Conference PaperPDF Available

Highlighting - Or Why Google Is That Successful.

Authors:
Highlighting – Or Why Google Is That Successful
Michael D. Kickmeier-Rust & Dietrich Albert
Department of Psychology
University of Graz, Austria
michael.kickmeier@uni-graz.at
dietrich.albert@uni-graz.at
Abstract:
In today’s information society, searching for information or for functions of a software
application is one of the most frequent tasks. Still, visual search tasks are reported to be
associated with low efficiency and user satisfaction. A simple method to improve search
performance is highlighting. With the current experiment we investigate if search
performance and user satisfaction can be significantly improved by presenting an optimal
density of highlighted items. As the results of the eye-tracking experiment yielded, presenting
15 percent of highlighted items on a display improved search speed, accuracy, as well as user
satisfaction in contrast to presenting no highlighted items or 49 percent. The results are
discussed regarding their practical impact, design implications, and regarding emerging
questions for future research.
Keywords: Highlighting, Search Performance, Visual Search, Visual Attention, Eye-tracking
1. Introduction
The basics of human visual perception and attention are a major foundation of software usability.
For instance in information design, legibility, brightness, contrast, color, saliency, or the structure
of information are crucial factors to enable successful and efficient information retrieval.
Searching (e.g., a piece of text on a website or a certain menu entry of a software
application) is probably the most common tasks of human-computer interaction. Efficiency,
accuracy, and user satisfaction of that interaction are strongly influenced by the visual properties
and the amount of visual stimuli of a display. We are facing a vast amount of electronically
available information, software applications providing countless functions, and searching is
reported to be oftentimes difficult, inefficient, and unsatisfying [2, 7]. The infinite number of
stored bits and the variety of software functions, however, are only valuable, if they can be
searched, found, and used by humans.
The question arises whether information or functionality can be displayed in a way that
performance, accuracy, and user satisfaction of visual search processes are possibly high; even
without knowing what someone is searching for.
Kickmeier-Rust, M. D., & Albert, D. (2005). Highlighting – Or why Google is that
successful. In A. Holzinger & K. H. Weidmann (Eds.), Empowering software quality:
How can usability engineering reach these goals? (pp. 45-52). Wien: Österreichische
Computer Gesellschaft.
A well-tried and successful method to facilitate visual search is highlighting; on paper as
well as on electronic displays. There is a variety of research concerning the basic features of
highlighting [14] and applied techniques of highlighting [8].
Highlighting is based on early visual processes. In a pre-attentive stage of vision the
visual field can be processed as a whole without limitations of capacity. For example, if one
searches a red circle among green circles, it can be detected immediately no matter how many
green circles surround the red target. In terms of Ulric Neisser [6], the red circle pops out of the
display. Generally, such pop-out effect occurs if the target is defined by single basic feature such
as color, shape, size, or motion [14]. According to the Guided Search model [13, 15, 16], these
early visual processes create an activation map. An activation map is a representation of the
visual field determined by continuous levels of activation. The activation for a certain location in
the visual field is determined by the visual properties of the item(s) in that location and
intentional control. For example, when someone is searching for a target, which is expected to be
red, a single red circle among blue ones provokes higher activation for its location than a single
green circle. In a later stage of vision, attention is directed towards the peaks of the activation
map. This guidance is determined by bottom-up aspects (e.g., saliency) and top-down aspects
(e.g., a searchers intentions or expectations). Attention acts like a filter for relevant information to
be passed to one or more limited-capacity processes of vision; only locations that are promising
for a certain goal are attentively processed.
So far, visual attention refers to focal fixations, e.g. fixating a letter. Information
processing, however, does not only occur for focally attended stimuli but also for the periphery of
a stimulus. While the fovea is amounting for 2 degree visual angle, which corresponds to about 2
per mille of the visual field, peripheral vision allows the processing of information from the
whole visual field. Due to the decrease of neurons related to retinal areas with increase
eccentricity, however, visual performance generally decreases with eccentricity [13]. Still, there
is evidence for the importance of peripheral vision in search tasks. Besides pre-attentive
processing of the whole visual field, peripheral vision contributes to feature detection [1] or letter
recognition [12] during search episodes.
