An experimental study of language intensity and response rate in e mail surveys.

Intense language, under many conditions, has been shown to substantially increase attitude change. However, the effect of language intensity on actual behavior has rarely been examined and the effectiveness of intense messages has never been tested in new media environments. The current study

empirically examined the effect of the level of intense language employed in an electronic mail survey by a university alumni association. Email records of 11,363 alumni at a large public southwestern university were extracted and assigned randomly to one of two alumni association surveys, a low language intensity survey or a high language intensity survey and emailed to the alumni database. Results supported the hypothesis that the survey introduced with higher levels of intense language yielded significantly higher response rates. Implications for both communication theory and for the more pragmatic concern of increasing survey responses are provided.

**********

Variations of language in persuasive messages have been studied since the days of Aristotle. In the heyday of persuasion research during the 1960's and early 1970's verbal message variables dominated (Cronhkite, 1969; Miller & Burgoon, 1978) although little of this research was incorporated into a single, unified, coherent theory of language. In the past decade message variables made a substantial comeback with active programs of research on a host of message variables including sequential message strategies (Burgoon, 1995; Dillard, 1991).

Over the years, one persistent research area has been on the effects of intense language. Indeed, in their review of important language variables Bradac, Bowers, and Courtright, (1979) reported that three variations in language have been demonstrated to influence receivers' responses; among these is language intensity. Studies have shown the positive effects of language intensity on attitude change in one-to-many communication, dyadic interaction and in mediated materials in health campaigns, however, no study has examined the persuasive effects of language intensity in on-line communication where billions of messages are sent daily through cyberspace. If language intensity is a robust persuasive variable it should be as effective in on-line communication as it is in more conventional forms. This study analyzes the effects of language intensity in cyberspace, more specifically, in seeking compliance by responding to an electronic email survey.

LANGUAGE INTENSITY

Early research conducted on language intensity was based around the widely accepted definition offered by Bowers (1963), "the quality of language which indicates the degree to which the speaker's attitude toward a concept deviates from neutrality" (p. 345). Similarly, Burgoon, Jones and Stewart (1975) defined language intensity as "language indicating degree and direction of distance from neutrality" (p. 241). These early definitions failed to differentiate between language intensity on the one hand and source discrepancy or source position on the other hand. Thus, more recently, language intensity has been defined as a feature of language conveyed through the properties of emotionality and specificity (Hamilton & Hunter, 1998; Hamilton, Hunter & Burgoon, 1990; Hamilton & Stewart, 1993). "Emotionality is the degree of affect expressed in the source's language. Specificity is the degree to which a source makes precise reference to attitude objects in a message" (Hamilton & Stewart, 1993, p. 231).

Communication researchers have sought to provide theoretically and empirically based advice to persuaders on the optimal degree of language intensity to employ in messages. In general, intense language has been shown to produce greater attitude than more neutral language (Buller, Borland & Burgoon, 1998; Hamilton et al., 1990; Hamilton & Hunter, 1998), but this in not without exception. A few studies even suggest that under some conditions more intense language leads to less attitude change due to interactions with source or situational variables (Bowers, 1963; Buller, et al., 1998; Burgoon et al., 1975; Burgoon & King, 1974). A meta-analysis reports that language intensity influences attitude change only for discrepant messages delivered by high credibility sources (Hamilton & Hunter, 1998). Buller et al. (1998) report that intense language is highly effective in motivating sun safety behavior in people that have already intended to engage in such behavior but produces a reactant response from those not intending to practice sun safety. Likewise Buller et al. (2000) found that language intensity is most effective in changing behavior when used in deductive rather than inductive messages.

Several theories have been employed to explain and predict the effects of language intensity on attitude change including information processing theory (Hamilton, Hunter & Boster, 1993; Hamilton & Hunter, 1998), language expectancy theory (Bradac, Bowers & Courtright, 1980; Burgoon, 1995; Burgoon, Denning, & Roberts, 2002; Burgoon & King, 1974; Burgoon et al., 1975; Burgoon, Dillard & Doran, 1984) and communication accommodation theory (Aune & Kikuchi, 1993). While information processing theory (Hamilton & Hunter, 1998) appears to be the most encompassing explanation for the persuasive effects of language intensity as suggested by meta-analytic research, recent persuasion campaigns based on language expectancy theory have been successful (Buller et al., 1998, 2000).

