TY - JOUR AU1 - Hill,, K AU2 - Monk,, A.F AB - Abstract Experiment 1 examined power of email communications, compared to equivalent printed communications, to influence a recipient's behaviour. 160 people were sent requests to volunteer for an experiment. Half received these requests through email and half on printed media through the internal mail. These groups were further subdivided into those who replied by internal or electronic mail. There was no evidence that an email message has less weight in persuading people to reply positively. The pattern of reply frequencies is completely explained by the relative effort required to reply. Experiment 2 examined recipients' reactions to email and print communications using rating scales. An email and printed text group were asked to rate two documents, a job application and a thank you note, on various qualities. While the content of the document had a significant effect on ratings, there was no evidence that recipients rate messages or their senders differently depending on the medium used. It is concluded that for this student population email has taken on many of the characteristics previously expected of print. 1 Introduction A recent issue of the You magazine from The Mail on Sunday (18 October 1998, p. 97) featured the following problem in their question and answer page, the “maze of modern manners”: Are you correct in advising that a thank you letter sent by e-mail is a less gracious means of communication than a penned missive on paper? With no set rules for this communication medium, people are naturally anxious as to how their messages are coming across. Recent years have seen a surge in the use of computers for communication purposes, namely through electronic mail (email). In 1994, there were an estimated 20–30 million email addresses in both public and private systems worldwide (Walther, 1996), and this figure has undoubtedly grown substantially since then. It appears, however, that this relatively young communication medium has yet to establish the rules of etiquette that we take for granted with other media such as the telephone, formal letter writing or face-to-face interactions. In such an uncertain area it seems of importance to understand just how email, and indeed the emailed messages themselves, are perceived and understood; and to determine whether there are any special differences between this and other communication media. The two experiments described in this paper compare the effects of the same message when it is sent either by email or by printed text. There is little in the way of precedence for this. Most of the previous experimental research on electronic communication has been concerned with comparing text with speech in synchronous communication tasks, i.e. where participants engage in some discussion task at the same time. In real life settings printed text and email are generally used for asynchronous communication tasks where messages are sent and read at different times (Daly-Jones et al., 1997) for a discussion of the advantages and disadvantages of synchronous and asynchronous communication). For completeness, this introduction will very briefly review the work comparing speech and text in synchronous and asynchronous communication. The conclusion to be drawn is that this work does not allow us to predict the results of our experiments though it does point to some important gaps in the experimental literature. There is to our knowledge no other published experimental work comparing email with printed text. Comparisons of speech and text in synchronous communication have centred on two issues: media choice and media effects. Media choice refers to why a person chooses a particular medium to send a message, when he or she could have used another medium instead. The most influential theory of media choice is ‘information richness’ theory (Daft and Lengel, 1984), which looks at how capable each medium is of conveying information and sharing meaning, regardless of the context. Email is presumed not to be a very rich medium when compared to speech. Print is presumably ranked similarly to email on this dimension. In contrast to this, ‘social influence theory’ (Fulk et al., 1990) highlights the influence of both social factors, such as work colleagues, and of organisational norms, which are lacking in the information richness theory. In line with the quote given at the start of this introduction Fulk et al. (1990) draw attention to the fact that not all media choice decisions are driven by considerations of efficiency, a particular medium may be chosen partly due to the perceived appropriateness of the medium and the social influences that may be at play. The topic of media effects is concerned with how the media, once selected, changes the nature of the conversation that ensues. Social presence (Short et al., 1976) measures “the extent to which one feels aware of the person one is communicating with” (Rudy, 1996). A related concept is the notion of reduced social cues. This theory promoted by Sproull and Kiesler (1986) highlights the diminishing levels of social context cues in textual communication media such as email. The Social Influence De-individuation (SIDE) model (Spears et al., 1990) accepts the diminished cues of email in comparison to speech, but argues that email affects social influence. The key notion here is that people have two identities, their personal identity and social identity, and they will switch to whichever one is most appropriate under a given circumstance, changing their behavioural style to match. Though these