Transitioning from the traditional classroom to the online learning environment.

Building student trust in online learning environments

Ye Diana Wang*

Department of Applied Information Technology, George Mason University, Fairfax, VA, USA

(Received 30 December 2013; final version received 19 June 2014)

As online learning continues to gain widespread attention and thrive as a legitimate alternative to classroom instruction, educational institutions and online instructors face the challenge of building and sustaining student trust in online learning environments. The present study represents an attempt to address the challenge by identifying the social and technical factors that can likely induce or influence students’ perception about the trustworthiness of an online course and integrating the factors into a socio-technical framework that can be empirically validated. The methodology used and the data obtained from a university-wide survey conducted in an American university are reported in this article. Feedback from students with disabilities was further investigated, and the result has important implications for our understanding of disabled students’ acceptance for online learning.

Keywords: trust; student perception; e-learning; distance education; disability; accessibility

Introduction and background

The need for building student trust in online learning environments

The tremendous advancements in information and telecommunication technologies and their adoption in education have opened up new avenues for educational institu- tions and students alike. As soon as the Internet became readily available to the pub- lic, distance education was transformed from old-fashioned correspondence schools using printed materials, telephone, TV, or videotape to today’s Internet-based e-learning or online learning (Bernard et al., 2009; Sims, 2008).

Although online learning continues to grow in both presence and importance, its potential is far from being fully utilized. There are still doubts about the effective- ness of online learning environments (see Hashem, 2011). In addition, a higher num- ber of e-learner dropouts, when compared with face-to-face courses, have been reported (Bell & Federman, 2013; Patterson & McFadden, 2009; Tyler-Smith, 2006). This is especially true in the United States, where strong and consistent national quality assurance measures are lacking for online educational institutions. According to a more recent report, the 2012 Sloan Survey of Online Education, nearly 90% of the surveyed academic leaders from more than 2800 colleges and uni- versities in the United States have identified lower retention rates for online courses as a barrier to the widespread adoption of online learning (Allen & Seaman, 2013).

*Email: [email protected]

© 2014 Open and Distance Learning Association of Australia, Inc.

Distance Education, 2014 Vol. 35, No. 3, 345–359, http://dx.doi.org/10.1080/01587919.2015.955267

From the student’s perspective, the decision to take an online course is not an easy one to make: The student must overcome the fear of potentially wasting time and money, disclosing sensitive information, and losing submitted work, and they must take such risks in the absence of face-to-face interactions. This is an even more serious issue for students with disabilities. Many countries have established legisla- tion to legally mandate that disabled individuals should receive equal educational opportunities. For instance, in the United States, the Americans with Disabilities Act of 1990 and Sections 504 and 508 of the Rehabilitation Act of 1973 require educa- tional institutions to provide reasonable accommodations to disabled students and make their electronic and information technology, including online learning applica- tions, accessible. In order to receive accommodations, however, disabled students are required to disclose their disability to the instructor and formally request the nec- essary services or accommodations. Consequently, unwillingness to disclose sensi- tive information is a significant obstacle to their adoption of online learning (Bowker & Tuffin, 2002).

As retention and self-disclosure are increasingly becoming challenges faced by today’s educational institutions and online instructors, as well as preventing the widespread adoption of online learning, it is necessary to seek answers from the topic of trust, which is the “firm belief in the competence of an entity to act depend- ably, securely and reliably within a specific context” (Grandison & Sloman, 2000, p. 4). Research shows that trust is vital for ensuring effective commitments and reducing the level of uncertainty (Kramer, 1999; Luhmann, 2000). It has been repeatedly identified as an important parameter in online learning (Anwar & Greer, 2012; Pelet & Papadopoulou, 2012; Xu & Korba, 2005) for several reasons: firstly, in the case of online learning, trust is involved in the decision-making process prospective students employ when enrolling in online courses. It is also a key factor in preventing current students from dropping out. As pointed out by O’Brien and Renner (2002), establishing trust is essential for online student retention. If prospec- tive students are trusting, they are more likely to enroll in online courses, thereby easing enrolment problems; if current students are trusting, they are less likely to drop out, thereby easing retention problems (Ghosh, Whipple, & Bryan, 2001). As a result, trust in the online learning environment is a large component in easing enrol- ment and retention problems. Secondly, disclosure of personal information depends on trust (Briggs, Simpson, & Angeli, 2004). Since trust reduces the perceived risks involved in revealing private information, it is a precondition for self-disclosure (Anwar & Greer, 2012; Steel, 1991). In other words, if disabled students are trust- ing, they are more likely to disclose their disability to the instructor and receive accommodations, thereby easing acceptance problems. Thirdly, although there are legal requirements for equal access to online education in many countries, there is no financial provision for this so the cost and problems of accommodating disabled students’ needs must be borne by the providers. The students therefore need to be able to entrust the providing institutions to use all the best means of provision at their disposal. Finally, as a student’s trust in a teacher determines the degree to which that student will be open to being taught by that teacher; trust is also a requi- site component of a student–teacher relationship for maximal learning to occur (Wooten & McCroskey, 1996). Therefore, building and maintaining students’ trust in online learning environments is crucial to the success and future of online learning.

