Perceptions of Graphics
Perceptions of Graphics Versus No Graphics on Web Sites
by
Rebecca J. Chamberlin
Doctoral Candidate
E.W. Scripps School of Journalism
Ohio University
109 N. Plains Rd.
Apartment 18
The Plains, Ohio 45780
740-797-2977
[log in to unmask]
Perceptions of Graphics Versus No Graphics on Web Sites
The World Wide Web is an evolving medium with traits very different from other
media. Learning how to best use this medium is one of the biggest challenges
facing communicators. John V. Pavlik, executive director of The Center for New
Media Studies at Columbia University's Graduate School of Journalism, wrote,
"For many of us in the field, the point (of online journalism) is to engage the
unengaged."1
Literature Review
Understanding how to best create online journalism is a major challenge. In the
September 1997 issue of Journalism and Mass Communication Monographs, Eric
Fredin proposed some ways of constructing news stories on the web.2 He suggested
options that would utilize many of the web's unique traits, such as hyperlinks
and non-linear writing.
Constructing great content is only half the battle of creating web journalism.
Mario Garcia, from the Pointer Institute for Media Studies, wrote, "The first
step in the design of a web site, or any other form of design for that matter,
is to create a visual environment that organizes the material to be presented
and that is suitable for the content."3 Once people click to a page, they must
feel an attraction to what they see and they must understand how they can get
the information they need from the site. That's where web site design comes in.
If a site is difficult to navigate, or difficult to read, the viewers are less
likely to read the content.
Mindy McAdams, a web consultant and AJR NewsLink columnist, wrote that web
designers must continually work on their sites to make them better.4 Successful
sites "experiment with new functionality and find out whether users react
positively," she wrote.
Jakob Nielsen, who leads web usability studies for Sun Microsystems, wrote,
"Web users are impatient. They want to get their answers immediately and do not
want to be slowed down by 'cool' features, mission statements, or self-promoting
grandstanding."5 Golding and White agree that many graphics may cause problems.
"On the web, what concerns us most is how quickly the file downloads."6 So
having graphics is not enough, but having fast-loading graphics is also
necessary.
Eric K. Meyer, a managing partner of the AJR NewsLink online research and
consulting firm, wrote that one myth of online publishing is that readers
"demand rich audio-visual presentation." In fact, he said readers, and
especially return visitors, want efficient presentation.7 Big advertisers and
big sites are even limiting their use of graphics.
When graphics in printed newspapers have been studied, researchers find that
you can have too many graphics, and too complicated graphics. James Tankard
warned graphic designers of several pitfalls of newspaper graphics, including
the "overly-complex graph" and the "unnecessarily three-dimensional graph."8
The question still remains: do graphics really help in the viewer's
understanding of the site? Cronin and Myers looked at the effects of using
visuals versus no visuals in the instruction of interactive media.9 They sited
examples of previous educational studies that found visuals helped people
understand the content. Visuals had also improved the recall of information. But
when Cronin and Myers performed their own study, they did not find evidence to
support the earlier research. The students who used visuals did not achieve
higher cognitive test scores and did not have significant gains on a listening
test. This study suggests that visuals are not as effective in learning as
earlier studies found.
Redundancy in television graphics allows viewers' to better understand the
story, according to an experiment by Drew and Grimes.10 They looked at both
images and sound and found that recall improved with redundancy of the two.
Mark Timney looked at visual referencing and found no improvement of recall for
high or low complexity graphics.11 Recall of information graphics on television
"appears to be related more to individual perceptual differences than graphical
treatment," he wrote.
Whether or not visuals do help in learning or recall, perhaps they do persuade
the viewers to read the content. But studies have not determined if visuals or
graphics really make the viewers more interested in the site, or if graphics are
preferred over content.
Research Questions
Many authors and web site designers have ideas about how they should design
their sites. But little is known about the viewer's perceptions and preferences
when it comes to web sites. If we understand what kind of sites people prefer,
we can create web sites that they enjoy. The main research question is:
How does the design of a web site affect people's perceptions of it?
There are six detailed research questions:
1. Can people find information easier on a high-graphic site or on a
low-graphic site?
2. Do people perceive a site with many graphics to be more appealing
than a
site with fewer graphics?
3. How do people judge a web siteDby looking at content or graphics?
4. Do people who focus on content v. graphics differ in their
perceptions
of a web site?
5. Do people of different demographics perceive web sites
differently?
6. Do heavy web users perceive sites differently than light web
users?