authors use email in their experiments, they use it in a synchronous way and thus they have little relevance to a comparison of print and email messages. There is some evidence that perceptions of email are changing as it becomes more widely used. The studies cited above that portray email as an impersonal and somewhat hostile medium were carried out in the 1970s and 1980s. More recently, Walther (1996) has argued that email is actually a warm and personal medium, carrying with it equivalent levels of information richness to face-to-face interaction. As pointed out above, another criticism of the work reviewed above is that it treats email as a synchronous communication medium. However, as Daly-Jones et al. (1997) point out, the requirements for effective communication when all participants are available at the same time (synchronous communication) are quite different from when they are not (asynchronous communication). Email can be used synchronously, as it is in the above experiments, but typically in real life it is used asynchronously and so should be compared with other messaging technologies and not using synchronous discussion tasks. A rare example of a study using a realistic asynchronous communication task is presented by Neuwirth et al. (1994). Reviewers looked over documents created by writers as part of their normal work. Reviewers could record their comments as speech or text annotations. The key manipulation in this experiment was that half of the spoken annotations were transcribed into writing, and half of the written annotations were transferred into speech. This allowed the authors to see, first, whether there were any differences in how the reviewers (the senders) compiled their annotations in each modality. Secondly, it showed whether there were any effects of receiving annotations, independently of how they had been produced. Effects were noted both of the medium in which a comment was composed and of the medium in which the comment was received. Interestingly, the latter effects were generally stronger than the former. Neuwirth et al.'s (1994) study is not a study of email per se but it is instructive in illustrating: (i) that it is possible to design experiments using realistic asynchronous communication tasks; and (ii) that it may be important to separate the effects of a medium on how a message is composed from the effects of a medium on how it is received. As Rudy (1996) (p. 208) points out, “past research on email is heavily biased towards the sender of messages and has almost completely neglected the recipient. This seems odd given that usually the purpose of sending a message is to communicate something to the receiver.” This study aims to fill this gap in current research by examining what happens when the same message is sent by email or as printed text. Experiment 1 examines power of email communications to influence a recipient's behaviour, specifically their probability of responding positively to a request to take part in Experiment 2. Experiment 2 examined a recipient's reactions to email and print communications of differing degrees of formality using rating scales. The messages in all cases are taken to be representative of realistic examples of asynchronous tasks that email is put to. The control condition, the same message in printed form delivered through the internal mail, is taken as a natural way of carrying out these tasks. 2 Experiment 1 Email appears to have less “weight” than a letter. Many people do not reply to emails, though most reply to letters. To some people an email request appears to be treated like a passing comment received from a colleague in the corridor. One might ignore or “forget” such a request when the same request stated in a letter is much more likely to be acted on. Assuming that this impression is correct, the concepts introduced by the experiments discussed above suggest three alternative explanations. It may be a matter of media choice. People may choose to use letters for more inherently important requests. If this is the case then the requests communicated by letter would be just as likely to be acted upon if they had been sent by email. It may be a media effect on the sender, i.e. the request may be formulated differently when it is sent by email than when it is sent in a letter. Finally, and perhaps most plausibly, it may be a media effect on the receiver. This experiment is to test this hypothesis, while controlling the other two possibilities. The same request to take part in an experiment was sent to students in a printed form via the University of York internal mail and by email. Half the participants received these requests through email and half through the internal mail. If the identically worded printed request did indeed have more weight than the email request (i.e. a media effect on the receiver) one would expect more replies to the former. As our dependent variable is the number of people replying to the request, a second possibility is that the probability of responding positively to this request depends on how easy it is to reply. For this reason, the electronic and internal mail groups were further subdivided. In each case, half were requested to reply by internal mail and half by electronic mail. Replying by email to an email request is the option requiring the least effort. Replying to an email request using the internal mail or to an internal mail request via email involve the most effort as they require the participant to change media. This experiment can thus be viewed as a comparison of the strength of two hypothetical effects, the persuasiveness of print versus the effort of replying. If persuasiveness of print is the major effect and effort of reply subsidiary, one would expect the two internal mail request conditions to have higher response rates than the two email request conditions. If the effort to reply is the major effect then the two same media response conditions should get higher response rates, particularly the email-request email-reply condition. The critical comparison then is of the probability of responding positively by email to either a request received by internal or electronic mail. If the persuasiveness of paper printed requests is most important then the former will result in more positive replies. If effort in replying is the critical factor, the latter will result in more positive replies. 