Electronic Commerce Research and Applications 9 (2010) 537–548

Contents lists available at ScienceDirect

Electronic Commerce Research and Applications

journal homepage: www.elsevier .com/locate /ecra

The importance of language familiarity in global business e-negotiation

Hsiangchu Lai a, Wan-Jung Lin a,*, Gregory E. Kersten b

a Department of Information Management, National Sun Yat-sen University, No. 70, Lienhai Rd., Kaohsiung 80424, Taiwan, ROCb Department of Decision Sciences and MIS, John Molson School of Business, Concordia University, 1455 De Maisonneuve Blvd. West, Montreal, Quebec, Canada, H3G 1M8

a r t i c l e i n f o a b s t r a c t

Article history:Received 23 July 2009Received in revised form 25 March 2010Accepted 25 March 2010Available online 2 April 2010

Keywords:E-negotiationLanguage familiarityLinguistic distanceMedia richnessOnline dispute resolution

1567-4223/$ - see front matter � 2010 Elsevier B.V. Adoi:10.1016/j.elerap.2010.03.005

* Corresponding author. Tel.: +886 7 5254776; fax:E-mail addresses: hclai@mail.nsysu.edu.tw (H. Lai),

edu.tw (W.-J. Lin), gregory@jmsb.concordia.ca (G.E. K

This study explores the influence of language familiarity on online persuasion behavior based on subjec-tive measurements and objective actual negotiation behavior. It was designed to test whether negotiatingin a non-native language decreases the negotiation self-efficacy, given that the increasing use of globale-marketplaces and popularity of international business trades make negotiation in a non-native lan-guage inevitable. An online experiment was conducted using a text-based asynchronous e-negotiationsystem, with two groups of subjects negotiating in native and non-native languages separately in pur-chasing negotiations. The analysis results show that language familiarity plays a critical role in inducingpersuasion behavior in e-negotiations, with a higher language familiarity leading to higher language self-efficacy and negotiation self-efficacy. However, only negotiation self-efficacy affects e-negotiation com-munication efficiency and effectiveness, both of which increase online persuasion behavior. Based onactual negotiation behavior, the results show that non-native language negotiators are less active thannative language negotiators in negotiations. However, the negotiation outcome did not differ significantlybetween the two groups, suggesting that the final outcome is also influenced by other factors. The resultsalso showed that language familiarity has a greater effect on the buyer than on the seller.

� 2010 Elsevier B.V. All rights reserved.

1. Introduction

The movement of the world into a global e-business age due totechnological and societal shifts has made international onlinenegotiation inevitable. E-negotiation is the negotiation process ofinformation exchange via electronic media (Bichler et al. 2003). Itcould be in the types of negotiation support systems, online disputeresolution, complaint handling, case appraisal, online mediation,online arbitration, etc. Tyler and Bretherton (2003) investigated76 online dispute resolution sites, including: ADRonline and Com-plain.com.au in Australia; Nova Forum and eResolution in Canada;Cybercourt, ECODIR, e-Mediater, and e-Settle.co.uk in Europe; andeNeutral, Cybersettle, JAMS, Mediation Arbitration Resolution Ser-vices (MARS), Online Resolution, WebMediate, SquareTrade, andSmartSettle in the USA. These resolution sites demonstrate the highdemand for online services that facilitate conflict resolution, sincethey avoid time-consuming and costly business travel, and areefficient.

The president and CEO of SquareTrade pointed out that Square-Trade’s surveys indicate that the access and commitment to onlinedispute resolution is a key factor influencing consumer willingnessto buy or bid on higher priced goods on eBay (Abernethy 2003).

ll rights reserved.

+886 7 5254799.d934020008@student.nsysu.

ersten).

Furthermore, after participating in SquareTrade’s dispute resolu-tion process, over 80% of buyers said that they would buy againfrom the marketplace or seller. Abernethy (2003) also reportedthat, as of mid-2003, SquareTrade handled over 50,000 new dis-putes a month and concurrently managed over 20,000 active cases.These statistics indicate that implementing online dispute resolu-tion using websites such as SquareTrade is a popular method forassisting consumers who experience problems related to onlinetransactions. Such problems often occur between participants indifferent cities or countries and in marketplaces where sellers areprimarily small and medium-sized businesses. Although Square-Trade discontinued the dispute resolution service it provided foreBay in June 2008 due to eBay stopping the submission of feedbackwithdrawal notices, its past success indicates the high demand foronline negotiation services. PayPal’s Resolution Center has re-placed SquareTrade in providing online dispute resolution for eBay.Different from the PayPal system, the Smartsettle system is a gen-eric tool for decision-makers with conflicting objectives, who wishto reach a formal agreement. This has led to Smartsettle beingadopted in e-commerce and other applications, such as for resolv-ing family issues, insurance issues, supply agreements, businessmergers, and acquisitions.

The inevitability of international online negotiations brings thechallenge of language differences, with negotiators from differentcountries possibly speaking different native languages. This canresult in either one side or both sides having to use a non-native

538 H. Lai et al. / Electronic Commerce Research and Applications 9 (2010) 537–548

language in a negotiation. However, clear and accurate communi-cation is the precondition for high-quality integrative negotiation(Lewicki et al. 2006). Since people generally find it easier to expressthemselves in their native language than in a non-native language,negotiating in a non-native language may decrease the communi-cation ability and hence the negotiation performance. Hutchinson(2005) showed that a gap in language understanding can signifi-cantly hinder business development by reducing the ability oftraders to develop cooperative relationships. The reason might befairly obvious that language is vital in the case of transaction nego-tiation. Stanley et al. (1990) pointed out that customers like to betalked to in their own language. Being a seller, talking to customersin customers’ own language will reflect in greater sales volumesbecause of the quicker user acceptance, repeat orders and refer-ence sales. It will also decrease vendors’ training costs, error costand time wasted on non-productive tasks (Stanley et al. 1990).The above arguments indicate that language ability affects thetransaction performance for both buyer and seller. Thus, sincenegotiating in non-native languages via the Internet is becomingmore common in the global e-business age, it is important tounderstand how language familiarity impacts negotiation behaviorand performance.

Negotiation is a form of decision-making involving two or moreagents who cannot make decisions independently, and hence con-cessions must be made to reach a compromise agreement (Kerstenet al. 1991). Persuasion is commonly used by every negotiator toachieve an agreement that is closer to his or her expectations.However, persuasion ability varies with communication ability,and language plays an essential role in communication. Negotiat-ing in a more familiar language makes it easier for negotiators toexpress themselves and understand their counterparts. Althoughmany studies have investigated the effects of persuasion behavioron the likelihood of reaching negotiated agreements and theirquality (Keough 1992, Pruitt and Carnevale 1993, De Dreu et al.2000), none of them have explored how language ability influencespersuasion behavior and performance.

Persuasion ability is influenced by both the actual words in amessage and non-verbal communications such as eye movements,facial expressions, gestures, spatial relations, and visual displays,which can add information, emphasis, subtlety, and rapport (Carrelland Heavrin 2007). The availability of non-verbal communicationimproves the media richness. In general, media richness is lowerwhen negotiating online than in the face-to-face situation (Daftet al. 1987). Since language is the fundamental tool of commu-nication and online communication media has less richness, theimportance of verbal language ability in online persuasion behav-ior increases in an e-negotiation context. In other words, if anegotiator has to use a non-native language in e-negotiations, hisor her language ability becomes a critical factor in the negotiationbehavior and performance.

