Help -
Help for Webmasters
« back to results for ""
Below is a cache of http://people.cs.vt.edu/~mccricks/papers/mcs.pdf. It's a snapshot of the page taken as our search engine crawled the Web.
The web site itself may have changed. You can check the current page or check for previous versions at the Internet Archive. Yahoo! is not affiliated with the authors of this page or responsible for its content. Evaluating Animation in the Periphery as a Mechanism for Maintaining Awareness Evaluating Animation in the Periphery as a Mechanism for Maintaining Awareness D. Scott McCrickard Richard Catrambone John T. Stasko Department of Computer Science School of Psychology College of Computing Virginia Polytechnic Institute and GVU Center and GVU Center and State University Georgia Institute of Technology Georgia Institute of Technology Blacksburg VA 24061-0106 USA Atlanta GA 30332-0170 USA Atlanta GA 30332-0280 USA mccricks@cs.vt.edu rc7@prism.gatech.edu stasko@cc.gatech.edu Abstract : Small animated displays such as tickers and faders are increasingly being used to convey information on computer screens. Relatively little is understood, however, about their use as peripheral displays, that is, tools
for communicating lower-priority awareness information to people. This article describes two experiments that
examine the tradeoff of communication capability versus distraction in peripheral displays. We found that the
presence of animated textual peripheral displays did not distract people from a central information browsing task,
and we identied particular animation and display characteristics that facilitate different information-centric tasks. Keywords : peripheral displays, awareness, monitoring, animation, empirical evaluation, dual-task evaluation 1 Introduction People naturally wish to stay continually informed of
ongoing events of interest. For instance, an ofce worker may want to stay appraised of the weather
outside, the trafc situation for the ride home, how
certain stocks are performing, or how well a favorite
team is playing. While people may want to maintain
awareness of such information, or perhaps even monitor
it intermittently, such awareness ideally should not
distract them from their primary work or task. A variety of information communication devices have been developed to help people maintain a sense
of casual awareness of interesting information. The
classic examples of these types of devices are email
alerts, load monitors, and stock tickers. More recently,
similar displays use visual and audio presentation
methods to show news, weather, sports, personal
data, and other information in a small portion of
the desktop (Greenberg, 1996; McCrickard, 1999; Zhao & Stasko, 2000). Also becoming prevalent are off-the-desktop interfaces that use objects in the
environment and changes in lighting or background
noise to communicate anything from network trafc to
trafc in the hallways (Ishii & Ulmer, 1997; Heiner
et al., 1999). Our focus in this article is a set of peripheral communication techniques used on computer displays
that we call peripheral displays. Typically, peripheral
displays use very little screen real estate, but they still
attempt to convey a fairly large amount of information.
Often, this translates into some use of animation to
cycle through items of interest via scrolling or fading
techniques. While animation has been shown to be a strong perceptive attention draw that consequently may
distract people from their primary task (Ware et al.,
1992), it has also proven to be an effective way to
show large amounts of information in a small space (Robertson et al., 1993). Researchers have speculated
that smooth animations would not be overly distracting
(Fitzpatrick et al., 2001), and organizations like
Yahoo, ESPN, and AOL provide tickering and fading
desktop displays that show continuously updated news
headlines, stock quotes, sports scores, weather reports,
and the computer activity of friends. There are even
toolkits that help enable programmers to include these
and similar techniques into their interfaces (Fitzpatrick
et al., 1998; McCrickard & Zhao, 2000). While numerous studies have examined peoples willingness to use peripheral displays in maintaining
awareness (for example (Parsowith et al., 1998;
McCrickard et al., 1999)), relatively little research has
been conducted to understand better the information
communication versus distraction tradeoff for different
techniques of peripheral communication. Our goal is
to explore the balance between distraction, reaction,
and comprehension for different animated peripheral
displays via empirical evaluations of realistic but
controlled situations. This paper describes several such
evaluations that asked participants to search hypertext
spaces for answers to a series of questions while
completing activities and answering questions based on
information in peripheral displays. 2 Related work Some of the earliest evaluations of constantly changing
displays examined the perceptibility and readability of
rapid serial visual presentations (RSVPs) of letters,
strings, and words. Foster found that participants could
correctly identify about four out of six words in a
sentence when rapidly presented a word at a time in
a single visual location (Foster, 1970). Juola also found