Past research on highlighting, primarily focused on valid highlighting. Fisher and
colleagues [3, 4] argued that a key variable of facilitating visual search through highlighting is
the frequency a target, as opposed to a distractor, is highlighted. The aim of the current
experiment is to investigate the effects of the density of highlighted items in a display.
On the basis of mechanisms of visual search, as introduced very briefly, we assume that
presenting an optimal density of highlighted items increases visual search performance. The
distribution of highlighted items facilitates the creation of an activation map in a pre-attentive
stage of vision and, subsequently, guides the gaze through the display. Due to mechanisms like
the inhibition of return [10] we can assume that these are attended only once along a serial path.
A target can be detected either if itself is highlighted or, by means of peripheral information
processing, if it surrounds a fixated (highlighted) item in a certain distance. This guidance is
important because the effort a searcher spends on surveying a display is limited by means of a
cost-benefit analysis, as outlined in the Information foraging Theory [9]. Not all objects within an
area to search are attended during a search episode. When people are searching information they
roughly scan through the visual field. However, if pre-attentive guidance corresponds to these
limitations, the full search capacity can be utilized for target detection.
With the current eye-tracking experiment we investigate the impact of the total density of
highlighted items in a display on search performance, accuracy, and user satisfaction.
2. An Eye-Tracking Experiment
2.1. Experimental Design
In an eye-tracking experiment participants were presented 48 panels consisting of 140 simple
nouns each. The task was to determine as fast and as accurate as possible, whether a given search
word was present among the nouns of the panel or not. In 26 tasks the target was present and in
22 tasks the target was absent. On the basis of a previous study on density effects of highlighting
[5] which reported highest search performance between 15 and 30 percent highlighting, three
different densities were realized: no highlighted items (0 percent) as a too small portion of
highlighting, 15 percent, which was supposed to be an optimal density, and 49 percent as too
high portion of highlighting. We realized 49 percent in order to avoid an inflection point at about
50 percent when the number of highlighted items starts to exceed the number of non-highlighted
items and non-highlighted items start becoming more salient than highlighted ones. The sequence
of tasks and the distributions of words and highlighting were created randomly for each
participant. As dependent variables response times, answer accuracy (correct and incorrect
responses), and eye movements were recorded.
2.2. Participants
The data of 36 participants were analyzed, thereof 27 females (75 percent) and 9 males (25
percent). The average age was 22.53 years (SD = 3.14), the youngest participant was 19 years of
age, the oldest 31. In total, 21 participants reported normal vision, 15 reported vision corrected to
normal. One participant of the “corrected to normal” group additionally reported a minor color
vision defect.
2.3. Material and Apparatus
We used the SMI Eyelink I eye-tracker (SensoMotoric Instruments GmbH, Germany;
www.smi.de) to record binocular eye-movements. The eye-tracking experiment was developed
using Microsoft Visual C++ 6.0 and the SMI Eyelink API templates. The test software presented
introduction and briefing screens and, subsequently, generated automatically random search
panels from a pool of simple nouns. To counter-balance position effects, the different search
tasks were automatically presented in random order for each participant. The experimental
sequence was presented on a 21” computer monitor with a screen resolution of 1280x1024. At a
distance of 58 cm from the screen a chin rest was mounted to stabilize head movements and the
distance between eyes and screen. Responses were given using the Y key (yes, the target is
present) and the N key (no, the target is absent) on a regular computer keyboard with German
layout. The keys were prepared with stickers that allowed haptic detection in order to prevent
participants from looking down to the keyboard during the experimental tasks.
2.4. Procedure
Participants were first asked to conveniently take place in front of the eye-tracker’s subject
monitor. Then the participants were briefed regarding the aims of the experiment and the eye-
tracking procedure. In a next step a paper-pencil data form was filled out, asking for age, sex,
education, vision (normal, corrected to normal, vision below 100 percent, and vision defects), and
Internet and search engine usage. Then participants were instructed to determine as fast and as
correct as possible whether a given word was present in the display or not. Additionally, the
participants were briefed to inform the experimenter about any events during search tasks, e.g.
forgetting the target word or usage of the wrong response key. After initial briefings the eye-
tracker’s head-mounted camera system was applied carefully and eye cameras were adjusted.