346 Y. D. Wang

The present study

The present study aims to address the challenge by identifying the social and techni- cal factors that can likely induce or influence students’ perception about the trust- worthiness of an online course and integrating the factors into a socio-technical framework of trust-inducing factors. The proposed framework is then validated through a survey conducted within an American university. In this study, student trust in an online course is defined as “the degree to which a student is willing to rely on the e-learning system and has faith and confidence in the instructor or the educational institution to take appropriate steps that help the student achieve his or her learning objectives.” This definition is consistent with the concept of trust found in the education literature (Bulach, 1993; Ghosh et al., 2001).

The rest of the article begins with an introduction to related work on trust in the online learning context, describes the proposed socio-technical framework, the research methodology, and the results of the survey, and, finally, ends with conclu- sions and directions for future explorations.

Related work on trust in online learning

Due to the fact that trust has existed as long as the history of human beings and the existence of human social interactions, trust, with all of its permutations, has been studied in numerous disciplinary fields, such as philosophy, psychology, manage- ment, and marketing (Wang & Emurian, 2005). Since the inception of the World Wide Web, a great deal of research focus has been on trust in electronic settings, for example, e-commerce (Cheung & Lee, 2006; Ha & Stoel, 2009; Jarvenpaa, Tractinsky, & Vitale, 2000) and virtual communities and online collaboration (Al-Ani et al., 2013; Ridings, Gefen, & Arinze, 2002; Ziegler & Lausen, 2004).

Compared to trust research in e-commerce and other fields, there is a dearth of literature on investigating trust in the context of online learning (Hashem, 2011). The earliest attempt was Xu and Korba (2002), who proposed a trust model, based on policy negotiation and public key cryptography, for solving security and privacy concerns inherent in distributed interactive e-learning applications. The paper pointed out the importance of maintaining a trustworthy e-learning environment as well as the need for trust evaluation mechanisms. Eight years later, Liu and Wu (2010) presented a survey, which aimed to give a “panoramic view” (p. 118) on trust and trustworthy online learning studies. However, the review covered only a limited number of studies at the intersection of trust and online learning. The authors further encouraged more proper trust models designed for constructing reliable online learning systems.

One line of research into trust and online learning is concerned with security and privacy issues of the online learning processes, platforms, or environments. The three most common approaches to trust establishment and evaluation found in the literature are as follows: (1) policy-based approaches are widely used in security and access control to change the behavior of large distributed systems. Trust-related poli- cies usually consist of authorization policies, which define the authorized and unau- thorized actions of a subject over an object (e.g., authorizing individual access), and obligation policies, which specify the positive and negative obligations of a subject toward an object (e.g., acquiring the user’s consent for collecting data) (El-Khatib, Korba, Xu, & Yee, 2003); (2) certificate-based approaches rely on the use of digital

Distance Education 347

certificates and signatures (e.g., X.509, PGP [Pretty Good Privacy]). A certification authority, which represents a trusted party, issues a digital certificate to identify whether or not a public key truly belongs to the claimed owner or to certify the party is authenticated to be trustworthy or not (El-Khatib et al., 2003); and (3) repu- tation-based approaches are based on one’s own past experience or the recommenda- tions from third parties. In this type of approach, reputation is used to measure trust for online learning. Many examples of formal methods for reputation assessment of a site or of a user can be found on the Web, such as eBay and Epinions (Anwar & Greer, 2012).

In most cases a combination of the aforementioned approaches to trust establish- ment and evaluation is used. For instance, El-Khatib et al. (2003) and Xu and Korba (2005) focused on both policy-based and certificate-based approaches when examin- ing privacy and security standards and issues associated with online learning. Anwar and Greer (2012) presented a new model for facilitating trust in online learning activities by integrating reputation (reputation is calculated on three dimensions) with policies (guarantor vouches for credentials based on reputation) in determining trust.

A notable number of studies focus on using trust mechanisms to collect suitable and trustworthy learning resources in online learning environments. Yang, Chen, Kinshuk, and Chen (2007) identified the difficulties in finding quality learning con- tent and trustworthy learning collaborators to be the major barriers to efficient and effective knowledge sharing in virtual learning communities. To overcome the afore- mentioned barriers, the authors applied peer-to-peer (P2P)-based social networks with trust-management mechanisms, which classified peers based on their content’s quality. In order to tackle the challenges faced by online learning providers in presenting the most suitable learning resources to learners, Carchiolo, Correnti, Longheu, Malgeri, and Mangioni (2008) and Carchiolo, Longheu, and Malgeri (2010) exploited the idea of trustworthiness associated with both learning objects and peers in a P2P e-learning scenario. They proposed a trust- and recommendation- aware framework for searching personalized and useful learning paths suggested by reliable or trusted peers. Similarly, Liu, Chen, and Sun (2011) presented a service- oriented architecture-based e-learning model with quality certification and trust evaluation based on trust computing formulas, which aimed to recommend high- quality and trustable education services to learners.