that comprehension of information was comparable
when presented as RSVPs and in multi-line paragraph
format (Juola et al., 1982). In some of the rst studies of smoother animated effects, Duchnicky and
Kolers performed a series of experiments examining the readability of text scrolled on visual display terminals
as a function of window size (Duchnicky & Kolers,
1983). They found that larger displays typically led to
faster performance on reading tasks. A study led by
Granaas found that in scrolled displays, larger jumps
(four to ten characters) led to better comprehension than
smaller jumps (one to two characters) (Granaas et al.,
1984). Kang and Muter, in comparing a tickering effect
to a non-animated RSVP effect, found no difference
in comprehension for a reading task (Kang & Muter,
1989). These experiments addressed many important
factors that we explore further in our research,
including different informational tasks (recognition and
comprehension), different sized displays, and different
ways to change the display. All the previously mentioned evaluations considered the reading of small animated displays as
the sole task of the participant. However, in the case of
peripheral displays, participants would be performing
some main task with attention to a small animated
display part of a secondary task. One experiment with
this type of dual-task scenario was conducted with
OwnTime, a peripheral timespace management system
that alerts people when visitors are waiting to meet with
them (Rodenstein et al., 1999). The study found that
OwnTime visitor interactions were less intrusive than
direct engagement for participants performing recall
and comprehension tasks. The research of Bartram et
al considered the effectiveness of using motion cues
to draw attention (Bartram et al., 2001). They found
that motion cues outperform static representations and
that certain types of motions are more distracting and
irritating than others. In other work, Maglio and Campbell performed a series of dual-task experiments in which participants
performed document editing tasks while a peripheral
display showed news headlines later used to answer
questions (Maglio & Campbell, 2000). The peripheral
displays included a continually scrolling display
that jumped ve pixels per step, a start-and-stop
scrolling display that briey paused when each headline
appeared on the screen, and a fading display that
increased the brightness of the text to make it visible.
They found no difference in the communication
abilities of different peripheral displays (as measured
by how well information is remembered). Also, all
of the animated peripheral displays were found to be
distracting to the main task of document editing, though
the start-and-stop display was the least distracting. Research on the effects of Instant Messaging (IM) notications on desktop computer tasks found that IM
typically was disruptive to primary tasks, particularly
so for fast, stimulus-driven search tasks similar to the
ones in the Maglio experiments (Cutrell et al., 2001).
However, IM does not use smooth animation in its
updates, which may have excaberated the distraction.
Our study examined whether a slower, semantic-based
search task is affected by various smooth peripheral
displays, and whether the peripheral displays can
effectively communicate information to users. 3 Experiments To examine whether animated displays impact information acquisition when maintaining awareness, two empirical evaluations were conducted. Participants
were asked to complete a series of browsing tasks
while simultaneously keeping abreast of a peripheral
display showing constantly changing news, weather,
stock, and sports information. We utilized three peripheral displays in these experiments: a tickering
display that horizontally moves information across the
screen, a fading display that gradually fades between
pieces of information, and a RSVP-style blast that
switches between items in the display without smooth
animation. For the tickering effect, we employed a smooth animation that repeatedly moves the text a pixel
at a time in an attempt to minimize distraction. The
early previously-described studies typically tickered
a display by several characters at a time (Duchnicky
& Kolers, 1983; Granaas et al., 1984; Kang & Muter, 1989), and even Maglio and Campbells 5-
pixel jump when scrolling creates a jerky effect that
may have resulted in unnecessary distraction (Maglio &
Campbell, 2000). Prior work has noted that people tend
to perform better on certain decision-making tasks with
smoother animations (Gonzalez, 1996). We suspect that smooth animations may prove to be less distracting
than the ones used in prior work. Figure 1: Layout of the experimental environment experienced by participants. At the center is the browser used by the participants in the experiment. At the top of the screen is the fade peripheral display that cyclically showed the state of several types of information. At the bottom is the area used for monitoring activities. After each round, the screen cleared except for a question area where the awareness questions were presented. Participants used the information presented in the peripheral display to complete short-term monitoring-
