Can
You Handle It?
Individual Management of Incoming
Information in the Workplace
James
Reffell
Sarah
Waterson
October 31, 2001
The purpose of this study is to identify whether individuals in the workplace feel overloaded with information and to characterize those who do and do not feel overloaded in terms of the amount and quality of information they take in, the sources from which they take in that information, and the strategies the use to handle that information, and other potentially related factors. In doing so, we hoped to identify whether information overload at the individual level should be a concern in the workplace, and, if so, to identify potential areas for addressing that overload. We hope to use these findings to better understand how businesses can aid individuals in managing their information.
We used an online survey clearinghouse to gather 166 responses to a 27 question survey over a five day period. With the survey, we attempted to ascertain:
We found that the majority of individuals in the workplace felt only somewhat overloaded with information, with smaller groups feeling either less than or more than somewhat overloaded. We also found that the majority of individuals reported receiving some or most of the information they needed to do their jobs. Frequency of feelings of information overload correlated negatively with frequency of receiving needed information, however, and while the frequent use of, high importance of, and frequent interruption by some specific information sources were linked with feelings of overload, the number of messages received did not. These findings collectively support the idea that quality and appropriateness of information received, rather than quantity, are linked to overload.
In addition we found that:
According to a 2000 Reuters study [3], 49% of managers feel that they are quite often or very frequently unable to handle the volume of information they receive. They also report that the amount of information required to execute their jobs effectively is large and growing, and they feel that much of the information they do receive is not important or unsolicited. In addition, 41% of managers agree that their working environment is extremely stressful on a day-to-day business.
How much of stress and overload is related to the information one receives at work? Does swimming around in large amounts of data – both useful and not useful information – lead to a sense of overload? There is a growing concern in the literature [4,5] and in the workplace about this sense of overload, whether or not it exists, where it comes from, and how to deal with it. However, in order to answer these questions, more work needs been done to identify the types and amount of information people deal with, how they receive and gather information, and what people do to manage it.
Pilot Study
Using online survey software from www.surveymonkey.com, we distributed two 43
question surveys consisting of a mix of quantitative, categorical and short
answer open-ended responses. These two surveys were nearly identical, except
for the questions regarding the amounts and types of email and voicemail. We
wanted to test two approaches, one asking the participant to quantify their
email and voicemail in the past day, and the other to describe their last 10
emails. We also wanted to know which questions we could remove to shorten the
survey and still get as much interesting data as possible. Six of the 43
questions were open-ended questions about the survey itself, asking for
feedback on the wording, confusing questions, etc. For each version, we had 5
trial participants, each taking approximately 20 minutes to fill it out
Based on the feedback from the pilot study, we decided to have users quantify
their email and voicemail for their last full workday, as that method proved easier
for people to calculate and seemed to give more useful data. We also found that
Instant Messaging (IM) was used surprisingly often by our sample population, so
we added questions about IM use similar to the Internet and Intranet questions.
Additionally, we reworded a number of questions for more clarity.
For the actual survey, we switched from surveymonkey.com to an online survey
clearinghouse, Zoomerang, which could gather an appropriate sample population
and distribute our survey. However, switching to Zoomerang required a number of
changes from our original design, as their software could not accommodate the
format for some questions that we initially desired and limited us to a 30
questions maxim. The combination of these two factors required considerable
modifications to our original survey. The final survey consisted of 27
questions, requiring quantitative responses, categorical and short answer
open-ended questions. Again, it took approximately 20 minutes to complete.
Survey Design
In designing our survey, we were attempting to both ascertain how overloaded by information professionals in the workplace felt and to identify those characteristics associated with the most overloaded individuals. Toward this end, we asked questions in the following areas:
Participants & Deployment
The survey that ran for 5 days from October 16th to October 20th, 2001. In this time period, we received 166 valid responses from Zoomerang. Zoomerang’s sample was randomly selected from individuals working at businesses that had some form of Internet connection.
Unless otherwise indicated, all significance results are reported at the two-tailed 0.05 level. All means are reported with standard deviation.
