)
The content design includes informational elements that are placed
on the website, either textual or graphical (e.g., correct information
or company logo). Furthermore, the use of seals of approval or
third-party certificates, a relevant domain name, links to security
and privacy policies as well as the use of comprehensive and correct information belong to this dimension. Finally, (4) social-cue
design refers to social cues that are integrated into the website such
as photographs and names of customer service agents, chat and
call-back opportunities, and photographs of the company.
2.6. Trust research applying critical incidents technique
A technique becoming increasingly important for trust research
is the critical incidents technique (CIT) (Münscher & Kühlmann,
2012). The CIT is a method of gathering facts (incidents) from users
of an existing system to gain knowledge of how to improve or
maintain the performance. According to Flanagan (1954, p. 338)
‘‘an incident is critical if it makes a ‘significant’ contribution, either
positively or negatively to the general aim of the activity.’’ Typically, critical incidents can be gathered by asking respondents to
tell a story about an experience they have had. Detailed analysis
of critical incidents enables researchers to identify similarities, differences and patterns, and to seek insight into how and why people
engage in the activity. Since its introduction by Flanagan (1954),
CIT has proven valuable in a number of research disciplines such
as education, service marketing and management (for an overview
see Münscher & Kühlmann, 2012). Uppvall (2009) showed through
the use of the CIT that maintaining trust is a key factor if two parties work together in product development. Moreover, Scarbrough,
Swan, Amaeshi, and Briggs (2013) used CIT to explore the role of
trust in the deal-making process for early-stage technology ventures and showed that the form of trust changes during the process. To the authors’ knowledge, there are no studies on trust or
distrust research applying the CIT in HCI. However, as Münscher
and Kühlmann (2012) have already noted, a joint look at critical
incidents enhancing trust and critical incidents causing distrust
can help to give a better understanding of the nature of trust and
distrust development.
2.7. Aim of the study and study rationale
The purpose of our study is to gain qualitative data on trustful
and distrustful experiences on different types of websites. Moreover, we also aim to supplement these experiences through quantitative data on the facets honesty, benevolence, and competence
for all experiences. The rationale behind this approach is to gain
insight into how and why people trust or distrust a website and
to gather information about specific website characteristics related
to trust and/or distrust. Our goal is to investigate whether certain
characteristics mainly evoke trust or distrust, or whether there
are characteristics that are relevant for trust and distrust or for
none of those. We want to outline important implications for
web designers and managers on how to enhance users’ trust or
to prevent distrust by optimizing specific website characteristics.
M. Seckler et al. / Computers in Human Behavior 45 (2015) 39–50 41
3. Method
We collected data using a web-based survey containing 27
questions (and further 41 questions for another research project
about privacy and security).
3.1. Design
A between-subject design was used for this study. The independent variable was the quality of the reported experience (trustful
vs. distrustful). Approximately half of the participants (n = 103)
were asked to describe an incident where they had felt exceptionally trustful about using a website, the other half (n = 118) were
asked to report on an incident where they had felt exceptionally
distrustful about using a website.
3.2. Questions
The questionnaire applied the critical incidents technique
(Flanagan, 1954) by beginning with the key item, which was an
open-ended question. The aim of the question was to receive
descriptions of trustful and distrustful web experiences:
‘‘Please think of an occasion where you felt exceptionally distrustful
using a website, for example with an information site, a social network or an online shop. Think of distrustful in whatever way makes
sense to you. Please try to describe your experience as accurately
and detailed as you remember it.’’
For the group that had to describe a trustful user experience, the
description was changed slightly by changing the word ‘‘distrustful’’ to ‘‘trustful’’. Questions about online user trust/distrust were
the same as used by Casaló et al. (2007) and Flavián et al. (2006).
Questions about the disposition to trust were taken from
McKnight et al. (2002). When answering these questions, participants were reminded to think of the critical incident. See Table 1
for detailed information about the questions.
3.3. Participants
All participants were recruited via Amazon Mechanical Turk. In
total, 367 participants started the study and 254 completed it. Out
of these 254, we did not accept the answers of 11 participants
because they described a distrustful incident instead of a trustful
incident; another 10 participants were excluded because they did
not refer to a specific incident. A further 12 participants had to
be excluded because they described the experiences too vaguely,
reducing the acceptable answers to 221 (49% female, 51% male).
All participants were from the U.S.A. The mean age was 29.4 years
(SD = 9.1; range: 18–62). All participants use the Internet daily and
in average for 13.6 years (SD = 3.9; range: 2–21).
3.4. Procedure
Participants were directed from Amazon Mechanical Turk to an
external questionnaire. All questions, except one on the participants’ age, were mandatory. For ratings of honesty, benevolence,
and competence, participants had the possibility of answering ‘‘I
don’t know’’. The order of the questions is shown in Table 1. On
average, completing the questionnaire took 18.5 min (for the full
questionnaire).
3.5. Data preparation, content and context analysis
The primary goal of the data preparation was to extract the key
website characteristic from the critical incidents that led to a
trustful or a distrustful experience. To categorize the critical incidents’ website characteristics, an affinity-diagramming workshop
was organized (also known as KJ method, see Scupin, 1997). An
affinity diagram is an organizing tool used to locate similar facts,
arguments, or other information together. The rationale behind
this technique is to reduce problems of variety and complexity
by categorizing information according to higher-level abstract concepts (Beyer & Holtzblatt, 1999). This technique enabled us to
focus on the website characteristics mentioned in each incident.
The process during the workshop consisted of the following six
steps: First (1), each story was printed on a notecard and four
researchers individually grouped the incidents based on the similarity in regard to the mentioned website characteristics. Researchers were told to focus on the website characteristic that was most
crucial for causing trust (or distrust). The grouping was done separately for trustful and distrustful experiences, as previous studies
have shown that trust and distrust could have different antecedents. To avoid overlooking these characteristics, we did not merge
trust and distrust reports in this first step. Then (2), we combined
the individual groupings of all researchers, again separately for
trust and distrust. If not all researchers agreed on the allocation
of an incident to a group, we put this incident aside and discussed
it at a later step. In total, 17.4% of all trust incidents and 17.8% of all
distrust incidents were sorted out for later discussion.