International e-negotiation will continue to increase, whichmakes it worthwhile to understand how language familiarity influ-ences the online persuasion behavior of a negotiator. If languagefamiliarity affects the persuasion behavior, what are the interven-ing variables? Since language is the fundamental tool in negotia-tion communication, is the impact related to communicationefficiency and communication effectiveness, and how are thesefactors related? To answer these questions, we conducted anexperiment to investigate how language familiarity impacts per-suasion behavior in e-negotiations. E-negotiation communicationscan now be performed using text, audio, and video, and these dif-ferent media have different degrees of richness. Considering timedifferences across the world and the facilities required to imple-ment e-negotiation, it is generally easier to adopt text-based asyn-chronous e-negotiation than text-based or audio- and video-basedsynchronous e-negotiation. In fact, most of the conflict resolution

sites listed above employ text-based asynchronous systems. Thisstudy, therefore, chose text-based asynchronous e-negotiation asthe research context. In addition, to elucidate the underlying cog-nitive processes, we also investigated the impact of languagefamiliarity on objective negotiation behavior, such as the numbersof offers and messages, and the outcome performance.

This paper is organized as follows. Section 2 discusses the the-oretical background and develops a series of hypotheses. Section 3describes the research methodology including the experimentaldesign and the questionnaire design. Section 4 and 5 present thedata analysis and discussions. Finally, conclusions are drawn inSection 6.

2. Theory and hypotheses

2.1. Language familiarity and self-efficacy

Language plays an essential role as the most important com-munication tool in our daily lives, and it has been evolvingcontinuously since ancient times. The thousands of different hu-man languages and dialects that are used to deliver messagesand to express thoughts and feelings have originated and devel-oped in different cultures, regions, and countries. People areusually more comfortable in expressing themselves and under-standing others when they use their own native language (alsoreferred to as the mother language, first language, or mothertongue) in contrast to a non-native language (a foreign or secondlanguage). Prior research has pointed out that people often feelanxious when they use a non-native language (Horwitz et al.1986, MacIntyre and Gardner 1991). Garcez (1992) investigatedthe negotiation behaviors of non-native and native English-speaking negotiators, and found that the style of a nativelanguage-speaking negotiator is more direct, locally coherent,and reiterated, whereas that of a non-native-speaking negotiatoris more indirect, summatively coherent, and conversational.These findings indicate that the use of a non-native languagenot only changes people’s mental state but also affects theircommunication performance and behavior.

Compared with the traditional face-to-face situation, onlinecommunication has less media richness and a lower social pres-ence, which changes how users communicate (Short et al. 1976,Daft and Lengel 1983). Some studies have shown that the ab-sence of non-verbal clues in the online environment increasessocial distance and the prevalence of task-oriented communica-tion relative to social-oriented communication. Moreover, com-petitiveness and tough attitudes are more likely in an onlinecontext (Burgoon and Hale 1987, Walther 1996, Purdy et al.2000, Thompson and Nadler 2002, Giordano et al. 2007). How-ever, the need for efficient business processes in a competitivebusiness environment makes it inevitable that negotiations willbe conducted online. Since online communication media haveless media richness, language ability is becoming more criticalto conducting successful online negotiations.

Bandura (1986, p. 94) defines self-efficacy as ‘‘. . .people’s judg-ments of their capabilities to organize and execute courses of action re-quired to attain designated types of performances. It is concerned notwith the skills one has but with judgments of what one can do withwhatever skills one possesses”. He pointed out that self-efficacy isa very important construct in social psychology because it refersto an individual’s assessment of his or her capabilities to organizeand perform a specific task. Self-efficacy is a broader construct thatcan be measured in various activities and situations (Bandura1986). Over the past few decades, its effects on a highly diverserange of human behaviors have been considered, including salesperformance (Barling and Beattie 1983), career development(Hackett and Betz 1981, Clement 1987), education (Shell et al.

H. Lai et al. / Electronic Commerce Research and Applications 9 (2010) 537–548 539

1989), research productivity (Taylor et al. 1984), and computerusage (Compeau and Higgins 1995, Marakas et al. 1998). Regard-less of the various definitions coming from the numerous studies,self-efficacy has two clear characteristics: (1) it is a strong predic-tor of the subsequent task-specific performance, and (2) it ulti-mately refers to what a people perceive their capabilities to bewith regard to a specific task (Marakas et al. 1998).

Self-efficacy can be influenced by many factors, including per-formance accomplishments, vicarious experience, verbal persua-sion, and emotional arousal (Bandura 1986). In the negotiationfield, O’Connor and Arnold (2001) found that the self-efficacy ofnegotiators increases with their experience of successful negotia-tions, and decreases when negotiators have experienced failures.Morin and Latham (2000) also confirmed that receiving trainingin interpersonal communication skills significantly increased theself-efficacy of individuals. In other words, a person’s perceptionof his or her ability to perform a particular task can be influencedby past performance and experiences.

Most people do not have sufficient experience of using a non-native language in their daily lives and therefore will have less con-fidence when using a non-native language than when using theirnative language. In other words, for a specific language, languageself-efficacy will be higher for a native language speaker than fora non-native language speaker. Since negotiation involves two-way communication aimed at reaching a joint agreement that sat-isfies different needs or ideas (Acuff 2008), and language plays akey role in communication, negotiators generally consider thatthey are more able to negotiate in their native language. In otherwords, a negotiator with a higher language self-efficacy will havea higher negotiation self-efficacy. For these reasons, we proposethe following hypotheses:

� H1a (Effect of language usage on language self-efficacy): Anegotiator with higher language familiarity will have higher lan-guage self-efficacy in an e-negotiation.� H1b (Effect of language usage on negotiation self-efficacy): A

negotiator with higher language familiarity will have higher nego-tiation self-efficacy in an e-negotiation.� H2 (Effect of language self-efficacy on negotiation self-effi-

cacy): In an e-negotiation, negotiators with higher language self-efficacy will have higher negotiation self-efficacy.

2.2. Self-efficacy and e-negotiation communication performance

Communication is the major activity in negotiation, with thenegotiation result being highly dependent on communicationperformance. Similar to the performance evaluations in many areas,the common constructs used to evaluate communication perfor-mance are efficiency and effectiveness, which are colloquiallydefined as ‘‘do the thing right” and ‘‘do the right thing”. Cleary(2004) formally defined communication efficiency as the time andmoney cost involved in sending a message, and communicationeffectiveness as the degree of correspondence between an idea ormessage initiated and intended by the sender and the messageperceived and responded to by the receiver. Yuan et al. (2003) inves-tigated e-negotiation communication efficiency in terms of the easeof use, information clarity, and response speed of the negotiationsystem, and e-negotiation communication effectiveness in termsof its expression power, memorability, and comfort. Lower languageself-efficacy can hinder the expression power and memorability ofexchanged messages. Therefore, the present study explores the ef-fects of using native and non-native languages on the communica-tion efficiency and effectiveness, where communication efficiencyrefers to the speed of and effort made responding and expressingone’s own feelings and opinions or understanding the counterpart’sfeelings and opinions, and communication effectiveness refers to

the accuracy of expressing one’s feelings and opinions or under-standing the counterpart’s feelings and opinions.

Self-efficacy is defined as the extent to which a person believesthat his or her capabilities to organize and execute courses of actionare sufficient to attain a designated performance (Bandura 1986,Compeau and Higgins 1995). In other words, perceived self-efficacyis a type of forethought that critically influences future intentions.Obviously, the lower language self-efficacy of non-native speakersresulting from them having less experience in using the languagemakes it more difficult for them to express information clearly andto reply rapidly. Lewicki and Litterer (1985) also claimed that peoplecan improve the accuracy of communication by questioning, activelistening and reflecting, and role reversal. That is what the commu-nication effectiveness means. The questioning and active listeningand reflecting abilities are obviously related to language ability.Although those authors only claimed that language and communica-tion effectiveness were related, we infer from their results that peo-ple with higher language self-efficacy can have more confidence andintentions to question and clarify what they have listened to, whichwill thereby improve their communication performance.