Few studies have examined the effects of language intensity on behavior change. One experimental study on language intensity in promoting sun safety behavior showed positive effects of more intense language (Buller et al., 1998). Likewise another recent experimental study did find an effect for greater language intensity in newsletters, tip cards and brochures on several sun safety behaviors for both children and adults (Buller et al., 2000).

In sum, the research on language intensity has shown that intense language has a generally positive but complex relationship with attitude change. The literature consistently shows far more positive effects of language intensity on attitude and behavior change if such messages originate from high credibility, high status, high similarity sources, if message discrepancy is high and counterarguing is minimal (Aune and Kikuchi, 1993; Bradac et al., 1979; Burgoon et al., 2002; Hamilton & Stewart, 1993; Hamilton & Hunter, 1998).

The present study extends research on language intensity in two important ways. First, unlike most studies it examines behavioral compliance rather than attitude change as the dependent variable. Second, it is the first study on language intensity in an on-line environment.

ELECTRONIC MAlL AND DATA COLLECTION

E-mail is becoming a commonplace form of communication for most Americans. Bachman, Elfrink, and Vazzana (2000) pointed out, at the turn of the century more than 140 million people worldwide can access the Internet, and more specifically, 81 million Americans are connected to the Internet and either send or receive an average of 26 e-mail messages daily. In this explosion of Internet technology, more organizations are turning to e-mail to persuade, to sell, to inform and to collect data.

E-Mail Surveys

Although traditional methods of data collection such as paper and pencil or telephone have been highly successful, electronic surveys are faster, more flexible and cheaper (Bachman et al., 2000; Cook, Heath & Thompson, 2000; Couper & Nicholls, 1998; Ramos, Sedivi and Sweet, 1998; Sheehan & McMillan, 1999) Although web-based surveys have become increasingly common (Manfreda, Batagelj, & Vehovar, 2002; Vehovar, Batagelj, Manfreda, & Zalarel, 2002), most electronic surveys are sent to respondents as a part of an email message or as an attached file to an email message (Ramos et al., 1998). Such surveys can be conducted by almost anyone since they do not need a server, just a computer and a list of potential respondents.

According to Krosnick (1999), new methods of survey research such as the email survey produce valid results. More importantly, the use of electronic mail for data collection offers a number of benefits that previous survey research techniques don't offer. For example, Parker (1992) claims that because people find the technology easy to use, they are willing to participate in the survey at their own pace and return it to the sender. Furthermore, Sheehan & Hoy (1999) explain that mailed surveys may often get lost as they are mislaid in the office, while electronic surveys with a potential respondent remains in place until purposefully deleted. Other benefits of electronic mail surveys include quick response time, more candid responses, ease of recontacting subjects, and subjects' willingness to respond to open-ended questions (Bachman et al., 2000; Litvin, 2001; Sheehan & McMillan, 1999). One study shows that email surveys produce a higher response rate than web-based surveys (Yun, 2000).

Response Rate

Obtaining a larger response rate on a survey is vital for three reasons: representativeness, perceived validly, and statistical power. Although increasing the response rate does not guarantee representativeness it generally makes a survey more representative particularly when a researcher is uncertain about representativeness (Cook et al., 2000; Groves, Cialdini, & Cooper, 1992). Moreover, any ballot or survey has greater perceived validity with the public if it is based on a larger sample of participants that approaches the actual population. Finally, attaining a larger response increases the sample size and contributes to statistical power. Unfortunately, response rates on surveys of all types have been falling since the 1950's (Krosnik, 1999).

Several factors have been shown to increase response rate in electronic surveys including personalized letters, issue salience, and a pre or post reminder notice, with a single follow-up producing the best results (Cook, et al., 2000; Groves et al., 1992; Sheehan & McMillan, 1999). While a plethora of factors have been investigated to increase response rate (Manfreda et al., 2002; Vehovar, et al., 2002) communication variables are conspicuously absent from these studies. No prior study has investigated language intensity, a key communication variable in increasing survey response rates.

HYPOTHESIS

Persuasive communication research and theory of language intensity has rarely been used to examine behavioral compliance in a field setting (see Buller et al., 1998, 2000 for exceptions) and has never been employed to examine increasing response rates to surveys. This study examines language intensity on one type of behavioral compliance: its impact on email survey response rates. As demonstrated in the above rationale, high credibility and high similarity sources have greater freedom to employ intense messages and these messages result in greater persuasive effectiveness. Thus, when a message is sent by a high credibility, high similarity organization, it is hypothesized that:

H: A message containing more intense language would achieve a higher compliance rate on an e-mail survey than a message containing less intense language.