2.1 Method (Experiment 1) Requests for participants to take part in an experiment were sent to 160 people, half of whom received this request via email, and half via internal mail. Participants were randomly assigned to groups, but care was taken to make sure that the distribution of males and females in each group was as equal as possible. Of the 80 people who received requests via email, half of them were asked to reply to an internal mail address, and half were asked to reply to an email address. Similarly, of the 80 people who received requests through internal mail, half were asked to reply to an internal mail address and the other half to an email address. All of the 160 people who received requests were first year undergraduates at the University of York. The email and internal mail addresses were obtained, with permission, from departmental records. The two departments used were Electronics and Language and Linguistic Science. In total there were 93 males and 67 females. The requests were designed to be as equal as possible in terms of content and structure. Participants responding by email to an email request were asked to respond to a different email address to the one from which the message had come so that they were not able to simply use the reply function. Participants responding to an internal mail request by internal mail were asked to write a note and send it back in the envelope that it came in. All requests asked people to reply both if they would, or would not, like to take part though only those agreeing to take part (all but two of those who responded) are included in the statistics presented below. All requests were sent out on the same day. The internal mail requests were sent out to departmental pigeon holes. A period of two weeks was allowed for all replies to come back before the responses from each group were counted. 2.2 Results (Experiment 1) Table 1 shows the total number of positive responses gained from each of the four groups. In addition, one participant in each of the email request groups responded negatively and three in the email request internal mail reply group responded incorrectly, i.e. by internal mail. The data are precisely as predicted by the degree of effort required to reply and there is no evidence that the printed form as sent by internal mail was more effective in gaining positive replies than the email request. Table 1 Total number of positive responses gained in each group out of 40 Group . Positive replies . Internal Mail request, Internal Mail response 10 Internal Mail request, Email response 4 Email request, Email response 17 Email request, Internal Mail response 1 Group . Positive replies . Internal Mail request, Internal Mail response 10 Internal Mail request, Email response 4 Email request, Email response 17 Email request, Internal Mail response 1 Open in new tab Table 1 Total number of positive responses gained in each group out of 40 Group . Positive replies . Internal Mail request, Internal Mail response 10 Internal Mail request, Email response 4 Email request, Email response 17 Email request, Internal Mail response 1 Group . Positive replies . Internal Mail request, Internal Mail response 10 Internal Mail request, Email response 4 Email request, Email response 17 Email request, Internal Mail response 1 Open in new tab The experiment is effectively a two-by-two factorial between subjects design that crosses medium of request (internal mail versus email) with medium of reply (same or different) and could have been analysed as two main effects and an interaction. However, the floor effect (only one response out of 40 for the least effective condition) makes the critical interaction effect in such an analysis difficult to interpret. Instead, the data were analysed for statistical significance through two 2×2 Chi Squared tests looking at the effect of medium of request on each response type, one for the email responses and the other for the internal mail responses. The critical comparison is of the email responses as here the hypothesis that printed requests were more persuasive works against the predicted effect of effort. The result (17 out of 40 versus 4 out of 40, Chi square (1)=9.30, p<0.01) shows that if the paper printed requests were more persuasive the effect is small compared to that of the additional effort required to reply in a different medium. The same comparison for the internal mail responses between same-medium versus different-medium requests (10 out of 40 versus 1 out of 40, Chi square (1)=6.75, p<0.01) is also significant. Here the predicted effect of persuasiveness of print and effort of replying are working together so one would expect responses in the hypothetically more persuasive and same-medium condition (internal mail request, internal mail response) to have the highest return rate. The fact that this is not the case gives further weight to the conclusion that effort of response is the major effect. The less striking difference between internal mail response rates is most simply interpreted as a lower response rate with internal mail replies, presumably due to the extra effort required, and a floor effect on the least effective condition (email request, internal mail response). Put simply, the reply rates are in order of effort to reply. The statistical comparison of email response rates to email and internal mail requests shows that there are significantly less responses to a printed request. There is no evidence in these data that print is more persuasive than email. 