Anda di halaman 1dari 34

Criteria for optimal web design (designing for usability)

By Michael L Bernard
Designing a website that takes into account the human element requires both an
understanding of our nature as well as our physiological limitations. sable websites
incorporate human tendencies and limitation into its o!erall design. "he questions below are
meant to address some of the more important human factors concerns in the design and
building of usable websites.
#ow should information be positioned in a typical website$
#ow can % make my website&s structure more na!igable$
#ow should te't be presented within a website$
#ow can % effecti!ely use images on my website$
(re frames e!er appropriate$
#ow can % design a !isually pleasing interface that follows usability principles$
#ow can % reduce the ma)or user annoyances on my site$
#ow can % make my site more accessible to children$
#ow can % make my site more accessible to older adults$
#ow can % make sure my site follows general *eb con!entions$
#ow can my website promote customer sales and loyalty$
#ow can % make my site more appealing to international users$

#ow should information be positioned in a typical website$
"he organi+ation of information within websites is !ital to its o!erall usefulness. %n fact, a
study by Morkes and -ielsen found that their e'perimental website .scored /01 higher in
measured usability when it was written concisely, 231 higher when the te't was scannable,
and 431 higher when it was written in an ob)ecti!e style instead of the promotional style
used in the control condition in many current web pages. (p. 5). "hat is, !iewers tend not to
spend too much time on a particular page. %nstead they usually scan for information that is of
direct interest to them. %t is therefore recommended that the te't within websites be !ery
succinct, with only one key idea per paragraph, as well as use highlighted keyword or phrases
and bulleted lists (Morkes 6 -ielsen, 5773).
sers ha!e grown accustomed to looking in certain areas on a screen to find specific items.
(naly+ing users& e'pectations of where they e'pect specific web ob)ects to be located
re!ealed that, 5) the web pages that link to pages within a site are generally e'pected to be
located on the upper left side of the browser window, 4) the web page links to other
websites are generally e'pected to be located on the right side or lower left side of the
browser window, 8) the .back to home. link is generally e'pected to be located at the top9left
corner and the bottom9center of the browser window, 2) the internal search engine is
generally e'pected to be located at the top9center of the screen, and /) ad!ertisement
banners are generally e'pected to be located at the top of the browser window (Bernard,
4::5).
%n follow9up study (Bernard, 4::4) that analy+ed participants who bought at least one item
online re!ealed that; <) the login=register button was e'pected to be located at the upper9left
corner of a web page, 3) the shopping cart button was e'pected to be located at the top9right
corner of a web page, 0) the help button was e'pected to be located at the upper9right side,
7) links to specific merchandise items were e'pected to be located at the left upper9center of
a web page, and 5:) the account=order button was e'pected to be located at the upper9right
of a web page. "he figure below shows the combined location e'pectations for the ten web
ob)ects.
Figure 1. Location for internal web page links Figure 2. Location for external website links
Figure 3. Location for "back to home" link Figure 4. Location for internal search engine
Figure 5. Location for adertisement banners

Figure !. Location for the login"register button
Figure #. Location for the shopping cart button

Figure $. Location for the help button
Figure %. Location for links to merchandise items