After dousing lights, eye cameras were focused on the pupils and cameras were calibrated and
validated. The camera adjustment, calibrations, and validations required about 10 to 15 minutes.
The experiment started with three training tasks. After successful completion of the
training, participants performed the first block of 24 search tasks. Each task was initiated with the
fixation of a center point on the screen. During this fixation the participants were told the target
word. The participants started the search task by pressing the space key. To terminate a search
task participants pressed either the Y key (yes, the target is present) or the N key (no, the target is
absent).
After the first block of 24 search tasks a ten minutes break followed. During this break the
uncomfortable head-mounted camera system was removed. The second block of search tasks was
performed identical to the first one. Finally, the camera system was removed and participants
rated three questions regarding their satisfaction with the different densities of highlighting.
Arising comments of participants were note by the experimenter.
3. Results
3.1. Response Times
As hypothesized, we found significantly shorter response times when presenting 15 percent
highlighted items in the search panel, in contrast to no highlighted items or 49 percent. In target-
present tasks mean response time for the 15-percent group was 12.62 seconds (SD = 7.25), in the
0-percent group 14.85 seconds (SD = 5.75), and in the 49-percent group 15.69 seconds
(SD = 7.35). In target-absent tasks mean response time for the 15-percent group was 23.6 seconds
(SD = 6.33), in the 0-percent group 26.01 seconds (SD = 7.84), and in the 49-percent group 23.23
seconds (SD = 6.70). As shown in Figure 1a, in total the 15-percent group yielded a mean
response time of 18.11 seconds (SD = 6.79), the 49-percent group 19.46 seconds (SD = 7.03),
and the 0-percent group 20.43 seconds (SD = 6.80). An analysis of variance (ANOVA) reported a
significant effect of the density of highlighted items (F(2, 1132) = 2.28, p = .013).
3.2. Accuracy
Focusing on target-present tasks only, also for answer accuracy (relative frequencies of correct
responses) we found expected results. Answer accuracy was significantly higher when presenting
15 percent highlighted items in the search panel, in contrast to no highlighted items or 49 percent.
As shown in Figure 1b, presenting 15 percent of highlighted items resulted in a relative frequency
of correct responses of .79; for the 0-percent group we found a relative frequency of .67 and in
the in the 49-percent group a relative frequency of .65. An ANOVA reported a significant
difference between the three highlighting densities (F(2, 1132) = 43.71, p < .001.).
(a) (b)
16,5
17
17,5
18
18,5
19
19,5
20
20,5
21
0% 15% 4 9%
De nsity of highlig hte d item s
Re spo ns e time s (s)
0,5
0,55
0,6
0,65
0,7
0,75
0,8
0,85
0,9
0% 15% 49 %
De nsity of h ighlighte d item s
Re lative frequ . of corre ct res po ns es
Figure 1. Panel (a) shows the response times for different densities of highlighted
items in the search panel; panel (b) the relative frequencies of correct responses.
3.3. Scan Paths
Analysis of eye movements yielded that the gaze was immediately shifted from the initial center
fixation to the item in top left corner of the search panel. This distinct saccade appeared in about
90 percent of search trials. Subsequently, scan paths were performed in reading-like direction
from the top left corner row-wise to the bottom right corner. As shown in Figure 2a the relative
frequency of saccades in reading-like direction (.62) was higher than the relative frequency of a
reverse directed saccade (.38). An ANOVA reported a significant difference between these
relative frequencies (F(1,2310) = 4702.65, p < .001). Eye-tracking data, moreover, reported that
fixations on highlighted items were significantly more frequent than fixations on non-highlighted
items. To assess the fixation ratio, we applied a coefficient Φ to obtain a meaningful value of the
portion of fixations on an item type in relation to the total number of fixations and the portion of
this type in the search panel. Φ combines the values of the number of total fixations φT, the
number of fixations on one item type φI, the total number of items in the display ΙT, and the
number of items of a type ΙI as
1
I I
T T
φ
φ
§ ·
Ι
Φ =
¨ ¸
Ι
© ¹ .
Φ results in values between 0 and 1, whereat 0 means a low Φ-frequency of fixations and 1 a
high frequency. The coefficient is 0 in two conditions: (a) If no fixations occur on an item type or
and/or (b) if the number of items of a type equals the number of total items.
