Examining the differences in gender perception in the use of speech recognition as a tool in group support systems.

Rebman, Carl M., Jr. ; Cegielski, Casey G. ; Prince, Brad 等

ABSTRACT

Group Support Systems (GSS) technology is recognized as a tool with advantages for facilitating, improving, and speeding communication between the different members in a meeting session. One innovation that could enhance the efficiency and effectiveness of GSS technology is speech recognition. While the concept of human speech interaction with computer-based information systems is not novel, Speech Recognition (SR) technology presents an opportunity to reduce the challenges in human-computer interaction. One of the advantages that Speech Recognition offers is the ability to capture a larger amount of text/data over traditional keyboard entry. This is of interest as one of the limitations of GSS performance and end user satisfaction has been attributed to keyboard-based comment entry challenges. Simply put, people talk and think faster than they can type. This paper reports the results of an experiment using a prototype voice-based Group Support System. Fourteen groups of five subjects each submitted speech-synthesized comments into the system and read others' comments visually on computer screens. This prototype seeks to maximize the relative efficiencies of speaking and reading to generate the most ideas in the shortest amount of time while maintaining accurate transcriptions. Results of this study illustrated some interesting differences between genders. First, male subjects reported spending too much time on idea generation while female subjects reported a higher level of understanding of decisions made. Females also reported a higher level of satisfaction with both the decision outcome and the process. These results provide evidence that Speech Recognition may be a viable tool for decision-making processes where gender issues play an important role.

INTRODUCTION

As organizations seek to engage in methods to improve collaboration and communication methods they must also analyze societal factors that can either advance or impact end-user adoption. One such factor that can play a major role in end-user adoption is gender. According to Nicovich, et. al, "the role of gender in reaction to communication stimuli has been of interest to both academics and practitioners for some time (Darley, 1995; Holbrook, 1986; Meyers-Levy & Maheswaran, 1991; Meyers-Levy & Sternthal, 1991; Prakash & Flores, 1984). Gender is often used as a segmentation variable in communication strategies, and interesting differences have been noted in academic literature over the past few decades." Furthermore, according to Wilson and Howcraft, "Within the IS literature, the issue of gender is largely under-theorized, partly due to the belief in the gender neutrality of technology" (Knights & Murray, 1994). Lastly, because there have been some reports of negative experiences with technology by women (Adam, 1997) it is important that research be conducted in the introductory stage of any new technology. This paper seeks to examine the viability of speech recognition as a new tool for current group support systems and more particularly its effects on gender acceptance.

OVERVIEW OF GROUP SUPPORT SYSTEMS

Group Support Systems (GSS) were originally envisioned by Huber (1984a), who extended the decision support systems (DSS) concept to group decision settings (Rebstock, et al., 1997). A GSS is a computer-based information processing system designed to facilitate group decision-making. It is a group technology, which represents a networked computer-based system that combines computing, communication, and decision support technologies (Aiken & Chrestman, 1995). A GSS supports a group of people involved in a common task through a shared environment. Its design is centered on groups and group productivity through idea generation, preference, and opinion exchange (Aiken & Chrestman, 1995; Aiken and Hassan, 1996). Group Support Systems allow for generating new ideas, developing new plans, and facilitating new research efforts (Anson, 1996). Overall group members in electronic meetings participate more, save more time, and are more satisfied than participants in traditional, verbal meetings (McLeod 1992).

A GSS typically includes: (1) a communicative interface component providing computer-mediated channels, (2) a decision modeling component-suggesting solutions, and (3) a documentation component recording the decision process (Pollack and Kanachowski, 1993; Alavi and Joachimsthaler, 1992, Huber, 1984b). In the past decade, numerous group support systems have been developed to support planning, brainstorming, negotiating, problem solving, and other decision making processes utilizing a variety of technologies (Teng and Ramamurthy 1993; Dennis et al., 1988; Huseman and Miles, 1988; Johansen, 1988; Kraemer & King, 1988; DeSanctis and Gallupe, 1987; Huber, 1984b). Improved computer-based information processing technologies and increased computer literacy have led to a constant impetus for further GSS development (Broome and Chen, 1992; Huber, 1984a). Consequently, GSS systems vary greatly today as a GSS can mean anything from electronic mail systems to sophisticated decision rooms and nearly anything in between (Rebstock, et al., 1997).