Figure 1&. Location for the account"order button
sers often miss important pieces of information simply because it is not seen. "his often
occurs because they forget or are unwilling to scroll in a particular direction (especially
hori+ontally), and thus do not see the information that is located outside of the primary
!iewing area. "o reduce this problem, important website information should always fit within
the typical hori+ontal !iewing area of the screen. "o do this, the rule is still to design for
lower resolution settings. (ccording to the most recent >? sur!ey (5770), o!er 541 of
users ha!e a resolution of <2: ' 20:. "he actual usable si+e to a!oid any scrolling at this
resolution is /7/ ' 47/ pi'els (the safe width for printing at this resolution is /8/ pi'els).
Most users howe!er ha!e their resolution set at 0:: ' <:: (851). "o a!oid scrolling here, the
usable si+e is 3/: ' 24/ pi'els. ( compromise would be to place the most important
information within areas that are !isible at lower resolution settings, while placing less
important information in areas !isible at higher resolution settings.
%n addition, when users do scroll, they may not see the information because it is placed in a
typically low information9priority area, such as the bottom of a page (-ielsen, 5777) or
placed in an area where users typically would not e'pect it to be placed.
@luid layouts are significantly preferred to both centered and left9)ustified layouts. %n a study
by Bernard and Larsen (4::5) participants indicated they percei!ed the fluid layout (which
the margins are not fi'ed at any particular width) as being the best suited for reading and
finding information, as well as ha!ing a layout that is most appropriate for the screen si+e
(for both small and large screens). "hey also indicated that the fluid layout looked the most
professional, and consequently preferred it to other layout conditions. Con!ersely, the
consistently least preferred condition was the left9)ustified layout. ( possible reason for the
lack of preference for this layout is that users had to hori+ontally scroll in order to see all the
information on the page. (s discussed abo!e, users particularly dislike to hori+ontally scroll.
Links with summaries are percei!ed as the most usable and are preferred to links without
summaries. ( study by Baker, Bernard, and Ailey (4::4) found no statistical differences in
search time across conditions with links with summary, links only, and full te't. #owe!er, the
summary condition was percei!ed as being the easiest in finding information, being !isually
pleasing, promoting comprehension, participants& satisfaction with the site, and looking
professional. "he summary condition was the most preferred, while the full te't condition
was the least preferred. "he full te't condition was percei!ed as being most difficult to find
information, not promoting comprehension, not being !isually pleasing, and not being
satisfying.
Barticipants reported that they preferred the summary condition o!er the Links only condition
because the brief summaries accompanying the links often guided them to the information
they were searching for. Barticipants commented that, in the links only condition, they
sometimes felt as if they were .)umping blindly. into the article. Ce!eral participants also
reported that they did not like ha!ing to scroll through all of the articles in the full te't
condition. "his study suggests that pro!iding a small amount of information about an article
on a page is superior to ha!ing long, scrolling pages filled with articles.
%n presenting a list of links, we found that it is best if they are bulleted. @or instance, as
discussed in sability -ews, Cpain compared the accuracy rates for three link conditions;
bulleted links, space between the links, and a no bullet=no space condition. %t was found that
the accuracy rate was 5::1 for bullets, 071 for spaces, and <31 for no spaces. (ll
participants preferred either the bullets or spacesD no one preferred the no space condition
(Cpain, 5777). %n support of this, Barkinson, Cisson, and Cnowberry (570/) found that menus
with spacing were searched 4/1 faster than menus without spacing.
#ow can % make my website&s structure more na!igable$
Beople often become lost within the structure. %n fact, /01 of users will make two or more
na!igational errors while searching for information (@orsythe, et al., 577<) and <<.01 of
users ha!e stated that one of the greatest problems about the *eb is .not being able to find
the information that % am looking for. (>?, 5770). >enerally there are four ma)or reasons
for this occurrence (@oss, 5707);
@irst difficulty is disorientation or .lost9in9hyperte't problems, which arises from an
unfamiliarity with the structure or conceptual organi+ation of the site. #ere, users ha!e
difficulty deciding which node (which is typically one web page) to !iew ne't because they
are unable to !isuali+e where the information they are looking for could be. "he decision
concerning which node to !iew ne't first in!ol!es understanding oneEs current location within
the site, then selecting the proper route. #owe!er, users may not e!en know their current
location within a site.
( proper way to reduce this problem is to organi+e the site according to the typical users&
mental model of how a site should be organi+ed. "his can be done by ha!ing representati!e
users sort cards into se!eral categorical piles in which each card represents the information
that would be placed on the actual website. Fach pile should indicate the information that
would be clustered within each category and subcategory. "his would gi!e the designer
knowledge on how users mentally organi+e the structure of a particular site 9a technique that
uses this method is discussed in sability -ews (Bernard, 4:::).
%n addition, the placement of submenu titles may also help reduce disorientation. @or
e'ample, >ray (570<) found that of the na!igation errors made within a hierarchy, 2:1 of
them were in the third and fourth le!els with submenu titles. *ithout submenu titles, /71 of
the errors were made in the third and fourth le!els. Moreo!er, according to Bransford and
Gohnson (5734), participants who ha!e read passages with titles recalled appro'imately twice
as many items from the passages and had higher comprehension than participants who did
not ha!e passages with titles.
"he use of na!igational aids such as color coding and consistent logos and banners should
also reduce disorientation (see the Hale Ctyle Manual for a discussion of this) and the use of
the .bread crumb. na!igation technique (see e'ample below) seems to work well in reducing
the disorientation problem as well (see -ielsen, 4:::, for a discussion of this).
#omepage I @irst le!el I Cecond le!el I presently !iewed le!el
"he second difficulty is the embedded digression problem. "his occurs when users pursue
digressi!e paths within websites and lose their place or forget to return to their original
document. "his can be lessened by reducing the number of links embedded in te't by placing
them instead at the end or on the side of the document. #owe!er, Jno!ed and Chneiderman
(570<) found that users preferred and were more accurate in answering information using
embedded links than an e'plicit grouping of links outside the te't. Het, they also stated that
embedded links could be disrupti!e in that the user .may be inclined to e'amine a particular
sub)ect or sub)ects in detail without first getting an appreciation of the o!erall conte't. (p.
85<).
( recent study by Bernard, #ull, 6 Drake (4::5) e'amined the effects of embedding
associati!e links with a document, as well as placing them at the bottom, at the top9left, and
left, at the same height in which they correspond with the document. -o significant
differences between the four link arrangements were detected in terms of search accuracy,
time, or efficiency. #owe!er, there were significant sub)ecti!e differences between the links
arrangements fa!oring the embedded links. "hat is, participants indicated that they belie!ed
that embedding the links within a document made it easier to na!igate, more easily recogni+e
key information, promoted comprehension, and was easier to follow the main idea of the
passages while searching for specific information. Moreo!er, participants significantly
preferred the embedded link arrangement to the other arrangements. Con!ersely, placing
links at the bottom of a document was percei!ed as being the least na!igable arrangement,
and was consequently least preferred.
"hus, while embedded digression may be a problem for some users, this should be weighed
against the sub)ecti!e perceptions that fa!or the embedded link arrangement.
"he third difficulty is the .art museum. problem. "his refers to the lack of memory for the
na!igational details of a significant part of the site because the !iewer is o!erwhelmed by the
sheer amount of information. @or instance, as when a patron !isiting a museum cannot hope
to remember the details of all the art work because of their great number, a large number and
!ariation of na!igational information (such as the !arious nodes they ha!e !isited) may
consequently o!erwhelm the user. "his often can ha!e the effect of reducing a person&s recall
of the pages they ha!e !isited.