We applied this coefficient on highlighted words (ΦH) and on standard, non-highlighted
words (ΦS). As shown in Figure 2b, the Φ-frequencies were higher at highlighted words
(Φ = .19) than at non-highlighted words (Φ = .14). An ANOVA reported a significant difference
between the Φ-frequencies of highlighted objects and standard, non-highlighted words
(F(1,2310) = 144.42, p < .001).
(a) (b)
0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0,8
Highlig hted Sta ndard
Re lative fixatio n freque nc y
0
0,05
0,1
0,15
0,2
0,25
Read ing-d ire ction Reverse direction
Re lative sac cad e fre qu en cy
Figure 2. Panel (a) shows the relative frequencies of fixations highlighted items in contrast.
to non-highlighted items. Panel (b) shows the relative Φ-frequencies of saccades in reading-
like and reverse directions.
3.4. User Satisfaction
In terms of software usability, user satisfaction is a crucial factor. To investigate not only
performance but also user satisfaction, we analyzed the participant’s rating of three questions at
the end of the test session. As summarized in Table 1, presenting 15 percent of highlighted items
in the search panel resulted in higher user satisfaction throughout. Summarized, the 15-percent
group yielded a rating of 4.00 (SD = 1.35), the 0-percent group a rating of 3.62 (SD = 1.10), and
the 49-percent group a rating of 2.75 (SD = 1.13). The differences between the groups was
significant (X2(10, N = 315) = 70.08, p<.01).
Highlighting Density
0% 15% 49%
In your opinion, how good could you determine whether
the target was present or not? 4.03 4.48 2.63
How convenient was the presentation of the search panels? 3.74 4.00 2.60
How easy was it to determine whether the target was
present or not? 3.09 3.54 3.03
Table 1. This table shows the average rating on a six-step rating scale;
1 corresponds to a negative rating, 6 is the best rating.
4. Discussion
It’s an important aim to increase search speed, accuracy, and user satisfaction of search tasks.
With the current experiment we could demonstrate that this can be achieved by presenting an
optimal density of highlighting. It’s a remarkable finding, since when measuring speed and
accuracy one would expect a speed-accuracy trade-off. As the results show, on average no such
trade-off was found. Longer response times did not result in higher accuracy. The reason for this
effect might by be the limited effort a searcher spends on surveying a display. Such limitations
are described in the Information Foraging Theory [9] which refers to cost-benefit analyses of
information’s value and the expected efforts to reach this information. Consequently, searching is
mostly a hasty and rough scanning of contents accompanied by fast decisions on the relevance of
contents. For the current experiment this means that participants clearly limited their efforts;
although participants were asked to search fast and accurate, none fixated all or at least almost all
word in the search panel. On average 65 fixations were applied to scan through the 140 words
within about 19 seconds. Within these limitations, a speed-accuracy trade-off does not
significantly impact on search performance.
In summary we found clear indications for the hypothesized mechanisms of visual search.
Highlighting facilitates - independent from its validity - the pre-attentive creation of activation
maps. The peaks of activation are fixated serially in an attentive stage of vision. As the results
show, fixations on highlighted items were significantly more frequent than fixations on non-
highlighted items. The deployment of fixations generally occurred in reading-like direction from
the top left row-wise to the bottom right. These results might be valid only for western culture. It
would be an interesting question for future research to compare search paths of different cultures.
Eye movements along the described search paths enable target detection either if the
target itself is highlighted or, by means of peripheral vision, if the target surrounds a fixated item.
Across target-present and -absent tasks in the 15-percent group response times were 1.35 seconds
faster than in the 0-percent group and 2.32 seconds faster than in the 49-percent group. For
target-present tasks, the 15-percent group yielded 2.23 faster response times than the 0-percent
group and 3.07 seconds faster response times than the 49-percent group. Moreover, the 15-
percent group reported 12 percent higher accuracy than the 0-percent group and 14 percent higher
accuracy than the 49-percent group.
The question remains, whether 15 percent highlighting is actually the optimal density for
the current search tasks. This value was arbitrary selected on the basis of a previous study, which
reported a distinct plateau of search performance between 15 and 30 percent highlighting. Future
research should address the question, which density is optimal for different kinds and different
amounts of material.