style awareness activities (monitoring activities) and
to answer longer-term knowledge-gain questions
(awareness questions). The experiments consisted of
several rounds (six in the rst experiment, eight in
the second), each consisting of four browsing tasks,
two monitoring activities, and up to ve awareness
questions. The layout of the information on the computer screen is in Figure 1. Motivations for our
experimental choices follow. 3.1 Browsing tasks In performing the browsing tasks, participants used
a simple browser and hypertext pages. The browser
consisted of a textual information area containing
a number of condensed pages from World Wide
Web sites. The text-only information area contained
highlighted, underlined links that pulled up other
pages when clicked with the mouse. The participants
navigated the information space by clicking on the
links and by using the forward and back buttons. The
browsing tasks were non-trivial: the participants had
to read and navigate through a hypertext space to nd
certain information in the pages, enter it into a box
connected with the browser, and press a button to
continue. To minimize the typing required, all solutions to browsing task questions were numerical (for example,
In what year was Mount Rushmore carved?) If an
incorrect answer was entered, the interface beeped and
the participant had to continue working on the problem
until the correct answer was entered. When the correct
answer was entered, the participant could proceed to the
next browsing task. The order in which browsing tasks
were presented was held constant for all participants. 3.2 Monitoring activities While performing the browsing tasks, the participants
used information in the peripheral display to complete
a set of monitoring activities and to answer a series of
awareness questions. The peripheral display cyclically
showed instances of different types of information,
such as a sports score, a stock quote, and a weather
report. Each instance was updated frequently but irregularly as it often is in real life. Participants were asked to press a button when the information
in the peripheral display matched some criteria (for
example, When the temperature drops below 35, press
OK1.) The information that was selected for display
was interesting but rarely vital, and the informational
occurrences that were selected were chosen because
they might spur a user to perform some real-life activity,
such as bringing in a plant that is outdoors or selling a
stock that is performing poorly. Each round included two such monitoring activities. The order in which monitoring activities were presented
was held constant for all participants. If the button
was pressed at the correct time (that is, after the
needed information was presented), it was greyed out
to alert the participant that the task had been completed
successfully. If the button was pressed too soon, the
interface beeped and the button remained active. 3.3 Awareness questions At the end of each round, the participants were
given awareness questions that asked them to recall
information that was shown in the peripheral display.
The questions were multiple-answer multiple-choice
questions that addressed both content and temporal
issues. Each question had four possible answers, all
initially unselected, and there was always at least one
correct answer. The rst question in each set listed four types of information and asked the participant to choose the
ones that had been displayed. If they correctly recalled seeing information, later questions asked about details
of it, such as which news stories appeared, which
stock quotes constantly increased, or which sports team
scored the most points. For example, if a participant
correctly noted that news headlines had been displayed,
later questions would present a list of headlines and
ask the participant to select the ones that had appeared.
All of the information was ctional but realistic, and
no attempt was made to intentionally deceive the
participants with slightly different information (for
example, a stock quote that almost always increased). 3.4 Data collection and evaluation To compare performance among groups, the dependent
variables were the times for all browsing tasks and
monitoring activities and the answers to the post-
round awareness questions. The results were analyzed
to determine whether differences in certain measures
occurred for participants in different conditions
(participants using different types of peripheral displays
in the rst experiment, and participants using different
sizes and speeds in the second). The browsing time is the time from which the browsing task and browser information appeared on
the screen to the time when the participant typed in
the correct answer and pressed the OK button. The
monitoring time is the time from when the information
was rst entered into the cyclic display until the
information was acknowledged as matching the current
criterion via a button press. For the awareness questions, the participants responses to each of the four answers were collected.
We considered each question as being worth four
points: correctly or incorrectly assessing each possible
response to a question. A number of different methods can be used to determine a participants ability to recall information.