"his can be lessened by reducing the amount of information presented at one time and
properly organi+ing the na!igational structure of the site. @or e'ample, in a study comparing
three types of structures; pure hierarchical (web pages at one le!el can only access by a web
page directly abo!e or below it), nonlinear (links could be connected to any number of other
web pages on the site), and mi'ed design (hierarchical structure with cross referential links)
researchers found that participants recalled more information with the mi'ed design. "he
pure hierarchical structure was found to be too restricti!e, and the nonlinear design
presented too much information at one time (McDonald 6 Cte!enson, 5770). "hus, sites
should present only the amount of links that are necessary for na!igation 9superfluous links
will increase the probability that the users will be confused and disoriented. (dditional
support for this conclusion can be deri!ed from the #ick9#yman law, which generally states
that the greater the number of options (in this case na!igational options), the longer it takes
to find the appropriate one because of greater uncertainty.
Kther aids that are beneficial to na!igation are the use of sitemaps. Citemaps may, if done
properly, present the structure of a site in a more cogniti!ely manageable way by showing a
site&s main structure and the !arious link to that structure. "his is discussed in the pre!ious
issues of sability -ews (Bernard, 5777).
"he fourth difficulty may be the structure itself. "hat is, it is generally found that people make
fewer mistakes if the hierarchical structure of the site is broader rather than deeper. %n fact,
research has generally found that ideally all information should be placed within three
hierarchical le!els from the initial homepage of the site. Cpecifically, the more le!els users
ha!e to take in order to get the information they want, the less chance they will find this
information. @or instance, in placing hyperlinks on a web page, Larson and C+erwinski (5770)
point out that a moderate le!el of breadth is optimal if it is preceded by a well9organi+ed
layout. %n their study, they reported that a two9le!el site beginning with 5< sequential links
on the first le!el, then 84 links on the other produced reliably faster searches for information
and produced less confusion than a three9le!el site with eight sequential links in all three
le!els. "he reasoning here is that the deeper the le!els, the more a user has to rely on short9
term memory. Deeper le!el sites also ha!e more general (and consequently more !ague) link
descriptions at the top le!el, which makes it more difficult for users to figure out and
remember the correct paths to a target (@or a good discussion of the breadth !ersus depth
issue see Larson 6 C+erwinski,5770).
#owe!er for sites that must ha!e deeper structures (2 or more le!els), -orman and Chin
(5770) found in their study of different menu tree structures that users browsing for specific
information will find this information faster if the
structure is conca!e (breadth of 0 ' 4 ' 4 ' 0 pages).
"hat is, it should be broad at the top le!el and at the
lowest or &base& le!el, while the interior of the web
structure should ha!e a narrower le!el of breadth (see
@igure 5 below). "hey argue that a broad top le!el
gi!es the user enough specific information to
formulate an idea as to the correct path to take, while
concentrating much of the information (and the
choices) at the base le!el will help the user find that
specific item. ( narrower breadth interior will, in turn,
reduce the likelihood of getting lost within the site
because the user will ha!e fewer choices, and
consequently less chance of being disoriented.
Figure 1. 'oncae ($ x 2 x 2 x $) menu tree
(s discussed by Bernard (4::4), depth alone may not be the sole, or e!en the greatest
determinate in predicting search performance. %n fact, as was shown, the shape of a
hyperte't structure had at least as much to do with search efficiency than its depth. %ndeed, a
(2 ' 2 ' 2 ' 2) structure was found to be not only less efficient than hyperte't shapes of the
same depth (i.e., a (< ' 4 ' 4 ' 54) structure), but structures that were deeper, such as a (8 '
4 ' 4 ' 4 ' 54) structure. (s discussed, much has been said about hyperte't depth, in that the
greater the depth, the less informationally efficient the structure should be (e.g., Gacko 6
Cal!endy, 577<D Cnowberry, et al., 5708). #owe!er, what seems to be occurring is that the
search efficiency is at least in part, determined by the properties related to the o!erall shape
of the hyperte't structure. "hese properties, then, act to either help facilitate or impede
hyperte't efficiency by altering the general comple'ity of the structure. (ccordingly, ha!ing
an inefficient shape will decrease a hyperte'tEs search efficiency.
Consequently, the goal should always be to reduce the comple'ity of the site as much as
possible. "hus as shown in @igure 4, the ideal structure of a website would ha!e much of the
site&s information accessible at the first le!el (shown as the
hori+ontal bar). Ctructures that ha!e multiple le!els should
concentrate the information at the first le!el when possible,
and at the le!el closest to the terminal nodes (at the
bottom of the pyramid).
Figure 2. *he ideal web structure with multiple leels
"he arrangement of links can ha!e a marked effect on search time and satisfaction. @or
e'ample, it has been found that search time is significantly faster when links are grouped in
columns rather than by rows (-ygren, 577<). %n addition, as discussed in sability -ews
(Bernard, 5777), we found that e'perienced and no!ice users find specific links faster and are
more satisfied with the structure of the site when the information is presented in columns
according to their respecti!e categories rather than when the links are presented in columns
according to an alphabetical listing of links. "his is belie!ed to occur because users ha!e a
difficult time trying to guess the appropriate link name in order to know where to initially
look within the alphabeti+ed column listing.
%n the same study we also found that users prefer to ha!e all the menu links presented on
one web page instead of initially showing only the link categories, which would then show the
sub9category menus on mouse9o!er. *e felt that the latter option of initially placing only the
link categories would reduce .link crowding. on the screen to a more manageable number
9thus impro!ing accuracy and satisfaction by making it easier to acquire the proper link.
#owe!er, no significant differences between the former and the latter category options were
found. %nterestingly, a large portion of users stated that they would prefer the latter option if
they were more familiar with the menu structure and the menu terms. "hus, one may want to
ha!e a full categorical link organi+ation, but also ha!e an option to initially show only the link
categories, which would show the sub9menus on mouse9o!er for frequent users of the site.
%n the specific placement of links within the website structure, Jim and Hoo (4:::) found in a
study of %nternet shopping mall sites that the combination of neighborhood links (links which
mo!e hori+ontally within the site), top links (links which mo!e the user upward to a
predetermined destination, such as the homepage), and inde' links (links which go to the
lowest le!el regardless of the the current position, such as information on a specific product)
significantly produce the greatest perception of na!igation ease as well as general
satisfaction. "hey also found that links which only mo!es up one le!el from its current
position and down one le!el in a site causes a significantly lower perception of ease of
na!igation, as well as generally lowering the le!el of satisfaction with the site.
#ow should te't be presented within a website$
F!idence suggests that the most commonly used fonts tend to be equally legible at the 5:9,
549, and 529point si+e. Comparing four sans serif fonts ((rial, Comic Cans MC, "ahoma, and
?erdana) and four serif fonts (Courier -ew, >eorgia, Century Cchoolbook, "imes -ew Aoman)
at a resolution of 5:42 ' 3<0 re!ealed no difference in effecti!e reading (font accuracy=speed
of reading) between font types (Bernard, Lida, Ailey, #ackler, 6 Gan+en, 4::4). "his finding is
supported by Bernard, Mills, @rank, and McJown (4::5), which did not find significant
differences as well.
"his is not to say there is no ob)ecti!e differences between the fonts. %n fact, there is some
e!idence that suggest that some serif fonts promote better comprehension than some sans
serif fonts. @or e'ample, a study by Boyarski, -euwirth, @orli++i, and Aegli (5770) found small
but significantly higher le!els of comprehension for the >eorgia font o!er the sans serif font
?erdana by people reading on a computer screen. #owe!er, it is really too early to draw any
definiti!e conclusions from this. Ctudies need to further e'amine the effect of different fonts
on reading comprehension.
Cignificant differences in reading time were found in that "imes -ew Aoman and (rial were
read faster than Courier -ew, Century Cchoolbook, and >eorgia. @onts at the 5:9point si+e
were read more slowly than fonts at the 549point si+e (see @igure 5). "he a!erage difference