Most incidents described concrete website characteristics that
either awoke the users’ distrust or caused trust. However, there
were some incidents on trustful experiences that mentioned prior
experiences with this site as the only reason. Other incidents
described how recommendations from friends influenced their
trust in a website; however, this happened mostly not on the
actual website but on a social media network. These incidents
stand out from the other incidents, but we could not find any reference to another website characteristic. Therefore, we defined one
separate group for prior experiences. Incidents concerning recommendation from friends were merged with incidents of recommendations from other users (website characteristics such as user
ratings and reviews). Then (3) all incidents where the researchers
had not agreed on a grouping were discussed until consensus
was reached and the incident could be assigned to one of the existing groups. At the end of this step, 13 different groups of distrust
and 11 different groups of trust incidents emerged.
Next (4), the 24 different distrust and trust groups were compared. Researchers looked at the different distrust groups and
wrote down keywords that characterize each group (see definitions in Table 2). Then, the same was done for the trust groups.
Based on the similarity of the keywords, the researchers merged
groups from the distrust and the trust condition (e.g., group with
incidents about a good usability and a group with bad usability
incidents). With this procedure, we managed to merge five trust
groups with five distrust groups, representing the following website characteristics: visual design, usability, security signs, privacy
and social proof.
Then (5) we looked at the remaining 14 groups (eight for distrust, six for trust). Three distrust and three trust groups did not
have as strong a connection as the groups defined in step 4 but
were described with related keywords. These groups were merged
and got rather broad titles: image/brand, expertise, customer service.
Five groups of distrust incidents were not mixable with any trust
groups. For these distrust groups, the following titles were defined:
pop-ups/ads, demands, web address, content, implausible promises.
Finally, three trust groups were not mixable with any distrust
group: policy, real-world link, prior experience.
We chose this bottom-up approach described so far because
there is little research on distrust characteristics. By applying this
approach, we wanted to ensure that we considered the characteristics that are important for the users; we did not want to just
42 M. Seckler et al. / Computers in Human Behavior 45 (2015) 39–50
adopt a predefined list of website characteristics from prior
research. This approach allowed us to identify 16 website characteristics. However, there were two groups – privacy and social proof
– that all researchers judged as too heterogeneous. Therefore, in a
next step (6), we decided to include a top-down step and evaluated
whether previous research could provide a more detailed subdivision of these two groups. Culnan and Armstrong (1999) present a
useful privacy classification; they differentiate between privacy
secondary use and privacy collection (see Table 2 for a precise definition). We applied this differentiation for our privacy group. Fuller,
Serva, and Benamati (2007) provided a framework about social
proof characteristics that differentiate between friends’ social proof
and users’ social proof (see Table 2), which we used to refine our
social proof group. For our remaining groups, there was no indication from literature to split them further. The whole process from
step one to step six resulted in 18 different website characteristics
(see Table 2 for an overview).
3.5.1. Classification of website characteristics to superior trust and
distrust dimensions
As the final step, the 18 website characteristics were grouped by
the classification from Wang and Emurian (2005). They argue that
website characteristics for trust can be described with four dimensions: graphic design, structure design, content design and social-cue
design. Three researchers independently assigned each of the 18
website characteristics to one dimensions in the framework
from Wang and Emurian (2005). An interrater agreement of
jFleiss = 0.588 (z = 7.85, p < .01) was achieved, indicating an
intermediate to good agreement between all three researchers.
The grouping worked well for all but three of the 18 characteristics.
These three characteristics did not target the design of the website
but (1) prior experience with a website, (2) social proof from other
users, and (3) social proof from friends. Therefore a new dimension
was defined, which was called ‘‘personal and social proof’’. The
five final dimensions and the corresponding characteristics are
presented in the results section.
3.5.2. Content of the experiences
Most descriptions of the participants’ web experiences not only
contained evaluative statements about the site but also included
narrative elements such as information about the context, the
users’ motivation to use the website, and their main action. In general, the structure of the incidents was similar to previous research
on self-reported user experiences (Tuch, Trusell, & Hornbæk,
2013). This is an example for a distrustful experience:
‘‘I had to ride in an ambulance to a hospital. They sent me a bill and
gave me a website where I could pay online if I wanted to. The website looked weird because it was a .info and I had never been to a
Table 1
Questions used in the survey.
Order of Questions
Critical Incident (Trustful or Distrustful Web Experiences)
Open-ended (text field) critical incident question: ’’Please think of an occasion where you felt exceptionally trustful (distrustful) using a website’’
Context of the Critical Incident
3 questions about the context of the experience
(1) How long ago did the event take place?
(2) Into what category does the website of your event fall?
(3) How often did you visit the website before the event occurred?
Honesty (Casaló et al., 2007)
5 questions answered as ‘‘strongly disagree’’(1) to ‘‘strongly agree’’ (7);
(1) I think that this website usually fulfills the commitments it assumes
(2) I think that the information offered by this site is sincere and honest
(3) I think I can have confidence in the promises that this website makes
(4) This website does not make false statements
(5) This website is characterized by the frankness and clarity of the services that it offers to the consumer
Benevolence (Casaló et al., 2007)
6 questions answered as ‘‘strongly disagree’’(1) to ‘‘strongly agree’’ (7);
(1) I think that the advice and recommendations given on this website are made in search of mutual benefit
(2) I think that this website is concerned with the present and future interests of its users
(3) I think that this website takes into account the repercussions that their actions could have on the consumer
(4) I think that this website would not do anything intentional that would prejudice the user
(5) I think that the design and commercial offer of this website take into account the desires and needs of its users
(6) I think that this website is receptive to the needs of its users
Competence (Casaló et al., 2007)
4 questions answered as ‘‘strongly disagree’’(1) to ‘‘strongly agree’’ (7);
(1) I think that this website has the necessary abilities to carry out its work
(2) I think that this website has sufficient experience in the marketing of the products and services that it offers
(3) I think that this website has the necessary resources to successfully carry out its activities
(4) I think that this website knows its users well enough to offer them products and services adapted to their needs
Disposition Trusting Stance (McKnight et al., 2002)
Three questions answered as ‘‘strongly disagree’’ (1) to ‘‘strongly agree’’ (7);
(1) I usually trust people until they give me a reason not to trust them
(2) I generally give people the benefit of the doubt when I first meet them
(3) My typical approach is to trust new acquaintances until they prove I should not trust them
Online Experience and Personal Background
Two questions about web usage and 3 demographic questions;
(1) How long have you been using the Internet?
(2) How many hours a week do you spend online (work and leisure time)?
(3) How old are you?
(4) Please indicate you gender
(5) What country do you live in?
M. Seckler et al. / Computers in Human Behavior 45 (2015) 39–50 43
website with that extension before. I decided to proceed anyway
but I was very wary. After double checking it was supposed to be
the right website. It had three links and that was to pay a bill or
ask a question or submit information I think. I decided to take
the risk and submitted an online payment and my insurance information. The website ended up being legit but the whole thing was
very strange to me.’’