The most obvious measure is to compare the percent
of correct responses for the awareness questions in
different situations. The percent of correct responses
is referred to as the correctness rate. The correctness rate measure potentially can misrepresent a participants awareness of information,
however. Note that a participant who did not remember
seeing anything in the peripheral display and left all
responses unchecked could have a correctness rate as
high or higher than a participant who remembered
seeing several items but was mistaken about what was
seen and checked the wrong box(es). An alternate measure for determining responsiveness is the hit rate, a term from signal
detection theory dened as the ratio of correctly
identied stimuli to the total number of times the
stimulus was presented. The hit rate is typically accompanied by the false alarm rate, the ratio of
incorrect stimuli responses to the total number of times
when the stimulus was not present. Since a typical goal
when using a peripheral display is to proactively recall
seeing information, it may be better for a participant
to be mistaken about seeing information that was
not displayed than to be mistaken about not seeing
information that in fact was displayed. Perhaps a person wants to remember that a story occurred, or that
a tornado watch is under way, or that a trafc bulletin appeared. The hit rate would reect the awareness potential of an animated display. In analyzing the results, analyses of variance (ANOVAs) were performed to check for statistical
signicance among different conditions of the experiments. If the ANOVA revealed a signicant difference, pairwise t-tests were performed to determine which conditions differed. 3.5 Experiment 1 The rst experiment compared relative performance
when using fading, tickering, and blasting displays as
well as when no peripheral display was present. 3.5.1 Method This experiment focused on three factors: the possibility for degradation in performance on a
browsing task when a peripheral display was present,
the speed in identifying and reacting to changes
in peripheral displays, and the ability to remember
information that appeared in a peripheral display. Seventy undergraduate students participated in this experiment for class credit. The experiment was conducted on identical workstations, each connected to
a 15-inch monitor with an optical mouse. Participants
were run in small groups, one participant per computer.
The experiment was explained to each group verbally
and again on the computer with examples. The participants performed six rounds of browsing tasks, monitoring activities, and awareness questions.
In each round, participants completed four browsing
tasks while performing two monitoring activities using
either a fade, ticker, or a blast animation. The speed with which the information was displayed
corresponded to the mean speeds for each device
selected by the participants in a previous study
(McCrickard, 2000). While this resulted in different
rates of information display for the animations, we felt
it was a more realistic and ecologically valid measure
of how people would use them. The ticker continually
shifted one pixel every 50 milliseconds, while the fade
and blast updated their entire contents every 2 seconds.
The fade required 500 milliseconds to fade between
items, while the blast updated instantaneously. At the end of each round, participants answered awareness questions about the information that appeared in the animated display. The rst question
asked which types of information appeared in the
display. For each correctly-identied instance of information appearing, two questions about the information were asked up to a total of ve questions. As a base case, one group of participants (n
The web site itself may have changed. You can check the current page or check for previous versions at the Internet Archive. Yahoo! is not affiliated with the authors of this page or responsible for its content. Evaluating Animation in the Periphery as a Mechanism for Maintaining Awareness Evaluating Animation in the Periphery as a Mechanism for Maintaining Awareness D. Scott McCrickard Richard Catrambone John T. Stasko Department of Computer Science School of Psychology College of Computing Virginia Polytechnic Institute and GVU Center and GVU Center and State University Georgia Institute of Technology Georgia Institute of Technology Blacksburg VA 24061-0106 USA Atlanta GA 30332-0170 USA Atlanta GA 30332-0280 USA mccricks@cs.vt.edu rc7@prism.gatech.edu stasko@cc.gatech.edu Abstract : Small animated displays such as tickers and faders are increasingly being used to convey information on computer screens. Relatively little is understood, however, about their use as peripheral displays, that is, tools
for communicating lower-priority awareness information to people. This article describes two experiments that
examine the tradeoff of communication capability versus distraction in peripheral displays. We found that the
presence of animated textual peripheral displays did not distract people from a central information browsing task,
and we identied particular animation and display characteristics that facilitate different information-centric tasks. Keywords : peripheral displays, awareness, monitoring, animation, empirical evaluation, dual-task evaluation 1 Introduction People naturally wish to stay continually informed of
ongoing events of interest. For instance, an ofce worker may want to stay appraised of the weather
outside, the trafc situation for the ride home, how
certain stocks are performing, or how well a favorite
team is playing. While people may want to maintain
awareness of such information, or perhaps even monitor
it intermittently, such awareness ideally should not
distract them from their primary work or task. A variety of information communication devices have been developed to help people maintain a sense
of casual awareness of interesting information. The
classic examples of these types of devices are email
alerts, load monitors, and stock tickers. More recently,
similar displays use visual and audio presentation
methods to show news, weather, sports, personal
data, and other information in a small portion of
the desktop (Greenberg, 1996; McCrickard, 1999; Zhao & Stasko, 2000). Also becoming prevalent are off-the-desktop interfaces that use objects in the
environment and changes in lighting or background
noise to communicate anything from network trafc to
trafc in the hallways (Ishii & Ulmer, 1997; Heiner
et al., 1999). Our focus in this article is a set of peripheral communication techniques used on computer displays
that we call peripheral displays. Typically, peripheral
displays use very little screen real estate, but they still
attempt to convey a fairly large amount of information.