Another line of research into trust and online learning is concerned with people’s trust and perception toward online learning. Jairak, Praneetpolgrang, and Mekhabunchakij (2009) investigated university instructors’ and students’ perception toward blended learning and fully online learning methods in Thailand, using trust as a predictor for online learning adoption in both delivery methods. Hashem (2011) examined Middle Eastern students’ attitudes toward online education and the role of new information technology in online education, based on various factors that may affect their trust. Both the aforementioned studies used a survey methodology to col- lect feedback from study participants. In addition, Pelet and Papadopoulou (2012) took a unique approach by investigating the effect of color on memorization and trust in online learning. The authors concluded that a careful selection of colors, which constitute an important variable for the design of online learning systems, can enhance students’ trust in the online learning environments and the available con- tent. This result reinforced the importance of interface design factors in inducing trust in online environments.

348 Y. D. Wang

As shown in reviewed work on trust in the online learning context, none of the work has thoroughly investigated the antecedents or determinants of trust in online learning environments. Therefore, a pressing need exists for a deeper and more com- prehensive understanding on the factors that influence student trust in online learn- ing. Without such understanding, it is difficult to build and manage student trust in an online environment, thereby making it more challenging for educational institu- tions and instructors to provide trustworthy, sustaining, and successful online courses.

Proposed socio-technical framework of trust-inducing factors1

A socio-technical framework of trust-inducing factors is proposed in an effort to synthesize existing literature on enhancing student or consumer trust in virtual envi- ronments. This framework is “socio-technical” because it includes both social and technical features of an online course (including its instructor and the system on which the course is built) that can likely induce or influence students’ perception about the trustworthiness of the online learning environment. Compared to other trust models or evaluation mechanisms proposed in the related studies that have been reviewed, the proposed framework is comprehensive since it takes into account not only different types of trust (i.e., policy-based, certificate-based, and reputation- based) but also the communicative styles of the instructor and the interface design of the online learning system. On the other hand, the framework is not exhaustive in the sense that it does not attempt to capture every possible trust-inducing factor that can be applied in an online course. Rather, it focuses on articulating the most promi- nent set of trust-inducing factors derived from numerous previous studies and pre- senting them as an integrated entity that can be evaluated empirically.

The framework classifies 12 trust-inducing factors into four broad dimensions: namely, (1) credibility, (2) design, (3) instructor socio-communicative style, and (4) privacy and security. The dimensions are identified on the basis of a semantic and functional grouping of factors obtained from the literature. Table 1 illustrates the framework in detail, including the dimensions, trust-inducing factors, and literature sources for each dimension.

Specifically, the credibility dimension refers to the cognition-based features, such as previous experience or reputation of the online learning system and the instructor, which are usually formed prior to the current course; the design dimension defines the overall design quality and accessibility of the informational and graphical com- ponents of the online learning system; the instructor socio-communicative style dimension refers to the patterns of communication and interaction behaviors of the instructor; And the last dimension, the privacy and security dimension relates to the privacy and security measures that can be included in the online learning system.

Methodology

Survey

After being reviewed and approved by the university’s Institutional Review Board, a Web-based survey was conducted to confirm the proposed framework of trust- inducing factors. To collect a sample that represents the point of view from the students, or the users of online learning, in higher education, the link to the survey was distributed to students in a four-year university in the United States through

Distance Education 349

various methods, including university listservs, online announcements, and faculty Twitter posts. To encourage participation, 12 $25 Barnes and Noble gift cards were used as incentives given to winners of a random drawing from the survey respon- dents. The data collection was anonymous with respondents’ express consent (by reading the informed consent form and checking the consent checkbox to proceed to the next Web page of the survey). The respondents were also informed that their personal information (i.e., university email address and ID) would not be linked to their submitted answers if they wanted to be included in the prize drawing by enter- ing their personal information on a separate Web page after completing the survey.

Table 1. A socio-technical framework of trust-inducing factors in online learning.

Dimensions Trust-inducing factors Literature sources

Credibility � Prior positive experience with the online learning system or the instructor

� Good reputation of the online learning system or the instructor

Anwar and Greer (2012); Song and Zahedi (2007)

Design � High information and design quality of the online learning system

� Good accessibility and usability of content and tools in the online learning system

� Display of contact details of the instructor or the physical entity behind the online learning system

Bansal, Zahedi, and Gefen (2008); Jaeger and Xie (2009); Nicolaou and McKnight (2006)

Instructor socio- communicative style

� Assertiveness of the instructor

� Responsiveness of the instructor

� A sense of care and community created by the instructor

Curzon-Hobson (2002); Wooten and McCroskey (1996)

Privacy and security

� Disclosure of understandable and adequate privacy and security policy statement

� Use of security mechanisms (e.g., the secure HTTP protocol, encryption, secured logging system)

� Compliance with third-party privacy assurance or standard (e.g., US-EU & US-Swiss Safe Harbor Frameworks, IEEE LTSC)

� Reliable and timely access to the online learning system

Akhter, Buzzi, Buzzi, and Leporini (2009); Bansal et al. (2008); El-Khatib et al. (2003); Raitman, Ngo, Augar, and Zhou (2005)

350 Y. D. Wang

The initial survey was first reviewed by four student counselors and one lan- guage expert for consistency, completeness, and readability. The objective of this step was to examine the validity of each item in the survey. As a result, several items were reworded to improve readability and clarity.