CLASSIFICATION OF GROUP SUPPORT SYSTEMS

Group Support Systems facilitate group decisions by allowing groups to meet in a single room or at remote sites, and the computer replaces chalkboards and projected video images. The interactive computer system has specifically designed hardware and software, which are designed to be easy to learn and use (Aiken, 1992). There are four possible GSS settings (Turban, 1993): Same Time-Same Place: Participants meet in the decision room. They are face-to-face, in one place at the same time. A brainstorming or voting session demonstrates this type of setting.

Same Time-Different Place: Participants are communicating at the same time, but they are in different places. This setting describes a meeting where participants are indifferent locations in the city.

Different Time-Same Place: Participants work in shifts in this type of setting. International trading is an example of this setting.

Different Time-Different Place: Participants are in different time zones or countries. A company may be located in Germany and its subsidiary is in Canada.

There are also three different types of GSS environments (Turrof et al., 1993):

 Single-computer systems are the simplest types of group support
 systems designed primarily for single users that are dispersed
 geographically. Each group member uses a video display system
 through which communication with other group members takes place.
 Members can observe how other group members approach problems and
 participate by presenting their ideas. Single-computer systems are
 relatively inexpensive, portable, and suitable for problems whereby
 confidentiality is not important (Gupta, 1996). One problem
 suitable for single-computer systems is making investment decisions
 based on standard weighted criteria. Group members can assign
 individual weights to the criteria, the system ranks the various
 alternatives, and members rank the criteria. Changes are suggested,
 while assigning and prioritizing the weights takes place until
 consensus is attained (Gupta, 1996).

 Keypad-response systems are systems where group members communicate
 with each other using hand-held keypads. In this type of system,
 group members are linked by networked personal computers. They are
 in a room referred to as the decision room. A projection screen at
 the front of the room is visible by all members in the decision
 room. All inputs, comments, or ideas are displayed on the screen.
 Each member uses the keypad to coordinate and communicate ideas
 with each other. The keypad has customized rating scales capable of
 processing, analyzing, ranking, and displaying the output, in text
 or graphical form. The advantages of keypad systems are anonymity,
 portability, and instantaneous summarization of input (Gupta,
 1996).

 Full-keyboard workstation systems are similar to the
 keypad-response systems, where a full-keyboard workstation system
 is also built in a single room containing a projection screen and a
 series of personal computer workstations networked through a
 facilitator workstation. Each workstation has a private display
 monitor. Because group member responses are displayed on the
 projector screen, as well as on the participant's monitors,
 anonymous communication is maintained. The system design is
 adaptable to a wide array of uses and informational requirements.
 The tools utilized for generating, evaluating, and ranking ideas
 result in efficient group sessions because time is not lost in
 tabulations, and because participants can monitor the progress of
 their work (Gupta, 1996; Aiken and Chrestman, 1995).

SPEECH RECOGNITION

Automated speech recognition technology has been around since the 1950's, but the trade press has begun only recently to focus attention on it because of recent advances in the quality of software, and personal computers have become fast enough to process it in real time (Lindquist, 1999; Miatkowski, 1999). SR technology digitizes spoken words, identifies individual sounds (phonemes), and uses mathematical models to selected discrete words or complete phrases.

Speech recognition can be used to enhance dictation by eliminating keyboard input and increasing productivity. Similarly, it can be used in telephony to replace cumbersome touch-tone menus or to reduce the number of screen prompts required to reach the desired person. Lastly, SR systems can be used to control machinery or recording inspection, allowing hands-free operation and greater concentration on more important tasks.

SPEECH RECOGNITION TECHNOLOGY

Speech Recognition can be categorized in terms of speaker input and speech mode (Rodman, 1999).

 Speaker Dependent or Independent Speaker dependent SR technology
 requires a user to train the program to recognize his or her voice
 (the process is referred to as "enrollment"), but this type of SR
 may be better for those with non-standard speaking (e.g., patterns,
 dialects, or foreign accents) (Rabiner, 1993). Speaker-independent
 programs are designed to interpret any user's voice with no
 enrollment, but this type of software is usually less accurate. If
 the software is speaker-independent, a default set of discrete
 sounds or phonemes is provided, otherwise, user enrollment user
 creates a personalized set of phonemes for improved accuracy.