And this is an example for a trustful experience:
‘‘XY.com is a web host that has gotten a lot of good reviews from
the tech community. I searched for reviews in forums, which I find
to be much more reliable than other channels. The many good
reviews with almost no bad reviews made me feel very trustful of
the website. Especially because I assume the people on forums have
no incentive to oversell the site, and generally the tech community
is very savvy when it comes to judging the quality of a web
service.’’
On average participants used 85 words to describe their experience. For the trustful web experiences, fewer words (77 words)
were used than for the description of the distrustful experiences
(92 words).
3.5.3. Context of the experiences
Twenty percent of all incidental experiences had happened
within the previous week, 16% between one week and one month,
14% between one and three months, 28% between three months
and one year, 21% happened between one and five years, and 3%
happened more than five years ago.
We also looked at the different website types (information site,
e-commerce, entertainment, finance/e-banking, social media, others) that were described in the experiences. We could not find
any significant difference between the different website types,
(v2 = 7.57, p = .181). See Table 3 for descriptive data.
4. Results
We begin this section by (1) presenting a short overview of the
frequencies of experiences in each dimension separately for trust
and distrust. In the second part of this section (2), we describe each
of the five dimensions and their corresponding website characteristics that emerged from our affinity diagramming. Doing so, we
provide concrete starting points for how to enhance trust and prevent distrust. Further (3), we analyze how often participants visited the website before the critical incident occurred. In the last
part (4), we compare the questionnaire ratings for trust and distrust facets to further investigate the differences between trust
and distrust.
4.1. Dimensions of website characteristics
The 18 website characteristics could be subordinated into the
four design dimensions described by Wang and Emurian (2005)
and into the additional dimension ‘‘personal and social proof’’
Table 2
Final 18 website characteristics mentioned in trust and/or distrust experiences with anonymized examples.
Website
characteristics
Definition Example of an experience N
Visual design Use of colors, site layout, layout complexity, photographs ‘‘When I visited their site it looked very cheaply put together and the overall
appearance of the site and products made me not feel safe shopping there.’’
31
Security signs Security aspects such as passwords, authentication
questions or the version of communication protocol
‘‘When I log in to my bank account, and it uses a secure connection as well as
requires me to have a personalized key, so I feel secure.’’
26
Privacy:
secondary
use
Users’ fears that their information might be used for
another reason than what it was collected for
‘‘When I learned that the social media website sells information to people.’’ 18
Usability Effectiveness and efficacy with the task flow, site
navigation, links
‘‘None of the coupon codes were working and a lot of items were out of stock or
mismatched. The links were broken and the images were not showing up.’’
17
Pop-ups/ads Pop-ups and visual or audio ads ‘‘As soon as I get to the site, I get pop ups, sounds and other annoying sensory
garbage that distract me. I felt as if the site wanted to manipulate me.’’
15
Privacy:
collection
Users’ concern that data could be collected by the
website operator
‘‘Although the sight is probably legitimate, it asked for me to enter information such
as my XY sign on and password which I would not like to share with any company.’’
15
Implausible
promises
Promises the participants felt could not be kept by
website operators
‘‘I started looking and comparing their computers to name brand computers and just
couldn’t believe how good the deals were. but in the end I didn’t buy from them
because I just couldn’t trust a site with deals like that.’’
15
Users’ social
proof
User ratings and reviews ‘‘The many good reviews with almost no bad reviews made me feel very trustful of
the website.’’
13
Customer
service
Availability of customer service agents, provision of
service to customers after a purchase
‘‘Whenever I buy from XY, I know I can trust them because their customer service is
forgiving and very well put together. If a package is broken, they’ll take my word for
it and send a new one.’’
13
Image/Brand Image or brand of the website operator ‘‘Of course, XY has a massive reputation. If they weren’t as well known, I would
probably have been more hesitant to give out the information they needed.’’
12
Expertise Competence and professional knowledge ‘‘The seller was very thorough in their explanation of the product and how it works,
as well as why it is more reasonable to purchase this way, as opposed to paying for
the same amount of less medication.’’
8
Prior
experience
Previous experience of the user on the same site ‘‘I feel company XY is a completely trustful website, I have never had any issues with
it.’’
8
Content Credibility of information, correct and up-to-date
information
‘‘The news I got from XY about Z turned out to be based on false speculations.’’ 7
Friends’ social
proof
Recommendations given by colleagues, friends and
family members
‘‘My friends and family did not report an unpleasant experience with company XY, so
I began to trust the site with my personal information.’’
7
Demands Demands to share a link, download a piece of software or
create an account to get access to a website or a service
‘‘Then the site seemed suspicious, as I had to answer multiple questions before I
could reach any kind of main page for the website.’’
6
Web address Domain name or website name ‘‘The website looked weird because it was a .info and I had never been to a website
with that extension before.’’
4
Policy Policy, general terms and conditions ‘‘They have a protection policy that convinced me.’’ 4
Real-world
link
A link to the life of the website owner or a link to the real
world such as a shop
‘‘So when I went to their website and found similar information set up in the same
basic format as the magazine I knew it was a legitimate site and trusted it implicitly.’’
2
44 M. Seckler et al. / Computers in Human Behavior 45 (2015) 39–50
(see Table 4). The most frequent descriptions of distrustful experiences focused on content design (51.7%) followed by structure
design (23.7%), and graphic design (20.3%). The more frequently a
dimension is mentioned, the more frequently these distrust incidents happen on the web.
In contrast, the most frequent trustful web experiences were
about content design (46.6%), personal and social proof (26.2%),
social-cue design (10.7%), and structure design (9.7%). To further
analyze whether there was a significant difference in the frequencies of experiences, we conducted a chi-square test. Results
showed that distrustful and trustful experiences significantly differ
in regard to the five dimensions of website characteristics,
(v2 = 46.00; p < .001), indicating that different characteristics are
important for trust and for distrust.
4.1.1. Graphic design
The dimension graphic design consists of a single website characteristic that we called visual design. Experiences concerning this
characteristic described the use of colors, the layout of the site such
as the complexity or the balance, and the use of photographs and
pictures. A typical incident of a distrustful experience is ‘‘The whole
site was poorly designed, with a static background and clashing colors.’’ An example for a trustful experience is the following: ‘‘I felt
comfortable because of how professional, clean and well-designed
the manufacturer’s site was. The pictures were crisp and clear.’’