Demographics
Our sample population of 166 was nearly evenly split across gender, with 79 female and 84 males respondents. 3 respondents did not indicate a gender. The age of the respondents was fairly normally distributed across the categories. (See Figure 1). Education levels of our respondents indicate most had at least a Bachelor’s or professional degree (See Figure 2).
Figure 1: Age

Figure 2: Education Level

Source Use, Importance, Interruptions
In order to begin understanding how the different information sources people used related to overload, we looked at (1) the frequency of the source use, (2) the level of importance, and (3) the frequency in which a source interrupted somebody during their work day.
Frequency of Information Source Use
The first question of the survey asked the respondent to indicate the information source they typically use the most. Figure 3 shows that e-mail and the phone are used the most out of all 13 sources we offered as selections in our study.
Figure 3: Of the following information sources, which do you use the most for your job?

While we intended this question to be the selection of a single source from the list, an ooopsie on our part allowed the respondent to select more than one source. Figure 4, the index of information source use, shows the number of sources each respondent indicated. Looking at this chart, we are inclined to believe that some people interpreted this question as asking for only one source, while others interpreted the question to mean they should indicate all important sources.
Figure 4: Source Use Index

Information Source Importance
Each respondent was asked to rank the importance of each 13 information source for getting their daily work done. The five point Likert scale options were not important, somewhat important, important, very important, critical, and an n/a option. For analysis, we grouped the ‘n/a’ results with the ‘not important’ results. Figure 5 shows the mean responses for each source.
Figure 5: Information Source Importance.

In order to create a more general measure of many sources people considered important at their job, we created an index of source importance, summing the numerical value of each source for each respondent. Figure 6 shows the distribution of this index. The index ranges from 0 (all information sources were not important) to 65 (all information sources are critical).
Figure 6: Source Importance Index.

Information Source Distraction
We were also interested in how frequently people were distracted from their work by the information sources. Figure 7 shows how frequently people felt the information sources interrupted their work. The five point Likert scale used here was never (1), rarely, sometimes, often, and always (5).
Figure 7: Source Interruption frequency.

Similar to the information importance index, we created an information distraction index by summing the numerical value for each source for each respondent. Figure 8 shows the distribution of this index. The index ranges from 0 (no information sources were ever distracting) to 65 (all information are always distracting).
Figure 8: Source Interruption Index.

Correlating Source Use, Importance and Interruption
Frequency
Why create all of these indices? We are interested in knowing how these three factors, use, importance, and interruption relate to one another. All three factors highly correlate with one another, as can be seen in Table 1. In general, these correlations show that if you use an information source frequently, you find it important and frequently distracting, if you find an information source important, you both frequently use it and find it distracting, and if you find an information source distracting, you use it frequently and is important to you.
Figure 8: Source Interruption Index.
|
Index |
Statistic |
Use (Spearman) |
Importance
(Pearson) |
|
Use |
C |
|
|
|
Sig. |
|
|
|
|
N |
|
|
|
|
Importance |
C |
0.212** |
|
|
Sig. |
0.006 |
|
|
|
N |
165 |
|
|
|
Interruption |
C |
0.267** |
0.573** |
|
Sig. |
0.001 |
0.000 |
|
|
N |
163 |
163 |
*Correlations significant at the 0.05 level
(2-tailed).
**Correlations significant at the 0.01 level (2-tailed).
While the indices give a general feel for the how these aspects relate to one another, taking a closer look at the correlations of individual sources amongst the three categories reveals more interesting relationships. Correlating each source against itself for use, importance, and interruption followed the same trend as described for the indices – all correlate very strongly and positively with one another. What this fact also means is that none of the sources listed, was it frequently used and not found distracting or unimportant. Nor do people frequently use information sources that they don’t find useful.
Many of the sources also have a positive, significant, correlation with other sources in expected fashions, e.g. if you use the internet a lot and find it important, you use email a lot and find it important as well. The interesting correlations here, however, are the negative ones. Table 2 shows the significant negative correlations of source importance between sources – people who find paper important, do not find e-mail or IM important, and vice versa.
