Differences Between Males and Females Concerning Perceived Electronic Mail Appropriateness

Electronic mail was once tangential software restricted to the mainframe computers of large universities. Today, electronic mail is one of the most common protocols on the Web accessible through practically every modern mobile device. Its role and relevance in organizations and in personal lives cannot be understated. Ranie and Horrigan (2005) state that electronic mail remains “the No. 1 activity and time consumer for the vast majority of Internet users” (p. 63). With its success, there have been vast amounts of scholarly activity devoted to the dynamics that shape electronic mail interaction.

One such dynamic is appropriateness, which has long been an important variable in face-to-face interaction (Rubin, 1982). Over the past 2 decades, investigators have observed that many of the norms of face-to-face appropriateness can extend to electronic communications. In both interpersonal (Jessmer & Anderson, 2002; Walther, Anderson, & Parks, 1994; Westmyer, DiCioccio, & Rubin, 1998) and organizational contexts (Allen, 1995; Markus, 1994; Rice, 1993; Sproull & Keisler, 1986), numerous researchers have uncovered the significance of appropriate electronic communication. Even further, investigations suggest that gender differences exist in the perception variables associated with appropriateness (Matheson, 2002; Savicki, Lingenfelter, & Kelley, 1996; Sussman & Tyson 2000). However, many of the aforementioned investigations relied upon dialectical analysis or observational techniques in order to draw their conclusions. Additionally, Spitzberg’s (2006) recent attempt to quantify the appropriateness of computer-mediated messages addresses a much smaller slice of the construct that proposed in the present report. Accordingly, the present investigation attempts to expand the range of variables that play a role in the perception of appropriate interaction in electronic mail. In addition, the author expects gender may influence the perception of these variables.

A variety of theoretical frameworks could potentially serve as a foundation for the electronic mail appropriateness measures presented in this report (see Spitzberg & Cupach, 1989). However, few theories of appropriateness have a history in the literature as the approach presented by John Wiemann and his colleagues (see Wiemann & Kelly, 1981; Wiemann & Knapp, 1975). The authors explain “appropriateness generally refers to the ability of an interactant to meet the basic … requirements of the situation” (Wiemann & Backlund, 1980, p. 191). Wiemann and his colleagues point to three specific requirements (verbal, relational, and environmental) that can be developed into variable form.

For Wiemann, an interactant must first meet the verbal context requirement by “making sense in terms of wording, of statements, and of topic” (p. 191). Norms are an immensely important aspect of Wiemann’s approach, and in this case, one must produce words and sentences that fit within the expectations of appropriate behavior. In electronic communications, the historic and current literature demonstrates that avoidance of behaviors referred to as flaming (an uninhibited sharp criticism or profanity that would not be expressed in face-to-face interaction) may fulfill the verbal context requirement (Kayany, 1998; Sproull & Kiesler, 1986). Wiemann and his colleagues also present relational context as an important requirement of appropriateness, in which messages serve to maintain the bond between the interactants. Past research has shown that the use of symbolic graphical accents or emoticons can serve relational maintenance functions. Walther’s social information processing perspective (see Walther, 1992;Walther et al., 1994) provides considerable support for the notion that appropriate interaction includes the careful use of emoticons. Wiemann and Backlund (1980) also emphasize environmental context or the constraints placed upon a message by “symbolic and physical” settings. The authors were attempting to characterize the influence that context has on communication interaction. Literature concerning social presence and immediacy has posited that the speed of message exchange plays a role in perceived appropriateness (see Rice, 1993; Tu, 2002), meaning that the quicker a person responds in online contexts, the more likely that interaction is considered appropriate. (Note: within the social presence literature, speed of message exchange is also referred to as interactivity.)

Lastly, gender has consistently appeared as a factor in electronic message exchange. For example, it has been found that females are more sensitive to verbal style (Sussman & Tyson, 2000; Thomson, Murachver, & Green, 2001) used in computer-mediated messages. Equally, researchers have identified that females pay more attention to perceived relational development online (McKenna, Green, & Gleason, 2002) than do males. And additionally, investigators have found that females use emoticons more than their male counterparts (Witmer & Katzman, 1997). Based on this literature, the present investigation is guided by the hypotheses that females will score higher than males in reference to verbal and relational appropriateness. This research report also inquires whether there is a gender difference concerning environmental appropriateness.

Addressing these issues was accomplished through the development of three measures. After careful consideration of the Wiemann approach to face-to-face appropriateness and the literature mentioned above, the researcher developed 18 Likert items (anchors of 1 = strongly disagree, 5 = strongly agree), which embody the variables of verbal, relational, and environmental appropriateness. Several levels of item reduction were applied to the items.

First, 14 student judges from a public liberal arts university in the northeast United States were provided with concept definitions (see  Appendix) and rated each item on a five-point Likert scale (anchors of 1 = strongly unfavorable to the concept, 5 = strongly favorable to the concept). By correlating the item with the summed score of the judges across all of the items respective to each measure, the researcher was able to discriminate or drop all items that fell below a .7 correlation threshold. With 17 items remaining, a convenience sample (university technology rules stipulated that the sampling frame could not be accessed directly for this investigation) of 144 respondents was drawn from the same university population in the northeast (35 males, 24.3%; 109 females, 75.7%; 1 American Indian, .7%; 2 Asian, 1.4%; 7 African American, 4.9%; 2 Hispanic, 1.4%; 128 Caucasian, 88.9%; 4 Biracial 2.8%). Cronbach’s alpha was then used to identify the three items for each measure that were the most stable (pretest verbal = .78; relational = .87; environmental = .80). Correlation analyses indicated that the three measures were uncorrelated due to low intermeasure association (all > .26); while within the measures, the items demonstrated significant cohesion (ranging between .50-.74).

