Advances In Management Err.

-1... "-. Vol. 5 (1) Jan. (2012)

Good Knowledge Management System, Bad Shared Knowledge: What happens to trust when experts share erroneous knowledge with novice KMS USERS?
Morrison Rodger
Dept. of Management, Troy University, Alabama 36082, LISA rmorrison@troy.edu

Abstract Expert locator systems are specialized forms of knowledge management systems used to create a social relationship between novice and expert users for the purpose of knowledge location, transfer and utilization. Ranging from simple directories of experts to complex systems that record entire conversations in a multimedia environment, they allow organizations to better capitalize on knowledge assets throughout their enterprise. These systems are especially useful in large organizations that utilize obsolete technologies, but have limited knowledge resources with which to maintain them. A key component of any successful system implementation project is user trust, which develops quickest in systems that provide routinely usable information. In this study, a research proposition is presented that the factuality of information provided by expert users to novice users has an effect on trust levels that novice users develop toward the expertise locator system itself. Keywords: Knowledge Management, Expert Locator, Trust, Novice User

aP^ faith has little to do with the person being trusted and more with the predisposition of the trusting person to develop trusting relationships in the first place. For example, a person who suffered at the hands of an abuser may be much slower to develop trusting relationships than a person that spent an entire life in a healthy, non-threatening environment. Consider also that dictionary definitions of faith generally describe it as a solidly accepted belief of a concept without any proof that the concept is true. Rempel et al's concept of faith is similar. It is the construct of trust in which the trusting person forms an opinion as to the trustworthiness of the other in the absence of any objective information. During interpersonal communications, perceived information about the constructs of trust, dependability, predictability and faith leads the communicating parties to develop a level of trust towards the other participants. In situations of high media richness such as face-toface, video and audio communications, a higher level of trust develops than when text alone is used.* This suggests that information not found in text-only environments also influences trust levels. That is, the non-verbal components of a communication environment influence the level of trust that develops. By perceiving facial expressions and body language, along with changes in vocal intonation, communicators can often perceive the subtle changes that may indicate dishonesty or a lack of confldence in information factuality.'" This information also influences the development of trust between communicators that use it to determine whether erroneous information was meant to deceive or was simply given in error. In text-only communications environments, information is not as rich as in face-to-face communications and such cues are typically not present. This can lead to a situation where the receiver of erroneous information is unable to determine whether or not the sender of the information purposed to be deceitful or simply made a mistake which may result in the receiver developing less trust toward the sender. Text-only chat environments are sometimes used in a specialized form of knowledge management system known as an expert locator system. These systems are used in organizations to assist novice users in locating expert users of a given knowledge domain. In addition to a directory of such expert users, these systems often contain a mechanism for the facilitation of text-based knowledge sharing. That is, they (9)

Introduction
Rotter'^ defined trust as "an expectancy held by an individual or a group that the word, promise, or verbal or written statement of another individual or group can be relied upon". In this context, trust has three primary components: dependability, predictability and faith.^* Dependability is related to the general infallibility of an entity such that decisions may be made based on information obtained from it. An entity is considered to have dependability if it is generally free from error. For example, a worker's estimate of the number of overtime hours worked on a Saturday may not be as accurate as the recorded punches by a time clock. The time clock, being less prone to error, would then be considered more dependable. Predictability, on the other hand, is related to an expectancy of outcomes given a particular circumstance. For example, if a driver turned on the radio of a car and instead of hearing the expected music, the windshield wipers began working; the outcome would not be what was predicted. Unlike two of the constructs of trust described by Rempel et

Advances In Management
often contain some form of a chat system by which novice users can dialog in a text only environment with expert users. When an information system is used to present information that may not be completely accurate, trust levels are not as high , even when the information was not generated by the system itself but the individuals that use it. There appears to be a gap in the knowledge management research body regarding the effects of erroneous information mistakenly introduced by expert users into such systems. To fill this research gap, researchers must understand the role that erroneous information provided by experts has on how much the novice users will trust the expert locator system itself. This information can then be used, to develop expert locator systems that will continue to be used even when erroneous information is mistakenly- presented by the experts that
provide it. T r u s t . , . .. ' , / ... ' ,. . . . ' . . . .