3 Experiment 2 The results of Experiment 1 cannot be interpreted as supporting the hypothesis that the recipients of email messages view them as more important than the same messages sent in printed form by internal mail. Rather, all the effects observed can be seen to be due to effects of the medium of reply, i.e. how easy it is to send a reply given the way the request was received. However, it is generally harder to change people's actual behaviour, in terms of a performance measure such as this, than to change their view of some communication medium as measured by rating scales (Monk et al., 1996). It remains possible that email messages are viewed differently from printed messages but that the effects are not large enough to change the dependent variable used in Experiment 1. Experiment 2 explores this possibility by obtaining subjective ratings, both of the message and the supposed sender of the message on a number of scales. Two kinds of message were used. They differ in formality, one was a job application and the other a thank you letter. Both were selected as realistic asynchronous tasks that require the sender to project their personal identity. It has been suggested that email is depersonalising. It “uses printed text, without even the texture of paper to lend it individuality, electronic communication tends to seem impersonal” (Kiesler et al., 1984, p. 1125). This idea, that email carries with it stable characteristics which could alter how messages are perceived, is of obvious importance to both communication at the organisational level and that of a personal nature. 3.1 Method (Experiment 2) Participants were divided randomly into two groups, one receiving printed text and one email versions of two letters. Both groups received their analogue rating scales on paper. The order in which they read and rated the two letters was counterbalanced. There were 44 participants, 24 females and 20 males, all of whom were in their first year of study at The University of York. These were the 34 individuals who responded positively to the requests made in Experiment 1 with another 10 recruited through further email requests. Of those participants who were recruited through Experiment 1, approximately the same ratio of ‘email-responders’ to ‘internal mail-responders’ were assigned to each group. The dependent variables in this experiment were the analogue rating scales. The scale used was a 10 cm line, with bipolar adjectives at each end. The lines were randomly alternated in their direction so that not all of the positive attributes were at one end of the scale. The formal letter was rated on 13 items, six regarding the candidate (Impolite–Polite, Sincere–Insincere, Unconfident–Confident, Enthusiastic–Unenthusiastic, Individual–Faceless, Inefficient–Efficient) and seven regarding the letter itself (Personal–Anonymous, Carefully prepared–Slapdash, Informal–Formal, Unclear–Clear, Well structured–Poorly structured, Concise–Wordy, Poorly presented–Well presented). The informal letter and its hypothetical writer were rated on the same items but with Grateful–Ungrateful in the place of Unconfident–Confident. The items had been generated through a brief survey that was completed by 22 people. The survey took the form of two open-ended questions, asking specifically for the kinds of qualities that are felt to be of importance in both the writing of a job application and an informal thank you letter. The results were pooled and those qualities that were mentioned most frequently were consequently used with addition of Personal–Anonymous and Individual–Faceless to assess de-individuation. The formal letter was a job application covering letter based on an extract from the ‘AGCAS’ Graduate Careers Information booklet on ‘Making Applications’ (p. 13) with minor changes to make it suitable for the participants in this experiment. Names were neutered. The informal letter was a thank you letter put together simply as a standard thank you note, in this case for a wedding. The letter was written from an adult to an adult. Again, names were neutered. The printed group received the complete experiment and a pre-addressed envelope for its return through internal mail. The email group received the two documents via email, and the rating scale (again with a pre-addressed envelope) through internal mail. For the printed group, the questionnaires contained step-by-step instructions taking participants through the experiment. The email group also received these same printed instructions and rating scales, the only difference being that this group was referred to documents on a screen rather than to the other side of the sheet containing the rating scales. Participants were asked to read the first letter and to “consider the document as a whole”. They were then instructed to fill in the rating scales for that document but not to refer back to it while doing so. The whole procedure was then repeated for the second presented letter. 