The resulting survey is described as follows. After the aforementioned informed consent form, the first section of the survey consisted of four radio- button groups gathering demographic information on a respondent’s class level, gender, weekly hours spent on the Internet, and experience with taking an online course. The second section of the survey included 12 items to rate, which corre- sponded to the 12 trust-inducing factors in the proposed framework. Respondents rated each item using a 10-point Likert-type scale, which allowed them to select a response indicating the trust-inducing importance of each factor. The scale anchors ranged from 1, representing that the factor was not important at all, to 10, indicating that the factor was extremely important. The third section con- sisted of four questions that are only relevant for students with disabilities. The last section of the survey was a feedback box providing for comments. As men- tioned previously, if the respondent wanted to be included in the prize drawing, he or she could provide personal information on a separate Web page following the survey.

Respondents

Although 398 students responded to the Web-based survey, a total of 361 respon- dents were included in the final analysis; the other 37 students were eliminated due to incomplete submissions. Table 2 presents the characteristics of the participants, based upon the information reported on the survey. Since the survey did not target specific individuals, there is no response-rate calculation. In addition, this approach did not yield a truly random sample from a population, but it did produce a repre- sentative pool of university students.

Among the respondents, 170 students (47%) were female, and 243 (67%) students reported that they had taken an online course. Most of them were undergraduate students (n = 221, 61%), and the rest were graduate and doctoral students. The large majority of the respondents were experienced with the Internet (n = 325, 90% spent more than 10 h per week online; n = 216, 60% spent more than 20 h per week online).

Table 2. Characteristics of survey respondents (N = 361).

Gender Online learning experience

Female 170 Yes 243 Male 191 No 118

Class level Weekly Internet hours

Freshman 35 1–10 h 36 Sophomore 34 11–20 h 109 Junior 87 21–30 h 100 Senior 65 >30 h 116 Graduate student 129 Other 11

Distance Education 351

Statistical analyses and results

To meet the research goals of the study, the data analysis had four parts: (1) validat- ing the proposed framework of trust-inducing factors and confirming the underlying dimensions; (2) evaluating the magnitudes of the ratings across the confirmed dimensions; (3) comparing trust ratings based on demographic and experiential sub- groups; and (4) investigating feedback from students with disabilities.

Validating the proposed framework

To create the classification of the four dimensions, the author initially applied a semantic grouping of the factors obtained from the literature. Additionally, the 12 trust-inducing factors were subjected to a confirmatory factor analysis (CFA) to assess the construct validity and internal reliability of the constructs. CFA is a pow- erful statistical tool for examining the nature of and relations among latent con- structs, for example, attitudes, traits, intelligence, and clinical disorders (Jackson, Gillaspy, & Purc-Stephenson, 2009). In the present study, it was used to validate the trust-inducing factors and determine the essential dimensions of the identified fac- tors. Before a CFA could be applied, however, two tests that indicated the suitability of the data for structure detection had to be run: The extremely high value (.908) from the Kaiser-Meyer-Olkin test, which measures sampling adequacy, indicated that a factor analysis would be useful with the data. The significant Bartlett’s test (p < .001), which examines whether the variables are related, indicated that the data were suitable for structure detection. Therefore, a CFA was performed.

The principal components analysis was used to analyze the raw matrix of 361 responses with the latent root criterion (eigenvalue = 1). Surprisingly, there were only two components with eigenvalues greater than 1 (i.e., eigenvalue = 6.65 and eigenvalue = 1.44); these two components accounted for 67% of the total variance of the data-set. The screen test, which showed that there were some bending points at two components, further verified the number of dimensions. Based on this initial analysis, the author tried several rotation methods to determine which factors loaded on each of the two dimensions. The Varimax rotation method, which best revealed the underlying relationship, was chosen eventually. As can be read from Table 3, all factor loadings reach the acceptable level of .3 (Nunnally, 1978), with most of them exceeding .7. This means that no factor in the proposed framework should be elimi- nated since every item fit into one of the two components (all factor loadings ≥.30). The analysis also showed that the items for each component loaded unambiguously.

The major difference between the analysis result and the proposed model was the number of dimensions: the 12 factors clustered into two components rather than four. Closer investigation indicated that the second component actually included all three factors but the last one in the privacy and security dimension, and the first component included all the other factors. Thus, the author named the first compo- nent the course instruction dimension, which related to different aspects (e.g., repu- tation, design quality, and instructor socio-communicative style) of the online course, and kept the last component as the privacy and security dimension. To examine the internal reliability of each dimension (i.e., course instruction, privacy, and security), Cronbach’s alpha was calculated on each dimension, and the alpha coefficients were .91 and .90, respectively. According to Nunnally (1978), an alpha of .50 or higher indicates a sufficient level of internal reliability. Therefore, based on

352 Y. D. Wang

these results, it may be concluded that these two dimensions represented different aspects or features of an online course or environment to promote student trust.

Evaluating relative importance of dimensions

To investigate the relative magnitudes in ratings among the survey items that fell within each of the two dimensions, the median rating across those items was deter- mined for each of the 361 respondents. The median is the appropriate index of cen- tral tendency for ordinal data (Sermeus & Delesie, 1996). Figure 1 presents boxplots of the medians of those ratings for each of the two dimensions. It shows that both medians exceed 5, but the median for the course instruction dimension is higher than that for the privacy and security dimension (9 vs. 8). The results of the Kruskal-Wallis test with pairwise comparisons show significant difference between the two dimensions (χ2 = 14.33, df = 1, p < .001). This data suggests that every

Table 3. Rotated component matrix of the trust-inducing factors (N = 361).