These results, furthermore, have significant practical impact. Referring to basic
mechanisms of human visual perception and attention, visual search speed as well as search
accuracy can be significantly improved with very simple methods. A crucial factor is that not
only performance can be increased but also a searchers satisfaction with a search episode and the
convenience with the presentation of contents.
This leads back to the sensationalism of this article’s title. It’s an interesting fact, that
Google, likely the most successful and popular Internet search engine, presents the search results
in a way that corresponds quite well to an optimal density of highlighting as reported in the
current and in previous [5] experiments. For example, the percentage of highlighting (bolding,
blue and green colors) of the first 10 results for when searching for “usability” is about 23%. Of
course, it’s far over-the-top to assume that Google’s design of search results, which by now is a
very popular design, is the only reason for the success of the company. But the results of the
current experiment suggest that highlighting contributed to that success to a certain extent.
5. References
1. Ben-Av, M. B., Sagi, D., & Braun, J., Visual attention and perceptual grouping. Perception and Psychophysics.
52 (1992).
2. Bullinger, H.-J., Ziegler, J., & Bauer, W., Intuitive Human-Computer Interaction - Toward a User-Friendly
Information Society. International Journal of Human-Computer Interaction. 14 (2002).
3. Fisher, D. L., & Tan, K. C.,Visual displays: The highlighting paradox. Human Factors. 31 (1989).
4. Fisher, D. L., Courey, B. G., Tengs, T. O., & Duffy, S. A., Minimizing the time to search visual displays: The
role of highlighting. Human Factors. 31 (1989).
5. Kickmeier, M. D., & Albert, D., The effects of scanability on information search: An online experiment, in:Ph.
Gray, H. Johnson, & E. O´Neill (Eds.), Proceedings of HCI 2003: Designing for Society, Volume 2, 2003.
6. Neisser, U., Cognitive Psychology, New York: Appleton, Century, Crofts 1967.
7. Padovani, S., & Lansdale, M., Balancing search and retrieval in hypertext: context-specific trade-offs in
navigational tool use. International Journal of Human-Computer Studies. 58 (2003).
8. Philipsen, G., Effects of six different highlighting modes on visual search performance in menu options.
International Journal of Human-Computer Interaction. 6 (1994).
9. Pirolli, P., & Card, S., Information foraging. Psychological Review. 106 (1999).
10. Posner, M. I., & Cohen, Y., Components of attention, in: H. Bouma & D. G.Bouwhuis (Eds.), Attention and
Performance X, Hillside: Erlbaum 1984.
11. Strasburger, H., Indirektes Sehen. Formerkennung im zentralen und peripheren Gesichtsfeld, in: A. Albert, M.
Oswald, K. Pawlik, & Stapf, K.-H. (Eds.), Lehr- und Forschungstexte Psychologie, Göttingen: Hogrefe 2003.
12. Wang, Q., Cavanagh, P., & Green M., Familiarity and pop-out in visual search. Perception and Psychophysics.
56 (1994).
13. Wolfe, J. M., Guided search 2.0: A revised model of visual search. Psychonomic Bulletin & Review. 1 (1994).
14. Wolfe, J. M., Visual Attention, in: K. De Valois (Ed.), Seeing, 2nd ed., San Diego: Academic Press 2000.
15. Wolfe, J. M., Guided Search 4.0: A guided search model that does not require memory for rejected distracters
[Abstract]. Journal of Vision. 1 (2001).
16. Wolfe, J. M., & Gancarz, G., Guided Search 3.0, in: V. Lakshminarayanan (Ed.), Basic and Clinical
Applications of Vision Science, Dordrecht: Kluwer Academic 1996.
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
One of the most important tasks of web-based inform ation retrieval and processing is searching. A large body of studies yielded that people, while foraging informa tion, do not read on-screen but rather scan. In the current study we try to elucidate the effects of salient items on re sponse time and answer accuracy in search tasks. In an online experiment we presented two text paragraphs, simila r to search engine results, asking the participants to d ecide as fast as possible whether the text is relevant for a given task or not. Each paragraph has been presented with a di fferent density of salient items (bold words). As dependent variables response time and answer accuracy have be en recorded. The results of this study show that the density of salient items has a clear impact on search efficien cy.