Method
To answer these research questions, an experiment was conducted using two web
publications. One publication contained articles about heart disease, and the
other contained articles about recycling. These topics were used because of
their neutrality. The web publications and articles were written and designed by
the author. Information for the articles was obtained using search engines on
the web. Then the web publications were designed to look like a newspaper
feature section or magazine publication. The sites did not look like the main
home page for a newspaper or magazine, because a main page would have included
many links and graphics. Instead, the publications looked like an "inside" page
of a site, where there are more articles and fewer links and graphics.
For each of the two publications, there was a low-graphic version and a
high-graphic version. The low-graphic version contained a simple nameplate as
the only graphic. The high-graphic version contained animation, advertisements,
photos and a nicer nameplate. Both versions contained identical text. Because
there were two publications, each with a low- and high-graphic version, there
were four sites to be viewed.
A pre-test was performed by eight undergraduate and graduate students to
determine if the links to the sites worked correctly and to verify that they
could see a distinct difference between the high-graphic and low-graphic sites.
The experiment was then conducted on five groups of students. The groups
included two Journalism on the Web classes (which contained undergraduate and
graduate students who were highly interested in the web), two beginning
communications classes (which contained undergraduate students with varied
interest in the web), and a secretarial class at a small career center (which
contained adult students with less web experience).
The respondents ranged in age from 19 to 52, with a mean of 24.3. Over half of
the respondents (58.7%) had been using the web between one and three years. Over
a quarter (25.8%) had been using the web less than a year, while the remaining
(15.5%) had been using the web over three years. When looking at the
respondents' weekly web use, nearly half (44.6%) used the web between two and
five hours per week. One-third (31%) were infrequent web users, getting online
one hour or less per week, while nearly a quarter (22.4%) were on the web six or
more hours per week.
Each student was asked to view two of the four web sitesDthe high-graphic of
one publication and the low-graphic of the other publication. The articles were
distributed so that half the students saw a high-graphic site first and half saw
the low-graphic site first. This eliminated any primacy effect. A total of 58
students viewed the sites. Because each student viewed two sites, the sites were
viewed a total of 116 times.
Results
When asked to name the kind of sites they view frequently, entertainment and
news sites led the respondents' lists, as shown in Table 1.
Table 1
Types of sites viewed frequently
Type N
Entertainment 42
News 39
Education 19
Other 13
Business 11
Government 5
Total 129*
*Total equals more than the number
of respondents because each respondent
could select more than one kind of site.
A Kolmogrov-Smirnov test for normality found the responses were not normally
distributed (p<.05); therefore, non-parametric procedures were used to answer
the research questions.
Research Question 1
The experiment questionnaire asked respondents to browse around the web
publications and answer three factual questions found on the site. They were
then asked how difficult it was to find the answers to the three factual
questions. These questions regarding difficulty were used to answer research
question 1: Can people find information easier on a high-graphic site or on a
low-graphic site? Table 2 shows there was no difference in difficulty between
the high- and low-graphic sites, no matter which of the two web publications
were being viewed. There was also no difference between the level of difficulty
of the high- or low- graphic sites when looking at the combination of the two
publications. Therefore, neither graphics, nor a lack of graphics, makes a
difference when searching for content on the web.
Table 2
Difficulty in answering questions
on high-graphic sites v. low-graphic sites
N mean rank sum of ranks
Publication A
Question 1 (high graphic) 28 26.3 737.0
Question 1 (low graphic) 30 32.5 974.0
Total* 58
Question 2 (high graphic) 28 29.8 834.5
Question 2 (low graphic) 30 29.2 876.5
Total** 58
Question 3 (high graphic) 28 26.9 752.5
Question 3 (low graphic) 30 32.0 958.5
Total*** 58
*Mann-Whitney U = 331.00; p = .117; not significant
** Mann-Whitney U = 411.50; p = .886; not significant
*** Mann-Whitney U = 346.50; p = .221; not significant
Publication B
Question 1 (high graphic) 30 29.2 876.5
Question 1 (low graphic) 28 29.8 834.5
Total* 58
Question 2 (high graphic) 30 29.3 876.5
Question 2 (low graphic) 28 29.7 832.5
Total** 58
Question 3 (high graphic) 30 30.7 920.5
Question 3 (low graphic) 28 28.2 790.5
Total*** 58
* Mann-Whitney U = 411.50; p = .884; not significant
** Mann-Whitney U = 413.50; p = .915; not significant
*** Mann-Whitney U = 384.50; p = .559; not significant
Table 2 (continued)
Difficulty in answering questions
on high-graphic sites v. low-graphic sites
N mean rank sum of ranks
Publications A & B combined
Question 1 (high graphic) 58 55.1 3196.0
Question 1 (low graphic) 58 61.9 3590.0
Total* 116
Question 2 (high graphic) 58 58.6 3399.0
Question 2 (low graphic) 58 58.6 3387.0
Total** 116
Question 3 (high graphic) 58 57.2 3319.5
Question 3 (low graphic) 58 59.8 3466.5
Total*** 116
*Mann-Whitney U = 1485.00; p = .226; not significant
** Mann-Whitney U = 1676.00; p = .972; not significant
*** Mann-Whitney U = 1608.50; p = .666; not significant
Research Question 2
Research question 2 asks: Do people perceive a site with many graphics to be
more appealing than a site with fewer graphics? Respondents were asked to judge
the sites for their level of attractiveness, ranging from very unattractive to
very attractive. A Mann-Whitney test found a significant difference between the
high- and low-graphic sites, as shown in Table 3. When looking at either web
publication, the high-graphic site was considered more attractive.