 Continuous or Discrete Speech Continuous speech SR allows the user
 to talk normally, in complete sentences, with no breaks between
 words, while discrete speech SR requires speakers to pause after
 each word. Continuous speech usually is considered to be more
 natural, less frustrating, and faster. In addition, being more
 complex, continuous speech SR can recognize individual words
 (discrete speech) as well as entire phrases. Although some studies
 have suggested that discrete speech SR is more accurate, others
 have shown the opposite to be true (Klevans, 1997; Niccolai, 1997).

The speech recognition process follows five steps (Markowitz, 1996):

 Audio Input The human voice is transmitted through a microphone
 connected to a PC with a standard sound card.

 Acoustic Processor The acoustic processor filters out background
 noise and converts captured audio into a series of phonemes.

 Word Matching The software attempts to match the sounds to the most
 likely words in two basic ways. First, it uses acoustical analysis
 to build a list of possible words that contain similar sounds.
 Then, it uses language modeling (the likelihood that a given word
 will appear between those coming before and after it) to narrow the
 list and come up with the best candidates. In addition, the word
 matching process draws on the domain (the set of vocabularies,
 pronunciations, and word-usage models, as well as a model of the
 user's speech and words) employed by the application. The user can
 extend the domain by adding new words or even creating multiple
 domains for different applications. Finally, continuous speech SR
 looks at contextual information to predict what words should come
 next. This also helps the system to distinguish homonyms.

 Decoder The decoder selects the most likely word based on the
 rankings assigned during word matching and assembles the words in
 the most likely sentence combinations. It then transfers the
 sentence to the word processing application.

 Text Output Some SR programs include their own word processors, but
 many also will allow text generation directly into a word processor
 or a text box in another application.

Several SR systems have been compared in the literature, and accuracies ranging from 80% to 99% have been reported (Alexander, 1999; Bethoney, 1999; Jang & Hauptmann, 1999; Wactlar, et al., 2000). Tests of Dragon Systems' Naturally Speaking, one of the leading SR program for personal computers, showed an accuracy of 89.1% at 74 wpm as compared to 87.8% at 69 wpm with IBM's ViaVoice SR software. Other tests of DragonSystems showed that text could be generated at 120 wpm with word accuracies of 80% to 95% (Pallett, Garofolo, & Fiscus, 2000).

In a comprehensive study of DragonSystems (Aiken, Wong, & Vanjani, 2001), the absolute accuracy was about 79% while an adjusted accuracy accounting for reader comprehension was 91%. That is, even though some words were transcribed incorrectly, an independent reviewer could correctly guess what was meant for many of the errors.

RESEARCH QUESTIONS

This study examines five issues related to the successful implementation of SR technology in a GSS with a particular focus on gender adoption. Specifically, the issues addressed in this research are: 1) background noise, 2) user training, 3) comment submission, 4) comment quality, and 5) transcription accuracy.

How does SR technology impact comment submission between genders in a GSS setting?

Many executives, in particular, are poor typists, and therefore, their productivity may suffer when they are restricted to using a keyboard (Aiken, et al., 2000). Given that many executives are male this question seeks to determine if SR technology can increase the ease of entering comments into the GSS application.

H1: Male subjects using a SR GSS system will report a higher ease-of-use than females using a GSS system.

How does SR technology impact GSS meeting satisfaction between genders?

Groups in GSS meetings utilizing keyboard input have reported higher satisfaction than group in verbal meetings. This question seeks to determine if SR technology can produce even higher satisfaction.

H2: Male subjects using an SR GSS will report higher satisfaction than their female counterparts.

Does SR Technology help females to express themselves better?

Most people speak faster than they can type. Past GSS studies have shown a higher number of total comments generated over verbal meetings. This question seeks to investigate the impact of SR input on total comment generation in the GSS application.

H3: Male subjects using a SR GSS system will generate more comments than female subjects.

How does SR technology impact conflict/non-task behavior between genders in a GSS Setting?

One of the drawbacks to any group decision-making process is the amount of irrelevant or flaming comments made that are not germane to the topic at hand. Prior research has shown that in an anonymous keyboard GSS session flaming increases over that of a face-to-face meeting. This question seeks to determine if SR technology can decrease the number of irrelevant comment submission and increase comment quality.