Article
Full-text available
Divides the study of human attention into 3 components: alertness, selectivity, and processing capacity. Experimental techniques designed to separate these components and examine their interrelations within comparable tasks are outlined. It is shown that a stimulus may be used to increase alertness for processing all external information, to improve selection of particular stimuli, or to do both simultaneously. Development of alertness and selectivity are separable, but may go on together without interference. Moreover, encoding a stimulus may proceed without producing interference with other signals. Thus, the contact between an external stimulus and its representation in memory does not appear to require processing capacity. Limited capacity results are obtained when mental operations, E.g., response selection or rehearsal, must be performed on the encoded information. (45 ref.) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Article
Full-text available
Information foraging theory is an approach to understanding how strategies and technologies for information seeking, gathering, and consumption are adapted to the flux of information in the environment. The theory assumes that people, when possible, will modify their strategies or the structure of the environment to maximize their rate of gaining valuable information. The theory is developed by (a) adaptation (rational) analysis of information foraging problems and (b) a detailed process model (adaptive control of thought in information foraging [ACT-IF]). The adaptation analysis develops (a) information patch models, which deal with time allocation and information filtering and enrichment activities in environments in which information is encountered in clusters; (b) information scent models, which address the identification of information value from proximal cues; and (c) information diet models, which address decisions about the selection and pursuit of information items. ACT-IF is instantiated as a production system model of people interacting with complex information technology. (PsycINFO Database Record (c) 2009 APA, all rights reserved)
Article
Full-text available
An important component of routine visual behavior is the ability to find one item in a visual world filled with other, distracting items. This ability to performvisual search has been the subject of a large body of research in the past 15 years. This paper reviews the visual search literature and presents a model of human search behavior. Built upon the work of Neisser, Treisman, Julesz, and others, the model distinguishes between a preattentive, massively parallel stage that processes information about basic visual features (color, motion, various depth cues, etc.) across large portions of the visual field and a subsequent limited-capacity stage that performs other, more complex operations (e.g., face recognition, reading, object identification) over a limited portion of the visual field. The spatial deployment of the limited-capacity process is under attentional control. The heart of the guided search model is the idea that attentional deployment of limited resources isguided by the output of the earlier parallel processes. Guided Search 2.0 (GS2) is a revision of the model in which virtually all aspects of the model have been made more explicit and/or revised in light of new data. The paper is organized into four parts: Part 1 presents the model and the details of its computer simulation. Part 2 reviews the visual search literature on preattentive processing of basic features and shows how the GS2 simulation reproduces those results. Part 3 reviews the literature on the attentional deployment of limited-capacity processes in conjunction and serial searches and shows how the simulation handles those conditions. Finally, Part 4 deals with shortcomings of the model and unresolved issues.
Article
Guided Search (GS) is a model of visual search implemented as a computer simulation. In GS2 (1994), items were ordered in terms of the amount of preattentive guidance and search proceeded from most likely to least likely until the target was found or until an adaptive "activation threshold" was reached and search was terminated. However, since then, our data have suggested that the search mechanism does not keep track of rejected distractors. This and other new results require a new GS implementation. Following in the tradition of Treisman's Feature Integration Theory, GS4 has two main stages. The first, preattentive stage generates a guidance map to direct the deployment of attention in the second stage. The preattentive stage processes a limited number of basic features in parallel across the visual scene. Color, size, and orientation are implemented in the simulation. There are two forms of preattentive guidance. Bottom-up guidance directs attention toward objects whose features differ from their neighbors. Top-down guidance directs attention toward objects that have target features. The guidance map is the noisy, weighted sum of top-down and bottom-up components. In GS4, target identification requires that preattentive features (e.g. red, vertical) must be bound by attention into a coherent object (e.g. a red, vertical bar) and that binding requires attention. In GS4's second, attentive stage, attention is deployed to the item with the greatest current preattentive activation, without regard to the previous history of deployments. GS4 successfully mimics human RT and error data for feature, conjunction, and "serial" search.It shows search asymmetries and eccentricity effects. It does not require memory for rejected distractors. Finally, it makes testable predictions about new search phenomena some of which we have been able to confirm.