Language is the fundamental communication tool in negotia-tions, and it is even more critical in the e-negotiation context.Especially in text-based asynchronous e-negotiation, which lacksnon-verbal cues, negotiators with higher language self-efficacy willhave more confidence in their communication performance (i.e.,their communication efficiency and effectiveness). Similarly, be-cause e-communication is a major activity of e-negotiation, nego-tiators with higher negotiation self-efficacy are more likely tocommunicate efficiently and effectively during the online negotia-tion process. Therefore, we propose the following hypotheses:

� H3a (Effect of language self-efficacy on e-negotiation com-munication efficiency): In an e-negotiation, negotiators withhigher language self-efficacy will have higher communicationefficiency.� H3b (Effect of negotiation self-efficacy on e-negotiation

communication efficiency): In an e-negotiation, negotiatorswith higher negotiation self-efficacy will have higher communica-tion efficiency.� H4a (Effect of language self-efficacy on e-negotiation com-

munication effectiveness): In an e-negotiation, negotiators withhigher language self-efficacy will have higher communicationeffectiveness.� H4b (Effect of negotiation self-efficacy on e-negotiation

communication effectiveness): In an e-negotiation, negotiatorswith higher negotiation self-efficacy will have higher communica-tion effectiveness.

Lewicki and Litterer (1985) claimed that an absence of feedbackcan hinder the communication accuracy. Butler (1999) also arguedthat the procedural efficiency of a negotiation process is signifi-cantly affected by the performance of the task. Thus, it can be con-cluded that higher e-negotiation communication efficiency resultsin higher e-negotiation communication effectiveness. Therefore,we propose the following hypothesis:

� H5 (Effect of e-negotiation communication efficiency one-negotiation communication effectiveness): In an e-negotia-tion, negotiators with higher communication efficiency will havehigher communication effectiveness.

2.3. Online communication performance and online persuasionbehavior

Persuasion behavior lies at the heart of negotiation (Sycara1990). Lewicki and Litterer (1985) defined persuasion as efforts

540 H. Lai et al. / Electronic Commerce Research and Applications 9 (2010) 537–548

designed to influence another person’s positions, perceptions, opin-ions, and attitudes. The goal of persuasion is to create leverage, unitethe behavior of the parties, and guide the process toward solutionconvergence (Sycara 1990). This requires negotiators to negotiateusing the correct arguments at the right time in order to convincetheir counterparts that what they offer is both valuable and reason-able. If negotiators perceive that they can communicate online moreefficiently and more effectively, this might stimulate them to per-form more online persuasion behavior. Therefore, we propose:

� H6 (Effect of e-negotiation communication efficiency ononline persuasion behavior): In an e-negotiation, negotiatorswith higher communication efficiency will exhibit more persuasionbehavior.� H7 (Effect of e-negotiation communication effectiveness on

online persuasion behavior): In an e-negotiation, negotiatorswith higher communication effectiveness will exhibit more persua-sion behavior.

2.4. Research model

Since persuasion is crucial to a successful negotiation (Sycara1990), it is worthwhile to elucidate how language familiarity influ-ences online persuasion behavior, and this forms the main purposeof the present study. The research model based on Hypotheses 1–7is shown in Fig. 1. The first concern is whether language familiarityaffects language self-efficacy and negotiation self-efficacy. Thesecond is whether higher language and negotiation self-efficaciesof negotiators result in higher communication efficiency andeffectiveness in e-negotiations. The third is whether the highere-negotiation communication efficiency and effectiveness inducenegotiators to perform more online persuasion behavior. In sum-mary, we will investigate how language familiarity affects onlinepersuasion behavior by examining language self-efficacy, negotia-tion self-efficacy, communication efficiency, and communicationeffectiveness in e-negotiations.

3. Research methodology

3.1. Experimental design and data collection

The following two negotiation language groups were designedin order to manipulate the level of language familiarity: native lan-guage (Chinese) and non-native language (English). In order toavoid interference from cultural differences, all subjects werenative Chinese speakers in Taiwan. There were two reasons forchoosing English as the non-native language. First, English is oneof the major languages commonly used in international communi-cation. It is very common for people from different countries with alarge linguistic distance to communicate in English. This choicetherefore increases the external validity of the research findings.Second, almost all of our subjects choose English as their first

Language familiarity♦ Native language ♦ Non-native language

Languageself-efficacy

Negotiationself-efficacy

H1a (Native > Non-native)

H2

H3a

H3b

H4a

H4b

H1b (Native > Non-native)

Fig. 1. Researc

foreign language, and so it was the best choice for implementingour experiment.

The negotiation scenario, role-play, and utility design were thesame in both groups – the only difference was the language used innegotiations. The negotiation scenario was a tire trade involving abuyer and a seller. The issues comprised the unit price, paymentterms, and delivery time, each of which had four options with dif-ferent utilities. The final utility of each subject depended on the le-vel of agreement on each issue (Table 1). If the subjects wereunable to reach agreement, they were assigned the utility levelof BATNA (best alternative to a negotiated agreement).

ANeGo, a Web-based negotiation system, was developed forthis e-negotiation experiment. This system allows negotiators tonegotiate with their counterparts asynchronously (see Fig. 2a andFig. 2b). The native language group was asked to negotiate in writ-ten Chinese, while the non-native language group was only al-lowed to negotiate in written English. An ‘‘English-only” functionwas designed in the English version of ANeGo, which ensured thatthe entire negotiation process was conducted in English only.

The subjects were recruited online. Because the e-negotiation inthe experiment was designed to be undertaken for 9 days and thesubjects had to be paired before the experiment started, the sub-jects first had to commit to participating in the experiment by reg-istering. All of the registered subjects were then paired into theroles of buyers and sellers. They were informed to start negotiationwith their counterpart on a particular day, using ANeGo to negoti-ate anytime and anywhere for 9 days. In order to encourage sub-jects to participate seriously in the experiment, we designed anincentive mechanism based on both the final utility and luck ofthe negotiators. Only subjects who achieved a sufficiently highutility were candidates to receive substantial prizes (e.g., an iPodNano and iPod Shuffle). After the winners of these prizes were con-firmed, the other subjects were candidates for the remaining prizessuch as convenience store coupons and memory-card readers. Allof the prize winners were decided by lottery. This two-stage incen-tive mechanism ensured that everyone had a chance to get a prizeas long as they finished the experiment, but only good performershad a chance to get a substantial prize.

3.2. Questionnaire design

The questionnaire included the following five constructs basedon the research model and hypotheses: language self-efficacy,negotiation self-efficacy, e-negotiation communication efficiency,e-negotiation communication effectiveness, and online persuasionbehavior. All of the construct items were measured on a seven-point Likert scale. In order to ensure the face validity and contentvalidity of the measurements, we initially developed the itemsbased on previous validated literature. However, the fit of the mea-surements to the specific context of our research model was im-proved by revising some of the items and creating some newitems (see Appendix).

E-negotiation communication efficiency

(efficiency)

E-negotiationcommunication effectiveness

(effectiveness)

Online persuasion behavior

H5

H6

H7

h model.

Table 1Utility table.