METHOD

The Research Setting

During the 2000-2001 academic year a group of Native American students and their allies attempted to change the name of the mascot and logo of San Diego State University which carried the names Aztecs, a Mezo-American Indian Tribe and their leader, Montezuma. Votes on replacing the mascot were taken by every constituency on campus: associated students, the faculty senate and the student body. To ascertain the position of the university alumni on this issue the associate director of the Alumni Association was asked to gauge the opinion of the alumni. This gave the authors of this study an opportunity to examine an important message variable, language intensity, in an electronic mail environment.

Participants

Email addresses of 11,876 alumni were extracted from the alumni association database located on the campus of San Diego State University. The participants included members of the alumni association living around the United States and in forty foreign countries.

Procedure

Email records of 11,876 alumni at San Diego State University were extracted and assigned randomly to one of two surveys, a high language intensity survey or a low language intensity survey using a randomization program called "SQL Script." The 11,876 emails were sent in batches of 1,000 starting with the low language intensity message over a period of three days between the hours of 5 p.m. and 7 a.m.

Since some addresses were obsolete, 493 emails were returned as "undeliverable" making the actual sample size 11,383 participants. The number of participants that received the high language intensity survey was 5887 while the number of subjects receiving low language intensity survey was 5989. Each respondent was sent an electronic mail message with the survey text below a short introduction. The surveys were not sent as an "attachment"; each recipient had equal ability to reply to the survey by simply using the "Reply" function of her or his respective email software program. Results of the survey were tallied after a one-week period passed after mailing the final email message. Electronic mail received after the results were compiled was not included in this study.

The Message Source

As demonstrated in the rationale, research consistently shows that linguistically intense messages are most effective when they come from high credibility and similarity sources. In the current experiment credibility and similarity were not manipulated, but controlled. Both the high and low language intensity messages originated from the Alumni Association, a high credibility organization that shares considerable similarity with the participants. All 11,383 participants were members of the Alumni Association and had both paid an Alumni Association membership fee and voluntarily provided their e-mail address. The behavioral act of joining an organization and of providing an ongoing link via email indicates that the members have high regard for the organization and suggests that it is highly credible in their eyes.

The Experimental Manipulation

To manipulate language intensity, two surveys were drafted: Survey A, the high intensity message (see appendix A) and Survey B, the low intensity message (see appendix B). The authors of this report and the executive director of the Alumni association of San Diego State University collaboratively designed the surveys. The surveys were virtually identical in all respects including length. Survey A was 329 words long whereas survey B was 325 words long. Both surveys focused on three key aspects regarding the issue at hand: the current name of the mascot, the logo, and the human portrayal of the mascot. Both surveys posed the same questions (see appendices A and B) and were virtually identical with the exception of seven locations where the language intensity manipulation occurred: 1) In the subject line the high intensity read "Important: SDSU Mascot" whereas the low language intensity condition read "SDSU Mascot." 2) In the first sentence the high language intensity message said, "We are sure that you heard ..." whereas the low intensity message said, "As you may be aware ..." 3) In the last sentence of the first paragraph the high intensity message said, "Our office has been swamped with calls" while the low intensity message stated, "Our office has received a high volume of calls." 4) In the second paragraph the high intensity condition included the word "vital" in the sentence, "I ask that you give me some vital input relating to three specific areas ... " whereas survey B did not. 5) In the third paragraph the high intensity manipulation read, "If you care at all about this issue, pro or con, please answer the three questions to help us understand the collective sentiment of 179,000 alumni" whereas the low intensity manipulation read, "Below are three questions to help us understand the collective sentiment of 179,000 alumni." 6) In the third paragraph respondents in the high intensity condition were told, "Please submit your crucial answers to me ..., whereas in the low intensity condition respondents were told, "You can submit your answers to me ..." 7) Last, in the final sentence before the survey, respondents in the high intensity condition were told, "The essential questions are divided into three sections," whereas respondents in the low intensity condition had the word essential omitted (see appendices A and B for the entire text of these messages).

Dependent Variable: Response Rate

The dependent variable was the response rate of participants during the week following distribution of the experimental messages across the two language intensity conditions.

Statistical Analysis

To test the hypothesis a chi-square test, a contingency coefficient, and a phi coefficient were calculated to ascertain the relationship between the language intensity manipulation and the response rate for the surveys. Alpha was set at .05 and power exceeded .99 for all effects: small, medium, and large.