-.;:;d Vol. 5(1) Jan. (2012) cause an event, that event should occur. If a different event occurs, perceptions of predictability will drop, as will overall trust levels. Finally, certain factors of an e-commerce website contribute to an individual's willingness to develop a trusting relationship.'" For example, one such factor is the presence of an e-commerce website's reputation. A good reputation assists in the development of trust without further evidence of trustworthiness. Some people are quicker than others in forming trusting relationships based on preconceived notions. The presence of a good reputation assists the development of trust by those individuals who do not develop trusting relationships quickly. The idea of differing levels of individuals' willingness to develop trust is similar to Rempel's notion of faith. Bos et al^ researched trust as it relates to cooperation and showed that in some cases, social forces may cause trust to develop more slowly in spme cases than in others. For example, in. the richer, environment of face-to-face communications, trust levels develop quicker and tend to be stronger than in an online text-only chat environment."' Richer communication environments generally allow for the communication of social cues such as facial expressions, intonation changes in voice patterns and gesturing, all of which are key to the development of interpersonal trust. However, by engaging in social activities early in online communication, trust between individuals develops quicker and stronger than between individuals who do not engage in such activities.-''* This supports the idea that text-only chat communication is still a social phenomenon in which social forces influence the speed at which trust develops between individuals. When individuals interact to accomplish a clearly defined goal in a limited amount of time, they must often rely on swift trust. Swift trust develops when individuals must temporarily put aside their negative suspicions about the dependability of individuals they do not know so that quick progress can be made toward accomplishing a common goal." For example, in text only interaction involving the sharing of expert-novice knowledge, the novice must overlook the absence of non-verbal cues about trustworthiness in order to develop trust in an expert, known only by reputation. Similarly, in asynchronous online instruction environments, where the experts are instructors and the novices are students, novices are typically unable to use non-verbal cues to trustworthiness that are associated with face-to-face communication environments." Novices must be willing to develop swift trust in the^ experts to share accurate knowledge, even though they may have no information as to the exerta'''trustworthiness such as the accuracy of knowledge previously shared with others. Erroneous I ;ihformation has a strong negative influence on an individual's willingness to develop trust in an information -system, as well. In the context of individuals

Researchers have noted that trust definitions differ, with context such as when comparing trust in an online chatroom with trust directed toward a website by a user or when comparing group trust to individual trust. For example, in a group context, definitions may include rhetoric that is not present in the context of interpersonal trust. Barber ^ defined trust in a group context as a "set of socially learned and socially confirmed expectations that people have of each other, of the organizations and institutions in which they live and of the natural and moral social orders that set the fundamental understandings for their lives." However, in an individual context, Komiak and Benbesat" define trust as having both an emotional and a cognitive context but no social context. Like the definitions of trust, rhetoric used tb describe the components of trust also differ based on the context in which they are given but is generally similar to the three described by Rernpel et al.^^ i Foi- example, Komiak and Benbesat'' describe one' of the' components of trust as a customer's feelings about whether or not one should rely on an entity. This is similar to Rempel's^^ description of the dependability or reliability, of an entity, one of the three components given for ^trust. The same concept can be found regarding components of trust with regard to trusting information systems. For example, the perceived reliability of the technology used on. an ecommerce website has been'shown to be a component of the level of trust that develops regarding visitor trust of the website. Visitors typically regard newer technology as more reliable and hence more trustworthy than severely dated technology.'" Users visiting website designed with text-only HTML menus may not view that website as being as dependable as one designed with java-enabled controls,' hit counters and active media.content. ' Next, it is important that the output- of an e-commerce website meet the user's expectations, similar to Rempel's idea of predictability. If a user clicks on a button to

(10)

Advances

In Management

>z^=2:2ni~=3xsm:=:=n;^isni::: ;

_,_

.,-.^,.....^~^,^..

.,, Vol 5 () Jan (2012)

trusting automated systems, trust levels tend to decline as system accuracy declines. Specifically, once automated systems do not present information as expected, once-high trust levels are diminished. Furthermore, if trust levels drop because of erroneous information, they may never fully recover to previous levels.'^ Similarly, as perceptions of an information system's general reliability go down, a user's willingness to trust that system also goes down.^' '* Individuals are not willing to develop trust in an information system that does not provide consistently reliable and predictable output. Although most studies involving trust differ widely in context and consequentially in rhetoric, the constructs under study usually fall into one of three main themes. The first addresses the factuality of the information itself. Naturally, information known to be correct leads to higher levels of trust than information known to be incorrect. Once someone discovers false information in the communication environment, trust levels drop quickly and are usually slow to recover. In fact, trust levels may never fully recover. Most researchers use the terms dependability or reliability to represent the truthfulness of the information. 15. 18.23 The second commonly expressed concept is that in a particular situation, the information presented will be what is expected. That is, a given circumstance should produce an outcome that fits within expected parameters. Researchers typically refer to this as predictability.'' '"* For example, if a user clicks a button on an online e-commerce website to place a single item in the shopping cart and the website adds items that the user does not intend to purchase, then the website is not producing the expected action and trust levels will most likely drop. Finally, many studies refer in some way to the idea that individuals or groups generally have some preconceived notion of the other party's trustworthiness that influences the development of trust.*' ' Similarly, some people are more prone than others to develop trusting relationships quickly in any given circumstance.'^ For example, an individual's ability to develop trusting relationships with the authority figures may be hindered if the individual was abused by an authority figure in the past.