Issue Options Buyer Seller

Unit price NT$ 4850 2100 600NT$ 5000 1800 800NT$ 5150 1500 1100NT$ 5300 1200 1320

Payment terms Pay on delivery 800 2100Pay 3 weeks after delivery 950 1800Pay 6 weeks after delivery 1100 1500Pay 9 weeks after delivery 1250 1200

Delivery time Deliver 1 week after order 1320 800Deliver 2 weeks after order 1100 950Deliver 3 weeks after order 800 1100Deliver 4 weeks after order 660 1250

BATNA (best alternative to a negotiated agreement) 3550 3490

H. Lai et al. / Electronic Commerce Research and Applications 9 (2010) 537–548 541

Taking language and negotiation self-efficacy constructs asexamples, the wide application area of self-efficacy resulted inhighly diverse instrument types. Bandura (1986) pointed out thatthe overall dimensions of a self-efficacy measurement could bemagnitude, strength, and generalizability. However, Lee and Bobko(1994) reported that there are five main methods of measuring self-efficacy. The most popular method is to measure the strength only,while the second one is to adopt magnitude measures only. Theremaining methods are the combination strength and magnitude

Fig. 2a. Chinese platform for the native language

Fig. 2b. English platform for the non-native l

measurements, and even a single-item measurement (Lee andBobko 1994). Such a wide diversity of measurements was also foundthrough a meta-research study in the field of computer self-effi-cacy. Marakas et al. (1998) found that almost 50% of classifiedstudies (19 out of 40) employed self-developed self-efficacy mea-surements, with the remaining studies using existing instrumentswith some modifications. Due to the lack of closely related studies,in the present study we referred to a few of the previous studies andalso developed some measurements based on the most popularstrength measurements.

This study used a total of 23 items. Subjects were asked to ratetheir strength of agreement on a seven-point Likert scale. Further-more, in order to avoid problems of misunderstanding or incorrecttranslation, the developed questionnaire was presented in Chineseto both the native language and non-native language groups. Allitems in language self-efficacy and negotiation self-efficacy con-structs had to be completed just prior to starting the e-negotiation,while the items in the other constructs had to be completed afterfinishing the experimental e-negotiation. To facilitate this, thequestionnaire was completed online.

4. Data analysis and results

4.1. Profiles of subjects

Initially 180 volunteer subjects were recruited, of which 125subjects finished the negotiation and completed the questionnaire.

group: read counterpart’s message and offer.

anguage group: send message and offer.

542 H. Lai et al. / Electronic Commerce Research and Applications 9 (2010) 537–548

As mentioned earlier, the subjects had to register first before wearranged them into pairs of buyers and sellers. After all registeredsubjects had been paired, they were told to start the e-negotiationon a specific day for 9 days. This long experimental period requireda significant commitment from the subjects, and although the de-sign of the e-negotiations allowed the subjects to negotiate any-time and anywhere, 45 subjects dropped out for the followingpossible reasons: (1) registered for the experiment but did notshow up during the e-negotiation period, (2) did not finish theexperimental process completely, and (3) finished the entireexperimental process but did not complete the questionnaire cor-rectly (e.g., answers to all items were the same). The 45 subjectswho dropped out comprised 19 people from the native languagegroup and 26 from the non-native language group. It seems thatthere was a higher dropout rate and higher proposition of em-ployed subjects in non-native language group. It could be becausethere were more students participating in the native languageexperiment, which resulted in fewer students participating in thenon-native language experiment, which was held after the nativelanguage experiment. Another reason might be that the non-nativespeaker language experiment was held during the end of thesemester.

The overall profiles of the final group of 125 subjects are pre-sented in Table 2. They comprised 45 females and 80 males, with73 subjects in the native language group and 52 in the non-nativelanguage group. The mean overall age was 29.1 years. 65 subjectswere graduate students or senior undergraduates, while 60 sub-jects were employed. In general, the native language group submit-ted more offers and messages and get higher final utility than didthe non-native language group. It seems that non-native languagenegotiators were less active in negotiations than native languagenegotiators due to the limited ability of using non-native language. Onthe other hand, negotiators with higher language familiarity weremore active.

To ensure that the substantial difference in occupation distribu-tion between the native language and non-native language groupsdid not reduce the validity of the analysis results, we compared thecommunication behavior and outcome between two occupationtypes: student and employee. The results are presented in Table 3.In the internal analysis we examined the difference betweentwo occupations within the same group, and in the external anal-ysis we examined the difference for the same occupation type be-tween the native language and non-native language groups. Theresults reveal that the only difference between the students andemployees in the internal and external analyses was a marginallysignificant difference in the number of messages submitted by stu-dents between the native language and non-native languagegroups. In other words, the students and employees exhibited sim-ilar negotiation behaviors, which might have been due to most ofthe student subjects being graduate students, and hence being suf-ficiently mature. This finding demonstrates that the difference inthe distribution of occupation status between the native languageand non-native language groups did not adversely affect the valid-ity of the analysis results.

Table 2Profiles of subjects.

Native group (Chinese)

Gender Female 24 32.88%Male 49 67.12%

Occupation Student 60 82.19%Employee 13 17.81%

Mean age (years) 26.3Number of offers 3.01Number of messages 3.95Utility 3697.67

4.2. Measurement model

We applied exploratory factor analysis (EFA) to examine thevalidity and reliability of the measurements. Table 4 presents theresults of the EFA as determined by SPSS, which reveals that allitems had higher loadings on their own construct than on otherconstructs. This confirms conformity with a simple structure solu-tion, with each variable having a single high loading on only onefactor (Hair et al. 1998). The result also indicates the capabilityto discriminate between constructs.

Table 5 lists the values of Cronbach’s a (Cronbach 1951), whichwere used to assess the reliability of all scale items of the con-structs. According to Kerlinger and Pedhazur (1973), the reliabilityis sufficiently high when Cronbach’s a is higher than 0.7, and henceour results indicate that all of the items and constructs were suffi-ciently reliable in our data analysis.

4.3. Hypothesis testing

4.3.1. Effects of language familiarity on language and negotiationself-efficacies

H1a and H1b were examined using the independent-samplest-test. The results of t-tests used to assess the significance of differ-ences between the two groups are summarized in Table 6, whichreveals that language self-efficacy (t = 6.72, p < 0.001) and negotia-tion self-efficacy (t = 5.31, p < 0.001) were significantly higher inthe native language group (6.05 and 6.09, respectively) than inthe non-native language group (4.52 and 4.79, respectively). Thisconfirms that negotiators exhibit higher language self-efficacyand negotiation self-efficacy when they negotiate in their nativelanguage. Therefore, H1a and H1b are supported by our researchdata.

4.3.2. Effects of language and negotiation self-efficacies ande-negotiation communication performance on online persuasionbehavior

We examined the mean difference of the dependent variable foreach hypothesis by using quintile analysis to transform the inde-pendent variables into categorical variables. We first ranked theobservations of each independent variable into order, and thenwe removed all of the observations within the middle 20% (thirdquintile group) before categorizing the top 40% and bottom 40%into the HIGH and LOW value groups, respectively. We removedthe third quintile group when forming the LOW and HIGH catego-ries in order to remove data that were marginally high and low.Moreover, the use of only two groups allowed the t-test rather thanANOVA to be used to examine H2–H7. Table 7 summarizes the re-sults for group differences, which reveal that H2, H3b, H4b, H5, H6,and H7 were supported, with only H3a and H4a not beingsupported.

The analysis results indicated that a negotiator with higher lan-guage self-efficacy has higher negotiation self-efficacy (supportingH2). In turn, a negotiator with higher negotiation self-efficacy hashigher e-negotiation communication efficiency (supporting H3b)

Non-native group (English) Total

21 40.38% 45 36.00%31 59.62% 80 64.00%5 9.62% 65 52.00%47 90.38% 60 48.00%33.1 29.12.13 2.652.63 3.403672.31 3687.12

Table 3Descriptive statistics: students vs. employees.