between the fastest and slowest read font was 77.2 seconds.
Figure 1. +eading time in seconds (longer bars indicated longer reading times)
"en9point "ahoma was percei!ed as more legible than 549point schoolbook. %n addition, 549
point ?erdana and Courier were significantly percei!ed as being more legible than 5:9point
Comic, Cchoolbook and ?erdana. (lso, 549point Courier was percei!ed as more legible than
549point Cchoolbook and "ahoma, as well as 529point Comic. %nterestingly, 5:9point >eorgia
had a significantly higher perception of legibility than 549point "ahoma and Cchoolbook.
@ourteen9point (rial was percei!ed as being more legible than529point Comic, and 5:9point
(rial was percei!ed as more legible than 549point "ahoma. Moreo!er at the 529point si+e,
only (rial was significantly percei!ed as being more legible than fonts at other si+es (5:9point
Cchoolbook and Comic). K!erall, (rial and Courier were considered the most legible fonts,
whereas Comic was percei!ed as the most illegible font (see @igure 4).
Figure 2. ,erceied font legibilit- (1 . /0ot at all1 and ! . /'ompletel-1)
K!erall analysis of the participantsE font preference re!ealed that "imes was significantly
less preferred to all fonts e'cept Cchoolbook. Cchoolbook was significantly less preferred to
?erdana. K!erall, ?erdana was the most preferred font, while "imes was the least preferred
font (see @igure 2).
Figure 4. Font preference (longer bars indicates more preferred font choice).
@or anti9aliased fonts, (rial may be the best font choice. Comparing aliased and anti9aliased
"imes -ew Aoman and (rial fonts at 5:9 and 549point si+es found that the 549point anti9
aliased (rial font tied for second in preference, as well as being )udged as the third to most
legible font presented (Bernard 6 Mills, 4:::). "he 5:9point "imes -ew Aoman font was not
ranked first or second by any participant (Cee @igure /).
Figure 5. *he percentage of times each font was chosen as the 1st or 2nd preference choice
#owe!er, one should use caution in creating anti9aliased te't that is based on a graphical
image, such as with a GFB> or >%@, because they are .fi'ed. at a particular font si+e that may
be too small to be read by a certain population of users (i.e., !ision impaired). %t is therefore
recommended that te't which cannot be si+e9ad)usted (because it is part of an image, etc.)
should be at least 8 mm in height. %n addition, all graphics should make use of the alternati!e
te't function (altL. . ) so that users who cannot see the te't images can ha!e the te't image
descriptions presented or read to them by a te't reader.
Klder (dults are more accurate with, and prefer larger font si+es. "hey also prefer sans serif
fonts o!er serif fonts. (s discussed by Bernard, Liao, and Mills (4::5) reading online
documents (about 4 pages), older adults
significantly preferred the larger, 529point
font si+e (see @igure 2 below). %n this
study, serif fonts (>eorgia and "imes -ew
Aoman) were compared to sans serif fonts
((rial and ?erdana) at 549 and 529points.
"he 529point fonts were found to be more
legible, promote faster reading, and were
preferred to the 549point fonts. (lso, at
the 529point si+e, serif fonts tended to
support faster reading (see @igure
<). F'amining participants& 5st and 4nd
preference choice further shows the
popularity of the 529point si+e (see @igure
3).
Figure !. 2ean reading time in seconds
"he sans serif fonts were, howe!er, generally more preferred than the serif fonts. "his finding
is supported by Corg (570/), which found that older adults preferred to read #el!etica, which
is a sans serif font similar to (rial,
compared to Century Cchoolbook,
which is a serif font.
Figure #. *he percentage of times each font was chosen as the 1st or 2nd preference choice. 3oth 124point
5eorgia and *imes were not selected as a 1st or 2nd choice.
"here is e!idence that children prefer sans serif fonts ((rial 6 Comic) o!er serif fonts. @or
instance, Bernard and Mills (4::5)
found that fourth and fifth graders
(mean age of 5:) significantly preferred
the 529point (rial and the 549point
Comic Cans MC font o!er the 549point
"imes -ew Aoman and Courier -ew
fonts. F'amining participants& 5st and
4nd preference choice further shows the
popularity of the Comic font (see @igure
0 below). -o difference in reading speed
or accuracy between the font types were
found, howe!er. "his may be due to the
fact that 549point font si+es and abo!e
tend to produce the same le!el of
performance, as long as the font types
are designed for legibility.
Figure $. *he percentage of times each font was chosen as the 1st or 2nd preference choice.
"he optimal te't line length is dependent upon se!eral factors. %t is commonly recommended
that shorter line lengths (about 55 words) should be used in place of longer, full9screen
lengths. "his is because longer line lengths require greater lateral eye mo!ements, which
make it more likely to loose one&s place within the te't (#orton, 5707D Mills 6 *eldon, 5703).
%t also has been pointed out by #orton (5707) that longer line lengths are more tiring to read.
#orton recommends that lines should be limited to lengths of around 2: to <: characters,
which is appro'imately 55 words per line. ( study by #uey (57<0) tends to support this in his
finding that shorter line lengths or appro'imately 2. (5: cm) are more accurate on the return
sweep than longer line lengths. Moreo!er, >regory and Bouton (573:) state that people with
poor reading ability performed better when the line length was appro'imately se!en words.
"his suggests that young readers who ha!e not mastered reading online, as well as readers
who ha!e !ision deficits, may be most benefited by ha!ing shorter line lengths.
(s far as reading time, a study by Houngman and Ccharff (5777) found that with :./ inch
margins, the fastest reaction times were for the shorter, 29inch (5: cm) lengths o!er the <9
and 09inch lengths (5/ and 4: cm, respecti!ely). "he 29inch lengths were also preferred o!er
the other lengths. #owe!er with no margin widths, the 09inch line lengths had the fastest
o!erall reaction times. ( study by Duchnicky and Jolers (5708), found that full screen lengths
resulted in 401 faster reading times o!er lengths of 5=8 of a screen. %t was also found that
full and 4=8 screen lengths were read significantly faster than the 5=8 screen lengths.
%n a recent study by Bernard, @ernande+, and #ull (4::4) that compared three links lengths
(42./, 52./, and 0./ cm, respecti!ely) for both children and adults supported the finding that
shorter line lengths are preferred more than full9screen line lengths. (s far as the perception
of reading efficiency, the results were mi'ed. @or adults, the @ull9length condition was
percei!ed as pro!iding the optimal amount of scrolling in comparison to the two other
conditionsMpresumably because this condition required the least amount of scrolling. "he
-arrow9length condition was percei!ed as promoting the highest amount of concentration,
while the Medium9length condition was considered to be the most optimally presented length
for reading.
K!erall these results suggest a trade9off between faster reading times of the longer lengths
and a more preferred reading arrangement of the shorter lengths. Bossibly the best
arrangement is somewhere between the two. More research needs to be done, howe!er.
Background te'tures and colors can affect the readability of te't. @or e'ample, #ill and
Ccharff (5777) found that plain backgrounds produce faster search times than medium
te'tured backgrounds. (n important determinant, though, is the contrast between the te't
and the background 99 the more te'tured the background, the greater the contrast should be
between them.
Moreo!er, te'tured backgrounds that are subtle at true9color (429bit) settings, often become
!ery noticeable at lower9color settings (i.e., 09bit), thereby reducing the contrast between the
te't and the background e!en further. "hus, if one is to use a te'tured background, it is
recommended to be !ery careful by testing it in different color settings.
(s for color, as long as there is sufficient contrast between the te't and the background,
many color combinations are possible. #owe!er, most studies ha!e shown that dark
characters on a light background are superior to light characters on a dark background (when
the refresh rate is fairly high). @or e'ample, Bauer and Ca!onius (570:) found that
participants were 4<1 more accurate in reading te't when they read it with dark characters
on a light background. Moreo!er, a sur!ey by Ccharff, et al. (577<) re!ealed that the color
combination percei!ed as being most readable is the traditional black te't on white
background. #owe!er, it is common for websites (such as this one) to ha!e an off9white
background in order to reduce the flicker and glare associated with white backgrounds.
%n the Ccharff et al. (577<) study, other color combinations that ranked high were white on
dark blue and red on yellow. #owe!er, one should be cautious in using colors such as red on
yellow that are pure or &saturated.& Caturated colors create !isual fatigue and make it difficult
to focus on the te't. %t is best to de9saturate colors by adding white or combining them with
other colors.