Table 2: Interesting Significant Correlations of Source Importance
|
Measures |
Pearson
Correlation |
Sig. (2-tailed) |
N |
|
|
Importance |
Importance |
|
|
|
|
Paper |
|
-0.259** |
0.001 |
153 |
|
Paper |
IM |
-0.204* |
0.022 |
126 |
*Correlations significant at the 0.05 level
(2-tailed).
**Correlations significant at the 0.01 level (2-tailed).
Use of email or IM does not correlate to use of paper significantly. However, use of email correlates negatively with finding paper important, as well as the converse of that: use of paper correlates negatively with finding email important. See Table 3. The use of paper does not correlate to finding IM or email distracting, nor does the use of email or IM correlate to finding paper distracting.
Table 3: Interesting Significant Correlations of Source Use to Source Importance
|
Measures |
Spearman Correlation |
Sig. (2-tailed) |
N |
|
|
Use |
Importance |
|
|
|
|
|
Paper |
-0.172* |
0.028 |
162 |
|
Paper |
|
-0.159* |
0.047 |
156 |
|
Paper |
Voicemail |
-0.225** |
0.006 |
147 |
*Correlations significant at the 0.05 level
(2-tailed).
**Correlations significant at the 0.01 level (2-tailed).
The use of paper also strongly correlates negatively to voicemail, so people that find voicemail important do not use paper. In addition, Table 4, shows significant negative correlations between the interruption frequency and importance for voicemail and paper.
Table 4: Interesting Significant Correlations of Source Interruption Frequency to Source Importance
|
Measures |
Pearson
Correlation |
Sig. (2-tailed) |
N |
|
|
Interruption |
Importance |
|
|
|
|
Voicemail |
Paper |
-0.158* |
0.046 |
160 |
|
Paper |
Voicemail |
-0.165* |
0.046 |
147 |
*Correlations significant at the 0.05 level
(2-tailed).
**Correlations significant at the 0.01 level (2-tailed).
More analysis needs to be done further identify relationships for use, importance and interruption across other variables, such as demographics, however a preliminary check shows very little other correlations of interest. This could, in part be due to the trouble in our survey with identifying most used information sources. The relationships and indices described in this section will be used in the next section to discuss how they relate to a sense of information overload.
We asked our respondents how often they felt overloaded with
information, on a 5-point scale (1=Never, 3=Sometimes, 5=Always). Our results
showed a mean frequency of feelings of information overload of 2.68 (S.D. = 0.98).
See Figure 9.
Figure 9: Distribution of Reported Frequency of Feelings of Information Overload

We tested this distribution using a Chi-square test, and found it significant (x2=78.319, S =0.000). This distribution shows a slight negative skew, as more individuals reported feelings of overload less than sometimes than more than sometimes. Feelings of information overload exist, but do not dominate in the workplace.
We first looked at how indices reflecting the overall usage, importance, and interruption levels for all information sources related to the frequency of feelings of overload. The usage and interruption indices correlated with information overload at the 0.05 level, while the importance index did not. This indicates that feelings of information overload were sensitive to overall usage of and frequency of interruption by the various information sources, but not to the importance ascribed to them. The three indices do correlate to each other, however, showing that the overall use of, importance of, and interruption by the various sources are related.
Table 5: Correlation between frequency of information overload and total usage, importance, and interruption indices
|
|
|
Usage Index |
Importance Index |
Interruption Index |
|
Information Overload |
Correlation(Pearson) N |
0.174* 163 |
0.120 163 |
0.186* 161 |
* Correlation is significant
at the 0.05 level (2-tailed).
Information Overload
& Information Sources (Specific)
We next investigated the relationship between the individual information sources and feelings of overload, to see if particular sources were associated with feelings of information overload.
We found that only four specific information sources—email, voicemail, database, and phone—correlated to information overload in terms of use, importance, or frequency of interruption. For these sources, all measures except amount of use of voicemail and frequency of interruption by databases correlated to overload. It seems likely that the bulk of the correlations found between information overload and the use, importance, and interruption indices are due to these four information sources.