RESULTS

Results supported the hypothesis that e-mail messages with greater language intensity resulted in more compliance than email messages with less language intensity. A chi-square test revealed that respondents were significantly more likely to respond to the survey in the high language intensity condition than in the low language intensity condition ([X.sup.2] = 185.02, p < .000001, C = .128, [phi] = .128). Although the overall response rate to the survey was 41.3%, the response-rate for the high language intensity condition was 47.6% and for the low language intensity condition it was 35% an increase from the low intensity to the high language intensity condition of 12.6%.

DISCUSSION

Summary

This study suggests that more language-intense email messages result in greater compliance than less intense e-mail messages. More specifically, the results of this study indicate that response rates to an electronic mail survey are significantly greater to a higher language-intensity message than a less intense one.

Theoretical Implications

Theoretically, the current investigation adds to a rich body of research on language intensity and extends it to the electronic mail survey environment. As with prior research based on several theories, messages generated by a high credibility source generate higher rates of compliance in a higher language-intensity condition than in a lower language intensity condition. To estimate the effect size of the language intensity effect on compliance, a contingency coefficient and a phi coefficient were computed. These coefficients are an analog of the correlation but since they cannot exceed .79, they underestimate the magnitude of the effect. The obtained contingency coefficient and the phi coefficient in the present study are .128, suggesting modest effects. However, another measure of the importance is that percentage increase in response rate to the survey based on the more and less intense message. The response rote in the high intensity condition exceeded the lower intensity condition by 12.6% suggesting a moderately large effect. Additionally, these results support the main findings of information processing theory in that the intense message originated from a high credibility, high similarity source with no counterarguing, though degree of message discrepancy was unknown and probably low.

Implications for survey researchers

The practical implications for researchers of this experiment are substantial. In a world characterized by an information glut, any message that can enhance compliance is a valuable message. A recent meta-analysis of 56 electronic surveys revealed a mean response rate of 34.6% very similar to the 35% in our low language intensity control condition. In our high language intensity condition 47.6% of subjects responded in the high intensity condition a difference of 12.6% and an increase in response rate of 39% compared to the control. The difference in compliance between that higher and lower language intensity conditions in the current experiment are nontrivial. Based on these findings an email survey will benefit greatly from more intense messages to increase response rate.

Suggestions for future investigation

Because this appears to be the first-ever electronic mail study involving language intensity, the next step in language intensity should be additional research that replicates this experiment using other forms of electronic communication including attachments and web sites.

Another test of language intensity that might be conducted is to manipulate the source similarity or credibility of the sender. This study was an opportunistic field study conducted by the alumni association where the Director of the Alumni Association was the sender of the message, so a credibility manipulation was not possible. Since prior theory and research has shown that source credibility interacts with language intensity, manipulating the source of the message by having the University President or a celebrity author send such a message, could produce even greater effects. Likewise using an unknown source or low credibility source may produce reduced effects. This sort of experiment could determine the relationship between language intensity and source credibility and could ascertain whether an electronic mail message would yield even higher response rates in the high-language intense condition if the source were perceived as more or less credible.

Appendix A

The High Intensity Message

Note: Underline portion differed between low and high intensity versions

Subject Line Said: Important: SDSU Mascot

Dear <<First_Name>>,

We are sure you have heard that on September 27, the Associated Students Council voted to retire the 'Aztecs' name, logo and the mascot, 'Monty' Montezuma. The decision is not final and will ultimately be decided by President Stephen L. Weber. Our office has been swamped with calls about this resolution; however, we are still looking for additional input from alumni.

The president has assured me that he will take into account input from many constituents including: alumni, students, faculty, and staff. In order to better represent the opinions of our alumni regarding this resolution, I ask that you give me some vital input relating to three specific areas: the name 'Aztecs', the logo, and the human mascot 'Monty' Montezuma.

If you care at all about this issue, pro or con, please answer the three questions to help us understand the collective sentiment of 179,000 alumni. Please submit your crucial answers to me by simply using the "Reply" function of your email program and identifying which answer you choose. These essential questions are divided in three segments:

Aztec Name:

1. I am in favor of retaining the name 'Aztecs' I am opposed to retaining the name 'Aztecs' I have no opinion

Aztec Logo:

2. I am in favor of retaining the 'Aztec' logo I am opposed to retaining the 'Aztec' logo I am in favor of modifying the 'Aztec' logo I have no opinion

Mascot:

3. I am in favor of retaining the human mascot, 'Monty' Montezuma I am opposed to retaining the human mascot, 'Monty' Montezuma I am in favor of modifying the human mascot, 'Monty' Montezuma I have no opinion

Sample images of the Aztec logo and mascot can be viewed on the SDSU Web site, www.sdsu.edu. I would like to commend and thank you for expressing your interest on the Associated Students resolution. As the alumni director, I assure you your voice will be heard.