Dimensions Features

Component

1 2

Course instruction C1 – Prior positive experience .515 C2 – Good reputation .733 C3 – High information and design quality .894 C4 – Contact details .810 C5 – Instructor assertiveness .632 C6 – Instructor responsiveness .599 C7 – A sense of care and community .801 C8 – Reliable and timely access .760

Privacy and security P1 – Privacy and security policy statement .835 P2 – Security mechanisms .897 P3 – Third-party privacy assurance or standard .883

Figure 1. Boxplot of the median ratings of the items within each dimension (the circles are outliers).

Distance Education 353

factor contributed to the value of the respondents’ evaluations, but the privacy and security dimension was rated as slightly less important than the course instruction dimension.

Comparing demographic and experiential subgroups

Based on the different characteristics of the respondents, the average ratings of the 12 trust-inducing factors were compared. The purpose of this part of the analysis was to investigate whether demographic characteristics and individual experiences (i.e., gender, class level, weekly hours spent on the Internet, and online learning experience) are related to students’ overall ratings of the trust-inducing factors in the socio-technical framework under consideration.

The Kruskal-Wallis test was conducted and showed no significant difference between female (n = 170) and male (n = 191) respondents in their average trust rat- ings (χ2 = 1.64, df = 1, p = .20). A comparison among the six class-level categories (i.e., freshman, sophomore, junior, senior, graduate student, and other) was not sig- nificant (χ2 = 4.68, df = 5, p = .46). The respondents selected one of four categories, based on their reported weekly hours spent on the Internet (i.e., 1–10 h, 11–20 h, 21–30 h, and >30 h). The result of the Kruskal-Wallis test across the four time inter- vals was not significant (χ2 = .93, df = 3, p = .82). A comparison was also made between the respondents who reported previous online learning experience (i.e., took at least an online course) (n = 243) and those who did not (n = 118). There was no significant difference in trust ratings between the two subgroups (χ2 = .029, df = 1, p = .865). Therefore, the results indicate that the demographic characteristics and individual experiences under investigation do not have a correlation to students’ overall ratings in the survey.

Investigating feedback from students with disabilities

Out of the 361 respondents, 15 students (9 females and 6 males) identified them- selves as having one or more disabilities and answered additional questions in the survey regarding self-disclosure and trust in online learning environments. Table 4 shows that the most common disability (60%) noted by the students was attention deficit disorder/attention deficit hyperactivity disorder, followed by learning disabil- ity (33%). Out of the 15 students, 7 (47%) reported more than one disability for a total of 35 disabilities. They were enrolled in a variety of class levels in both under- graduate and graduate schools.

As explained previously, a disabled student is required to send a formal request in order to receive the necessary services or accommodations in an online course. At the university where the study took place, this is initiated by a disabled student sub- mitting the faculty contact sheet (FCS), which is issued from the Office of Disability Services and discloses the type(s) of disability that the student has and the specific accommodations that the student needs, to the instructor of the online course. Out of the 15 students with disabilities, 10 students (67%) indicated that they would pro- vide the FCS to the instructor only when they need accommodations, and the other 5 students (33%) indicated that they would do so during the first week of the semester or before class starts. In responding to the question ‘If an online course is perceived to be trustworthy to you, will you provide the Faculty Contact Sheet to your instructor before or during the first week of the semester?’, 11 students (73%)

354 Y. D. Wang

indicated ‘yes’, including 6 out of the 10 students who only wanted to provide the FCS to the instructor as needed. Therefore, the results show that trust, or perceived trustworthiness of an online course, does have a positive influence on the level of self-disclosure of students with disabilities.

Discussion

In this article, the underlying dimensions of the proposed socio-technical framework of trust-inducing factors were confirmed, and the relative magnitudes of respon- dents’ ratings of the confirmed dimensions were further evaluated. The results of the CFA suggest that two underlying dimensions, course instruction and privacy and security, exist among the 12 trust-inducing factors. Although all 12 factors were found to contribute to the respondents’ perception of the trustworthiness of an online course, the two identified dimensions differed in terms of their relative importance to inducing student trust. The course instruction dimension was rated about 10% higher than the privacy and security dimension. This suggests that the social and course design factors (e.g., reputation, design quality, and instructor socio-communi- cative style), when used effectively, can help overcome students’ privacy and secu- rity concerns for an online course.

A further step was taken to investigate whether demographic characteristics and individual experiences (i.e., gender, class level, weekly hours spent on the Internet, and online learning experience) were related to students’ average ratings of the 12 trust-inducing factors. The results show that none of the demographic and experien- tial factors have significant relationships with trust ratings in the current study. The lack of significant differences in the results might be attributable to the sample size and the similar background of the respondents, who are all students in a four-year university in the US. In addition, a small number of the respondents might be unfa- miliar with some of the terminology (e.g., security protocols or mechanisms) used in the trust-inducing factors to assess. More explicit examples or visual demonstration of features can be used to assist respondents in assessing the perceived trustworthi- ness of certain factors, and thus, more thorough examination and elaboration of the results may be needed in the future.