Descriptive data show that this dimension is frequently mentioned in distrust experiences, but not in trust experiences. To analyze whether this dimension is significantly more important for
causing distrust than for enhancing trust, we conducted a configural frequency analysis with Eye (Grüner, 2008). This test is able to
detect patterns in the data that occur significantly more or less
often than expected by chance. Results showed that there is a significant difference between expected and effective frequency for
‘‘graphic design’’ for distrust (z = 1.90, p = .028) as well as trust
(z = 2.03, p = .021), indicating that graphic design is especially relevant for distrust but less relevant for trust. This means that users
often do not explicitly appreciate when a website has a clear
design; however, as soon as there are some deficits (e.g., grammar
issues, pixelated photographs, high visual complexity), users will
focus on those deficits and experience a website as distrustful.
4.1.2. Structure design
This dimension appeared in 19% of all experiences, making it
the second most frequent dimension. It consists of four website
characteristics: usability, pop-ups and ads, and demands. Whereas
pop-ups and ads are most frequently mentioned in distrust experiences, a good usability is often reported in trust experiences. Two
typical incidents of this dimension are the following. ‘‘I was very
impressed with how user-friendly the site’s interface was and I felt
secure’’ (trustful experience) and ‘‘As soon as I get to the site, I get
pop ups, sounds and other annoying sensory garbage that distract
me’’ (distrustful experience).
Descriptive data show that in total, there are more distrust
experiences that could be assigned to this dimension. Configural
frequency analyses with Eye (Grüner, 2008) again showed that
trust and distrust differ significantly (distrust: z = 1.80, p = .036;
trust: z = 1.91, p = .028). A good structure design therefore is able
to lower distrust but not to enhance trust in a website.
4.1.3. Content design
Appearing in almost half of all experiences, content design was
the most prominent dimension. We found eight website characteristics that are part of this dimension, making this dimension more
heterogeneous than the other dimensions. Incidents mentioning
security signs of the website were the most frequent in trust experiences (e.g., ‘‘I felt more secure because it requires 2 passwords, and a
secret word in a certain order in order to gain access to the account.’’).
Furthermore, incidents focusing on the website operators’ image or
brand were also more frequently mentioned in trust experiences
(e.g., ‘‘I felt like it was secure because it is a well known, big
company.’’).
The most mentioned website characteristic in distrustful experiences was privacy. The experiences focusing on these issues
could be further divided into incidents that focused on users’ concerns that their data could be collected by the website operator
(privacy: collection) such as ‘‘I recently felt very distrustful of one
of these sites that retains all your personal info like name, address,
phone, E-mail, social networks…’’ and users’ fears that their information was being used for another reason than what it was collected
for (privacy: secondary use) for instance ‘‘I felt like they just wanted
to verify my information to steal my identity’’. In total, 13% of the distrust but none of the trust experiences concerned promises on the
Table 3
Number and percentage of distrust and trust experiences for different website types.
Website type Examples Distrust N (%) Trust N (%)
Information site Sport, travel, health, news 20 (16.9) 13 (12.6)
E-Commerce Clothing, electronics, jewelry 44 (37.3) 49 (47.6)
Entertainment Movie/video sites, online gaming, music streaming 14 (11.9) 5 (4.9)
Finance/e-banking Banking websites, online money transfer services 11 (9.3) 15 (14.6)
Social media Facebook, Twitter 19 (16.1) 16 (15.5)
Others Surveys, work tasks 10 (8.5) 5 (4.9)
Total 118 (100) 103 (100)
Table 4
Number and percentage of website characteristics and their corresponding dimension
mentioned in distrust and trust experiences.
Dimensions/website characteristics Distrust N (%) Trust N (%)
Graphic design 24 (20.3) 7 (6.8)
Visual design 24 (20.3) 7 (6.8)
Structure design 28 (23.7) 10 (9.7)
Usability 7 (5.9) 10 (9.7)
Pop-ups/ads 15 (12.7)
Demands 6 (5.1)
Content design 61 (51.7) 48 (46.6)
Security signs 4 (3.4) 22 (21.4)
Image/brand 1 (0.8) 11 (10.7)
Expertise 2 (1.7) 6 (5.8)
Privacy: collection 13 (11.0) 2 (1.9)
Privacy: secondary Use 15 (12.7) 3 (2.9)
Content 7 (5.9)
Web address 4 (3.4)
Implausible promises 15 (12.7)
Policy 4 (3.9)
Social-cue design 4 (3.4) 11 (10.7)
Customer service 4 (3.4) 9 (8.7)
Real-world link 2 (1.9)
Personal and social proof 1 (0.8) 27 (26.2)
Users’ social proof 1 (0.8) 12 (11.7)
Friends’ social proof 7 (6.8)
Prior experience 8 (7.8)
M. Seckler et al. / Computers in Human Behavior 45 (2015) 39–50 45
website that the participants felt could not be kept by operators
(implausible promises). The website characteristics content, policy,
expertise and web address were mentioned in less than 5% of all
experiences.
No significant difference was found for this dimension between
expected and effective frequency for trust and distrust. As this
dimension is rather heterogeneous, this result is not surprising.
Within the dimension, however, different website characteristics
were mentioned in trustful and distrustful incidents.
4.1.4. Social-cue design
Social-cue design is the least mentioned dimension; its website
characteristics were found in only 7% of all experiences. It consists
of customer service (e.g., ‘‘they [customer service agents] informed
me that they would be returning my money’’) and real-world links
(e.g., ‘‘I discovered that the people who worked on X and who ran
the website, had their own profiles’’). No significant difference was
found for this dimension between expected and effective frequency for trust and distrust. This dimension therefore is not as
important as the other dimensions for enhancing trust or causing
distrust.
4.1.5. Personal and social proof
Personal and social proof mainly appeared in trust experiences.
Out of a total of 28 experiences concerning personal and social
proof, just one was an incident causing distrust. Personal and social
proof is made up of (1) users’ social proof and (2) friends’ social
proof, and (3) the participants’ prior experience with the reported
website. The difference between users’ and friends’ social proof is
that users’ social proof focuses on user rating and reviews (e.g., ‘‘I
looked at the star ratings of the seller and also looked at past comments from previous buyers to assess whether or not this person
was trustworthy’’) and friends’ social proof on information given
by colleagues, friends, and family members (e.g., ‘‘Many friends
were active on the site’’). Configural frequency analyses with Eye
(Grüner, 2008) showed significant differences between expected
and effective frequency for personal and social proof for distrust
(z = 3.74, p < .001) as well as trust (z = 3.98, p < .001), indicating
that this dimension is especially important in enhancing users’
trust (but not their distrust) in a website.