Often, this translates into some use of animation to
cycle through items of interest via scrolling or fading
techniques. While animation has been shown to be a strong perceptive attention draw that consequently may
distract people from their primary task (Ware et al.,
1992), it has also proven to be an effective way to
show large amounts of information in a small space (Robertson et al., 1993). Researchers have speculated
that smooth animations would not be overly distracting
(Fitzpatrick et al., 2001), and organizations like
Yahoo, ESPN, and AOL provide tickering and fading
desktop displays that show continuously updated news
headlines, stock quotes, sports scores, weather reports,
and the computer activity of friends. There are even
toolkits that help enable programmers to include these
and similar techniques into their interfaces (Fitzpatrick
et al., 1998; McCrickard & Zhao, 2000). While numerous studies have examined peoples willingness to use peripheral displays in maintaining
awareness (for example (Parsowith et al., 1998;
McCrickard et al., 1999)), relatively little research has
been conducted to understand better the information
communication versus distraction tradeoff for different
techniques of peripheral communication. Our goal is
to explore the balance between distraction, reaction,
and comprehension for different animated peripheral
displays via empirical evaluations of realistic but
controlled situations. This paper describes several such
evaluations that asked participants to search hypertext
spaces for answers to a series of questions while
completing activities and answering questions based on
information in peripheral displays. 2 Related work Some of the earliest evaluations of constantly changing
displays examined the perceptibility and readability of
rapid serial visual presentations (RSVPs) of letters,
strings, and words. Foster found that participants could
correctly identify about four out of six words in a
sentence when rapidly presented a word at a time in
a single visual location (Foster, 1970). Juola also found
that comprehension of information was comparable
when presented as RSVPs and in multi-line paragraph
format (Juola et al., 1982). In some of the rst studies of smoother animated effects, Duchnicky and
Kolers performed a series of experiments examining the readability of text scrolled on visual display terminals
as a function of window size (Duchnicky & Kolers,
1983). They found that larger displays typically led to
faster performance on reading tasks. A study led by
Granaas found that in scrolled displays, larger jumps
(four to ten characters) led to better comprehension than
smaller jumps (one to two characters) (Granaas et al.,
1984). Kang and Muter, in comparing a tickering effect
to a non-animated RSVP effect, found no difference
in comprehension for a reading task (Kang & Muter,
1989). These experiments addressed many important
factors that we explore further in our research,
including different informational tasks (recognition and
comprehension), different sized displays, and different
ways to change the display. All the previously mentioned evaluations considered the reading of small animated displays as
the sole task of the participant. However, in the case of
peripheral displays, participants would be performing
some main task with attention to a small animated
display part of a secondary task. One experiment with
this type of dual-task scenario was conducted with
OwnTime, a peripheral timespace management system
that alerts people when visitors are waiting to meet with
them (Rodenstein et al., 1999). The study found that
OwnTime visitor interactions were less intrusive than
direct engagement for participants performing recall
and comprehension tasks. The research of Bartram et
al considered the effectiveness of using motion cues
to draw attention (Bartram et al., 2001). They found
that motion cues outperform static representations and
that certain types of motions are more distracting and
irritating than others. In other work, Maglio and Campbell performed a series of dual-task experiments in which participants
performed document editing tasks while a peripheral
display showed news headlines later used to answer
questions (Maglio & Campbell, 2000). The peripheral
displays included a continually scrolling display
that jumped ve pixels per step, a start-and-stop
scrolling display that briey paused when each headline
appeared on the screen, and a fading display that
increased the brightness of the text to make it visible.
They found no difference in the communication
abilities of different peripheral displays (as measured
by how well information is remembered). Also, all
of the animated peripheral displays were found to be
distracting to the main task of document editing, though
the start-and-stop display was the least distracting. Research on the effects of Instant Messaging (IM) notications on desktop computer tasks found that IM
typically was disruptive to primary tasks, particularly
so for fast, stimulus-driven search tasks similar to the
ones in the Maglio experiments (Cutrell et al., 2001).