"he least readable combination were green on yellow, white on fuchsia, red on green, and
fuchsia on blue. (lso, for all combinations, the lighter backgrounds with darker te't was
considered to be more readable than darker backgrounds with lighter te't.
(ppro'imately 01 of males and a little less than :./1 of females ha!e a color deficit of some
kind. %n fact, one study found that around 21 of %nternet users are !isually impaired in some
way (>?, 5770). "hus, it is important to note that different font si+es and font color
combinations can ha!e a dramatic effect on the readability of a site.
@or te't colors, it is important to ha!e a good contrast difference between colors that need to
be distinguished. Come color combinations generally frustrate users and make it !irtually
unreadable for color deficit or .colorblind. users (-ielsen, 577<). "hat is, for many color
deficit users, red, green, brown, or purple may look the same if these colors ha!e the same
contrast. Cince color deficit users cannot distinguish between a large spectrum of colors, it is
therefore ad!ised to strongly contrast the colors (make sure one color is darker than the
other) between the foreground and background, as well as between other colors that need to
be distinguished (see *olfmaier, 5777, for a good description of the proper font9color
mi'ture).
#ow can % effecti!ely use images on my website$
sers often may .instincti!ely. ignore the graphics that are presented on a website. @or
e'ample, according to the Boynter %nstitute, users are twice as likely to fi'ate on te't than on
the images in their initial !isit to a site. %n fact, they found that users did not look at the
images until the second or third !isit to the site. "his
effect has been found to ha!e the greatest impact on
effecti!eness of banners. @or instance, the Boynter
%nstitute found that only 441 of ads and promotional
icons were !iewed, 2/1 of the banner ads presented
were !iewed, and <21 of photos that were presented
were !iewed (Boynter.org, 4:::).
"o determine where users generally e'pect them to be
located on a typical web page, 8:2 participants were
e'amined. %t was found that they e'pect ads to be
located at the top9half of a web page. (s seen in @igure
5, the darker shades of blue indicate higher e'pectations
for them to be located in that particular area.
Figure 1. 6here people expect ads to be located. *he darker the shade of blue7 the greater the expectation.
#owe!er, an important factor in increasing the effecti!eness of banner ads may be to place
them where they are generally not e'pected, since people tend to ignore these areas. "hat is,
it is possible that ads may be more effecti!e if they are placed in an area where they are
generally not anticipated. "his is because indi!iduals may tend to ignore areas where they
belie!e ad!ertisements are typically placed. Cupporting this argument, Benway (5770) found
that banners located at the top of a web page tended to be ignored more often than banners
located lower down on the page. (nother study found a higher click9through rate for
ad!ertisements placed 5=8 of the way down the page as opposed to the top of the page
(*ebAeference.com, 5773). "hus, it is generally recommended to place ads lower down on
the web page 9 particularly at the middle9center of the page.
Benway and Lane (5770) also found that participants identified graphical banners only /01
of the time, compared to 721 for the te't9based link alternati!e. %t also took more time to
find the banners than the te't9based links. %nterestingly, Benway&s (5770) study also showed
that e'tremely colorful and ob!ious banners tend to be ignored by users. Aeasons for this
may be that !iewers ha!e learned to ignore these types of graphics.
(nother consideration is the type of browsing that is taking place. (ccording to a study by
Bagendarm and Cchaumburg (4::5), recall and recognition of banners were higher when
!iewers were browsing aimlessly than if they were searching for specific information. "hus, it
is possible that with aimless or unstructured browsing (which is often done with online
maga+ines), !iewers are more prone to percei!e banners and respond to them.
%n addition, banners should be much larger than the surrounding te't for !iewers to initially
notice them. %n fact, @araday (4::5) found that if the si+e of the te't and image is
appro'imately equal, then the te't will be attended to first and will be more preferred.
%t should be noted that graphical images can easily be distracting and may increase load time,
which is the greatest complaint of %nternet users (>?, 5770). #owe!er, it has been shown
that !isual information such as graphics is generally more persuasi!e for simple messages
than te'tual information, while te'tual information is more persuasi!e for more comple'
messages (see Jing, Dent, 6 Miles, 5775).
(nimating the graphics may not help as well. %n a study by Benway and Lane (5770),
animated graphics showed no ad!antage o!er non9animated graphics. Moreo!er, there is
some e!idence that animated graphics may e!en reduce te't retention by ser!ing to distract
the user from attending to the te'tual information around the graphic (*right, Milroy, 6
Lickorish, 5777). Ctudies ha!e also been mi'ed about whether animated graphics are
preferable to only te't9based interfaces. %t has been suggested that animated graphics should
be kept at a minimum in order not to distract the user from the main points of the page, as
well as to reduce the download time.
%mportantly, graphics that look like banners should normally not ser!e as important
links. "his is because users tend to ignore animated graphic because they are generally
associated with ad!ertisements. "he graphics that are presented should con!ey a simple
message to portray the intended mood of the site or to catch the &eye& of the user for a brief
moment. (ny animation that is presented should animate only for se!eral seconds in order
not to annoy and distract the user.
Colored images are more easily remembered than black and white images. "hat is, >ilbert and
Cchleuder (577:) compared black and white to colored ads and found that the colored ads
were more readily recalled and were processed with greater speed.
(re frames e!er appropriate$
@rames ha!e the potential to confuse users by breaking the user&s model of a website. @or
instance, instead of the concept of a node as being a single unit of information, framed pages
may consist of many units that can go in any direction, which may make it difficult to later go
back to the original node. (lso, !iewers cannot bookmark frames, frames are not accessible to
many users who employ assisting technology such as screen readers, and some search
engines re)ect the framed pages outright, so what is left is only the main or dominant frame
(#owe!er, there are some ways around this, see http;==searchenginewatch.com=
webmasters=frames.html)D (nd, of course, users hate poorly designed, framed web pages
(-ielsen, 577<).
Broper and parsimonious use of frames may, howe!er, be appropriate for websites if it
promotes easier na!igation. (see Briestley, 5773, for a more detailed discussion of this). Kne
way to use frames is by using a na!igational menu frame (also called an inline frame, which is
a frame dedicated to displaying the main na!igational links within a site). ( menu frame can
sol!e the problem of the disappearing menu when users scroll down a non9framed page
because the menu frame will always be !isible (these are normally placed on the left side or
top of the screen). Drawbacks of using menu frames is that the amount of information placed
in the menu page must be rather small in order for the entire menu to fit within the frame.
"hat is, forcing the user to scroll to see the entire menu frame defeats the purpose of ha!ing
one 9which is to always ha!e a !isible menu.
@rames might also be used to allow users to follow an e'ternal link while keeping the original,
initial site in !iew. "o do this, typically the top frame shows the initial site as a reminder to
return to that site and the lower frame shows the sites that are e'ternal to the initial site
(hotmail.com uses this technique). #owe!er, there should be an option to completely lea!e
the initial site. %n addition, links that e'it the site should use the "(A>F"L.Ntop. tag to
ensure that users can lea!e the initial site without being embedded within the frameset of the
initial site.
%n a study we e'amined performance and preference of framed !ersus non9framed pages in
which participants were presented with four documents, each with a different link
arrangement. @or each arrangement they were instructed to search for specific information
pertaining to ten questions related to that document (such as, .*ho found e!idence linking
tribes from Ciberia to the (mericas$.). %n one condition the links were embedded within a
document, as would be found with many online documents (see @igure 5). "his was
accomplished by using an original online article with embedded links. ( second condition
placed links at the upper9left of the document (see @igure 4). (nother condition placed links
within a hori+ontal, top frame abo!e the document (see @igure 8) and a fourth condition
placed the links within a !ertical frame at the left of the document (see @igure 2).