Table 3
Attractiveness of high graphic sites v. low graphic sites
N mean rank sum of ranks
Publication A (high graphic) 30 35.5 1063.5
Publication A (low graphic) 28 23.1 647.5
Total* 58
Publication B (high graphic) 28 35.2 986.5
Publication B (low graphic) 30 24.2 724.5
Total** 58
* Mann-Whitney U = 241.50; p<.05
** Mann-Whitney U = 259.50; p<.05
Research Question 3
Research Question 3 asks: How do people judge a web siteDby looking at content
or graphics? The respondents were asked how they determine if a site is good or
badDby looking at content, graphics, or a combination. As shown in Table 4, the
responses were almost evenly split between "both, focusing on graphics," "both,
focusing on content" and "both graphics and content evenly." The respondents
seemed to recognize that graphics and content are both important when looking at
web sites.
Table 4
How people judge a web site: by graphics or content?
Type %
Content only 0.0%
Graphics only 3.5%
Both, focusing on graphics 29.3%
Both, focusing on content 37.9%
Both evenly 29.3%
N=58
Research Question 4
Research Question 4 asks: Do people who focus on content versus graphics differ
in their perceptions of a web site? A crosstab looked at the respondents'
preferences for graphics or content and their views of the attractiveness of the
web sites. There were fewer than five case in many of the cells, so the
categories were collapsed. The preferences for "content only" and "graphics
only" were combined with "both, focusing on content" and "both, focusing on
graphics," respectively. On the scale of attractiveness, "very unattractive" and
"unattractive" were combined, and "very attractive" and "attractive" were
combined. Please note that though there were 58 respondents, they each viewed
two sites, so the total number of responses is 116. As shown in Table 5, the
crosstab found no significant differences. Therefore, people who focus on
content versus graphics do not differ in their perceptions of a web site.
Table 5
Preference for content or graphics by attractiveness of sites
focus on both focus on
graphics evenly content
% % %
unattractive 15.8% 11.8% 13.6%
neutral 50.0% 41.2% 40.9%
attractive 34.2% 47.1% 45.5%
N=38 N=34 N=44
chi square=1.55; df=4; p>.05; not significant
Research Question 5
Research question 5 asks: Do people of different demographics perceive web
sites differently? Crosstabs looked at gender and age against respondents'
preferences for graphics or content and their perceptions of site
attractiveness. Because of too few categories in some cells, the collapsed
categories that were used in research question 4 were used again. The age groups
were created by dividing the respondents into three nearly equal groups. The
groups are as follows: ages 20 and under (34.5%), ages 21 to 24 (36.2%), and
ages 25 and over (29.3%).
There was no difference between men's and women's preferences toward content or
graphics, but there was a difference in the age groups (see Table 6). The age 25
and older group had very few people who focused on graphics (17.7%), while the
age 21 to 24 group had more than half of the respondents selecting graphics
(52.4%). The 20 and under group was more evenly distributed, but it still had
only a quarter of its respondents selecting graphics (25%).
Table 6
Gender, Age by preference for content or graphics
Gender male female
% %
content 33.3% 40.5%
graphics 42.9% 27.0%
both 23.8% 32.4%
N=42 N=74
chi square=3.10; df=2; p=.213; not significant
Age 20 and under 21 to 24 25 and over
% % %
content 45.0% 23.8% 47.1%
graphics 25.0% 52.4% 17.7%
both 30.0% 23.8% 35.3%
N=40 N=42 N=34
chi square=12.38; df=4; p<.05
To further answer research question 5, crosstabs looked at the gender and age
groups' view on site attractiveness. Both found significant differences. The
females rated nearly half the sites either attractive or neutral (47.3% each),
with few (5.4%) rated as unattractive. Males, on the other hand, considered more
sites to be unattractive.