Finally, the feedback from 15 students with disabilities was investigated. The majority of the students initially held reservations against disclosing their disabilities to, and requesting accommodations from, the instructor in an online course before the class starts or during the first week of the semester; however, should the online course be perceived to be trustworthy, 60% of these students indicated their desire to withdraw their reservations and reach out to the instructor. This shows that trust,

Table 4. Percentage of respondents indicating various disabilities/impairments (N = 15).

Disability/Impairment Percentage of respondents (%)

ADD/ADHD 60 Learning disability 33 Mobility 20 Medical impairment 20 Speech and language impairment 13 Deaf and hard of hearing 13 Asperger/Autism 13 Psychological/Emotional impairment 13

Distance Education 355

or perceived trustworthiness of an online course, does have a positive influence on the level of self-disclosure of students with disabilities. Despite the small sample size, this observation has important implications for our understanding of disabled students’ acceptance for online learning.

The results of the survey are supportive of implementation of trust-inducing fac- tors in all aspects of an online learning environment, including the delivery system or platform, the course content, and the instructor. Instructional designers and online instructors would be well advised not to neglect the contributions of all these aspects, as these social and technical factors act together to promote student trust. There was no direct comparison among the trust-inducing factors in the framework with respect to which factor has the highest strength of correlational relationship with the overall perceived trustworthiness of an online course. Rather, we suggest that all factors serve as antecedent variables that might be influential in the rated ingredients constituting the dimensions, with the course instruction dimension show- ing the most robust correlational relationship with the students’ trust ratings.

Conclusion

As online learning continues to gain widespread attention and thrive as a legitimate alternative to classroom instruction, educational institutions, and online instructors face the challenge of building and sustaining student trust in online learning environ- ments. The present study represents an attempt to address the challenge by identify- ing the social and technical factors that can likely induce or influence students’ perception about the trustworthiness of an online course and integrating the factors into a socio-technical framework that can be empirically validated.

The contributions that the present study brings to the research field are threefold. First, the study identifies 12 trust-inducing factors from the literature and provides empirical evidence and indicative support for their importance in affecting students’ trust in online learning. It fills the gap in research by focusing on the antecedents or determinants of student trust in online learning environments. Second, the study extends the CFA to a new application area of trust evaluation in online learning. Although online learning evaluation has been traditionally limited to the evaluation of teaching effectiveness, CFA offers an accessible analysis method for researchers to investigate and promote online learning from a unique angle. Last, but not least, the results of the study contribute to the growing literature, which suggests that trust is a precondition for disclosing of sensitive information by students with disabilities not only in face-to-face situations but also in online environments. By implementing strategies and features that enhance the trustworthiness of online learning environ- ments, online instructors can be more effective in meeting their responsibilities under the law by practicing inclusive instruction and helping students with disabili- ties succeed in online learning.

One line of future research is in relation to the additional factors that could be continuously added to the socio-technical framework. For example, the respondents in the survey have repeatedly identified that the inclusion of a face-to-face opportu- nity and interactive social media could help foster their trust in an online course, in addition to the identified trust-inducing factors. Future research may continue to val- idate the framework in a more controllable experimental setting and examine the issues of online trust in regard to gender, ethnicity, or culture. Another aspect of trust that can be worthy of investigation is the student-to-student trust in the online

356 Y. D. Wang

learning context, as the present study exclusively focuses on student trust in an online course, including the e-learning system, the instructor or the educational insti- tution. With the increasing use of social media and tools for collaborative learning in online courses, the interaction and trust between students may play a role in the course’ sustainability and the students’ performance.

Future research may also explore the intersection of online learning and disabil- ity in depth, especially investigating how to establish a trustworthy online learning environment for students with disabilities, and ultimately improving the online learn- ing experience of students with a wide variety of barriers.

Acknowledgments Funding for this research was provided by the Office of Distance Education, George Mason University.

Note 1. A briefer version of the framework was presented at the 15th Annual ATINER Interna-

tional Conference on Education, 2013, and appeared within the proceedings.

Notes on contributor Ye Diana Wang is an associate professor in the Department of Applied Information Technology at George Mason University. She has taught information technology (IT) for over 10 years, at undergraduate and graduate levels and in classroom and online settings. Her recent research focus is on pedagogy, IT education, and distance education.

References Akhter, F., Buzzi, M. C., Buzzi, M., & Leporini, B. (2009). Conceptual framework:

How to engineer online trust for disabled users. In P. Boldi, G. Vizzari, G. Pasi, & R. Baeza-Yates (Eds.), Proceedings of the 2009 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology (Vol. 3, pp. 614–617). New York, NY: IEEE.

Al-Ani, B., Redmiles, D., de Souza, C. R., Prikladnicki, R., Marczak, S., Lanubile, F., & Calefato, F. (2013). Trust in virtual teams: Theory and tools. In A. Bruckman & S. Counts (Eds.), Proceedings of the 2013 Conference on Computer Supported Cooperative Work Companion (pp. 301–306). New York, NY: ACM.