H4: Male participants using a SR GSS system will generate less irrelevant and flaming comments than female participants using a keyboard GSS system.

How does SR technology impact comment accuracy in a GSS setting?

One of the challenges with any transcription method is the number of errors or misspelled words that limit the comprehension of the comment. This question seeks to determine if SR technology can more be more accurate in the elimination of misspelled words.

H5: Male subjects using a SR GSS system will generate less misspelled words and have an overall higher accuracy than their female counterparts.

EXPERIMENTAL STUDY

A group of 70 Business school students participated in the experiment for extra credit. Five groups of seven each used electronic gallery writing to type comments while five groups of seven each used DragonSystems Naturally Speaking SR software to generate the text. The DragonSystems software is a speaker-dependent program that accepts continuous and discrete speech and increases its speech recognition with each use of the software. An a-priori power test indicated a level of .80 for the study (Cohen, 1988). Of the total 70 subjects, 44 were male, 26 were female, 16 were seniors, 41 were juniors, and 13 were sophomores. All participants were trained in a specialized computing facility that minimized background interference.

All subjects were provided with an explanation of the purpose of the study and each participant spent 20 minutes in the training phase of the software. The training time was determined based on the results of previous research (Rebman, 2001). The subjects also spent 10 minutes discussing promoting tourism, a topic used in many prior experiments. All comments were automatically recorded in a separate file with user number, group number, and time written. In addition all comments were anonymous save for the possibility of hearing others speak. Upon completion of the training exercise each participant was asked to vote and select the format of his or her final exam. Following the meeting the subjects completed a questionnaire that was adopted from previous voice-GSS studies (see Appendix A).

RESULTS

Table One presents the summary statistics between the male and female participants. Overall, there were very little significant differences in gender. Actually, there were no significant differences between males and females in the keyboard session, yet there were a few noticeable differences between males and females in the voice sessions. It appeared that both males and females had approximately the same perceived efficiency, and ease of use of the GSS application. The few items that came close to being statistically significant were keyboard experience (females reported a higher number of hours of keyboard experience), and females found the voice GSS system a little easier to submit comments. Two items that were statistically significant were satisfaction of the process and outcome, and importance of decision, which is illustrated in Table 2. Interesting enough, females in the voice groups reported that not enough time was spent on generating ideas. T-Tests were conducted on 46 variables in both voice and keyboard groups and only 6 variables were found to be statistically significant. These differences are only applicable to male and female subjects in the voice sessions and they are listed and discussed below; there was no significant difference between male and female keyboard participants.

Idea generation--Male subjects reported feeling that they had spent too much time on idea generation during the group discussion as opposed to the female participants.

Group understanding of decisions made--Females reported a higher level of understanding of the decisions made.

Consensus on Decisions made--Females reported a stronger impression that they felt the group decision were made by consensus as opposed to some other type of decision-making style.

Satisfaction with Outcome--Females reported a much higher satisfaction level (5.88) then their male counterparts (4.72).

Importance of Decision--Females apparently felt the decision they were engaged in was more important than the males.

Satisfaction with Process--In addition to satisfaction with the outcome, females also reported higher satisfaction with the decision-making process.

LIMITATIONS

This study made attempts to analyze gender differences, however the study was conducted with unequal sample sizes. Although, the sample sizes were different, it did become evident that gender differences did exist in the speech-computer interface. More research is required to determine the full depth and implications of these differences. Likewise, research is necessary to determine other demographic differences stemming from ethnicity and age.

CONCLUSIONS

For several decades, system developers intensely explored opportunities to integrate speech recognition into the human-computer interface. Due to previous inefficiencies in the technology, the promise of SR in the human-computer interface has yet to be fully realized. However, evolutions in speech recognition software during the last 10 years again have prompted systems developers and researchers alike to examine the feasibility of speech recognition technology as a method of human-computer interface. Although the latest evolutions of SR technology provide greater technical opportunities for system integration, there are still many human user variables associated with the technology that remain to be examined. We offer the findings presented in this paper as an initial point of investigation from which researchers may begin to explore the human user variables associated with speech recognition. Gender, as used in this paper, is only one user variable that requires examination in the context of speech recognition-enabled systems. Additionally, researchers should consider the larger scope and dynamic relationship among other human user variables both on an individual and a group level and examined these variables in similar contexts used in other studies. Speech recognition technology and SR research must move beyond the technical assessments of the systems and focus upon the users of the technology.

