Cognitive tasks in information analysis: Use of event dwell time to characterize component activities Page: 4 of 7
This article is part of the collection entitled: Office of Scientific & Technical Information Technical Reports and was provided to Digital Library by the UNT Libraries Government Documents Department.
The following text was automatically extracted from the image on this page using optical character recognition software:
are all routine components. However, because of the increasingly heavy reliance intelligence
analysts place upon Internet resources, we have chosen to focus our efforts in this arena.
The GBAE collects and time stamps workstation data at a very high rate. In order to develop an
approach to reducing this volume and discerning high-information yield behaviors in the data,
we initially carried out three model tasks. The model tasks were executed by a senior researcher
with extensive experience in conducting information searches via the Internet.
The first model task involved conducting an open source Web-based search for items of interest
in the area of Psychological Operations in information warfare. The task was conducted using
the Google search engine and Internet Explorer as a browser. The search task lasted 40 minutes
and yielded 308 discrete Glass Box events. In order to better understand the relationship
between events that comprise "search result filtering" versus "reviewing items of interest," the
second model task involved seeking a single item of interest on the general topic of "central
nervous system effects of biological weapons." The total session time for this model task was
7.6 minutes, yielding a total of 93 Glass Box data events. To evaluate dwell time patterns in a
somewhat different type of task, the analyst used the Web pages of interest that were found in
Model Task 1 as the basis for a 3/a page summary of the findings. The total session time for this
model task was 18 minutes and generated 136 Glass Box data events. To determine the extent to
which similar patterns might be observed in actual intelligence analyst activity, we selected a
segment of data from a single government-contracted intelligence analyst who has been
conducting his work in the GBAE for the past six months. This session sample was 48 minutes
in duration and generated 437 events.
Time Measure Computation
GBAE data were filtered to eliminate system-generated events, and arrayed in an Excel
spreadsheet for analysis. The time stamps were used to calculate event dwell time as:
Event Dwell Time (Row N) = Time in Row (N + 1) -Time in Row N.
This metric can be used to evaluate the relationship between events according to their duration,
patterns of event duration, and relationships between application usage and time.
The principal data analysis methods for this task analytic and case study approach involved
visual inspection of dwell time distributions, analysis of time interval frequency data, and
analysis of temporal patterns in application usage. The SPSS software package was used to
construct the plots and time interval distributions.
Figures 1 - 4 show the dwell time data across events arranged chronologically. The results
depicted in Figure 1 show a cyclic pattern of alternation between events that have durations of
less than 10 seconds, and those that exceed 10 seconds. Figure 2 shows the dwell time data for
the single item search. This plot also shows the preponderance of data events at the sub-10
second duration, with cyclic activities at the 10-35 second duration, and clear peaks at events 40
and 43. Evaluation of the specific event data indicates that the user was reading on-screen Web
pages related to the target search topic. The remaining events are associated with launching
Here’s what’s next.
This article can be searched. Note: Results may vary based on the legibility of text within the document.
Tools / Downloads
Get a copy of this page or view the extracted text.
Citing and Sharing
Basic information for referencing this web page. We also provide extended guidance on usage rights, references, copying or embedding.
Reference the current page of this Article.
Sanquist, Thomas F.; Greitzer, Frank L.; Slavich, Antoinette L.; Littlefield, Rik J.; Littlefield, Janis S. & Cowley, Paula J. Cognitive tasks in information analysis: Use of event dwell time to characterize component activities, article, September 28, 2004; Richland, Washington. (digital.library.unt.edu/ark:/67531/metadc874907/m1/4/: accessed November 18, 2018), University of North Texas Libraries, Digital Library, digital.library.unt.edu; crediting UNT Libraries Government Documents Department.