Performance Occupation Native group (Chinese) Non-native group (English) Significance (external)b

Number of offers Student 3.10 2.20 0.280Employee 2.62 2.13 0.169Significance (internal)a 0.295 0.885

Number of messages Student 3.97 2.20 0.090*

Employee 3.85 2.68 0.212Significance (internal)a 0.880 0.562

Utility Student 3694.5 3582.0 0.545Employee 3712.3 3681.9 0.774Significance (internal)a 0.879 0.548

* p 6 0. 1.a t-test results for the difference between students and employees within the native language or non-native language group.b t-test results for the students’ or employees’ differences between the native language and non-native language groups.

Table 5Reliability results.

Construct (Number of Items) Cronbach’s a

Language self-efficacy (4) 0.972Negotiation self-efficacy (4) 0.969E-negotiation communication efficiency (6) 0.931E-negotiation communication effectiveness (4) 0.914Online persuasion behavior (5) 0.886

Table 6Results of t-test of language and negotiation self-efficacies.

Construct Group Mean t

Language self-efficacy Native language 6.05 6.72***

Non-native language 4.52Negotiation self-efficacy Native language 6.09 5.31***

Non-native language 4.79

*** p 6 0.01.

Table 4Exploratory factor analysis (EFA) results developed using SPSS.

Language self-efficacy (LS)

Negotiation self-efficacy (NS)

E-negotiation communicationefficiency (efficiency)

E-negotiation communicationeffectiveness (effectiveness)

Online persuasionbehavior (PB)

LS1 0.894 0.283 0.095 0.021 0.085LS2 0.934 0.232 0.064 0.031 �0.027LS3 0.914 0.301 0.023 0.017 �0.014LS4 0.908 0.333 0.063 0.061 �0.003NS1 0.270 0.851 0.177 0.101 0.049NS2 0.418 0.837 0.128 0.097 0.138NS3 0.398 0.857 0.136 0.052 0.100NS4 0.359 0.857 0.147 0.060 0.152Efficiency1 0.057 0.097 0.834 0.190 0.107Efficiency2 0.072 0.207 0.866 0.168 0.132Efficiency3 0.102 0.236 0.858 0.237 0.099Efficiency4 0.134 0.310 0.798 0.195 0.164Efficiency5 �0.007 �0.173 0.585 0.339 0.279Efficiency6 �0.037 �0.047 0.636 0.452 0.302Effectiveness1 0.106 0.326 0.350 0.608 0.364Effectiveness2 0.178 0.318 0.334 0.644 0.294Effectiveness3 �0.003 0.009 0.350 0.837 0.255Effectiveness4 0.002 0.019 0.372 0.832 0.201PB1 0.014 0.114 0.042 0.293 0.750PB2 0.001 0.127 0.044 0.320 0.822PB3 0.016 0.099 0.078 0.230 0.853PB4 0.034 �0.017 0.302 0.056 0.756PB5 0.003 0.102 0.353 �0.024 0.763

H. Lai et al. / Electronic Commerce Research and Applications 9 (2010) 537–548 543

and e-negotiation communication effectiveness (supporting H4b),but this is not the case for a negotiator with higher languageself-efficacy (H3a and H4a are not supported). In addition, a nego-

tiator with higher communication efficiency also has higher com-munication effectiveness (supporting H5). Finally, a negotiatorperforms more online persuasion behavior when he or she hashigher communication efficiency (supporting H6) and effective-ness (supporting H7).

The above analysis results indicate that when people negotiatein their native language, they will be more confident in using thelanguage and in the overall negotiation. Furthermore, a highernegotiation self-efficacy but not a higher language self-efficacy willresult in higher e-negotiation communication efficiency and effec-tiveness. Either higher communication efficiency or higher com-munication effectiveness will induce more online persuasionbehavior. Negotiation self-efficacy is more critical than languageself-efficacy in influencing online persuasion behavior. This resultis consistent with Bandura’s (1989) statement that the predictivecapability of a self-efficacy estimate is stronger and more accuratewhen it is determined by specific domain-linked measures ratherthan by general measures. The present study found that althoughlanguage self-efficacy impacts negotiation self-efficacy, it doesnot significantly impact e-negotiation communication efficiencyand effectiveness. On the other hand, negotiation self-efficacy im-pacts both e-negotiation communication efficiency and effective-ness, which in turn impact online persuasion behavior.

Table 7t-Test results for Hypotheses 2–7.

Construct Group Mean t

H2 Negotiation self-efficacy Language self-efficacy: low

4.66 �5.75***

Language self-efficacy: high

6.17

H3a E-negotiationcommunication efficiency

Language self-efficacy: low

5.10 �1.50

Language self-efficacy: high

5.44

H3b E-negotiationcommunication efficiency

Negotiation self-efficacy: low

5.01 �2.71***

Negotiation self-efficacy: high

5.61

H4a E-negotiationcommunicationeffectiveness

Language self-efficacy: low

5.39 �1.47

Language self-efficacy: high

5.68

H4b E-negotiationcommunicationeffectiveness

Negotiation self-efficacy: low

5.28 �3.16***

Negotiation self-efficacy: high

5.85

H5 E-negotiationcommunicationeffectiveness

Efficiency: low 4.88 �8.34***

Efficiency: high 6.14

H6 Online persuasion behavior Efficiency: low 4.92 �5.57***

Efficiency: high 5.86H7 Online persuasion behavior effectiveness: low 4.82 �6.07***

effectiveness: high 5.85

*** p 6 0.01.

Table 9t-Tests for the differences in offers, messages and final utility between the buyer andseller groups.

Role Nativegroup(Chinese)

Non-nativegroup(English)

Significance(external)b

Number of offers Buyer 3.22 2.00 0.001***

Seller 2.68 2.26 0.241Significance(internal)a

0.195 0.378

Number of messages Buyer 4.04 2.48 0.003***

Seller 3.78 2.78 0.105Significance(internal)a

0.657 0.542

Utility Buyer 3765.11 3676.00 0.337Seller 3589.31 3668.89 0.410Significance(internal)a

0.053* 0.942

* p 6 0.1.*** p 6 0.01.

a t-Test results for the differences between buyers and sellers within the nativelanguage or non-native language group.

b t-Test results for the buyers’ or sellers’ differences between the native languageand non-native language groups.

544 H. Lai et al. / Electronic Commerce Research and Applications 9 (2010) 537–548

4.3.3. Objective behavior and outcome analysisAfter examining H1–H7 based on subjective questionnaire-

based measurements, we further objectively examined the actualbehaviors and outcomes (i.e., the numbers of submitted offersand messages) and the final utility. We first examined differencesbetween the native language and non-native language groups,and then those between buyers and sellers.

Table 8 lists the results of t-tests of differences between thenative language and non-native language groups, which indicatesignificant differences in the submission behavior for offers(p = 0.001) and messages (p = 0.001). They indicated that nativelanguage negotiators submitted more offers and messages thannon-native language negotiators, which indicates that negotiatorswith higher language familiarity are more active in negotiations.However, the final utility did not differ significantly between thetwo groups, which might be due to the final utility also being influ-enced by other factors such as the negotiator’s reservation value,BATNA, and negotiation strategy.

The results of t-tests of the differences between buyers and sell-ers are presented in Table 9, which reveals that the numbers of of-fers and messages submitted by buyers were significantly higher inthe native language group than in the non-native language group.However, this was not the case for sellers. Table 9 indicates thatnegotiating in their native languages may make buyers more activeand aggressive due to their better language efficacy. Moreover,

Table 8t-Test results for the differences in offers, messages and final utility between thenative language and non-native language groups.

Native group(Chinese)

Non-native group(English)

Significance

Number of offers 3.01 2.13 0.001***

Number ofmessages

3.95 2.63 0.001***

Utility 3697.67 3672.31 0.704

*** p 6 0.01.

Table 9 indicates that the buyer’s utility was higher than the sell-er’s utility in the native language group but not in the non-nativelanguage group. On the other hand, there was no difference inthe final utility when negotiating in a non-native language. Thebuyer generally has a higher negotiation power, and such powermight be supplemented by the higher language familiarity whennegotiating in the native language.