4.2. Site visits before the critical incident occurred
We further analyzed how often the participants visited the
website before the critical incident occurred (see Table 5). There
was a significant difference between the distrustful and trustful
experiences and their visits, (v2 = 13.34, p = .004). Descriptive data
show that in almost half of all cases for distrustful experiences, the
critical incident happened at the participants’ first visit to a site. In
these cases, once visitors to a website distrusted the site, they
would likely not use that site again. Nonetheless, distrustful experiences could also happen when a site had been visited more than
100 times previously. This was especially the case for incidents
happening on social networks (e.g., chance of privacy policy).
Trustful experiences, on the contrary, occurred more often after
several previous visits. Particularly for the category of e-commerce
sites, most of the trustful experiences happened on websites that
had been visited (many times) before. For example, one participant
reported having had good, but not extraordinary experiences with
an online shopping website several times. Then the trustful critical
incident was a very positive experience with the customer service
of this online store.
4.3. Ratings for trust and distrust experiences
We compare the ratings of honesty, benevolence, and competence to further investigate potential differences between these
three facets for distrustful and trustful experiences. Reliability
analyses for all subscales show good internal consistency with
Cronbach’s a between 0.84 and 0.94 (see Table 6). The authors of
the scale (Casaló et al., 2007) showed good ratings for the construct
validity. They assessed the convergent as well as the divergent
validity. Convergent validity analyses showed that the factor loadings of the confirmatory models were all statistically significant on
the 0.01 level and loaded substantively on each of the constructs.
Analyses of Casaló et al. (2007) regarding the discriminatory validity showed values less than 0.8.
Descriptive data (see Table 6) shows that distrustful incidents
can be characterized by low ratings of honesty and benevolence.
Perceived competence of a website was not rated as low as one
might expect. It seems that competence is not as strongly associated with a distrustful experience as the other facets.
In contrast, trustful experiences show another pattern of the
three facets, which is characterized by high honesty and competence ratings. Benevolence, however, is rated slightly lower than
the other two facets. A trustful experience is therefore rather associated with perceived honesty and competence of a website.
Benevolence seems to be slightly less important than the other
two facets.
Statistical analyses support this interpretation. We examined
these differences with two within-subject ANOVAs (one for distrust, one for trust). Results showed that there is a significant difference within the facets for distrust (p < .001, N2 p ¼ :29) as well
as trust (p = .005, N2 p ¼ :07). Post-hoc tests by pairwise comparison
and Bonferroni correction revealed that in the distrust condition,
competence significantly differs from the other two facets (both
p < .001) whereas in the trust condition, benevolence significantly
differs from the other two facets (p = .021 resp. p = .011).
4.4. Ratings for the five different website dimensions
To analyze whether there is a difference in honesty, benevolence, and competence between the different website dimensions
within trust and distrust, we conducted for each facet and for trust
and distrust a one-way ANOVA for independent samples (the five
website dimensions as independent and three facets as dependent
variables). However, results showed that there are no significant
differences for any of the facets and the five dimensions within distrust as well as within trust. Thus we looked at differences at the
level of the website characteristics.
4.5. Ratings for the most mentioned website characteristics
Of the three most frequently mentioned website characteristics
in distrust experiences, privacy secondary use got the highest and
implausible promises the lowest ratings on all three facets (see
Table 7). We conducted for each facet a one-way ANOVA for independent samples to analyze whether the three website characteristics (independent variable) significantly differ in terms of
honesty, benevolence, and competence (dependent variable). The
three website characteristics differ significantly in regard to
Table 5
Number and percentage for the number of visits before the critical incident occurred.
Visits before Distrust N (%) Trust N (%)
First time visit 54 (45.8) 26 (25.2)
2–9 visits before 33 (28.0) 28 (27.2)
10–99 visits before 17 (14.4) 28 (27.2)
More than 100 visits before 14 (11.9) 21 (20.4)
Total 118 (100) 103 (100)
46 M. Seckler et al. / Computers in Human Behavior 45 (2015) 39–50
honesty (p = .018, N2 p ¼ :15), benevolence (p = .032, N2 p ¼ :13), and
competence (p = .002, N2 p ¼ :21), further indicating that implausible
promises are evaluated as significantly more distrustful than the
other two facets. Although privacy secondary use and visual design
are mentioned more often and therefore are likely more common,
if users encounter critical incidents due to implausible promises
these are rated as more distrustful.
Table 8 shows that trustful experiences mentioning users’ social
proof were rated highest on all three facets. In contrast to the distrust condition, the three most frequently mentioned characteristics in trustful experiences significantly differ only on
benevolence (p = .047, N2 p ¼ :14) but no significant difference was
found for the other two facets. These results show that (1) there
are smaller differences in the evaluation of the most mentioned
website characteristics for trust than for distrust. Furthermore
(2), for image the facet benevolence is rated the lowest.
5. Discussion
The main goals of this research were (1) to identify website
characteristics that influence trust and/or distrust and (2) to show
whether and how trustful and distrustful web experiences differ in
terms of perceived honesty, competence, and benevolence. In the
following (3) we discuss our findings and the implication for
trust conceptualization as well as the implications for designing
websites.
5.1. Website characteristics and dimensions
First, our findings highlight that web trust and web distrust do
not have the same website characteristics as antecedents. Website
characteristics associated with graphical design and structure
design were significantly more often reported in distrust than trust
experiences. On the contrary, personal and social proof was associated with trust rather than distrust incidents. For content design
and social-cue design, no significant differences between trust
and distrust were found. However, content design is a very
heterogeneous dimension; descriptive data show some differences
between trust and distrust at the level of its website characteristics.
Security signs and image/brand were often mentioned within
trustful experiences, whereas implausible promises and privacy
concerns were the most frequent topics of distrustful experiences.
5.1.1. Graphic design dimension
We were able to show that the website characteristic visual
design is especially relevant for distrust but of little relevance for
trust. Andrade et al. (2012), however, found an effect of unpleasant
design both on distrust and on trust. In contrast to our study, these
authors used a single-item scale, which asked to rate ‘‘unpleasant
design’’ of six preselected websites. As Gliem and Gliem (2003)
were able to show, single-item questions are less reliable than
multi-item scales and should not be used in drawing conclusions.
For our study, we used a qualitative approach without a predefined
scale for graphic design.
5.1.2. Structure design dimension
Regarding the ease of browsing the site (or usability in general),
the present study is in line with Andrade et al. (2012) who found
only an effect on trust (but not on distrust).