Knowledge, therefore, is frequently used to describe an individual's insight or know-how concerning a particular subject area. Knowledge assets are generally held to be among the most critical to corporate success. It is knowledge that separates the expert user from the novice and justifies the expert's higher compensation. However, proper utilization of certain types of knowledge has proven to be a challenge. Many authors have drawn distinctions between different types of knowledge. For example, deep knowledge is knowledge typically found in technically complex areas or in scientific areas and differs from shallow knowledge, which might be found in social areas where understood theory and organized learning is not as common."' Another description of the knowledge found in scientific domains is esoteric meaning that it applies to a very narrow domain. The opposite form of knowledge from esoteric is exoteric, applying to a broader domain which is sometimes referred to as common sense. Within organizations, information is frequently recorded for later use in decision-making. Facts that are recorded for later use are collectively referred to as declarative knowledge while recorded methodologies or ways of doing things, are referred to as procedural knowledge.^' Similarly, explicit knowledge is knowledge that can be easily recorded for later use. Tacit knowledge, on the other hand, usually exists only in the minds of individuals and. is generally not recordable. An example of tacit knowledge includes an individual's insight, or expertise in a given subject area. 24.25

Expertise Locator Systems

An expertise locator system is a specialized form of knowledge management system that is designed to facilitate the search for, location of and sharing of tacit knowledge within an organization. These systems are typically comp; rised of a directory of individuals with expertise in one or. more subject areas, along with a way for system users to search for, locate, contact and sometimes communicate with those individuals to help fill some need. The idea behind such systems is to facilitate the sharing of tacit knowledge between expert and novice users. This sharing is a social process and is categorized as organizational knowledge transfer.'^ . Some researchers have focused on the social aspect of knowledge sharing to suggest that knowledge management in general deals less with managing knowledge and more with managing workers whose output is a product of what they know and learn.'' '*' ^* Because of the social forces that coincide with the use of these systems, some researchers have used social exchange theory as a lens through which to study them.^' ''' ^^ Social exchange theory"* would suggest that individuals use such systems because they expect a balance to be maintained in the sharing of information. This balance

Trust and Knowledge Management Systems

The ability of an organization to leverage its assets against those of its competitors is one of the key factors that lead to success. One corporate asset that can be both difficult to obtain and difficult to leverage is knowledge. Knowledge can be referred to as the product of human learning, that is, the interpretation of information.'^ Although the definitions found in extant literature are similar for both information and for knowledge, many authors tend to regard information as recorded concepts and knowledge as a set of beliefs concerning the interrelationships between those concepts.^^' ^^' '"'

(11)

Advances In Management
typically takes the form of an expectation of reciprocity, both in giving and in receiving, such that using a knowledge system incurs an unwritten obligation to contribute to it at a later time. Conversely, individuals contribute to these systems, expecting future returns in their favor, , Social exchange theory provides much insight into how often individuals use knowledge management systems, Watson and Hewett^^ use social exchange theory to describe four main issues affecting the frequency in which these systems are used including (1) how fast an individual is promoted, (2) how long an individual has been with the company, (3) how effective the shared knowledge was in terms of contributing to project success and (4) the frequency that shared knowledge is used. This work suggests that when shared information contributes to project performance, the knowledge sharing system will likely be used more frequently. This concept is based on the idea that as valuable knowledge is shared by means of a knowledge management system, individuals will use that system more often because they believe that further use will continue to result in positive outcomes.^' However, information that is shared may be factual but may be viewed differently by experts and novices. Researchers have shown that experts store and process information and knowledge differently than novice users. One difference is related to the ways in which experts and novices solve problems,^ For example, novices in mathematics often .realize the value of graphically representing a problem in geometry but are slow to apply the method of graphic modeling to other areas of mathematics, Fxperts in mathematics, on the other hand, quickly use graphic modeling as a problem solving aid in many areas of mathematics,^^ Another difference is related to the manner in which users specify information-". Novice users may have difficulty expressing the necessary details of a problem when seeking the advice of expert users who may likewise have difficulty choosing rhetoric suitable to novice users. Assuming that the expert was able to communicate valuable knowledge effectively, trust levels should increase toward the knowledge sharing system, eveh though the knowledge was not generated by the system but by the expert. Further, the inverse should hold true. Erroneous knowledge that was effectively communicated should lead to lower levels of trust development towards the knowledge management system, even though the system did not generate the,.information. Thus, a positive relationship should exist between the accuracy of shared knovvledge and the trust level that develops to ward the knowledge sharing system.