Figure 1. Links embedded in the document Figure 2. Links at top4left

Figure 3. Links within a hori8ontal frame Figure 4. Links within a ertical frame
(n analysis of the results re!ealed no significant differences in search performance
(accuracy, search time, and search efficiency) or preference between the four conditions.
%nterestingly, the participants had rather strong preferences towards the top9left and
hori+ontal framed layouts 99which, in effect, canceled each other out.
%n a follow9up study (Bernard 6 #ull, 4::4) that only compared the top9left with the !ertical
frame conditions in terms of preference did re!eal significant differences that fa!ored the
frame condition O+ L 98./0, p P .::5Q. F'amining the number of participantsE ranking either
the @rame or "op9Left (no fame) conditions as their first choice further illustrates the
preference for the @rame condition (see @igure below).
#ow can % design a !isually pleasing interface that follows usability principles$
Much has been said about the design process of websites, such as establishing the proper
mood or .feel. to create user interest or e!en e'citement with the site. "his is a !ery
important concern, but ultimately users tend to be far more satisfied and stay with websites
that are designed for their use in mind (see "edeschi, 5777). Considering this, three core
principles concerning interface design are presented;
5) Jeep the interface simple 9 "o quote Mies !an der Aohe, .less is more.. Krgani+e the
interface by reducing un9needed !isual elements as much as possible. "hat is, remo!e all
unnecessary !isual .noise.. "his will make the important ob)ects that are there stand out
e!en more. Moreo!er, as Fdward "ufte stated, .it is not how much space there is, but rather
how it is used. %t is not how much information there is, but rather how effecti!ely it is
organi+ed. ("ufte, 577:, p. /:). "he use of open space is generally more effecti!e in
organi+ing and grouping information than using imposed, artificial structures such as !isually
nested frames or bars. %t is also more aesthetically pleasing. %n fact, we found empirical
support for the notion that the proper use of open space can increase satisfaction with a
website (Bernard, Chaparro, 6 "homasson, 4:::).
4) Make action9ob)ects !isible 9 (ccording to Donald -orman, a design should a) make it easy
to determine what actions are possible at any moment, b) make things !isible, including the
conceptual model of the system, the alternati!e actions, and the results of actions, and c)
make it easy to e!aluate the current state of the system. Kn a web interface, one of the chief
mechanisms to do this is the proper use of percei!ed affordances (-orman, 5700).
(n affordance refers to the .properties of the thing, primarily those fundamental properties
that determine )ust how the thing could possibly be used. (p. 7). (ffordance pro!ides us with
clues as to the operations of things. More importantly for interfaces, howe!er, are the
percei!ed affordances which pro!ide !isual feedback that ad!ertise affordances. @or e'ample,
a link button may be percei!ed to afford clicking because of its &89D& or &raised& appearance.
Consequently, it is often helpful to gi!e link buttons the physical appearance of a button, or
any ob)ect that affords clicking, in order for them to be seen as a button to be clicked
(-orman, 5700). "hus, it is important to make na!igation buttons look like they should be
clicked as well as follow the con!ention of underlining links when they are te't9based links.
Con!ersely, non9na!igation ob)ects should not look like they could be clicked in order not to
&trick& the user into thinking they are links.
>enerally, buttons ser!e as primary ob)ect for initiating actions, such as submitting or
confirming information. Buttons also can act as the primary link for mo!ement to other web
pages, usually within the same website. *hen this occurs, te't9based links often ser!e as a
less important, secondary or supplemental link for the buttons. -ormally, howe!er, te't9
based links are the primary link to other internal web pages.
Moreo!er, physical appearance of ob)ects such as icons can significantly affect na!igational
performance. @or e'ample, Aogers (5703) found that icons with abstract but simple symbols
that represented concrete ob)ects resulted in the fewest number of errors and requests for
help. %n addition, Byrne (5778) found that large and simple icons outperformed comple' ones
by a significant margin. Byrne suggests that icons need to be simple, large, and easy to
discriminate in order to be effecti!e. Comple' icons tend to clutter the screen with
unnecessary information. Moreo!er, -orman (570/) suggests that icons are best used to
represent graphic tools and ob)ects. ?erbal labels, such as .to sa!e. are best for formal
commands (for a checklist on designing effecti!e icons, see #orton, 5773).
8) Balance and unify the interface 9 Balance and unity has always been a key component in
good design. #umans on a preconscious le!el seek structure in the things they see. %f there is
no intentional structure, we will impose our own. Ceeking the appropriate balance among
things, as well as unifying those things that are related will generate structures that are not
only pleasing to the eye, but will make the interface more understandable (see Mullet 6 Cano,
577/, for an e'cellent discussion on design and !isual interfaces). Fmpirical studies ha!e
supported this claim by finding that the position within a plane (as well as si+e and contrast)
to be one of the most perceptually important !ariables in !isual search tasks (Cle!eland,
570/).
Kne of the fundamental concepts of balance is the notion of the >olden Cection. "he >olden
Cection is a ratio of a rectangle in which the smaller side to the larger is the same as that of
the larger to the sum of both 9which is a ratio of appro'imately :.<50 to 5.::: or a standard
0./ ' 55 page. F'amples of the >olden Cection are almost ubiquitous in art as well as in
nature (from the Barthenon to a nautilus shell). ( web page that structures its graphical
layout according to this ratio will look more appealing and will ha!e a greater impact than
other ratios, such as a ratio of 5 to 5.
(lso, when placing se!eral ob)ects on a web page, one should take into account the .!isual
mass. of these ob)ects (its si+e and presence). @or e'ample, ideally the placement of ob)ects
should be positioned in the same way as you would balance solid ob)ects on a fulcrum. "hat
is, a larger ob)ect should be placed closer to the center of the screen to offset the smaller
ob)ect(s). "his will create an equilibrium between the ob)ects, and will be more appealing
(see "ufte, 577:).
"he unity of the interface is important because it has the potential to link concepts and
ob)ects together that belong together. @or e'ample, *ickens& (570<) compatibility of
pro'imity principle states that tasks that necessitate mental integration of information should
be in close pro'imity. #owe!er, tasks that require focused attention on specific !ariables will
be harmed by this close pro'imity. "his can be applied to things such as the organi+ation of
links. @or instance, care should be made to group links that belong together, as well as
separate those that do not belong.
2) Be aware of @itt&s Law 9 @ormally, it states that pointing time is a function of the distance
and the width of a target (@itts, 57/2). >enerally speaking, it states that the smaller and
farther away an ob)ect is, the longer it will take to reach that ob)ect. Ce!eral researchers
ha!e argued that important buttons should be placed on the right side of the screen because
the mouse arrow pointer is usually resting ne't to the scroll bo', and thus it would take less
time to click the ob)ect. #owe!er, what is important here is that knowledge that if there are
se!eral different buttons that need to be clicked in succession, the smaller and the farther
apart they are, the longer it will take to click them. (@or more information related to this topic
see, Blamondon 6 (limi, 577<).
#ow can % reduce the ma)or user annoyances on my site$
@or online shoppers, ad!ertisements tend to be the most frustrating factor in using the *eb.
(ccording to Aetail @orward (4::4), the top fi!e online shopping frustrations are;
Bop9up bo'es with !isiting=shopping a site /41
Banners ad!ertisements /:1
Congested web pages (e.g., too many ads, images, etc.) 8/1
Clow load times 4<1
Difficult to find a specific product 4:1
>eneral *eb user sur!eys ha!e found that most dissatisfying web e'periences are, a) not
being able to find specific information, b) using websites that are confusing, and c) websites
with slow download time, respecti!ely (5:th >? sur!ey, 5770). "he fist two ma)or
annoyances are addressed in the navigation and the positioning of information discussion,
respecti!ely.
"he third annoyance, slow downloads, are a !ery common complaint, which ha!e been known
to negati!ely affect user satisfaction, and ultimately sales. %n fact, it has been estimated that
as much as R2.8/ billion in e9commerce sales ha!e been lost each year due to user frustration
related to slow downloads (Sona Aesearch, 5777). %n fact, SD-et reported that a sur!ey of
54,::: online customers found that 201 of them ga!e up trying to purchase an item online
because the web pages took too long to load (SD-et, 4:::). Moreo!er, it has been found that
slower web pages were significantly )udged as being less interesting than their faster
counterparts (Aamsay, 5770), and were thought to ha!e lower quality products, as well as
ha!ing compromised security (Bouch, et al., 4:::).
Currently, the a!erage connection speed is appro'imately /Jbps (kilobytes per second). "hus,
a 2: Jb web page will take appro'imately eight seconds to download. "his )ust fits within the
so9called &eight second rule&, which is considered the recommended loading time for web
pages. "hat is, it has been suggested that users will tolerate no more than around eight
seconds for a web page to download.
"here are some studies which back up this suggestion. @or e'ample, a study by Bouch et al.
(4:::) found that participants& a!erage load9time tolerance was 0./3 seconds. #owe!er, the
standard de!iation was /.7 seconds. (nother study by Sona Aesearch (5777), which
e'amined a website with a homepage of 2: Jb, had a bailout rate of 8:1, while other pages
with 8498/ Jb ranges in the same site had bailout rates of <1 to 01. Aeducing the load time
for the home page to 82 Jb decreased bailout rate to the same range as the other pages.
(ccording to Dillart and Jahn (5777), the factors that determine users& frustration with a
downloading a website are;
#ow long they ha!e to wait