3.2 Results (Experiment 2) Fig. 1 presents mean ratings and standard deviations for each medium on each scale. The email and printed documents result in very similar ratings. Each of the 12 rating scales that were applied to both letters were entered in turn into a 2×2×2 split-plot analysis of variance. The three factors being analysed were: (i) medium (between subjects) with the two levels ‘printed’ and ‘email’; (ii) order (between subjects) with the two levels ‘formal letter first’ and ‘informal letter first’; and (iii) letter type (within subjects) with the two levels ‘formal’ and ‘informal’. None of the 12 items analysed showed a statistically significant main effect of medium. However, there were several significant main effects of letter type. These are detailed in Table 2. The person who wrote the job application was viewed as significantly more Polite and Efficient but less Individual than the person who wrote the thank you letter. The job application itself was rated as significantly more Carefully prepared, Clear, Well structured, Concise and Well presented but less Personal. Most importantly, the job application was rated as very much more Formal, thus validating this manipulation of content type. Table 2 Significant main effects of letter type Scale . Mean for job application . Mean for thank you letter . F(1,40) . p . Polite 7.360 6.484 8.32 <0.01 Individual 5.962 6.819 6.88 <0.05 Efficient 7.149 6.176 10.16 <0.01 Personal 5.687 7.308 15.63 <0.01 Carefully prepared 7.297 4.550 45.89 <0.01 Formal 7.465 2.403 183.78 <0.01 Clear 7.623 6.712 s8.73 <0.01 Well structured 7.187 5.417 15.09 <0.01 Concise 7.149 6.228 5.46 <0.05 Well presented 7.204 5.275 19.20 <0.01 Scale . Mean for job application . Mean for thank you letter . F(1,40) . p . Polite 7.360 6.484 8.32 <0.01 Individual 5.962 6.819 6.88 <0.05 Efficient 7.149 6.176 10.16 <0.01 Personal 5.687 7.308 15.63 <0.01 Carefully prepared 7.297 4.550 45.89 <0.01 Formal 7.465 2.403 183.78 <0.01 Clear 7.623 6.712 s8.73 <0.01 Well structured 7.187 5.417 15.09 <0.01 Concise 7.149 6.228 5.46 <0.05 Well presented 7.204 5.275 19.20 <0.01 Open in new tab Table 2 Significant main effects of letter type Scale . Mean for job application . Mean for thank you letter . F(1,40) . p . Polite 7.360 6.484 8.32 <0.01 Individual 5.962 6.819 6.88 <0.05 Efficient 7.149 6.176 10.16 <0.01 Personal 5.687 7.308 15.63 <0.01 Carefully prepared 7.297 4.550 45.89 <0.01 Formal 7.465 2.403 183.78 <0.01 Clear 7.623 6.712 s8.73 <0.01 Well structured 7.187 5.417 15.09 <0.01 Concise 7.149 6.228 5.46 <0.05 Well presented 7.204 5.275 19.20 <0.01 Scale . Mean for job application . Mean for thank you letter . F(1,40) . p . Polite 7.360 6.484 8.32 <0.01 Individual 5.962 6.819 6.88 <0.05 Efficient 7.149 6.176 10.16 <0.01 Personal 5.687 7.308 15.63 <0.01 Carefully prepared 7.297 4.550 45.89 <0.01 Formal 7.465 2.403 183.78 <0.01 Clear 7.623 6.712 s8.73 <0.01 Well structured 7.187 5.417 15.09 <0.01 Concise 7.149 6.228 5.46 <0.05 Well presented 7.204 5.275 19.20 <0.01 Open in new tab Fig. 1 Open in new tabDownload slide Mean ratings and standard deviations for printed and email versions: (a) ratings of the author of the job application; (b) ratings of the job application letter; (c) ratings of the author of the thank you letter; and (d) ratings of the thank you letter. Fig. 1 Open in new tabDownload slide Mean ratings and standard deviations for printed and email versions: (a) ratings of the author of the job application; (b) ratings of the job application letter; (c) ratings of the author of the thank you letter; and (d) ratings of the thank you letter. There was a statistically significant main effect of order for three of the items: Personal (F(1,40)=4.31, p<0.05), Formal (F(1,40)=10.16, p<0.01) and Confident (F(1,40)=5.78, p<0.05), and a significant 3-way interaction on Individual (F(1,40)=6.92, p<0.05). No other statistically significant results were obtained. The 3-way interaction is due to the printed version being rated as less individual than the email version but only for the second presented letter. This does not relate meaningfully to any of the hypotheses considered here, neither is it mirrored in the results for theoretically similar scale Anonymous. It can probably be dismissed as a chance result. 4 Discussion and conclusions The two experiments described above set out to determine whether the medium a message is sent in can affect either the behaviour or perceptions of its recipient. The messages, a request to take part in an experiment, a job application and a thank you letter, all could have been plausibly sent by email or print. There was no apparent effect of the medium in which the message was received in either experiment. This very much contrasts with the results of studies of sender behaviour. Sender studies have shown that email provokes more candid and impersonal writing — even ‘flaming’ where flashes of anger and frustration are included (Baron, 1998). However, from the results of the present two experiments it appears that once a message has been sent there is no difference between the two media studied. It is always difficult to interpret a null result. The results of Experiment 1 could be explained as due to the size of the differences in ease of reply. Perhaps if replying by email could have been made harder, or replying by internal mail could have been made easier, an advantage for print could have shown through. In defence, it can be said that the conditions