5.1.3. Content design dimension
Most previous studies focused on website characteristics within
the dimension content design. Contrary to our results, Chang and
Fang (2013) investigated privacy policy and security indications
and found no effects on trust or distrust. Andrade et al. (2012) concluded that privacy and security signs influence both trust and distrust, whereas our study reveals that privacy concerns are
associated with distrust and security indications with trust. A possible explanation for this difference between our results and the
results from other authors may be due to the different methodological approaches. By imagining a critical incident, users might
remember different aspects than predefined questionnaires could
measure. Regarding security, Chang and Fang (2013) only asked
for signs or symbols from third-party companies; however, our
participants also mentioned additional authentication questions
or the version of communication protocol (e.g., https or http).
Andrade et al. (2012), moreover, only used a single item for privacy
and a single item for security. With our approach, we received
Table 6
Descriptive statistics for distrust and trust ratings and their three facets.
Facets Distrust Trust
M (SD) (N = 76) Cronbach’s alpha M (SD) (N = 78) Cronbach’s alpha
Honesty 2.5 (1.4) 0.92 6.5 (0.8) 0.89
Benevolence 2.6 (1.5) 0.94 6.3 (0.9) 0.84
Competence 3.6 (1.9) 0.91 6.5 (0.9) 0.84
Table 7
Descriptive statistics for distrust ratings regarding the most frequently mentioned website characteristics.
Facets Privacy secondary use M (SD) (N = 22–24) Visual design M (SD) (N = 19–20) Implausible promises M (SD) (N = 11–13)
Honesty 3.0 (1.2) 2.9 (1.8) 1.6 (0.8)
Benevolence 3.0 (1.5) 2.8 (1.8) 1.5 (1.0)
Competence 4.8 (1.6) 3.5 (1.9) 2.7 (1.8)
Table 8
Descriptive statistics for trust ratings regarding the most frequently mentioned website characteristics.
Facets Security signs M (SD) (N = 18–21) Users’ social proof M (SD) (N = 17–19) Image M (SD) (N = 9–11)
Honesty 6.5 (0.9) 6.7 (0.4) 6.6 (0.7)
Benevolence 6.4 (0.8) 6.8 (0.4) 6.0 (1.0)
Competence 6.4 (0.9) 6.8 (0.3) 6.5 (0.8)
M. Seckler et al. / Computers in Human Behavior 45 (2015) 39–50 47
information not only about the characteristic ‘‘data collection’’ and
‘‘privacy policy’’ but also about users’ fears that their information
might be used for another reason than what it was collected for
(‘‘secondary usage’’).
5.1.4. Social-cue design dimension
In our study, customer service was only mentioned a few times
and therefore is not likely an important antecedent for trust or distrust. However, customer service (Ou & Sia, 2010), particularly
order fulfillment (Chang & Fang, 2013), had a significant effect on
web trust in prior studies. The difference between the present
study and prior studies is probably due to the website types that
were used. Ou and Sia (2010) as well as Chang and Fang (2013)
focused on online shops as study material and customer service
might be more important for this specific website type.
5.1.5. Personal and social proof dimension
Finally, Ou and Sia (2010) were not able to find an effect of elements of social proof (such as ratings) on trust or on distrust. Our
results, however, suggest that these elements have an important
influence on trust. These differences could be based on the fact that
Ou and Sia (2010) only used two different websites as stimuli and
social proof characteristics might not have been relevant or prominent characteristics on these sites. A contribution of our paper is
that by applying the CIT, we were able to find website characteristics in distrustful incidents (demands, implausible promises, users’
social proof, friends’ social proof, prior experience, real world link,
web address, pop-ups/ads and policy) that users actually experienced and that were overlooked by earlier work on website characteristics’ effect on distrust (Chang & Fang, 2013; Ou & Sia, 2010).
5.2. Facets web trust and web distrust
Besides the website characteristics, we analyzed differences in
the facets of web trust and web distrust. Our results suggest that
different facets characterize a trustful and a distrustful experience.
Distrustful experiences are based on the lack of honesty and
benevolence of a website but to a lesser extent on competence.
In contrast, for web trust high competence and high honesty of a
website are needed, although significantly less benevolence. This
implies that to prevent distrust, resources should be invested to
enhance the honesty and benevolence of the website, whereas to
enhance trust, one should rather focus on competence and honesty. These results contradict the findings of Cho (2006), who
showed that trust is primarily driven by benevolence whereas distrust is based on a lack of competence. In contrary to the present
study, Cho (2006) focused on existing online shops and this may
explain the different results. Another explanation could be the
eight-year difference between our research and Cho’s (2006) study.
There are many more tools available nowadays that makes it easier
for a company to design a website that has a competent appearance. More research is needed to further clarify the facets of trust
and distrust.
Depending on the reported website characteristic, the facets of
web trust and web distrust receive different ratings. For the three
most mentioned website characteristics for web trust experiences
(security signs, users’ social proof, and image), we found significant
differences in benevolence between all three. The highest ratings
for all three facets were reached by the website characteristic
users’ social proof. This implies that to receive high web trust,
users’ social proof is important. For the three most mentioned
characteristics in the distrust condition, implausible promises
received low ratings for all facets, which suggests that implausible
promises have a large effect on distrust.
5.3. Trust and distrust concepts
Previous literature argues that trust and distrust are two distinct constructs that coexist and that different website characteristics may need to be managed in order to elevate trust and to reduce
distrust (Ou & Sia, 2010). Our results support Lewicki et al.’s (1998)
statement that it would be misleading to assume that the positive
predictors of trust would necessarily be negative predictors of distrust or vice versa. Our research findings provide support that web
trust and distrust are affected by different antecedents (Chang &
Fang, 2013; Ou & Sia, 2010). Therefore, efforts to build trust may
not always eliminate distrust (Chang & Fang, 2013). However, like
Schoorman et al. (2007), we still raise concerns about the deliberation of trust and distrust as distinct constructs. It may be possible
to experience trust as well as distrust at the same time due to different attribution factors. Users may trust a website because of
good reviews and a good brand image, but at the same time experience distrust due to a bad visual design and privacy concerns. We
cannot support the statements from Ou and Sia (2010), who argue
that if trust and distrust are found to be the same construct, then
users would note the same website characteristics in a positive
or negative way. It still might be possible that trust and distrust
are the same construct but have different antecedents. Our study
provides a more detailed insight into different website characteristics; however, more studies are needed to investigate trust and distrust to conclude whether they are the same or two distinct
constructs.