Table 6: Correlation between
frequency of feelings of overload and use, importance, and interruption
measures for specific information sources
Measure |
Source |
Correlation |
Significance |
N |
|
|
|
(Spearman) |
|
|
|
Use |
|
0.161* |
0.040 |
163 |
|
Use |
Voicemail |
0.085 |
0.280 |
163 |
|
Use |
Database |
0.190* |
0.015 |
163 |
|
Use |
Phone |
0.163* |
0.038 |
163 |
|
|
|
(Pearson) |
|
|
|
Importance |
|
0.243** |
0.002 |
154 |
|
Importance |
Voicemail |
0.170* |
0.040 |
146 |
|
Importance |
Database |
0.280** |
0.001 |
138 |
|
Importance |
Phone |
0.187* |
0.018 |
161 |
|
Interruption |
|
0.236** |
0.003 |
161 |
|
Interruption |
Voicemail |
0.211** |
0.008 |
160 |
|
Interruption |
Database |
0.071 |
0.383 |
161 |
|
Interruption |
Phone |
0.186* |
0.018 |
161 |
** Correlation is significant at the 0.01 level (2-tailed).
* Correlation is significant at the 0.05 level (2-tailed).
Information Overload
and Demographic Factors
We next looked at how feelings of information overload were distributed by various demographic measures, including gender, age, job category, and level of education attained. We found no significant differences except in the case of gender, where we found that women reported feeling overloaded by information at a slightly higher rate than men:
Table 7: Information Overload by Gender (Equal variances assumed T-test)
|
|
N |
Mean |
Std. Dev. |
t |
Sig. (2-tailed) |
|
Female |
78 |
2.85 |
0.97 |
2.110 |
0.036 |
|
Male |
84 |
2.52 |
0.98 |
2.110 |
0.036 |
(Overload frequency:1=Never,
3=Sometimes, 5=Always)
Information Overload and Incoming Information
We looked at a number of measures of incoming information from specific information sources (email, voicemail, Internet, Intranet, and instant messaging). We found no relationship between the total number of emails received in a day and information overload, but when we looked at the number of emails received in various categories, we found a positive correlation between the number of internal emails received and frequency of feelings of overload. (C=0.195, S=0.015, N=155) This could indicate that intracompany communication by email is more of a factor than intercompany communication information overload.
We also looked at the frequency with which people reported checking their email and voicemail and using the Internet, Intranet, and instant messaging. Two of these measures correlated with feelings of information overload: the frequency of checking voicemail (C=0.159, S=0.044, N=161) and the frequency of using the Internet (C=0.171, S=0.030, N=161).
In terms of work outside the traditional office and office hours, we found that of information overload positively correlated with both the amount of time people spent on work outside the office or normal workday (C=.158, S=.045, N=162) and the number of days a week they spent telecommuting (C=.164, S=.037, N=162). When we looked at telecommuters (those who reported telecommuting one or more days a week) and non-telecommuters separately, however, both correlations disappeared for both groups. This suggested the possibility that most of the increase in information overload was associated with the difference between telecommuters and non-telecommuters, but an independent samples T-Test did not find a significant difference in the mean information overloads of the two groups (t=-1.704, sig. (two-tailed)=0.90, mean difference= -0.28).
Finally, we found no significant relationship between information overload and the information sources people most wished to rid themselves of or information overload and the sources used to find job-related information in specific categories.
As well as measuring the frequency that people felt overloaded with information, we also measured how frequently they felt they were receiving all the information they were required to take in for their job, how frequently they felt they were receiving all the information they would ideally like to take in for their job, and how frequently they felt they were provided by others with all the information they needed to do their job. These were measured on a 5-point scale (1=Never, 3=Sometimes, 5=Always).