There were also significant differences when the age groups were crosstabulated
with site attractiveness. Half of the 20 and under group considered the sites to
be neutral (50%), while over half of the 25 and over group considered the sites
to be attractive (52.9%). The middle age group, like the younger group, also
considered more sites to be neutral (45.2%) than attractive (40.5%). There
could be several reasons why the older age group considered sites to be more
attractive than the younger age groups. Perhaps the older group had less
experience viewing graphics and considered all graphics to be attractive. The
younger people may have taken the graphics for granted or may have had higher
expectations for their images. Another explanation can be found by looking at
Table 6 again. Recall that the middle age group focused more on graphics than
the other groups. Perhaps because they focused on graphics, they were more
critical. The older age group that focused on content may have been less
critical of graphics and attractiveness.
Table 7
Gender, age by attractiveness of site
Gender male female
% %
unattractive 28.6% 5.4%
neutral 38.1% 47.3%
attractive 33.3% 47.3%
N=42 N=74
chi square=12.18; df=2; p<.05
Age 20 and under 21 to 24 25 and over
% % %
unattractive 15.0% 14.3% 11.8%
neutral 50.0% 45.2% 35.3%
attractive 35.0% 40.5% 52.9%
N=40 N=42 N=34
chi square=12.38; df=4; p<.05
Research Question 6
Research question 6 asks: Do heavy web users perceive sites differently from
light web users? The preference for content or graphics and the level of site
attractiveness were collapsed the same way as in research question 5. To
determine the levels of heavy or light web users, the respondents were asked how
many hours per week they use the web and how many years they have been using the
web.
For the weekly web use, the respondents selected categories of "less than one
hour" (31.0%), "2 to 5 hours" (46.6%), "6 to 10 hours" (13.8%) and "more than 10
hours" (8.6%). The respondents were collapsed into two categories: "5 hours or
less" and "6 or more hours."
Respondents categorized their length of web use by selecting from "6 month or
less" (15.5%), "7 months to a year" (10.3%), "1 to 2 years" (32.8%), "2 to 3
years" (25.9%), "3 to 4 years" (10.3%), and "over 4 years" (5.2%). The first two
and last two groups were combined to make four relatively even categories.
There was no difference in the preference for content or graphics when looking
at the weekly web use, as shown in Table 8. However, there was a significant
difference when looking at the number of years they had been using the web.
People using the web less than a year focused on "content" (46.7%), while people
using the web 1 to 2 years tended to select "both content and graphics evenly"
(52.6%). In each of the "2 to 3 years" and "over 3 years" categories,
respondents selected "content" and "graphics" rather than selecting "both."
There is no obvious explanation for these differences. Perhaps, when people
first use the web, they are discovering new sites and feel like they focus on
content. As they use the web longer, they recognize the fact that they look at
both content and graphics. Then, as they become more experienced, they tend to
find their own niche and move toward either content or graphics. Another
explanation is that the respondents in this study just happened to fall into
these categories, but the rest of the population would not necessarily
categorize themselves in the same way.
Table 8
Weekly web use, length of web use
by preference for content or graphics
Weekly web use 5 hours or less 6 hours or more
% %
content 35.6% 46.2%
graphics 33.3% 30.8%
both 31.1% 23.1%
N=90 N=26
chi square=1.08; df=2; p=.582
Length of web use
under 1 year 1 to 2 years 2 to 3 years over 3 years
% % % %
content 46.7% 26.3% 40.0% 44.4%
graphics 33.3% 21.1% 40.0% 44.4%
both 20.0% 52.6% 20.0% 11.1%
N=30 N=38 N=30 N=18
chi square=15.87; df=6; p<.02
To further answer research question 6, a crosstab looked at the weekly web use
and length of web use by the site attractiveness. There were no significant
differences in either crosstab, as shown in Table 9.
Overall, there was only one significant difference in research question 6Dlight
versus heavy web users' perceptions of web sites. And, as stated earlier, that
one difference can not be easily explained. Therefore, there does not seem to be
any explainable difference in light versus heavy users' perceptions of the web.