APPENDIX A EXPERIMENTAL QUESTIONNAIRE Overall ease of use

1. In general using the GSS program was

1 Extremely difficult 2 Difficult 3 Fairly Difficult 4 Neither easy or difficult 5 Fairly easy 6 Very easy 7 Extremely easy

2. Overall, compared with traditional meetings (verbal), using the GSS program appears to be

1 Very much easier 2 Easier 3 A little easier 4 About the same 5 A little harder 6 Harder 7 Very much harder

Perceived Efficiency

1. My impression of the GSS program is that it is

1 Extremely inefficient 2 Very inefficient 3 Inefficient 4 Neither efficient nor inefficient 5 Efficient 6 Very efficient 7 Extremely efficient

Previous Keyboard Experience

1. I estimate that before today's session my keyboard experience was

1 More than 1,000 hours 2 500 to 1,000 hours 3 250 to 499 hours 4 100-249 hours 5 50 to 99 hours 6 1 to 49 hours 7 0 hours

GSS Questions

Please Circle your answers below.

1. I was afraid that others in my group would criticize my comments.

1 2 3 4 5 6 7
Disagree Neutral Agree

2. It was easy to submit and read comments.
1 2 3 4 5 6 7
Disagree Neutral Agree

3. I know the members of my group well.

1 2 3 4 5 6 7
Disagree Neutral Agree

4. I was able to express my opinion easily.

1 2 3 4 5 6 7
Disagree Neutral Agree

5. I prefer this type of meeting to a traditional, verbal meeting.

1 2 3 4 5 6 7
Disagree Neutral Agree

6. The program was able to accurately print out (on the screen) my
comments.

1 2 3 4 5 6 7
Disagree Neutral Agree

7. I was able to read and understand everyone's comments.

1 2 3 4 5 6 7
Disagree Neutral Agree

REFERENCES

Adam A (1997) 'What should we do with cyberfeminism?' Women in Computing Intellect Books, Exeter

Aiken, M.; Rebman, C.; and Paolillo, J.; "Lessons Learned with a Voice-based Group Support System: Proceedings of the 32nd Annual Decisions Sciences Institute, November 2001, San Francisco, California.

Aiken M.; Rebman, C.; and Vanjani, M.; "A Voice-based Group Support System," Proceedings of the 32nd Annual Southwest Decision Sciences Institute Conference, New Orleans, Louisiana, February 2001,

Aiken, M., Wong, Z., and Vanjani, M. (2001). An analysis of errors using a speech recognition system. Proceedings of the 32nd Annual Southwest Decision Sciences Institute Conference, February 2001, New Orleans, LA.

Aiken, M., Sloan, H., Paolillo, J., & Motiwalla, L. (1997). The use of two electronic idea generation techniques in strategy planning meetings. Journal of Business Communication, 34(4), October 1997, 370-382.

Aiken, M.; and Hassan, B. (1996). Total Quality Management: A group Decision Support System Approach, Information Systems Management, 13(1), January 1996, 73-76.

Aiken, M.; and Chrestman, M. (1995). Electronic Meeting Systems" Journal for Quality and Participation, 18(4), April 1995, 98-102.

Aiken, M., Vanjani, M., and Krosp, J. (1995). Group decision support systems. Review of Business, 16(3), Spring 1995, 38-42.

Aiken, M. Kim, D., and Singleton, T. (1994). Future developments of group decision support systems. 1994 Southeast Decision Sciences Institute Conference, Williamsburg, VA, March 1994.

Aiken, M. (1992). Using a Group Decision Support System as an Instructional Aid: An Exploratory Study. International Journal of Instructional Media, 19(4), April, 328-329.

Aiken, M., Liu Sheng, O., and Vogel, D. (1991). Integrating Expert Systems with Group Decision Support Systems. ACM Transactions on Information Systems, 9(1), January, 75-95.

Alavi, M.; and Jachimsthaler, E.(1992). Revisiting DSS implementation research: A meta-analysis of literature and suggestions for researchers, MIS Quarterly, 16(1) March, 95-116.