However, IM does not use smooth animation in its
updates, which may have excaberated the distraction.
Our study examined whether a slower, semantic-based
search task is affected by various smooth peripheral
displays, and whether the peripheral displays can
effectively communicate information to users. 3 Experiments To examine whether animated displays impact information acquisition when maintaining awareness, two empirical evaluations were conducted. Participants
were asked to complete a series of browsing tasks
while simultaneously keeping abreast of a peripheral
display showing constantly changing news, weather,
stock, and sports information. We utilized three peripheral displays in these experiments: a tickering
display that horizontally moves information across the
screen, a fading display that gradually fades between
pieces of information, and a RSVP-style blast that
switches between items in the display without smooth
animation. For the tickering effect, we employed a smooth animation that repeatedly moves the text a pixel
at a time in an attempt to minimize distraction. The
early previously-described studies typically tickered
a display by several characters at a time (Duchnicky
& Kolers, 1983; Granaas et al., 1984; Kang & Muter, 1989), and even Maglio and Campbells 5-
pixel jump when scrolling creates a jerky effect that
may have resulted in unnecessary distraction (Maglio &
Campbell, 2000). Prior work has noted that people tend
to perform better on certain decision-making tasks with
smoother animations (Gonzalez, 1996). We suspect that smooth animations may prove to be less distracting
than the ones used in prior work. Figure 1: Layout of the experimental environment experienced by participants. At the center is the browser used by the participants in the experiment. At the top of the screen is the fade peripheral display that cyclically showed the state of several types of information. At the bottom is the area used for monitoring activities. After each round, the screen cleared except for a question area where the awareness questions were presented. Participants used the information presented in the peripheral display to complete short-term monitoring-
style awareness activities (monitoring activities) and
to answer longer-term knowledge-gain questions
(awareness questions). The experiments consisted of
several rounds (six in the rst experiment, eight in
the second), each consisting of four browsing tasks,
two monitoring activities, and up to ve awareness
questions. The layout of the information on the computer screen is in Figure 1. Motivations for our
experimental choices follow. 3.1 Browsing tasks In performing the browsing tasks, participants used
a simple browser and hypertext pages. The browser
consisted of a textual information area containing
a number of condensed pages from World Wide
Web sites. The text-only information area contained
highlighted, underlined links that pulled up other
pages when clicked with the mouse. The participants
navigated the information space by clicking on the
links and by using the forward and back buttons. The
browsing tasks were non-trivial: the participants had
to read and navigate through a hypertext space to nd
certain information in the pages, enter it into a box
connected with the browser, and press a button to
continue. To minimize the typing required, all solutions to browsing task questions were numerical (for example,
In what year was Mount Rushmore carved?) If an
incorrect answer was entered, the interface beeped and
the participant had to continue working on the problem
until the correct answer was entered. When the correct
answer was entered, the participant could proceed to the
next browsing task. The order in which browsing tasks
were presented was held constant for all participants. 3.2 Monitoring activities While performing the browsing tasks, the participants
used information in the peripheral display to complete
a set of monitoring activities and to answer a series of
awareness questions. The peripheral display cyclically
showed instances of different types of information,
such as a sports score, a stock quote, and a weather
report. Each instance was updated frequently but irregularly as it often is in real life. Participants were asked to press a button when the information
in the peripheral display matched some criteria (for
example, When the temperature drops below 35, press
OK1.) The information that was selected for display
was interesting but rarely vital, and the informational
occurrences that were selected were chosen because
they might spur a user to perform some real-life activity,
such as bringing in a plant that is outdoors or selling a
stock that is performing poorly. Each round included two such monitoring activities. The order in which monitoring activities were presented
was held constant for all participants. If the button
was pressed at the correct time (that is, after the
needed information was presented), it was greyed out
to alert the participant that the task had been completed
successfully. If the button was pressed too soon, the
interface beeped and the button remained active. 3.3 Awareness questions At the end of each round, the participants were
given awareness questions that asked them to recall
information that was shown in the peripheral display.