5.4. Implications
Our findings imply that to avoid distrust, a website should focus
on improving the graphic and structure design, as well as the content design in terms of enhancing privacy and avoiding implausible
promises. On the other hand, to achieve more trust a website
should provide good usability and use security sign cues such as
lock symbols. Furthermore, social-cue design and personal and
social proof enhance trust in a website.
It should be noted that distrust can be prevented more easily by
website operators because changes in visual design, avoiding secondary use of users’ data, and making sincere promises do not
involve third parties. All these issues are under a company’s own
control. However, changing how users perceive a company’s image
and getting good ratings from users to enhance social proof is more
difficult to achieve. Enhancing users’ trust is therefore more difficult. In Table 9 we used the most frequently mentioned website
characteristics for trust and distrust to provide guidance to
enhance trust and avoid distrust. Furthermore, we supplement
these characteristics with references from previous studies.
5.5. Limitations and further research
Although there are positive aspects of the CIT, it leads to some
limitations in our research. First, participants have to be capable
of verbalizing the experienced incident. As the participants have
to recall a past event, we have to rely on participants’ memory.
Experiences that took place far back in the past may not be remembered with the same accuracy as newer incidents. Memory biases
may have influenced the participants’ answers. Furthermore, it is
important to highlight that we focused on incidents that are critical
and not everyday experiences.
There are also some limitations concerning the affinity diagram
process and the coding procedure. Because this is a group process,
it is important that there is a shared understanding of all the characteristics and dimensions. We tried to eliminate any uncertainties
during consolidation with all researchers (step two); however,
there is no guarantee that there were no differences between the
48 M. Seckler et al. / Computers in Human Behavior 45 (2015) 39–50
four researchers. A further limitation is the possibility that not
each researcher had the same influence and that there might have
been an individual dominating the group decisions. Finally, in step
four and five we reduced our set of data by creating self-written
keywords describing the incident groups. The groups with related
keywords were then merged. Comparing keywords instead of the
incidents themselves and the consequential merging may lose
some of the actual meaning; however, at some point in the evaluation process, a reduction of the qualitative data must take place.
The focus of this study, moreover, was on websites in general.
Further studies should analyze the different types of websites separately to learn more about the differences between the various
website types and to provide more inferences to the practice of
specific website types.
To statistically support our findings on website characteristics,
in future research larger sample sizes should be applied. Further
research is needed to explore whether the findings from this study
can be replicated by other studies using other methods or participants from different countries. Additionally, from an economic
standpoint it would be interesting to know how our findings
may not only influence the trust or distrust of a website but also
result in higher conversion rates.
6. Conclusion
This paper contributes to the growing body of literature on web
trust in two ways. First, we show that distrust is mostly an effect of
graphical (e.g., complex layout) and structural (e.g., pop-ups)
design issues of a website, whereas trust is based on social factors
such as reviews or recommendations by friends. The content of
websites affects both trust and distrust: privacy issues had an
effect on distrust and security signs enhanced trust. Second, our
results showed that trustful experiences can be characterized by
high honesty and competence, whereas a distrustful experience
is based on missing honesty and missing benevolence.
References
Almadhoun, N. M., Dominic, P. D. D., & Woon, L. F. (2011). Perceived security,
privacy, and trust concerns within Social Networking Sites: The role of
information sharing and relationships development in the Malaysian Higher
Education Institutions’ marketing. In 2011 international conference on control
system, computing and engineering (pp. 426–431). Penang: IEEE.
Andrade, A. A., Lopes, V. V., & Novais, A. Q. (2012). Quantifying the impact on
distrust of e-commerce trust factors: A non-parametric study. In 2012
international conference on internet technology and secured transactions
(pp. 356–361). London: IEEE.
Bart, Y., Shankar, V., Sultan, F., & Urban, G. L. (2005). Are the drivers and role of
online trust the same for all web sites and consumers? A large-scale exploratory
empirical study. Journal of Marketing, 69(4), 133–152.
Beldad, A., De Jong, M., & Steehouder, M. (2010). How shall I trust the faceless and
the intangible? A literature review on the antecedents of online trust. Computers
in Human Behavior, 26(5), 857–869.
Benamati, J., Serva, Mark A., & Fuller, Mark A. (2006). Are trust and distrust distinct
constructs? An empirical study of the effects of trust and distrust among online
banking users. In Proceedings of the 39th annual Hawaii international conference
on system sciences, HICSS’06 (Vol. 6, pp. 121b). IEEE.
Benamati, J. S., Serva, M. A., & Fuller, M. A. (2010). The productive tension of trust
and distrust: the coexistence and relative role of trust and distrust in online
banking. Journal of Organizational Computing and Electronic Commerce, 20(4),
328–346.
Beyer, H., & Holtzblatt, K. (1999). Contextual design. Interactions, 6(1), 32–42.
Casaló, L. V., & Cisneros, J. (2008). An empirical test of the multiplicative effect of
usability on consumer trust and satisfaction. In DEXA’08 19th international
workshop on database and expert systems application (pp. 439–443). IEEE: Turin.
Casaló, L. V., Flavián, C., & Guinalíu, M. (2007). The role of security, privacy, usability
and reputation in the development of online banking. Online Information Review,
31(5), 583–603.
Chang, Y. S., & Fang, S. R. (2013). Antecedents and distinctions between online trust
and distrust: Predicting high-and low-risk internet behaviors. Journal of
Electronic Commerce Research, 14(2), 149–166.
Chen, S. C., & Dhillon, G. S. (2003). Interpreting dimensions of consumer trust in ecommerce. Information Technology and Management, 4(2–3), 303–318.
Cho, J. (2006). The mechanism of trust and distrust formation and their relational
outcomes. Journal of Retailing, 82(1), 25–35.
Culnan, M. J., & Armstrong, P. K. (1999). Information privacy concerns, procedural
fairness, and impersonal trust: An empirical investigation. Organization Science,
10(1), 104–115.
Dimoka, A., Pavlou, P. A., & Davis, F. D. (2007). NEURO-IS: The potential of cognitive
neuroscience for information systems research. Information Systems Research,
22(4), 687–702.
Doney, P. M., & Cannon, J. P. (1997). An examination of the nature of trust in buyerseller relationships. Journal of Marketing, 61(2), 35–51.
Flanagan, J. C. (1954). The critical incident technique. Psychological Bulletin, 51(4),
327–358.
Flavián, C., Guinalíu, M., & Gurrea, R. (2006). The role played by perceived usability,
satisfaction and consumer trust on website loyalty. Information and
Management, 43(1), 1–14.