"his is ob!iously a !ery important fact. Moreo!er, the frustration associated with waiting
tends to be highest when there is a lot of gratuitous graphics. Blacing images that do not add
to the site will decrease rather than increase user satisfaction with the site. #owe!er, there is
no hard9fast rule about download speed and annoyance. Gust factoring in the time it takes to
load a web page does not take into account other !ariables that determine users& frustration
with the site.

*hether the consumer e'periences uncertainty about the wait

-ot knowing how long to wait for a page to load is often !ery frustrating to users. %n fact, #ui
and "se (577<) found that users reported negati!e effects on mood if they had no information
concerning how long they ha!e to wait, which affected their o!erall ser!ice e!aluation. "hus,
for longer waits it is recommended to inform the user how long the wait should generally
take.

"he kind of information that is pro!ided about the wait

Bro!iding adequate information about their e'pected wait time is important. "o do this, it is
recommended to specify by the link the Jbs for those pages (especially large BD@ pages). @or
!ery long waits, it is recommended to do this, plus use countdown information at regular
inter!als that indicated how much of the site has been downloaded, and thus how much
longer they ha!e to wait (Dillart 6 Jahn, 5770).

*here the waiting occurs within the %nternet episode

(ccording to Dillart and Jahn (5770), consumers can separate the download e'perience from
their retrospecti!e e!aluation of sites if the the waiting occurs in e'pected times, such as the
initial download stage, then the negati!e carry9o!er effects of waiting will be minimal.