The questions were multiple-answer multiple-choice
questions that addressed both content and temporal
issues. Each question had four possible answers, all
initially unselected, and there was always at least one
correct answer. The rst question in each set listed four types of information and asked the participant to choose the
ones that had been displayed. If they correctly recalled seeing information, later questions asked about details
of it, such as which news stories appeared, which
stock quotes constantly increased, or which sports team
scored the most points. For example, if a participant
correctly noted that news headlines had been displayed,
later questions would present a list of headlines and
ask the participant to select the ones that had appeared.
All of the information was ctional but realistic, and
no attempt was made to intentionally deceive the
participants with slightly different information (for
example, a stock quote that almost always increased). 3.4 Data collection and evaluation To compare performance among groups, the dependent
variables were the times for all browsing tasks and
monitoring activities and the answers to the post-
round awareness questions. The results were analyzed
to determine whether differences in certain measures
occurred for participants in different conditions
(participants using different types of peripheral displays
in the rst experiment, and participants using different
sizes and speeds in the second). The browsing time is the time from which the browsing task and browser information appeared on
the screen to the time when the participant typed in
the correct answer and pressed the OK button. The
monitoring time is the time from when the information
was rst entered into the cyclic display until the
information was acknowledged as matching the current
criterion via a button press. For the awareness questions, the participants responses to each of the four answers were collected.
We considered each question as being worth four
points: correctly or incorrectly assessing each possible
response to a question. A number of different methods can be used to determine a participants ability to recall information.
The most obvious measure is to compare the percent
of correct responses for the awareness questions in
different situations. The percent of correct responses
is referred to as the correctness rate. The correctness rate measure potentially can misrepresent a participants awareness of information,
however. Note that a participant who did not remember
seeing anything in the peripheral display and left all
responses unchecked could have a correctness rate as
high or higher than a participant who remembered
seeing several items but was mistaken about what was
seen and checked the wrong box(es). An alternate measure for determining responsiveness is the hit rate, a term from signal
detection theory dened as the ratio of correctly
identied stimuli to the total number of times the
stimulus was presented. The hit rate is typically accompanied by the false alarm rate, the ratio of
incorrect stimuli responses to the total number of times
when the stimulus was not present. Since a typical goal
when using a peripheral display is to proactively recall
seeing information, it may be better for a participant
to be mistaken about seeing information that was
not displayed than to be mistaken about not seeing
information that in fact was displayed. Perhaps a person wants to remember that a story occurred, or that
a tornado watch is under way, or that a trafc bulletin appeared. The hit rate would reect the awareness potential of an animated display. In analyzing the results, analyses of variance (ANOVAs) were performed to check for statistical
signicance among different conditions of the experiments. If the ANOVA revealed a signicant difference, pairwise t-tests were performed to determine which conditions differed. 3.5 Experiment 1 The rst experiment compared relative performance
when using fading, tickering, and blasting displays as
well as when no peripheral display was present. 3.5.1 Method This experiment focused on three factors: the possibility for degradation in performance on a
browsing task when a peripheral display was present,
the speed in identifying and reacting to changes
in peripheral displays, and the ability to remember
information that appeared in a peripheral display. Seventy undergraduate students participated in this experiment for class credit. The experiment was conducted on identical workstations, each connected to
a 15-inch monitor with an optical mouse. Participants
were run in small groups, one participant per computer.
The experiment was explained to each group verbally
and again on the computer with examples. The participants performed six rounds of browsing tasks, monitoring activities, and awareness questions.
In each round, participants completed four browsing
tasks while performing two monitoring activities using
either a fade, ticker, or a blast animation. The speed with which the information was displayed
corresponded to the mean speeds for each device
selected by the participants in a previous study
(McCrickard, 2000). While this resulted in different
rates of information display for the animations, we felt
it was a more realistic and ecologically valid measure
of how people would use them. The ticker continually
shifted one pixel every 50 milliseconds, while the fade
and blast updated their entire contents every 2 seconds.
The fade required 500 milliseconds to fade between
items, while the blast updated instantaneously. At the end of each round, participants answered awareness questions about the information that appeared in the animated display. The rst question
asked which types of information appeared in the
display. For each correctly-identied instance of information appearing, two questions about the information were asked up to a total of ve questions. As a base case, one group of participants (n
download Evaluating Animation in the Periphery as a Mechanism for Maintaining ...