Table 9
Weekly web use, length of web use by attractiveness of site
Weekly web use 5 hours or less 6 hours or more
% %
unattractive 13.3% 15.4%
neutral 42.2% 50.0%
attractive 44.4% 34.6%
N=90 N=26
chi square=.800; df=2; p=.670
Length of web use
under 1 year 1 to 2 years 2 to 3 years over 3 years
% % % %
unattractive 10.0% 18.4% 6.7% 22.2%
neutral 40.0% 42.1% 56.7% 33.3%
attractive 50.0% 39.5% 36.7% 44.4%
N=30 N=38 N=30 N=18
chi square=5.37; df=6; p=.497 (25% of cells have expected counts less than 5)
Conclusions
Decisions made in this study have, of course, impacted the results, and readers
must be cautious about them. The sites that the respondents viewed were created
by the author and did not contain some of the Java scripting and other graphic
elements that are found in many professional web sites today. Because the high-
and low-graphic sites needed to contain the same content, it was impossible to
make the high-graphic site look like many of the well-made newspaper sites
today. Though students did see a difference between the high- and low-graphic
sites, there still could have been much more added to the high-graphic sites to
make them even more realistic of today's journalism on the web.
Another of the sites' traitsDspeedDcould have also influenced the results. Even
on the high-graphic sites, the images were created in small file sizes, and thus
loaded quickly. The experiment was also conducted on a T1 web connection.
Therefore, the slow download time that many modem-users dislike was eliminated.
Perhaps if the high-graphic sites had loaded slowly, the results would have been
different.
Yet, this study provides a clear indication that though the high-graphic sites
were considered more attractive than the low-graphic sites, there was no
difference in the level of difficulty between finding information on a high- or
low- graphic site. In addition, respondents used both content and graphics to
judge the quality of a web site. They were almost evenly split between "both,
focusing on graphics," "both, focusing on content" and "both graphics and
content evenly." Also, people who focused on content versus graphics did not
differ in their perceptions of a web site.
When looking at gender and age, there was no difference between men's and
women's preferences toward content or graphics, but there was a difference in
the age groups. The older age group had very few people who focused on graphics.
When looking at attractiveness of a site, females considered sites to be more
attractive, while males considered them more unattractive, and the older age
group considered sites to be more attractive than the younger age groups did.
When comparing light versus heavy web users, there does not seem to be any
explainable difference in their perceptions of the web.
The overall research question asked: How does the design of a web site affect
people's perceptions of it? There may not be any difference between finding
information in a high-graphic site versus low-graphic site. However, different
demographic groups do have different perceptions of attractiveness and different
preferences for graphics or content. So even though they could adequately find
information on any well-organized site, people may prefer certain types of
sites. Some people will prefer the graphics, while others will focus on content.
It is important to continue to evaluate web sites and people's attitudes toward
those sites. The media have performed audience analyses for years, and the web
audience should be no exception. Future study could include more detailed
evaluation of the kind of graphics people prefer on web sites. It could also
include content analyses of the web designers' continual struggle between
content and graphics. End Notes
1 John V. Pavlik, "The Future of Online Journalism," Columbia Journalism Review,
July/August 1997, <http://www.cjr.org/html/97-07-08-online.html> (March 1998).
2 Eric S. Fredin, "Rethinking the News Story for the Internet: Hyperstory
Prototypes and a model of the User," Journalism and Mass Communication
Monographs, 163 (September 1997).
3 Mario R. Garcia, Redesigning Print for the Web, (Indianapolis, IN: Hayden
Books, 1997).
4 Mindy McAdams, "Back to the Drawing Board," AJR NewsLink, March 10-16, 1998,
<http://www.newslink.org/mmcol3.html> (March 1998).
5 Jakob Nielsen, "Changes in Web Usability Since 1994," Alertbox, December 1,
1997_ <http://www.useit.com/alertbox/9712a.html> (March 1998).
6 Mordy Golding and Dave White, Web Designer's Guide to Color, (Indianapolis,
IN: Hayden Books) 1997.
7 Eric K. Meyer, "The 10 Myths of Online Publishing," AJR NewsLink, March 10-16,
1998, <http://www.newslink.org/emcol3.html> (March 1998).
8 James W. Tankard, Jr., "Quantitative Graphics in Newspapers," Journalism
Quarterly 64 (summer/fall 1987).
9 Michael W. Cronin and Sharon L. Myers, "The Effects of Visuals Versus No
Visuals on Learning Outcomes from Interactive Multimedia Instruction," Journal
of Computing in Higher Education 8 (spring 1997).
10 Dan G. Drew and Thomas Grimes, "Audio-Visual Redundancy and TV News Recall,"
Communication Research 14 (August 1987).
11 Mark C. Timney, "Television News Infographics and Variables Affecting
Information Recall," dissertation, Ohio University, November 1997.
Perceptions of Graphics Versus No Graphics on Web Sites
An experiment was conducted to better understand how the design of a web site
affects the viewers' perceptions of it. High-graphic and low-graphic versions of
web sites were compared by five groups of viewers. There was no difference in
how difficult the viewers felt it was to find information on the sites. However,
different demographic groups had different perceptions of attractiveness and
different preferences for content or graphics.
|