Sincerely,

Jim Herrick

Appendix B

The Low Intensity Message

Subject Line Said: SDSU Mascot

Dear <<First_Name>>,

As you may be aware, on September 27, the Associated Students Council voted to retire the 'Aztecs' name, logo and the mascot, 'Monty' Montezuma by the end of the current academic year. The decision is not final and will ultimately be decided by SDSU President Stephen L. Weber. Our office has received a high volume of calls about this resolution; however, we are still looking for additional input from alumni.

The president has assured me that he will take into account input from many constituents including: alumni, students, faculty, and staff. In order to better represent the opinions of our alumni regarding this resolution, I ask that you give me some input relating to three specific areas: the name 'Aztecs', the logo, and the human mascot 'Monty' Montezuma.

Below are three questions that will help us understand the collective sentiment of our 179,000 alumni. You can submit your answers to me by simply using the "Reply" function of your email program and identifying which answer you choose. Questions are divided in three segments:

Aztec Name:

1. I am in favor of retaining the name 'Aztecs' I am opposed to retaining the name 'Aztecs' I have no opinion

Aztec Logo:

2. I am in favor of retaining the 'Aztec' logo I am opposed to retaining the 'Aztec' logo I am in favor of modifying the 'Aztec' logo I have no opinion

Mascot:

3. I am in favor of retaining the human mascot, 'Monty' Montezuma I am opposed to retaining the human mascot, 'Monty' Montezuma I am in favor of modifying the human mascot, 'Monty' Montezuma I have no opinion

Sample images of the Aztec logo and mascot can be viewed on the SDSU Web site, www.sdsu.edu. I would like to commend and thank you for expressing your interest on the Associated Students resolution. As the alumni director, I assure you your voice will be heard.

Sincerely,

Jim Herrick

REFERENCES

Aune, K. R., & Kikuchi, T. (1993). Effect of language intensity similarity on perceptions of credibility, relational attributions and persuasion. Journal of Language and Social Psychology, 12, 224-237.

Bachman, D. P., Elfrink, J., & Vazzana, G. (2000). E-mail and snail mail face off in rematch. Marketing Research, 11, 10-15.

Bowers, J. W. (1963). Language intensity, social introversion and attitude change. Speech Monographs, 30, 345-352.

Bradac, J., Bowers, J. W., & Courtright, J. A. (1979). Three language variables in communication research: Intensity, immediacy, and diversity. Human Communication Research, 5, 257-269.

Bradac, J., Bowers, J. W., & Courtright, J. (1980). Lexical variations in intensity, immediacy and diversity: An axiomatic theory and causal model. In R. N. St. Clair & H. Giles (Eds.), The social and psychological contexts of language (pp. 193-223). Hillsdale, NJ: Erlbaum.

Buller, D. B., Borland, R., & Burgoon, M. (1998). Impact of behavioral intention on effectiveness of message features: Evidence from the family sun safety project. Human Communication Research, 24, 433-453.

Bullet, D. B., Burgoon, M., Hall J. R. Levine, N., Taylor, A. M., Beach, B. H., Melsher, M. A., Buller, M. K., Bowen, S. L., Hunsaker, F. G. & Bergen, A. (2000). Using language intensity to increase the success of a family intervention to protect children from ultraviolet radiation: Predictions from language expectancy theory. Preventive Medicine, 30, 103-114.

Burgoon, M. (1995). Language expectancy theory: Elaboration, explication, extension. In C. R. Berger and M. Burgoon (Eds.) Communication and social influence processes, (pp. 29-51). East Lansing, MI: Michigan State University Press.

Burgoon, M., Denning, V. P., & Roberts, L. (2002). Language expectancy theory. In J. P. Dillard and M. Pfau (Eds.). The persuasion handbook: Developments in theory and practice, (pp. 99-138). Thousand Oaks, CA: Sage.

Burgoon, M., Dillard, J. P., & Doran, N. (1984). Friendly or unfriendly persuasion: The effects of violations of expectations by males and females. Human Communication Research, 10, 283-294.