Article
Highlighting (HI), for example of menu options, is usually considered as improving information processing efficiency. Particularly relevant are the attention effects of valid (target) highlighting, when contrasted with the effects of highlighting distracting information (invalid HI). A visual search experiment was conducted to analyze the requirements of such an overall performance gain under valid HI. The efficiency of six different modes of highlighting (brightness increase, reverse video, color [red], reverse color, blue‐on‐yellow, and red‐on‐green) on a black screen background was determined. Sixty‐two subjects indicated the presence of a prespecified menu option in a list of CAD menu options, of which a subset was highlighted on every trial. Also varied was familiarity of highlighting. Dependent measures were the absolute performance scores (search time and accuracy) and highlighting benefits. Significant absolute performance differences between the modes were found. They were accompanied by a marked dissociation between the absolute and the benefit measures. The HI benefits were calculated as the individual performance differences between target‐highlighted trials (valid HI) and target‐standard (invalid HI) conditions. The bright white mode, which caused the shortest absolute search time, did not result in a benefit. In contrast, the red color condition produced a large benefit, accompanied by only mid‐range absolute performance. The widely used reverse video modes generally led to the poorest performance. Familiarity of highlighting had a differential effect, depending on the dependent measure preferred. Recommendations for the use of highlighting in display design are given.
Article
This article is an extended version of a keynote address presented at the 9th International Conference on Human-Computer Interaction held in New Orleans, August 5 through 10, 2001. The article presents an overview of issues pertaining to the development of a user-friendly information society, and introduces 3 concepts for improving the intuitiveness of human-technology interaction in the future. Dynamic visualizations, multimodal interaction, and cooperative exploration are considered as 3 major approaches to improving human-technology interaction. These concepts and interaction techniques based on them are currently being investigated in a large joint research effort, the project INVITE (Intuitive Human-Technology Interaction in the Information Society). An overview of INVITE's objectives and achievements is given. Selected application areas are presented in which the concepts and techniques are applied. Finally, the issue of the relation between the physical work environment and creative activities is addressed. A spatial environment with various technical devices is described that offers a range of techniques for stimulating creativity and intuition.
Article
We compare user navigation performance using two hypertext information sites of identical node structure but embedded in different metaphors. The first is based upon the layout of a house and is consistent with Euclidean space. The second represents social links between people for which a spatial metaphor is not apparent. Search for targets within the structures, and the speed of their subsequent retrieval on a second search, is compared in a 2 × 4 × 2 factorial design manipulating: metaphor (spatial or non-spatial); navigation, tools (participants have both a site map and bookmark tool, one of these, or no tools at all) and the time pressure under which navigation is carried out (paced or unpaced). A strong main effect is found in which the spatial metaphor produces higher performance under all conditions. Similarly, time pressure has the general effect of trading-off a faster initial search with less efficient retrieval later. However, navigation tool use is highly context dependent and sometimes counterintuitive: certain conditions show poorer performance with two navigation aids than one. We argue that navigation tools are mediating structures for activities, such as bookmarking and learning the structure of the site, which represent cognitive investment for future retrieval. In this view, user performance is optimized by the balance of two potentially antagonistic conditions. First, the usability of tools and metaphor must free cognitive resources for planning; but also, the difficulty of the task and the need for planning must remain visible to the user. The implications for design are discussed.
Article
Perceptual organization is thought to involve an analysis of both textural discontinuities and perceptual grouping. In earlier work, we found that textural discontinuities were detected normally even when visual attention was engaged elsewhere. Here we report how perceptual grouping is affected when visual attention is engaged by a concurrent visual task. To elicit perceptual grouping, we used the Gestalt demonstrations of grouping on the basis of proximity and of similarity. Four tasks were investigated, some requiring the observer to discriminate between horizontal and vertical grouping, and some requiring the observer to merely detect the presence or absence of grouping. Visual attention was engaged at the center of the display by a form identification task. The detection of a textural discontinuity served as a control task. Concurrent form identification conflicted with all four grouping tasks, resulting in a significant reduction of grouping performance in each case. No performance reduction was observed when either form identification or grouping discrimination was combined with the detection of a textural discontinuity. These results suggest that perceptual grouping and form identification compete for visual attention, whereas the detection of a textural discontinuity does not.