Allen, I. E., & Seaman, J. (2013). Changing course: Ten years of tracing online education in the United States. San Francisco, CA: Babson Survey Research Group and Quahog Research Group LLC.

Anwar, M., & Greer, J. (2012). Facilitating trust in privacy-preserving e-learning environ- ments. IEEE Transactions on Learning Technologies, 5, 62–73. doi:10.1109/TLT.2011.23

Bansal, G., Zahedi, F., & Gefen, D. (2008). The moderating influence of privacy concern on the efficacy of privacy assurance mechanisms for building trust: A multiple-context investigation. Proceedings of the Twenty-ninth International Conference on Information Systems (pp. 1528–1546). Atlanta, GA: Association for Information Systems.

Bell, B. S., & Federman, J. E. (2013). E-learning in postsecondary education. The Future of Children, 23, 165–185. doi:10.1353/foc.2013.0007

Bernard, R. M., Abrami, P. C., Borokhovski, E., Wade, C. A., Tamim, R. M., Surkes, M. A., & Bethel, E. C. (2009). A meta-analysis of three types of interaction treatments in dis- tance education. Review of Educational Research, 79, 1243–1289. doi:10.3102/ 0034654309333844

Distance Education 357

Bowker, N., & Tuffin, K. (2002). Disability discourses for online identities. Disability and Society, 17, 327–344. doi:10.1080/09687590220139883

Briggs, P., Simpson, B., & Angeli, A. D. (2004). Personalisation and trust: A reciprocal rela- tionship? In C. Karat, J. O. Blom, & J. Karat (Eds.), Designing personalized user experi- ences in eCommerce (pp. 39–55). Dordrecht: Kluwer Academic.

Bulach, C. (1993). A measure of openness and trust. People and Education, 1, 382–392. Carchiolo, V., Correnti, D., Longheu, A., Malgeri, M., & Mangioni, G. (2008). Reliable per-

sonalized learning paths: the contribution of trust to e-learning. In M. D. Lytras, J. M. Carroll, E. Damiani, R. D. Tennyson, D. Avison, G. Vossen, & P. Ordóñez de Pablos (Eds.), The open knowledge society: A computer science and information systems mani- festo (pp. 221–228). Berlin: Springer.

Carchiolo, V., Longheu, A., & Malgeri, M. (2010). Reliable peers and useful resources: Searching for the best personalised learning path in a trust-and recommendation-aware environment. Information Sciences, 180, 1893–1907. doi:10.1016/j.ins.2009.12.023

Cheung, M. K., & Lee, M. K. O. (2006). Understanding consumer trust in Internet shopping: A multidisciplinary approach. Journal of the American Society for Information Science and Technology, 57, 479–492. doi:10.1002/asi.20312

Curzon-Hobson, A. (2002). A pedagogy of trust in higher learning. Teaching in Higher Edu- cation, 7, 265–276. doi:10.1080/13562510220144770

El-Khatib, K., Korba, L., Xu, Y., & Yee, G. (2003). Privacy and security in E-learning. Inter- national Journal of Distance Education Technologies, 1, 1–19. doi:10.4018/ jdet.2003100101

Ghosh, A. K., Whipple, T. W., & Bryan, G. A. (2001). Student trust and its antecedents in higher education. Journal of Higher Education, 72, 322–340. doi:10.2307/2649334

Grandison, T., & Sloman, M. (2000). A survey of trust in Internet applications. Communica- tions Surveys & Tutorials, IEEE, 3, 2–16. doi:10.1109/COMST.2000.5340804

Ha, S., & Stoel, L. (2009). Consumer e-shopping acceptance: Antecedents in a technology acceptance model. Journal of Business Research, 62, 565–571. doi:10.1016/j.jbus- res.2008.06.016

Hashem, M. E. (2011). The role of new information technology in education and develop- ment: A case study of building trust in distance education. International Journal of Arts & Sciences, 4, 387–398.

Jackson, D. L., Gillaspy, J. A., & Purc-Stephenson, R. (2009). Reporting practices in confir- matory factor analysis: An overview and some recommendations. Psychological Methods, 14, 6–23. doi:10.1037/a001469

Jaeger, P. T., & Xie, B. (2009). Developing online community accessibility guidelines for persons with disabilities and older adults. Journal of Disability Policy Studies, 20, 55–63. doi:10.1177/1044207308325997

Jairak, R., Praneetpolgrang, P., & Mekhabunchakij, K. (2009). An investigation of trust in e-learning for instructors and students in private and public universities. In S. Charmonman (Ed.), Proceedings of the 6th eLearning for Knowledge-Based Society Conference (pp. 17–18). Bangkok: Assumption University Press.

Jarvenpaa, S. L., Tractinsky, N., & Vitale, M. (2000). Consumer trust in an internet store. Information Technology and Management, 1, 45–71. doi:10.1111/j.1083-6101.1999. tb00337.x

Kramer, R. M. (1999). Trust and distrust in organizations: Emerging perspectives, enduring questions. Annual Review of Psychology, 50, 569–598. doi:10.1146/annurev.psych. 50.1.569

Liu, Y., Chen, D., & Sun, J. (2011). A trustworthy e-learning based on trust and quality eval- uation. Proceedings of 2011 International Conference on E-Business and E-Government (pp. 1–4). New York, NY: IEEE.