After item reduction, a factor analysis was used to confirm that there were only three factors in the dataset. The researcher wanted to ensure that each item would load upon the correct factor with no cross-loading over .3, with a cumulative variance explained above 70%. With another convenience sample of 166 respondents, a specified three-factor principal component analysis with Varimax rotation revealed that the items converged on the correct factors. The rotated solution indicated that the first factor (eigenvalue = 2.7, 30.1%) was composed of the relational context items 4, 5, and 6 (.88, .88, and .90, respectively), with the second factor (eigenvalue = 2.1, 24.3%) composed of the environmental items 7, 8, and 9 (.87, .86, and .79, respectively). The last factor (eigenvalue = 1.6, 18.3%) was maintained by the verbal context items 1, 2, and 3 (.80, .82, and .79, respectively). These values demonstrate that the items clustered around their correct measures with little cross loading, all of which lead to a substantial 72.8% total variance explained by these factors.

To assess the hypotheses, a balanced sample of 276 respondents was acquired (135 males, 48.9%; 141 females, 51.1%; 2 Asian, .7%; 11 African American, 4.0%; 4 Hispanic, 1.4%; 1 Pacific Islander, .4%; 247 Caucasian, 89.5%, 11 biracial, 4.0%). Internal consistency of the measures revealed moderate to strong alpha coefficients (verbal context = .79; relational context = .89; environmental context = .81). An ANOVA indicated that females (M = 13.7, SD = 1.9) scored significantly higher than males (M = 13.1, SD = 2.6) in reference to verbal context (p < .05, F = 4.44, η² = .016). Equally, a significant difference was observed between females (M = 9.8, SD = 3.0) and males (M = 6.9, SD = 3.6) for the relational context variable (p < .01, F = 50.39, η² = .155). And finally, the environmental context variable also yielded a significant difference (p < .01, F = 14.63, η² = .051) between females (M = 10.6, SD = 2.8) and males (M = 9.2, SD = 3.1).

As mentioned above, electronic mail remains the primary activity for all Internet users. Thus, an examination of the perceived appropriateness of those messages becomes a critical concern. The results presented here must be interpreted with caution because a convenience sample, necessitated because of university privacy regulations, always increases potential errors in statistical reasoning. The results of this investigation extend previous findings that females have greater sensitivity to a variety of variables related to the appropriateness of electronic mail interaction, which is consistent with prior computer-mediated communication research (Sussman & Tyson, 2000; Thomson, Murachver, & Green, 2001). Females appear to spend more time in the construction of the electronic communications to ensure that they fit with social norms. Future researchers in both the interpersonal and organizational arenas can benefit from the present findings.

Spitzberg (2006) demonstrates the utility of scaling variables related to electronic interaction, especially appropriateness. The measures presented in this report expand the knowledge base beyond previous scales, while demonstrating significant differences between males and females. Interpersonal researchers have often linked appropriateness to effectiveness. Combined, the two are often considered the pillars of competent interaction (Canary & Spitzberg, 1987). Future research may center on examining whether females are more efficacious in their electronic mail interactions. In other words, it would an interesting investigation to ascertain whether females achieve better interpersonal outcomes with their electronic communications than do males. Although this would require a structurally different investigation, it would allow researchers to examine both sides of the communication interaction.

Organizational researchers and practitioners are still searching for effective tactics to combat the inappropriate exchange of electronic mail messages and avert potential litigation. Lictash (2004) explains that “given the absence of clear guidelines as to what constitutes a misuse of e-mail … arbitrators and courts will continue to outline the contours of these issues” (p. 35). The literature has shown that scholarship, such as the present measures, can be a cornerstone for training and mediation programs to correct inappropriate organizational behavior (King & Xia, 1997). One approach could be for an organization to distribute these measures across the company and then conduct systematic training and intervention activities to bring more awareness to the issue of inappropriate electronic communication. Thus, these measures present both useful lines of theoretical and applied research in the area of interpersonal electronic communication.

These items measure a person’s perceived capacity to use language in an electronic mail that fits within social norms defined by the situation.

These items measure a person’s perceived ability to use elements of the electronic mail medium (emoticons, text, etc.) to construct electronic mails that keep personal relationships in harmony.

These items measure a person’s perceived capacity to use electronic mail because it takes advantage of the unique characteristics of electronic mail to communicate (i.e., speed, inexpensive).

• 1VC: *I use rude language in my e-mails.

• 2VC: *I offend people with the language in e-mails.

• 3VC: *I think the language that I use in my e-mails is impolite.

• 4RC: In my e-mails, I use special characters like the ;-) to show my feelings.

• 5RC: In my emails, I use special characters like ;-) to help maintain my relationships.

• 6RC: In my relationships, the special characters like :-( are helpful to convey my feelings.

• 7EC: E-mail is the fastest way to get a message to someone.

• 8EC: I use e-mail because messages arrive almost instantly

• 9EC: I use e-mail when information needs to be received quickly.

*Asterisk denotes reversal items.