3 Vol. 5 (1) Jan. (2012) main purpose of this paper is to propose the possibility that the presence of shared erroneous knowledge influences the level of trust that novice users develop toward knowledge sharing types of expert locator systems. Based on the literature, accurate knowledge is expected to have a positive influence on the amount of trust that novices develop toward knowledge sharing types of expert locator systems. However, what if the information is not generated by the system, but by another user? The proposition presented here deals specifically with novice trust levels directed at knowledge sharing types of expert locator systems rather than toward other users of those systems. If this research proposition holds true, the ramifications are significant. First and most importantly, it would suggest that the accuracy of information presented by expert users is critical to the long-term acceptance of the knowledge management system by the novice users. An implementation of such a system in an environment where there are no, true expert users may be counterproductive, causing widespread reluctance by the novices to continue using it. ... In some instances, delaying implementation of such a knowledge management system may be warranted, until a sufficient number of experts can be accurately identified. Such a delay can be significant in terms of strategic impact and other business factors. Second, it may suggest that more research may be warranted toward innovations that can assist in the identification of experts that are more trustworthy than others, at least in terms of information accuracy. Additional research in this area may lead to a more effective way to identify experts such that knowledge management system implementation can be more strategically timed. This, in turn, may also result in the ability to more accurately identify individuals claiming to be experts that are, in actuality, novices in terms of the information they present. Lastly, it may be possible to identify possible gaps in an organization's body of knowledge based on continued usage of a knowledge management system by novices. Although such a possibility may require a large amount of data to be effective, the ability to identify an area of knowledge that is lacking is critical to strategic analysis and may well be worth the expense.

Acknowledgement
An earlier version of this paper was originally presented at the Southern Association for Information Systems Conference in Richmond, VA, USA March 13'"-15*, 2008.

References
1. Alvesson M. and Karreman D., Odd couple: making sense of the curious concept of knowledge management. Journal of Management Studies, 38, 995-1018 (2001) 2. Anderson K.C. and Lelnhardt G., Maps as Representations: Expert Novice Comparison of Projection Understanding, Cognition

Conclusion
The. accuracy of, knowledge shared by experts through an expert locator system will be positively related to trust development by novices toward that system itself The

(12)

Advances In Management
& Instruciion, 20 (3), 283 (2002) 3. Barber B., The Logic and Limits of Trust, New Brunswick, NJ, Rutgers University Press (1983) . 4. Blau P., Exchange and power in social life. New York, Wiley and Sons (1964) 5. Bock G. and Young-Gul K., Breaking the myths of rewards: An exploratory study of attitudes about knowledge sharing. Information Resources Management Journat, 15, 14-21 (2002) , 6. Bond G.D. and Lee A.Y., The darkest side of trust, validating the generalized communication suspicion scale with prison inmates, ^Personality & Individual Differences, 38 (6), 1429 (2005) . 7. Bonsall P.W. and Parry T., Using an interactive route-choice simulator to investigate driver's compliance with rout choice advice. Transportation Research Record, 1306, 59-68 (1991) 8. Bos N. et al. Effects of four computer-mediated communications channels on trust development, in conference on Human Factors in_ Computing Systems (CHI), Minneapolis MN, ACM Press (2002) 9. Feng J. et al, Empathy and online interpersonal trust: A fragile relationship. Behavior & Information Technology, 23 (2), 97 (2004) 10. Friedman B., Kahn P.H. and Howe D.C., Trust Online, Communications ofthe ACM, 43 (12),.34 (2000) 11. Hiltz S.R. and Turolf M., What Makes Leaming Networks Effective? Communications ofthe ACM, 45 (4), 56 (2002) 12. Huber G., Organizational learning: The contributing processes and the literatures. Organization Science, 2 (1), 88-115 (1991) 13. Hudson B., Trust: towards conceptual clarification, Australian Journal of Political Science, 39 (1), 75 (2004) 14. Jiun-Yin J., Bisantz A.M. and Drury CG., Foundations for an Empirically Determined Scale of Trust in Automated System, International Journal of Cognitive Ergonomics, 4 (1), 53 (2000) 15. Johns J.L., A concept analysis of trust. Journal of Advanced g, 24 (1), 76 (1996) 16. Kantowitz B.H. and Hanowski R.J., Driyer acceptance of unreliable traffic information in familiar and unfamiliar settings. Human Factors, 39 (2), 164 (1997) 17. Komiak S.Y.X. and Benbasat I., The Effects Of Personalization And Familiarity On Trust And Adoption Of Recornmendation Agents, MIS Quarterly, 30 (4), 941 (2006) 18. Lerch F.J. and Prietula M.J., How do we trust machine advice?, in Designing and Using Human-Computer Interface and Knowledge Based Systems, Salvendy G. and Smith M.J., Editors, Elsevier Science Publishers, North-Holland, 410-419 (1989) 19. Meyerson D., Weick K.E. and Kramer R.M., Swift trust and temporary systems. Trust in Organizations, 166-195 (1996) 20. Mitroff 1. et al. The unbounded mind: Breaking the chains of traditional business thinking, Oxford University Press, (1993) 21. Morgan R. and Hunt S., The commitment-trust theory of relationship marketing. Journal of Marketing, 58 (July), 20 (1994) 22. Muir B.M. and Moray' N., Trust in automation: Part II, Experimental studies of trust and human intervention in a process