Liu, Y., & Wu, Y. (2010). A survey on trust and trustworthy e-learning system. In F. L. Wang, Z. Gong, X. Luo, & J. Lei (Eds.), Proceedings of 2010 International Conference on Web Information Systems and Mining (pp. 118–122). Berlin: Springer-Verlag.

Luhmann, N. (2000). Familiarity, confidence, trust: problems and alternatives. In D. Gambet- ta (Ed.), Trust: Making and breaking cooperative relations, 6 (pp. 94–107). New York, NY: Basil Blackwell.

358 Y. D. Wang

Nicolaou, A. I., & McKnight, D. H. (2006). Perceived information quality in data exchanges: Effects on risk, trust, and intention to use. Information Systems Research, 17, 332–351. doi:10.1287/isre.1060.010

Nunnally, J. C. (1978). Psychometric theory (2nd ed.). Englewood Cliffs, NJ: McGraw-Hill. O’Brien, B. S., & Renner, A. L. (2002). Online student retention: Can it be done? In

P. Barker & S. A. Rebelsky (Eds.), Proceedings of the World conference on educational multimedia, hypermedia and telecommunications (pp. 1479–1483). Waynesville, NC: Association for the Advancement of Computing in Education.

Patterson, B., & McFadden, C. (2009). Attrition in online and campus degree programs. Online Journal of Distance Learning Administration, 12. Retrieved from http://www.west ga.edu/~distance/ojdla/

Pelet, J., & Papadopoulou, P. (2012). Investigating the effect of color on memorization and trust in e-learning: The case of KMCMS.net (knowledge management and content man- agement system). In Information Resources Management Association (Ed.), Organiza- tional learning and knowledge: Concepts, methodologies, tools and applications (pp. 1501–1527). Hershey, PA: Business Science Reference.

Raitman, R., Ngo, L., Augar, N., & Zhou, W. (2005). Security in the online e-learning envi- ronment. In P. Goodyear, D. Sampson, D. Jin, T. Yang, Kinshuk, T. Okamoto, R. Hartley, & N.-S. Chen (Eds.), Proceedings of the Fifth IEEE International Conference on Advanced Learning Technologies (pp. 702–706). Washington, DC: IEEE Computer Soci- ety.

Ridings, C. M., Gefen, D., & Arinze, B. (2002). Some antecedents and effects of trust in vir- tual communities. The Journal of Strategic Information Systems, 11, 271–295. doi:10.1016/S0963-8687(02)00021-5

Sermeus, W., & Delesie, L. (1996). Ridit analysis on ordinal data. Western Journal of Nurs- ing Research, 18, 351–359. doi:10.1177/019394599601800309

Sims, R. (2008). Rethinking (e)learning: A manifesto for connected generations. Distance Education, 29, 153–164. doi:10.1080/01587910802154954

Song, J., & Zahedi, F. (2007). Trust in health infomediaries. Decision Support Systems, 43, 390–407. doi:10.1016/j.dss.2006.11.011

Steel, J. L. (1991). Interpersonal correlates of trust and self-disclosure. Psychological Reports, 68, 1319–1320. doi:10.2466/PR0.68.4.1319-1320

Tyler-Smith, K. (2006). Early attrition among first time eLearners: A review of factors that contribute to drop-out, withdrawal and non-completion rates of adult learners undertaking e-Learning programmes. Journal of Online Learning and Teaching, 2, 73–85. Retrieved from jolt.merlot.org/

Wang, Y. D., & Emurian, H. H. (2005). An overview of online trust: Concepts, elements, and implications. Computers in Human Behavior, 21, 105–125. doi:10.1016/ j.chb.2003.11.008

Wooten, A. G., & McCroskey, J. C. (1996). Student trust of teacher as a function of socio- communicative style of teacher and socio-communicative orientation of student. Commu- nication Research Reports, 13, 94–100. doi:10.1080/08824099609362075

Xu, Y., & Korba, L. (2002). A trust model for distributed e-Learning service control. In M. Driscoll & T. Reeves (Eds.), Proceedings of World Conference on E-Learning in Corporate, Government, Healthcare, and Higher Education (pp. 2419–2434). Chesapeake, VA: AACE.

Xu, Y., & Korba, L. (2005). Building trust for interactive e-learning. In C. Howard, J. Boettcher, L. Justice, K. Schenk, P. Rogers, & G. Berg (Eds.), Encyclopedia of dis- tance learning (pp. 192–195). Hershey, PA: Information Science Reference.

Yang, S. J. H., Chen, I. Y. L., Kinshuk, & Chen, N.-S. (2007). Enhancing the quality of e-learning in virtual learning communities by finding quality learning content and trust- worthy collaborators. Educational Technology & Society, 10, 84–95. Retrieved from http://www.ifets.info/

Ziegler, C. N., & Lausen, G. (2004). Analyzing correlation between trust and user similarity in online communities. In C. Jensen, S. Poslad & T. Dimitrakos (Eds.), Proceedings of the 2nd International Conference on Trust Management (pp. 251–265). Berlin: Springer.

Distance Education 359

Copyright of Distance Education is the property of Routledge and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.