For what purposes do people use their laptops during meetings? Recent press articles and academic studies describe a pervasive tendency to use computers for multiple purposes during lectures and workplace meetings.2,3,4 Participants may use their laptops to take notes in electronic format, follow presentation slides at their own pace, or look up related information on the Internet while the meeting is in progress to enhance their acquisition and processing of information.1 These are compliant uses because they are related to the meeting's objectives. Participants, however, may use their laptops to handle email, play computer games, or review unrelated documents. These are distracting uses because they indicate deviation from the meeting's objectives. Since computers allow users to engage in more than one task concurrently, participants' laptop use during meetings may combine compliant and distracting uses. A better understanding of the prevailing nature of computer-based multitasking in the context of professional meetings is necessary. Such understanding would help establish effective management responses to control multitasking and leverage it for the purposes of the meeting. It will also inform current debates over multitasking's appropriateness in terms of social mores and its efficacy in terms of performance implications.1,3
Multitasking consists of performing at least two tasks at the same time. Our research interest lies with computer-based multitasking behavior that occurs in a face-to-face (F2F) meeting context, particularly on assessing compliant and distracting uses in such context. Thus, the purpose of this study is to empirically examine the type and level of computer-based multitasking activity for a group of subjects who attend face-to-face meetings. Based on Wasson,6 we expect that participants' computer-based multitasking levels would vary across meeting activities as mediated by attention requirements (Figure 1).
We collected objective data from computer monitoring logs of laptop-equipped participants who were attending lecture meetings of an undergraduate interdisciplinary core course at a small private business college. Since the focus of the study is on computer-based multitasking, we concentrate our analyses on subjects who were using their laptops for multiple concurrent uses, rather than those who chose not to use their laptops.
A monitoring program was installed on the laptops of the subjects who agreed to participate. The software unobtrusively logged computer use only during each 80-minute session of the course. Laptop usage outside the lecture period was not logged by the software. Recording occurred in stealth mode, and the log file was periodically copied from each subject's computer to a file server by a "harvest" program that operated inconspicuously.
Research procedures were designed to mitigate the potential influence of monitoring awareness. Participants' answers to a survey, administered after the monitoring period finished, suggested little influence of monitoring awareness. Eighty-four percent of participants disagreed with the survey item"I changed how I used my computer because I knew that I was being monitored."
We collected monitoring data on 67 subjects (28% of 240 attendees) and examined their computer-based multitasking behavior over twenty-eight meeting sessions. We measured multitasking activity by counting the number of times the user changed between active windows of open (resident) applications during the meeting sessions. This variable measures task switching levels among concurrent computer-based activities.
Obtaining the multitasking measure required three steps. First, each computer-based activity was coded by two independent raters with respect to task-type (inter-rater reliability 82%). Task-types included on-task (if activity is related to the meeting), off-task (if unrelated), and neither. For example, an activity was coded as on-task, if the participant used the web browser to look up information about "Break Even Formulas" during the class session about this topic. Alternatively, an activity was coded as off-task, if the participant looked up information about "Spring Break Destinations." When an activity was system-generated, or there was not enough information to evaluate its nature, it was coded as neither.
After coding, a custom software program assigned task switching types based on task-type code changes between two consecutive monitoring log records. There are nine possible switching types, given the three task-type codes (Figure 2). The third step involved classifying switching types into four broader categories of multitasking: distracting (off-to-off), compliant (on-to-on), oscillating (off-to-on and on-to-off) and indeterminate (others). Typical distracting task switching involved changing the active window from an IM window to an email window, or between two IM windows, with correspondence unrelated to the topic of the meeting. A common compliant task switching involved concurrently taking notes in a word-processing window while reviewing slides at a self-regulated pace in a presentation window.
An analysis of multitasking shows that 76% of computer-based task switching centered on distracting activities while only 13% focused on compliant activities (Figure 3). This finding is consistent with previous anecdotal and survey-based reports about a preponderance of computer use directed towards personal or leisure interests during face-to-face meetings.2,4 Oscillating multitasking (switching between two computer-based activitiesone related to the meeting and one not) accounted for about 8% of multitasking, and 3% of activity was indeterminate. A statistical test of differences between proportions shows that the percentage of distracting multitasking is significantly higher than all the other types combined (p<.0001).
Further analyses across participants and sessions yielded additional insights. For instance, participants switched between computer-based activities about 37.5 times on average per meeting session. Although switching level was highly variable across subjects (32.7 standard deviation), only one percent of our observations-22 out of 1,876 data points (67 subjects * 28 sessions)-correspond to instances of single-tasking (i.e., computer usage for a single task during a session).
Additionally, there were wide fluctuations in task switching levels of all types across sessions. For instance, while there were noticeable peaks and valleys in distracting multitasking throughout most of the monitoring period (28 sessions), there was a gradual falling off and comparative stability of distracting multitasking activity during the latter sessions. Compliant multitasking also fluctuated, but unlike distracting multitasking, it was practically non-existent in some sessions. As a result, oscillating multitasking occurred at a much lower level as well.
We infer from these results that meeting activities, and other meeting characteristics such as content and pace, create varied demand for attention, which then influences both the types and levels of participants' computer-based multitasking activities. For instance, slow paced meetings for information sharing may generate more distracting multitasking activity than fast paced ones. It is also possible that during periods that create need for compliant activities, subjects are tempted to concurrently engage in distracting activities because they can conceal them with compliant activities.
Since this research was conducted in an academic setting with subjects attending lecture meetings, any generalization to a workplace environment should be cautioned. However, given the pressure to increase employee productivity at all levels and to balance competing demands for time, there is reason to expect that these types and levels of multitasking activities occur at workplace meetings as well.
Our results are helpful to advance managers' understanding of computer-based multitasking behavior during meetings. Using objective data from actual computer usage, instead of participants' self-reports, this study provides evidence for a prevalence of distracting multitasking activities and the limited existence of compliant ones. In light of these findings, we present the following recommendations for managers.
2. Chudoba, K.M., Wynn, E., Liu, M. and Watson-Maneim, M.B. How virtual are we? Measuring virtuality and understanding its impact in a global organization. Information Systems Journal 15, (2005), 279306.
5. Truman, G. E. The Effectiveness of Classroom Control Systems in a Traditional Learning Environment. IV International Conference on Multimedia and Information and Communication Technologies in Education, (Seville, Spain, 2006)
©2009 ACM 0001-0782/09/0200 $5.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2009 ACM, Inc.