Vol. 5 (1) Jan. (2012)

control simulation. Ergonomics, 39, 429-460 (1996) 23. Mullin A., Trust, Social Norms and Motherhood, Journat of Social Philosophy, 36 (3), 316 (2005) 24. Nonaka L, A dynamic theory of organizational knowledge creation. Organizational Science, 5 (1), 14-37 (1994) 25. Nonaka I. and Takeuchi H., The knowledge creating company. New York, Oxford University Press (1995) 26. Owen J.M., Knowledge management and the information professional. Information Services & Use, 19 (1), 7 (1999) 27. Paiadice D. et al. Organizational knowledge management. Information Resources Management Journal, 2 (3), (1989) 28. Rempel K.J. et al. Trust in close relationships. Journal of Personality and Social Psychology, 49 (1), 95-112 (1985) 29. Rocco E., Trust breaks down in electronic contexts but can be repaired by some initial face-to-face contact, in Conference on Human Factors in Computing Systems'(CHI), ACM Press (1998) 30. Ross C.E., Pribesh S. and Mirowsky J., Powerlessness And The Aniplification Of Threat; Neighborhood Disadvantage, Disorder and Mistrusti American Sociological Review, 66 (4), 568 (2001) 31. Rotenberg K.J. and Sullivan C , Children's Use of Gaze, and Limb Movement Cues to Infer Deception, Journal of Genetic Psychology, 164 (2), 175 (2003) ' 32. Rotter J.B., A new scale for the measurement of interpersonal trust. Journal of Personality. 35 (4), (1967) 33. Rowley J., What is information? Information Services & Use, 18 (4), 243 (1998) 34. Smoliar S., Interaction management: The next (and necessary) step beyond knowledge management. Business Process Management Journal, 9. 33il (2003) 35. Stylianou D.A. and Silver E.A., The Role of Visual Representations in Advanced Mathematical Problem Solving: An Examination of Expert-Novice Similarities and Differences, Mathematical Thinking & Learning, 6 (4), 353 (2004)' 36. Tiwana A. and Bush A., A social exchange architecture for distributed web communities, .Journal of Knowledge Management, 5, 242-248 (2001) 37. Trumpower D.L. et al. Goal specificity and knowledge acquisition in statistics problem solving: Evidence for attentional focus. Memory & Cognition, 32 (8), 1379 (2004) 38. Watson S. and Hewett K., A multi-theoretical model of knowledge transfer in organizations:. Determinants of knowledge contribution and knowledge reuse. Journal of Management Studies. 43(2), 141-173(2006) 39. Zheng J. et al. Trust without touch: jumpstarting long-distance trust with initial social activities, in Conference on Human Factors in Computing Systems, Minneapolis MN, ACM Press (2002) 40. Zimmermann H.J., Knowledge Management, Knowledge Discovery and Dynamic Intelligent Data Mining, Cybernetics & Systems, 37 (6), 509 (2006).

(Received 8* August 2011, accepted 10'" December 2011)

(13)

Copyright of Advances in Management is the property of Advances in Management 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.