Alexander, S. (1999). Speech recognition. Computerworld 33(45), November 8, 1999, 65.

Anson, A. (1996). Distinguishing the effects of functional and dysfunctional conflict on strategic decision making: Resolving a paradox for top management teamsAcademy of Management Journal, 39(1) January 1996, 123-148.

Bethoney, H. (1999). Speech products: The talk of the town. PC Week, August 16, 1999, 5.

Broome, B.; and Chen, M. (1992). Guidelines for computer-assisted group problem solving Small Group Research, 23(2) February 1992, 216-236.

Darley, W. K., & Smith, R. E. (1995). Gender differences in information processing strategies: An empirical test of the selectivity model in advertising response. Journal of Advertising, 24 (1), 41-56.

Dennis, A.; George, J.; Jessup, L.; Nunamaker, J.; and Vogel, D. (1988). Information technology to support electronic meetings. MIS Quarterly 12(4) December 1988, 591-624.

DeSanctis, G.; and Gallupe, B. (1987). A foundation for the study of group decision support systems Management Science, 33(8) August 1987, 589-609.

Gupta, U. (1996). Management Information Systems. St. Paul, MN: West Publishing Company.

Holbrook, M. B. (1986). Aims, concepts, and methods for the representation of individual differences in esthetic responses to design features. Journal of Consumer Research, 13 (3), 337-347.

Huseman, R.; and Miles, E.(1998). Organizational communication in the information age: Implications of computer-based systems Journal of Management, 14(4), April, 181-204.

Huber, G. (1984a). A theory of the effects of advanced information technologies on organizational design, intelligence, and decision making Academy of Management Review (15(1) January 1984a, 47-71.

Huber, G. (1984b). Issues in the design of group decision support systems MIS Quarterly, 8(3) September 1984b .195-204. Jang, P. and Hauptmann, A. (1999). Learning to recognize speech by watching television. IEEE Intelligent Systems, 14(5), May 1999, 51-58.

Johansen, R. Groupware: Computer support for business teams. New York, NY. Free Press.

Jones, D.; Hapishi, K.; and Frankish, C. (1991). Automated Speech Recognition in Practice Behavior and Information Technology, March/April 1991, 47-52.

Klevans, R.; and Rodman, R. (1997). Voice Recognition. Boston, Artech House.

Knights D & Murray F (1997) Markets, managers, and messages: Managing information systems in financial services In: B P Bloomfield, R Coombs, D Knights & D Littler (eds.) Information Technology in Organizations: Strategies, Networks, and Integration Oxford: Oxford University Press

Kraemer, K.; and King, J. (1988). Computer based systems for cooperative work and group decision making ACM Computing Surveys, 20(2) February1988, 115-146.

Lindquist, C. (1999). Speak easy PC World, October 1999, .185-195.

Markowitz, J. (1996). Using Speech Recognition. Upper Saddle River NJ, Prentice Hall

McLeod, P. (1992). An assessment of the experimental literature on electronic support of group work: Results of a metaanalysis Human-Computer Interaction, July 1992, 257-280.

Miastkowski, S. (1999). Latest speech software gets you up and running faster. PC World, November 1999, . 63-66.

Meyers-Levy, J., & Maheswaran, D. (1991). Exploring differences in males' and females' processing strategies. Journal of Consumer Research, 18 (1), 63-70.

Meyers-Levy, J., & Sternthal, B. (1991). Gender differences in the use of message cues and judgments. Journal of Marketing Research, 28 (1), 84-96.

Niccolai, J. (1997). First speech-recognition email announced InfoWorld, November 24, 1997, . 57.

Nicovich, S. G., Boller, G. W., and Cornwell, T. B. (2005). Experienced presence within computer-mediated communications: Initial explorations on the effects of gender with respect to empathy and immersion. Journal of Computer-Mediated Communication, 10(2), article 6.

Nunamaker, J.F., Dennis, A.R., Valacich, J.S., Vogel, D.R., and George, J.F. "Electronic Meeting Systems to Support Group Work," Communications of the ACM, (34:7), July 1991, pp. 40-61.

Pallett, D.S., Garofolo, J.S., Fiscus, J.G. (2000). "Measurements in Support of Research Accomplishments". Communications of the ACM. (43:2_, February 2000, pp. 75-79.