5. Discussion

The analysis results confirm all of the study hypotheses exceptfor H3a and H4a, which were based on cognitive measurements.This shows that there are significant differences in objective nego-tiation behavior, in terms of the numbers of submitted offers andmessages. The results further indicate that language familiarity af-fects not only the negotiators’ perceptions but also their actualnegotiation behavior. Regarding the cognitive measurements, first,the results reveal that the language familiarity influences the nego-tiators’ language self-efficacy and negotiation self-efficacy. In otherwords, if negotiators negotiate in their native language, they willhave more confidence in using the language and negotiating. Fur-thermore, a negotiator with higher language self-efficacy will havehigher negotiation self-efficacy. Second, the negotiation self-effi-cacy but not the language self-efficacy influences the communica-tion efficiency and effectiveness, with a higher communicationefficiency producing a higher communication effectiveness. Third,a higher communication efficiency and a higher communicationeffectiveness will result in more persuasion behavior.

The differences in the actual behaviors between the native lan-guage and non-native language groups are consistent with the re-sults of hypothesis testing. Native language negotiators are moreactive in negotiations than are non-native language negotiators,but the final utility does not differ significantly between them. Thiscould be due to the final utility being influenced by other factorssuch as the negotiator’s reservation value and negotiation strategy.The numbers of offers and messages submitted differed signifi-cantly between the native language and non-native languagegroups for buyers but not for sellers. The final utility differed sig-nificantly between buyers and sellers in the native language groupbut not in the non-native language group.

Four major findings can be concluded from the above analysisresults. First, language familiarity plays a critical role in e-negotia-tion persuasion behavior, with negotiators with higher language

H. Lai et al. / Electronic Commerce Research and Applications 9 (2010) 537–548 545

familiarity performing more active negotiation behavior. Second,negotiation self-efficacy rather than language self-efficacy affectsthe communication performance that can induce persuasionbehavior. Third, the final negotiation outcome is not dominatedby the language familiarity. Finally, language familiarity is morecritical to the buyer’s negotiation behavior than to that of theseller.

The analysis results have implications for both practice and aca-demic research. First, since language familiarity plays a critical rolein e-negotiation persuasion behavior, how to choose an appropriatenegotiation language is very important in international e-negotia-tions. Obviously the native language is the best option, but alterna-tively the negotiator should choose the one with the smallestlinguistic distance from his or her own native language. Linguisticdistance is a measure of how much languages differ from each other(Chiswick and Miller 2004), and has been shown to greatly influ-ence business trade. Hutchinson (2005) found that the amount oftrading by the U.S. with another country has been lower whenthe official language of that country has a larger linguistic distancefrom English. The research findings may explain this, since a lowerlanguage familiarity will reduce the communication performanceand hence also the negotiation persuasion behavior.

Second, since only the negotiation self-efficacy affects the nego-tiation communication performance, improving negotiation self-efficacy is more important than improving language self-efficacyfor inducing persuasion behavior. Possible methods for improvingnegotiation self-efficacy include training on the use of e-negotia-tion systems and providing decision support to e-negotiation(e.g., the Inspire e-negotiation system allows people to practicenegotiation online), since more experiences will provide negotia-tors with more confidence in e-negotiations. E-negotiation systemsthat teach negotiators about the ‘‘negotiation dance” or allow themto rank all possible alternative offers will give them more confi-dence in exchanging messages and offers. In fact, this can also im-prove communication efficiency and effectiveness because itmakes it easier for the negotiators to determine the negotiationstatus and compare the alternatives, and therefore to submit mes-sages and offers correctly and quickly.

Third, in addition to improving language familiarity, it is alsoimportant to pay attention to other factors that have an impacton the negotiation outcome. Since the negotiation outcome resultsfrom the dynamic interactions between the negotiators, it will beinfluenced by many factors such as the negotiators’ reservation va-lue, BATNA, and negotiation strategy. Language familiarity isimportant because language usage is the fundamental skill incommunication.

Fourth, the importance of language appears to vary with the po-sition of negotiators in e-negotiations. The analysis results suggestthat language familiarity plays a more important role for buyers,allowing them to exhibit more active behaviors when negotiatingin their native language than in a non-native language. This alsomeans that the negotiation outcome will be better for buyers thanfor sellers when they are negotiating in their native language.

Fifth, IT support should be developed that improves negotiationself-efficacy, communication efficiency, and communication effec-tiveness in order to induce persuasion behavior. For example,Schoop et al. (2003) developed Negoisst, which provides differentmessage types for negotiators to increase the feedback speed andtherefore improve the communication efficiency. Furthermore, asmentioned above, an e-negotiation system also can provide a deci-sion support function to improve both communication efficiencyand effectiveness. For example, Inspire provides a negotiator withan updated diagram of the ‘‘negotiation dance” (Kersten andNoronha 1999) during the negotiation process, which can helpthe negotiator to quickly judge the negotiation status and thusprovide better responses.

Finally, this research was performed in a text-based asynchro-nous e-negotiation context. It is of interest to consider whetherthe analysis results would be the same for other e-negotiation con-texts, such as text-plus-audio- or text-plus-audio- and video-basedsynchronous e-negotiation. Media richness theory indicates thatthe media richness is determined by four criteria: feedback imme-diacy, language variety, non-verbal cues, and personal focus (Daftand Lengel 1986). In general, a richer communication medium willallow more effective communication. Feedback immediacy will al-low negotiators to exchange messages more quickly, and languagevariety and non-verbal cues will provide negotiators with morealternatives for expressing what they want to say. In other words,feedback immediacy will increase communication efficiency, andlanguage variety and non-verbal cues will increase communicationeffectiveness. Furthermore, feedback immediacy may also increasecommunication effectiveness since it allows negotiators to makethe communication clearer by increasing the speed of questioning,responding, and reflecting. On the other hand, language variety andnon-verbal cues may also increase communication efficiency sincethey make it easier for negotiators to express what they want toexpress or understand what their counterparts are trying to ex-press. Therefore, a richer media will also improve the communica-tion efficiency and effectiveness.

However, Yuan et al. (2003) argued that both text-plus-audiocommunication and text-plus-audio- and video-based communi-cation were significantly better than text alone, but that the addi-tion of video to text-and-audio-based communication not only didnot significantly improve communication efficiency, effectiveness,or positive social-emotional communication, but actually dis-tracted negotiators from focusing on the negotiation task. On theother hand, Lai et al. (2008) demonstrated that feedback immedi-acy has a positive effect on message clarity, deception–detectionability, and control ability, which in turn exert a positive effecton decision quality in e-negotiations employing instant messen-ger-like systems. However, the effect might be different for therelation between feedback immediacy and language familiarity.Poor language familiarity will impair the response ability due tothe inability to provide immediate feedback. Therefore, whetherthe analysis results of this research can be applied to other e-nego-tiation contexts with different degrees of media richness needs tobe investigated.

6. Conclusions

The increasing use of global e-marketplaces and popularity ofinternational business trades make negotiating in non-nativelanguages inevitable. Language is the basic tool in negotiationcommunication, and negotiating in a non-native language will hin-der the language self-efficacy and negotiation self-efficacy due tothe lower language familiarity. A previous study found that lan-guage influences business trade development (Hutchinson 2005),and the present study explored the influence of language familiar-ity on e-negotiation behavior in terms of online communicationperformance and online persuasion behavior. An e-negotiationexperiment was conducted with two groups who used native andnon-native languages separately in negotiations.