In each case, we found a negative correlation between the frequency of information reception and the frequency of information overload. This shows that those people who most often receive the information they need or would ideally have to do their job feel overloaded the least often, and that those who most often are provided the information they need to do their job feel overloaded the least often. This indicates a potentially surprising relationship between the quality and quantity of incoming information and feelings of information overload. If overload were strictly related to the amount of information received, and those who received the information they needed received more information than their less well-informed peers, then one would expect a positive correlation between incoming information and overload. This could mean that information overload is as much associated with the quality of the information they take in as the quantity.
That the required and ideal measures correlate more strongly with information overload than the provided measure does may suggest that overload is related to the information that individuals seek out as part of their job as well as the information that is given to them.
Table 8: Correlation between frequency of feelings of overload and information reception measures
Measure
|
Correlation
(Pearson) |
Significance |
N |
|
Required |
-0.303** |
0.000 |
163 |
|
Ideal |
-0.323** |
0.000 |
163 |
|
Provided |
-0.159* |
0.043 |
163 |
** Correlation is significant at the 0.01 level
(2-tailed).
* Correlation is significant at the 0.05 level (2-tailed).
As could be expected, we also found that the measures of information reception correlated very highly to each other.
Table 9: Correlation between information reception measures
Measure
|
Correlation
(Pearson) |
Significance |
N |
|
Required and Ideal |
0.692** |
0.000 |
163 |
|
Required and Provided |
0.313** |
0.000 |
162 |
|
Ideal and Provided |
0.436** |
0.000 |
162 |
** Correlation is significant at the 0.01 level
(2-tailed).
We next looked at how measures of information reception were distributed by various demographic measures, including gender, age, job category, and level of education attained. As with information overload, we found no significant differences except in the case of gender, where we found that women reported receiving all the information they required to do their job at a slightly lower rate than men.
Table 10: Required Information Received by Gender
(Equal variances assumed T-test)
|
|
N |
Mean |
Std. Dev. |
t |
Sig. (2-tailed) |
|
Female |
78 |
3.35 |
0.92 |
-1.976 |
0.050 |
|
Male |
84 |
3.62 |
0.83 |
-1.976 |
0.050 |
(Required reception
frequency:1=Never, 3=Sometimes, 5=Always)
The measures of information reception did not, on the whole, correlate with the measures of the amount of incoming information for specific sources. One exception was the number of voicemail messages received that were considered redundant, which correlated negatively with the provided measure (C=0.202, S=.013, N=151). This indicates that people who receive a large number of redundant voicemail messages do not feel they are usually provided with all the information they need.
While we found no significant differences in overload between groups who used various strategies to organize their incoming email, we did find that the mean level of required information received for those who used one particular strategy—filtering—was significantly lower than for those that did not use that strategy.
Table 11: Required Information Received by Use of Email Filtering
(Equal variances assumed T-test)
|
|
N
|
Mean |
Std. Dev. |
t |
Sig. (2-tailed) |
|
Used filtering |
29 |
3.03 |
0.94 |
3.147 |
0.02 |
|
Did not use filtering |
134 |
3.59 |
0.84 |
3.147 |
0.02 |
(Required reception
frequency:1=Never, 3=Sometimes, 5=Always)
One possible explanation for this finding is that imperfect email filtering algorithms are discarding or misfiling emails in such a way that users are not able to locate and read all the ones that contain needed information.
Our preliminary findings relating source use, importance, and interruptivity show that the three factors are very closely related to each other, not only in terms of each specific information sources, but across most different information sources. That is, individuals who use, find important, and are interrupted by one sources are likely to use, find important, and be interrupted by other sources. The exception seems to be with paper: the negative correlations between these factors in terms of paper versus email, voicemail, and IM could indicate either that paper is being replaced with those sources in some situations or that high-paper use individuals simply work differently than low-paper use individuals.
Overload
We found that the majority of individuals in the workplace felt only somewhat overloaded with information, with smaller groups feeling either less than or more than somewhat overloaded. We also found that the majority of individuals reported receiving some or most of the information they needed to do their jobs. Frequency of feelings of information overload correlated negatively with frequency of receiving needed information, however, and while the frequent use of, high importance of, and frequent interruption by some specific information sources were linked with feelings of overload, the number of messages received did not. These findings collectively support the idea that quality and appropriateness of information received, rather than quantity, are linked to overload.