were realistic and that if the effects of medium on recipients are that small then they are not of practical importance. Similarly, one could question the discriminatory power of the rating scales used in Experiment 2 and the weakness of a role-play context. However, the same rating scales and context do provide significant effects of the content of the message (thank you letter versus job application) in a way that very much conforms to common sense predictions as to what the differences should be. Perhaps a larger experiment could have detected effects of medium on subjective ratings, but again, if such a difference cannot be detected with a total N of 44 is it really of practical importance? Given the apparent disparity between the results of studies of senders and receivers, a more sensible question than “is there a difference?” may be “how does the size of the effect on receivers compare with the size of the effect on senders?” In this respect it would be interesting to compare the effects of medium on sender and receiver by using the experimental design devised by Neuwirth et al. (1994). Senders would compose letters to be sent by email or print and then half of those letters would actually be sent using the other medium. The prediction is that the recipient's ratings of the letters and their senders would be much more affected by the medium in which it was composed than the medium in which it was received. Finally, a note of caution is required when comparing these data with those collected in previous studies. The two experiments reported above were carried out in 1999, on a user population consisting of undergraduates in university departments that depend on email for day-to-day administrative tasks. These participants are young and familiar with the medium of electronic mail in a way that previous experimental samples may not have been. As little as five years ago email was not commonly used except by staff in universities and computer related industries, now it accounts for a large part of the communications received by many people at work and at home. Arguably, our user population may have learned to work with email in a way that previously studied populations may not have done. A better case for the generalisability of experimental findings could be made by first determining what communication tasks email is used for and by what user populations. One could then devise realistic experimental tasks and recruit convincing samples of experimental participants to examine how email affects senders and receivers relative to comparable media used for those same real world tasks. While such a programme of research is well beyond the scope of this short paper the concepts introduced here do at least suggest how it might be approached. References Baron, 1998 Baron N.S. , Letters by phone or speech by other means: the linguistics of email , Language and Communication 18 ( 1998 ) 133 – 170 Google Scholar Crossref Search ADS WorldCat Daft and Lengel, 1984 Daft R.L. Lengel R.H. , Information richness: a new approach to managerial information processing and organization design , Research in Organizational Behaviour 6 ( 1984 ) 191 – 233 OpenURL Placeholder Text WorldCat Daly-Jones et al., 1997 Daly-Jones O. et al. , Multimodal messages: the pen and voice opportunity , Interacting With Computers 9 ( 1997 ) 1 – 25 Google Scholar Crossref Search ADS WorldCat Fulk et al., 1990 Fulk J. Schmitz J. Steinfield C.W. , A social influence model of technology use Fulk J. Steinfield C.W. Organizations and Communications Technology 1990 Sage , Newbury Park, CA 117 – 140 OpenURL Placeholder Text WorldCat Kiesler et al., 1984 Kiesler S. Siegal J. McGuire T.W. , Social psychological aspects of computer-mediated communication , American Psychologist 39 ( 1984 ) 1123 – 1134 Google Scholar Crossref Search ADS WorldCat Monk et al., 1996 Monk A.F. McCarthy J.C. Watts L.A. Daly-Jones O. , Measures of process Thomas P. CSCW Requirements and Evaluation 1996 Springer , Berlin 125 – 139 OpenURL Placeholder Text WorldCat Neuwirth et al., 1994 Neuwirth, C.M. et al. (Eds.), 1994. Distributed collaborative writing: a comparison of spoken and written modalities for reviewing and revising documents. Proceedings of ACM CHI'94 Conference on Human Factors in Computing Systems, p. 202. Rudy, 1996 Rudy I.A. , A critical review of research on electronic mail , European Journal of Information Systems 4 ( 1996 ) 51 – 57 Google Scholar Crossref Search ADS WorldCat Short et al., 1976 Short J. Williams E. Christie B. , The Social Psychology of Telecommunications 1976 Wiley , London Spears et al., 1990 Spears R. Lee S. Lea M. , De-individuation and group polarisation in computer-mediated communication , British Journal of Social Psychology 29 ( 1990 ) 121 – 134 Google Scholar Crossref Search ADS WorldCat Sproull and Kiesler, 1986 Sproull L. Kiesler S. , Reducing social context cues: electronic mail in organizational communication , Management Science 32 ( 1986 ) 1492 – 1512 Google Scholar Crossref Search ADS WorldCat Walther, 1996 Walther J.B. , Computer-mediated communication: impersonal, interpersonal and hyperpersonal interaction , Communication Research 23 ( 1996 ) 3 – 43 Google Scholar Crossref Search ADS WorldCat Elsevier Science B.V. TI - Electronic mail versus printed text: the effects on recipients JF - Interacting with Computers DO - 10.1016/S0953-5438(00)00040-0 DA - 2000-12-01 UR - https://www.deepdyve.com/lp/oxford-university-press/electronic-mail-versus-printed-text-the-effects-on-recipients-eqaUwqTMRh SP - 253 EP - 263 VL - 13 IS - 2 DP - DeepDyve ER -