Burgoon, M., Jones, S. B. & Stewart, D. (1975). Toward a message-centered theory of persuasion: Three empirical studies of language intensity. Human Communication Research, 1, 240-256.

Burgoon, M. & King, L. B. (1974). The mediation of resistance to persuasion strategies by language variables. Human Communication Research, 1, 30-41.

Cook. C., Heath, F. & Thompson, R. L. (2000). A meta-analysis of response rates in web or Internet-based surveys. Educational and psychological measurement, 60, 821-836.

Couper, M. P., & Nicholls, W. L. (1998). The history and development of computer assisted survey information collection methods. In M. F. Couper, R. P. Barker, J. Bethlehem, C. Z. F. Clark, J. Martin, W.L., Nicholls, & J. M. O'Reily (Eds.). Computer assisted survey information collection, (pp. 1-22). New York. John Wiley & Sons, Inc.

Cronkhite, G. (1969). Persuasion: Speech and behavior change. Indianapolis: Bobbs-Merrill.

Dillard, J. P. (1991). The current status of research on the sequential-request compliance techniques. Personality and Social Psychology Bulletin, 17, 283-288.

Groves, R. M, Cialdini, R. B. Cooper, M. P. (1992). Understanding the decision to participate in a survey. Public Opinion Quarterly, 56, 475-495.

Hamilton, M. A. & Hunter, J. E. (1998). The effect of language intensity of receiver evaluations of message source and topic. In M. Allen & R. W. Preiss (Eds.), Persuasion: Advances through meta-analysis (pp. 99-138). Cresskill, NJ: Hampton Press.

Hamilton, M. A., Hunter, J. E., & Boster, F. J. (1993). The elaboration likelihood model as a theory of attitude formation: A mathematical analysis. Communication Theory, 3, 50-65.

Hamilton, M. A., Hunter, J. E. & Burgoon, M. (1990). An empirical test of an axiomatic model of the relationship between language intensity and persuasion. Journal of Language and Social Psychology, 9, 235-255.

Hamilton, M. A., & Stewart, B. L. (1993). Extending an information-processing model of language intensity effects. Communication Quarterly, 41, 231-246.

Krosnick, J. (1999). Survey research. Annual Review of Psychology, 50, 537-567.

Litvin, S. W. (2001). E-surveying for tourism research: Legitimate tool for a researcher's fantasy? Journal of Travel Research, 39, 308-314.

Manfreda, K. L. Batagelj, Z., & Vehovar, V. (2002). Design of web survey questionnaires: Three basic experiments. Journal of Computer Mediated Communication, 7, http://www.ascusc.org/jcmc/vol7/issue3/vehovar.html

Miller, G. R. & Burgoon, M. (1978). Persuasion research: Review and commentary. In B. D. Ruben (Ed.). Communication Yearbook (Vol. 2, pp 29-47). New Brunswick, NJ: Transaction.

Parker, L. (1992). Collecting data the email way. Training and Development (July), 52.

Ramos, M., Sedivi, B. M., & Sweet, E. M. (1998). Computerized self-administered questionnaires. In M. P. Couper, R. P. Barker, J. Bethlehem, C. Z. F. Clark, J. Martin, W. L. Nicholls, & J. M. O'Reilly (Eds.) Computer assisted survey Information collection. (Pp. 389-401). New York, John Wiley & Sons, Inc.

Sheehan, K. B., & Hoy, M G. (1999). Flaming, complaining, abstaining: How online users respond to privacy concerns. Journal of Advertising, 28 (3), 37-51.

Sheehan, K. B. & McMillan, S. J. (1999). Response variation in email surveys: An Exploration. Journal of Advertising Research, 39, 45-54.

Vehovar, V., Batagelj, Z., Manfreda, K. L. & Zalarel, M. (2002). Nonresponse in web surveys. In R. M. Groves, D. A. Dillman, J. L. Eltinge & R.J.A. Little (Eds.) Survey Nonresponse, (pp. 229-242) New York: Wiley.

Yun, G. W. (2000). Comparative Response to a Survey Executed by Post, E-mail, & Web Form. Journal of Computer Mediated Communication, http://www.ascusc.org/jcmc/vol6/issue1/yun.html

Peter A. Andersen (Ph.D., FLorida State University, 1975) is a Professor of Communication At San Diego State University. Tammy R. Blackburn (M.A., San Diego State University, 2001) is the Associate Director SDSU Alumni Association.

An earlier version of this manuscript was presented at the Western States Communication Association Convention, Long Beach, CA, February 2002