In addition we found that:
Also interesting was the lack of some expected correlations. In general, the number of emails received and voicemails received, total or in any category, did not relate significantly to overload, with the notable exception of the number internal emails received. It might be worth probing more deeply into the differences between internal and external email flow and information overload in the future. Since the frequency with which individuals checked their voicemail and used the Internet (though not email, Intranet,, or IM) did relate to overload, it may be that individuals work habits may be more of an issue than the actual quantity of information received. On the other hand, at least for email use, the various methods used to organize information flow did not seem to be a factor, and indeed, proved to be a negative factor in some cases when looking at whether all required information was received.
Altogether these findings do not give us a clear picture of what characterizes individuals who feel overloaded with information. They do, however, begin to show us that the quality of information received, the context in which it is received, and general work habits may prove more indicative of overload than sheer quantity of information received.
Running our survey online limits our study to people that use the internet and email, which means we do not represent the greater working population. Additionally, we intentionally limited our information sources to a list of thirteen. Other sources, such as television and radio were not included, both of which are used frequently in certain professions.
A large amount of data was collected with survey, and will require much more in-depth analysis to better identify specific relationships of workplace information intake to information overload. Specifically, it will be interesting to do a closer look at characterizing the various demographics and types of information users. Additionally, we asked a number of questions about the sources people used for specific types of information, and asked people to characterize information for certain sources – how do people use these different channels for different types of information, and how do they feel about them? This aspect needs more investigation, as does a better understanding of distraction and interruption frequency.
The survey and data is available online at:
http://www.sims.berkeley.edu/academics/courses/is271/f01/projects/Individual_Overload/
James Reffell is a Master’s student in the School of Information Management and Systems at the University of California, Berkeley. His research interests are human-computer interaction, information visualization, usability, the contemporary Scottish novel, and legumes.
Sarah Waterson is a graduate student in Computer Science at the University of California, Berkeley. Her research interests are human-computer interaction, information visualization, and shiny metallic objects.
This survey was sponsored and funded by BackWeb
Technologies (www.backweb.com).
[1] Ahmed, Z. “The Demise of Digital Dysfunction.” The Next Big Thing, 2001. (http://tnbt.com/jsp/TNIP.jssp?x=1&y=0&p=TNAsset.jsp&a=79692)
[2] Bezroukov, Nickolai. “Information/Work Overload Annotated Webliography.” 2001. (http://www.softpanorama.org/Social/overload.shtml)
[3] “Dying for Information? An Investigation into the Effects of Information Overload Worldwide,” Dow Jones Reuters Business Interactive Limited, 2000. (http://about.reuters.com/rbb/research/overloadframe.htm)
[4] “More Online, Doing More.” Pew Internet & American Life Project, 18 February 2001.
[5] Owen, Jim. “Coping with Information Overload: Too much of a good thing can hurt your job performance.” CareerBuilder, 2001. (http://www.careerbuilder.com/wl_work_9905_overload.html)
[6] Zoomerang (http://www.zoomerang.com)
[7] Survey Monkey (http://www.surveymonkey.com)
(27 Questions)
If you could remove one information source from your
working day, what would it be?
“My supervisor. Why? Obviously, you've never dealt with him.”
“bosses.”
“assistant supervisor.”
“too many chiefs.”
If your company could do one thing to help you manage
your information, what would it be?
“Clone me...” (2)
“Provide matches to burn all the unwanted magazines! Otherwise, nothing.”
“Give me more of it.”
“leave me alone. Because non techs ask a lot of dumb questions”
“honesty. It makes things quicker, easier, more beneficial to work and wellness of workers”
“I don't know. I'm the President. Tough question.”
“put a woman in charge :)”
“Put in my hand directly rather than let it die a slow death going through the proper channels”
“Get off my back because I know my job better than the bosses do.”
“strap on an extra brain to my head”