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 2500
participants. In Proceedings of the 2003 conference on designing for user
experiences (pp. 1–15). NY: ACM.
Fuller, M. A., Serva, M. A., & Benamati, J. (2007). Seeing is believing: The transitory
influence of reputation information on E-commerce trust and decision making.
Decision Sciences, 38(4), 675–699.
Gliem, J. A., & Gliem, R. R. (2003). Calculating, interpreting, and reporting Cronbach’s
alpha reliability coefficient for Likert-type scales. In Midwest research-topractice conference in adult, continuing, and community education.
Grüner, H. (2008). Configural frequency analysis, <http://gruener.userpage.fuberlin.de/spss-dialogs.htm>.
Jarvenpaa, S. L., Tractinsky, N., & Saarinen, L. (1999). Consumer trust in an Internet
store: A cross-cultural validation. Journal of Computer-Mediated Communication,
5(2), 1–5.
Lewicki, R. J., McAllister, D. J., & Bies, R. J. (1998). Trust and distrust:
New relationships and realities. Academy of Management Review, 23(3),
438–458.
McKnight, D. H., & Choudhury, V. (2006). Distrust and trust in B2C e-commerce: Do
they differ? In Proceedings of the 8th international conference on electronic
commerce: The new e-commerce: innovations for conquering current barriers,
obstacles and limitations to conducting successful business on the internet
(pp. 482–491). NY: ACM.
McKnight, D. H., Choudhury, V., & Kacmar, C. (2002). Developing and validating
trust measures for e-commerce: An integrative typology. Information Systems
Research, 13(3), 334–359.
Table 9
Implications for trust and distrust.
How to enhance trust How to avoid distrust
– Make your site easily accessible and easy to use (Wang & Emurian, 2005) – Take care of your visual design (Andrade et al., 2012)
– Provide security sign cues such as https encryption, lock symbols or third-party
certificate (Bart et al., 2005)
– Do not abuse user data and state clearly what you are going to do with
data provided (Bart et al., 2005)
– Show your brand prominently and care for your image (Bart et al., 2005) – Do not force your users to do something they do not want to do (Bart
et al., 2005)
– Highlight your expertise (Wang & Emurian, 2005) – Avoid pop-ups and ads (Fogg et al., 2003)
– Provide links to your policy and make it easy to understand (Shneiderman, 2000) – Use a well-known web address and domain name (Wang & Emurian,
2005)
– Provide a helpful and friendly customer service (Chang & Fang, 2013) – Provide comprehensive, correct, and up-to-date information (Bart et al.,
2005)
– Care for good reviews and ratings (Ou & Sia, 2010) – Do not make implausible promises (Sher & Lee, 2009)
– Care for your users to enhance the possibility for further visits (prior experience with a
site enhances trust) (Fogg et al., 2003)
– Do not ask for unnecessary input and explain what the data are used
for (Bart et al., 2005)
M. Seckler et al. / Computers in Human Behavior 45 (2015) 39–50 49
Münscher, R., & Kühlmann, T. M. (2012). Using critical incident technique in trust
research. In G. Lyon, G. Möllring, M. Saunders, & J. Hatzakis (Eds.), Handbook of
research methods on trust (pp. 161–172). Cheltenham, UK: Edward Elgar.
Ou, C. X., & Sia, C. L. (2010). Consumer trust and distrust: An issue of website design.
International Journal of Human-Computer Studies, 68(12), 913–934.
Ridings, C. M., Gefen, D., & Arinze, B. (2002). Some antecedents and effects of trust in
virtual communities. The Journal of Strategic Information Systems, 11(3),
271–295.
Riegelsberger, J., Sasse, M. A., & McCarthy, J. D. (2005). The mechanics of trust: A
framework for research and design. International Journal of Human–Computer
Studies, 62(3), 381–422.
Scarbrough, H., Swan, J., Amaeshi, K., & Briggs, T. (2013). Exploring the role of trust
in the deal-making process for early-stage technology ventures.
Entrepreneurship Theory and Practice, 37(5), 1203–1228.
Schlosser, A. E., White, T. B., & Lloyd, S. M. (2006). Converting web site visitors into
buyers: how web site investment increases consumer trusting beliefs and
online purchase intentions. Journal of Marketing, 70(2), 133–148.
Schoorman, F. D., Mayer, R. C., & Davis, J. H. (2007). An integrative model of
organizational trust: Past, present, and future. Academy of Management Review,
32(2), 344–354.
Scupin, R. (1997). The KJ method: A technique for analyzing data derived from
Japanese ethnology. Human Organization, 56(2), 233–237.
Shankar, V., Urban, Glen L., & Sultan, F. (2002). Online trust: A stakeholder
perspective, concepts, implications, and future directions. The Journal of
Strategic Information Systems, 11(3), 325–344.
Sher, P. J., & Lee, S. H. (2009). Consumer skepticism and online reviews: An
elaboration likelihood model perspective. Social Behavior and Personality: An
International Journal, 37(1), 137–144.
Shneiderman, B. (2000). Designing trust into online experiences. Communications of
the ACM, 43(12), 57–59.
Sledgianowski, D., & Kulviwat, S. (2009). Using social network sites: The effects of
playfulness, critical mass and trust in a hedonic context. Journal of Computer
Information Systems, 49(4), 74–83.
Tuch, A. N., Trusell, R., & Hornbæk, K. (2013). Analyzing users’ narratives to
understand experience with interactive products. In Proceedings of the SIGCHI
conference on human factors in computing systems (pp. 2079–2088). ACM: NY.
Uppvall, L. (2009). The collaborative challenge of product development: Exploring
sustainable work systems through critical incidents in R&D alliances. Stockholm,
SE: Department of Industrial Economics and Management, Royal Institute of
Technology.
Wang, Y. D., & Emurian, H. H. (2005). An overview of online trust: Concepts,
elements, and implications. Computers in Human Behavior, 21(1), 105–125.
Wathen, C. N., & Burkell, J. (2002). Believe it or not: Factors influencing credibility on
the Web. Journal of the American Society for Information Science, 53(2), 134–144.
50 M. Seckler et al. / Computers in Human Behavior 45 (2015) 39–50

YOU MAY ALSO READ ...  BPL 5100 Chapter 7 – A firm implements a corporate diversification
Order from Academic Writers Bay
Best Custom Essay Writing Services

QUALITY: 100% ORIGINAL PAPERNO PLAGIARISM – CUSTOM PAPER