The analysis results show that language familiarity plays acritical role in inducing persuasion behavior in e-negotiations.Negotiators with higher language familiarity exhibit higher lan-guage self-efficacy and negotiation self-efficacy. Moreover, higherlanguage self-efficacy will result in higher negotiation self-effi-cacy, but only negotiation self-efficacy affects both e-negotiationcommunication efficiency and effectiveness. Furthermore, com-munication efficiency affects communication effectiveness. High-er communication efficiency and effectiveness will induce more

546 H. Lai et al. / Electronic Commerce Research and Applications 9 (2010) 537–548

online persuasion behavior. In addition to the analysis based onsubjective measurements made using a questionnaire, we alsoobjectively examined the actual behavior and outcomes; thatis, the numbers of submitted offers and messages, and the finalutility. In terms of the actual negotiation behavior, negotiatorswith higher language familiarity were more active in e-negotia-tions. However, the negotiation outcome did not differ signifi-cantly between native language and non-native languagenegotiators. This suggests that the negotiation outcome is alsorelated to other factors such as the negotiator’s reservation valueand negotiation strategies, and the interaction process. Finally,the analysis also revealed that language familiarity has a greatereffect on the buyer than on the seller.

The above research findings were obtained in an experimentperformed in a text-based asynchronous e-negotiation context.Media richness theory states that richer media will improve thecommunication performance of negotiators, and this has been dem-onstrated previously (Purdy et al. 2000, Yuan et al. 2003, Lai et al.2008). However, this conclusion may not be valid in the contextof negotiating in a non-native language because language familiar-ity plays a critical role in e-negotiations, as mentioned above. Theinteraction between language familiarity and richness of communi-cation media is a critical research issue because an understandingof this will help people to choose appropriate communication med-ia based on the language used in e-negotiations. This also consti-tutes one of the limitations of the present research, since only thetext-based asynchronous e-negotiation context was considered.

The second limitation is in language familiarity, which wemanipulated by Chinese as a native language versus English as anon-native language. Chinese is a non-Indo-European languagewhile English is an Indo-European language, and the linguisticdistance theory indicates that the distance between English andChinese is larger than for other combinations of languages, suchas Chinese versus Japanese or English versus French (Chiswickand Miller 2004). The research findings may therefore not applywhen using other languages. Furthermore, the perceived linguisticdistance might differ with the subjects’ language environment.English is the second foreign language taught in most schools,and it might be the only foreign language for some individuals.These people would have a higher language familiarity whenusing English than when using Japanese, even though Chinesehas a smaller linguistic distance from Japanese than from English.In addition, the increasingly global e-business activities may meanthat people will have to learn a third language in order to conduct

Appendix A. Summary of scale items

Construct Measures

Language self-efficacy I consider it easy to read and undenewspapers and magazinesI consider it easy to write EnglishI consider it easy to communicate(Chinese)Overall I consider it easy to read a(Chinese)

Negotiation self-efficacy I consider it easy to negotiate in EI consider that I have the ability to(Chinese)I consider that I can negotiate in EOverall I have confidence in negot

E-negotiation communicationefficiency

I could respond to my counterpartI could respond to my counterpart

international negotiations. In such a situation, language learningis impacted by factors additional to the linguistic distance fromthe native language (Cenoz et al. 2001), and hence these factorsalso have to be considered when choosing an appropriate negoti-ation language or when determining how to improve languagefamiliarity. How to utilize IT to reduce the limitations of languagefamiliarity in e-negotiations is another important research topicfor future studies.

The third limitation is that the effect of language familiaritymight be mediated by the different types of communication anddecision support provided by e-negotiation systems and thedegree of task complexity, which were not considered in this re-search. It is worthwhile to explore how information support canreduce the limitations of language familiarity and increase negoti-ation self-efficacy, communication efficiency, and communicationeffectiveness in the e-negotiation context. Finally, the experimen-tal e-negotiation was performed over 9 days in asynchronousmode, and hence external interruptions might have influencedthe results of the negotiation process. However, this is also the casefor asynchronous e-negotiations in the real world. Indeed, it couldbe advantageous to adopt asynchronous e-negotiations since thiswould allow the negotiators to get help from others during thenegotiation process.

Acknowledgments

An earlier version of this article was presented at the 2009Hawaii International Conference on Systems Science. We thankthe co-chairs of the Negotiation Support Systems Mini-Track,Tung Bui and Mel Shakun, and the anonymous reviewers at HICSSfor their helpful comments. We benefited from value commentsby the three reviewers, Byungtae Lee, Ting-Ping Liang and DavidWeber and input from the other participants at the conference.In addition, we appreciated Juliana Tsai and Shu-Chun Ho forrecording comments for us during our presentation. Further, wethank Chris Westland and Rob Kauffman for their commentsand suggestions, especially for the statistical methods, throughthe review process, as well as the anonymous reviewers at ECRA.Hsiangchu Lai acknowledges the support for this research fromthe National Science Council (Grant #96-2416-H-110-013-MY2)and the ‘‘Aim for the Top University Plan” of National Sun Yat-sen University and the Ministry of Education, and Taiwan, Repub-lic of China. All errors and omissions are the sole responsibility ofthe authors.

Main references

rstand English (Chinese) Cheng and Yuh-show (2001),Coronado-Aliegro (2006),and this study(Chinese) letters

well in text using English

nd write in English

nglish (Chinese) Cheng and Yuh-show (2001)and this studynegotiate in English

nglish (Chinese) fluentlyiating in English (Chinese)

rapidly during this negotiation Dennis and Kinney (1998)and this studyeasily during this negotiation

Appendix A (continued)

Construct Measures Main references

I could express my feelings rapidly during this negotiationI could express my feelings easily during this negotiationI could understand the feelings of my counterpart rapidly duringthis negotiationI could understand the feelings of my counterpart easily duringthis negotiation

E-negotiation communicationeffectiveness

I could express my opinions correctly during this negotiation This studyI could express my feelings correctly during this negotiationI could understand the opinions of my counterpart correctlyduring this negotiationI could understand the feelings of my counterpart correctlyduring this negotiation

Online persuasion behavior I explained my reasons proactively This studyI provided information proactivelyI persuaded my counterpart to accept my offersI described the issues, goals, and priorities that concerned meOverall I used persuasion during this negotiation

H. Lai et al. / Electronic Commerce Research and Applications 9 (2010) 537–548 547

References

Abernethy, S. The SquareTrade experience in online and offline disputes. InProceedings of the UNECE Forum on ORD 2003, Palais des Nations, Geneva,Switzerland, June 30–July 1, 2003.

Acuff, F. How to Negotiate Anything with Anyone Anywhere Around the World.Amacom, New York, NY, USA, 2008.

Bandura, A. Social Foundations of Thought and Action. Prentice-Hall, Englewood Cliffs,NJ, 1986.

Bandura, A. Regulation of cognitive processes through perceived self-efficacy.Developmental Psychology, 25, 5, 1989, 729–735.

Barling, J., and Beattie, R. Self-efficacy beliefs and sales performance. Journal ofOrganizational Behavior Management, 5, 1, 1983, 41–51.

Bichler, M., Kersten, G., and Strecker, S. Towards a structured design of electronicnegotiations. Group Decision and Negotiation, 12, 4, 2003, 311–335.

Burgoon, J. K., and Hale, J. L. Validation and measurement of the fundamentalthemes of relational communication. Communication Monographs, 54, 1, 1987,19–41.

Butler, J. K. Jr., Trust expectations, information sharing, climate of trust, andnegotiation effectiveness and efficiency. Group and Organization Studies, 24, 2,1999, 217–238.

Carrell, M. R., and Heavrin, C.. Negotiating Essentials: Theory. Skills and Practices.Pearson Prentice Hall, 2007.