Pollack, C.; and Kanachowski, A. "Application of theories of decision making to group support systems (GDSS)" International Journal of Human-Computer Interaction, (5:1) January 1993, pp. 71-94.

Pollard, C. Electronic meeting systems: Specifications, potential, and acquisition strategies. Journal of Systems Management, (33:3) May/June 1996, 33, pp. 22-28.

Prakash, V., & Flores, R. C. (1984). A study of psychological gender differences: Applications for advertising format. Advances in Consumer Research, 12 (1), 231-237.

Rabiner, L.; and Biing-Hwang, J. Fundamentals of Speech Recognition. Englewood Cliffs NJ, Prentice Hall, 1993.

Rebman, C. "An Exploratory Study of the Impact of Training Times on User Acceptance of Speech Recognition Systems," Proceedings of the 32nd Annual Southwest Decision Support Institute Conference, New Orleans, Louisiana, February 2001, pp.54-57.

Rebstock, S.; Williams, S.; and Wilson, R. "Group support systems, power and influence in an organization: A field study" Decision Sciences, (28:4), December 1997, pp. 911-937.

Rodman, R. Computer Speech Technology. Boston, Artech House, 1999.

Teng, J., and Ramamurthy, K. "Feedback as a source of control in decision support systems: Clarifying the concept and establishing a functional taxonomy" INFORMS, (41:3), March 1993, pp.166-185.

Turban, E. Decision Support and Expert Systems: Management Support Systems, 3rd Edition. New York, NY: MacMillan Publishing, 1993.

Turoff, M., Hiltz, R.S., Ahmed, N.F., and Bahgat, A.R. Distribution Group Support Systems. MIS Quarterly, (17:2), June 1993, pp. 399-405.

Wactlar, H., Hauptmann, A., Christel, M., Houghton, R., and Olligschlaeger, A. Complementary video and audio analysis for broadcast news archives. Communications of the ACM, (43:2), February 2000, pp. 42-47.

Wilson, M., and Howcraft, D. "Gender and User Resistance in Nursing Information Systems Failure" Working Paper Manchester School of Management http://www.sm.umist.ac.uk/wp/Papers/wp2013.htm

Carl M. Rebman Jr., The University of San Diego

Casey G. Cegielski, Auburn University

Brad Prince, Auburn University

Table 1

 Std.
Variable Gender Mean Deviation

Ease of Use Male 3.37 1.24
 Female 3.79 1.44
Tkcompar Male 5.24 1.35
 Female 5.29 1.12
Peffic Male 3.57 1.46
 Female 3.88 1.03
Keyexp Male 1.96 1.21
 Female 1.54 0.66
Afraid Male 1.76 1.35
 Female 1.63 1.31
Submit Male 3.41 1.82
 Female 4.33 2.24
Knowgroup Male 3.43 2.09
 Female 3.17 2.1
Express Male 4.72 2
 Female 5.08 1.82
Prefer Male 2.06 1.79
 Female 2.92 1.59
Accuracy Male 2.87 1.5
 Female 2.79 1.44
Reak ok Male 4.41 1.48
 Female 4.75 1.67
Ease of comm Male 2.80 1.36
 Female 3.08 1.61
Satisfied Male 3.41 1.64
 Female 3.88 1.65
Work in future Male 5.26 1.48
 Female 5.17 1.40
Importance of Dec Male 4.59 1.51
 Female 5.29 1.16
Satisfied w/Dec Male 5.61 1.39
 Female 6.21 1.18
Satified w/Process Male 4.52 1.64
 Female 5.58 0.97

Table 2
 Sig.
Variable t stat (2-tailed)

Ease of Use -1.22 0.23
Tkcompar -0.17 0.86
Peffic -1.03 0.31
Keyexp 1.86 0.07
Afraid 0.41 0.69
Submit -1.74 0.09
Knowgroup 0.51 0.61
Express -0.77 0.44
Prefer -0.74 0.46
Accuracy 0.21 0.83
Reak ok -0.83 0.41
Ease of comm -0.72 0.47
Satisfied -1.11 0.27
Work in future 0.26 0.79
Importance of Dec -2.17 0.03 *
Satisfied w/Dec -1.90 0.06
Satified w/Process -3.39 <.001 *

df=65, male=46, female=24

* significant at 0.05