Cenoz, J., Hufeisen, B., and Jessner, U. Cross-linguistic Influence in Third LanguageAcquisition: Psycholinguistic Perspectives. Multilingual Matters, Clevedon, UK,2001.

Cheng and Yuh-show Learners’ beliefs and second language anxiety. Concentric:Studies in English Literature and Linguistics, 27, 2, 2001, 209–223.

Chiswick, B. R., and Miller, P. W. Linguistic distance: a quantitative measure of thedistance between English and other languages. Journal of Multilingual &Multicultural Development, 26, 1, 2004, 1–11.

Cleary, S. The Communication Handbook: A Student Guide to Effective Communication.Juta & Co. Ltd., 2004.

Clement, S. The self-efficacy expectations and occupational preferences of femalesand males. Journal of Occupational Psychology, 60, 3, 1987, 257–265.

Compeau, D. R., and Higgins, C. A. Computer self-efficacy: development of ameasure and initial test. MIS Quarterly, 19, 2, 1995, 189–211.

Coronado-Aliegro, J. The Effect of Self-assessment on the Self-efficacy of StudentsStudying Spanish as a Foreign Language. Ph.D. dissertation, Foreign LanguageEducation, University of Pittsburgh, Pittsburgh, PA, 2006.

Cronbach, L. J. Coefficient alpha and the internal structure of tests. Psychometrika,16, 3, 1951, 297–334.

Daft, R. L., and Lengel, R. H. Information richness: a new approach to managerialbehavior and organization design. Research in Organizational Behavior, 6, 1983,191–233.

Daft, R. L., and Lengel, R. H. Organizational information requirements,media richness and structural design. Management Science, 32, 5, 1986,554–571.

Daft, R. L., Lengel, R. H., and Trevino, L. K. Message equivocality, media selection, andmanager performance: implications for information systems. MIS Quarterly, 11,3, 1987, 355–366.

De Dreu, C. K., Weingart, L. R., and Kwon, S. Influence of social motives onintegrative negotiation: a meta-analytic review and test of two theories.Interpersonal Relations and Group Processes, 78, 5, 2000, 889–905.

Dennis, A. R., and Kinney, S. T. Testing media richness theory in the new media: theeffects of cues, feedback, and task equivocality. Information Systems Research, 9,3, 1998, 256–274.

Garcez, P. M. Point-making styles in cross-cultural business negotiation: amicroethnographic study. English for Specific Purposes, 12, 2, 1992, 103–120.

Giordano, G. A., Stoner, J. S., Brouer, R. L., and George, J. F. The influences ofdeception and computer-mediation on dyadic negotiations. Journal ofComputer-Mediated Communication, 12, 2, 2007, 362–383.

Hackett, G., and Betz, N. E. A self-efficacy approach to the career development ofwomen. Journal of Vocational Behavior, 18, 3, 1981, 326–339.

Hair, J. F., Anderson, R. E., Tatham, R. L., and Black, W. C. Multivariate Data Analysis.Prentice Hall, Englewood Cliffs, NJ, 1998.

Horwitz, E. K., Horwitz, M. B., and Cope, J. Foreign language classroom anxiety. TheModern Language Journal, 70, 2, 1986, 125–132.

Hutchinson, W. K. ‘‘Linguistic distance” as a determinant of bilateral trade. SouthernEconomic Journal, 72, 1, 2005, 1–16.

Keough, C. Bargaining arguments and argumentative bargainers. In L. L. Putnam andM. E. Roloff (eds.), Communication and Negotiation, Sage Publications, NewburyPark, CA, 1992, 109–127.

Kerlinger, F. N., and Pedhazur, E. J. Multiple Regression in Behavioral Research. Holt,Rinehart and Winston, New York, NY, 1973.

Kersten, G. E., Michalowski, W., Szpakowicz, S., and Koperczak, Z. Restructurablerepresentations of negotiation. Management Science, 37, 10, 1991, 1269–1290.

Kersten, G. E., and Noronha, S. J. WWW-based negotiation support: design,implementation, and use. Decision Support Systems, 25, 2, 1999, 135–154.

Lai, H., Lin, H.C., Wu, Y.C. Impact of feedback immediacy on e-negotiation:Synchronous vs. asynchronous modes. Group Decision and Negotiation 2008,Coimbra, Portugal, 2008.

Lee, C., and Bobko, P. Self-efficacy beliefs: comparison of five measures. JournalApplied Psychology, 79, 3, 1994, 364–369.

Lewicki, R. J., Barry, B., Saybders, D. M., and Minton, J. W. Negotiation. McGraw Hill,New York, NY, 2006.

Lewicki, R. J., and Litterer, J. A. Negotiation: Readings, Exercises, and Cases. IrwinProfessional Publishing, Burr Ridge, IL, 1985.

MacIntyre, P. D., and Gardner, R. C. Methods and results in the study of anxiety andlanguage learning: a review of the literature. Language Learning, 41, 1, 1991, 85–117.

Marakas, G. M., Yi, M. Y., and Johnson, R. D. The multilevel and multifacetedcharacter of computer self-efficacy: toward clarification of the construct and anintegrative framework for research. Information Systems Research, 9, 2, 1998,126–163.

Morin, L., and Latham, G. P. The effect of mental practice and goal setting as atransfer of training intervention on supervisors’ self-efficacy andcommunication skills: an exploratory study. Applied Psychology: AnInternational Review, 49, 3, 2000, 566–578.

O’Connor, K. M., and Arnold, J. A. Distributive spirals: negotiation impasses and themoderating role of disputant self-efficacy. Organizational Behavior and HumanDecision Processes, 84, 1, 2001, 148–176.

Pruitt, D. G., and Carnevale, P. J. Negotiation in Social Conflict. Taylor & Francis Group,Pacific Grove, CA, 1993.

Purdy, J. M., Nye, P., and Balakrishnan, P. V. The impact of communication media onnegotiation outcomes. The International Journal of Conflict Management, 11, 2,2000, 162–187.

548 H. Lai et al. / Electronic Commerce Research and Applications 9 (2010) 537–548

Schoop, M., Jertila, A., and List, T. Negoisst: a negotiation support system forelectronic business-to-business negotiations. Data and Knowledge Engineering,47, 3, 2003, 371–401.

Shell, D. F., Murphy, C. C., and Bruning, R. H. Self-efficacy and outcome expectancymechanisms in reading and writing achievement. Journal of EducationalPsychology, 81, 1, 1989, 91–100.

Short, J., Williams, E., and Christie, B. The Social Psychology of Telecommunications.Wiley, London, UK, 1976.

Stanley, J., Ingram David and Chittick, Gary. The Relationship between InternationalTrade & Linguistic Competence. Australian Government Publishing Service,Canberra, Australia, 1990.

Sycara, K. P. Persuasive argumentation in negotiation. Theory and Decision, 28, 3,1990, 203–242.

Taylor, M. S., Locke, E. A., Lee, C., and Gist, M. E. Type A behavior and faculty researchproductivity: What are the mechanisms? Organizational Behavior and HumanPerformance, 34, 3, 1984, 402–418.

Thompson, L., and Nadler, J. Negotiating via information technology: Theory andapplication. Journal of Social Issues, 58, 1, 2002, 109–124.

Tyler, M. C., and Bretherton, D. Seventy-six and counting: An analysis of ODR sites.Proceedings of the Workshop on Online Dispute Resolution at the InternationalConference on Artificial Intelligence and Law, Edinburgh, UK, June 28, 2003, 13–28.

Walther, J. B. Computer-mediated communication: impersonal, interpersonal, andhyperpersonal interaction. Communication Research, 23, 1, 1996, 1–43.

Yuan, Y., Head, M., and Du, M. The effects of multimedia communication on web-based negotiation. Group Decision and Negotiation, 12, 2, 2003, 89–109.