#ow long the wait is compared to the consumers& e'pectations

"he old e'pression, &a watched pot ne!er boils& seems appropriate for users perceptions of
waiting time. "hat is, users more often than not )ust stare at the screen until the page is
loaded. #owe!er, Dillart and Jahn suggest that the duration of the download often has little
effect on retrospecti!e e!aluation of the site if the wait is not salient to the user. "o lessen
the saliency of the wait, it is recommended not to place large (Jb) images at top of the screen
or imbed important te't within the images.
(nother cause of slow browsing is due to a poorly defined structure. ( lack of careful
consideration as to where information should be placed often produces unnecessary or
une'pected delays ultimately slowing down browsing within the site 9sometimes to a point
where it would be faster to find the information by other means 9 which often causes the user
to lea!e the site. @or e'ample, Cel!idge found that the a!erage time to complete an
information search task across certain airline sites was about 53 minutes. "ime to complete
the same task by calling the airline by telephone a!eraged 8 minutes (Cel!idge, 5777). -ot
surprisingly, these types of delays ha!e a strong negati!e effect on the o!erall satisfaction of
these sites (see Briley 6 Ctolt+, 5777).
#owe!er, it does seem as though people are more tolerant in waiting than they were in the
past (see 7th >? sur!ey, 5773). (s discussed in sability -ews, Cel!idge found that users
were frustrated by 8: and <: second delays in page loading time, but would tolerate the 4:
second delays (Cel!idge, 5777). (gain, many factors are at play, such as the pro!ided
information, the e'pected waiting time, their moti!ation, and their general tolerance for
delays.
sers also consistently rate non9working hyperlinks as one of the most annoying aspects of
web browsing. (n easy way to reduce this problem is to use free ser!ices such as
websitegarage.com, which tests websites for such things as download time, non9working
links, and browser capability.
#ow can % make my site more accessible to children$
Children tend to e'plore websites because they seek to ha!e fun as well as to learn. "hus,
websites should try to to be playful and e'ploit their general curiosity by making the site&s
content attention9grabbing and, to a small degree, challenging in order to entice them to go
through the site. @or e'ample, children may seek to e'plore certain areas within a site
because e'ploring it satisfies a curiosity need that is enticed by the content of the site.
MaMaMedia.comT did this by offering children the opportunity to make their own
customi+able cartoon character, thereby satisfying both the children&s playful and curiosity
needs.
Children often seek out interacti!ity with the site, again to be creati!e and to ha!e fun. "his
can be in the form of coloring pages that can be printed out. %n fact coloring pages can be
more popular than games on a site (@ishler, 5770). Kther popular methods are to allow
children to create music or draw pictures.
%n addition, children are much more attracted to animation than adults. (nimation adds a
great deal to the .fun factor. if it is done well, but too much animation will distract and
disorient them (Culli!an, -orris, Coloway, 6 Beet, 4:::). Moreo!er, children do not ha!e the
attention span for longer downloads. (ccording to Da!e Lewn+i, an author of an online
children&s site, .( lot of children e'pected to see a picture when they hit a button, and they
were disappointed if they didn&t see one.. "hus, he recommends placing many multi9colored
pictures that load relati!ely quickly (@ishler, 5770).
%n another study by Culli!an, et al. (4:::) children tended to wait for images to completely
load on a page before na!igating to another in the belief that a complete loading was
mandatory. "his wait produced signs of frustration.
"he interest in the site howe!er will last as long as the contents are inno!ati!e and fun. "hus,
it is ad!isable that websites geared toward children frequently ha!e new features in order for
the site to be in!enti!e and interesting to the children. %t is also recommended that children
be rewarded with different and interesting features at each le!el within the site. "his will help
draw them deeper within the site.
Houng children generally ha!e !ery short attention spans and do not think in abstract terms.
Aesearch has consistently shown that most children under the age of fi!e will ha!e an
attention span of only around 0 to 5/ minutes. Many children will ha!e e!en less. "hus, the
layout and content of a site should reflect the low attention span of children by being
designed to be accessed quickly. Directions should be !ery short, uncomplicated, and easily
read (children tend not to read directions !oluntarily), and games should not take longer than
their attention span.
%n addition, children under 0 usually find it difficult to think in abstract terms. @or e'ample,
children may find it easier to recogni+e actual pictures of ob)ects than to use symbols to
represent them. %t is important to be as &concrete& as possible when e'plaining directions.
Children can make the distinction between a &fun& website and a usable one. Buttons should
represent familiar things to children, easily con!ey their purpose, and should be fairly large 9
the si+e of a quarter to accommodate their poorer hand9eye coordination. Buttons should
indicate when they are being moused o!er, such as being highlighted (see #anna, Aisdan,
C+erwinski, 6 (le'ander, 5770).
(nother important concern is the actual language of the site. @or instance, the website&s
language should lie somewhere between being understandable 9don&t use computer lingo 9
and fun, but not too childish for the intended audience or they will resent it. (lso, the younger
the intended audience, the more concrete the language should be. "hat is, young children
ha!e difficulty in understanding
abstract concepts. %n addition, the
content should be free of words
that may cause the website to be
blocked by web filtering agents.
F!en statements such as, .Jing
(rthur wore a sil!er breast9plate
to protect him in battle. could
cause the site to be blocked from
children because it has the word,
.breast.. "herefore, it is
recommended that the content&s
language be e'amined for words
that would cause the site to be
blocked from !iewing.
Figure 3. *he percentage of times each font was chosen as the 1st or 2nd preference choice.
"he font type of the te't may also be a factor in the attracti!eness of a website. @or instance,
Bernard and Mills (4::5) found that fourth and fifth graders significantly prefer the 529point
(rial and the 549point Comic Cans
MC font o!er the 549point "imes
-ew Aoman font and 549point
Courier -ew. F'amining
participants& 5st and 4nd preference
choice further shows the popularity
of the Comic font (see @igure 8).
Figure 4. 0umber of time each line length was chosen as the 1st7 2nd7 or 3rd preference choice.
Children prefer narrower to broader line lengths. ( study by Bernard, @ernande+, and #ull
(4::4) that compared three online links lengths (42./, 52./, and 0./ cm, respecti!ely) found
that children significantly prefer narrower lengths to the other compared line lengths (see
@igure 2). -o difference was found for reading speed. Moreo!er, assessing children&s
perception that the amount of scrolling was optimal for a particular line length condition, that
a particular line length was preferable or, a particular length was percei!ed as promoting
easier concentration also re!ealed no significant differences between any of the line length
conditions.
"eenage boys and girls largely surf the *eb for different reasons. (ccording to a sur!ey of
5,/4: respondents by Gupiter Communications (4:::), boys are more interested in playing
games (<41 to 251), building web pages (8/1 to 421), downloading software (321 to
201), and downloading music (<01 to 321) than teenage girls.
"eenage girls on the other hand, tend to read more online maga+ines (8<1 to 571), send e9
greetings (3/1 to 201), and do homework online (321 to <81) more than teenage
boys. %n appealing to this audience, it thus may be beneficial to pro!ide one or more of these
features.
#ow can % make my site more accessible to older adults$
Klder users generally take more steps than younger users in finding the same amount of
information. @or e'ample, Meyer, et al. (5773) found that this occurs because older adults
tend to re!isit already !iewed pages. @urthermore, Mead, et al. found that older users&
success at finding information declined significantly when they had to follow more than 8 to <
hyperlink paths to get to the desired information. %t is belie!ed this decline in performance
occurs because older users tend to ha!e greater trouble remembering their location within a
site because of a decrease in working memory efficiency. "herefore, if the site is geared
towards an older population it is especially important not to ha!e a !ery deep hierarchy. %t is
also important that the site&s contents be e'plicit at the beginning of the site (at the
homepage le!el) instead of relying on e'ploration. "his could be done by placing a site inde'
or sitemap at the homepage le!el. %t is also recommended to use clearly !isible na!igational
aids to help them !isuali+e the structure of the site.
%n another study, >roff, Liao, Chaparro, and Chaparro (5777) found that older users were
slower than college9age users in finding information. %n general, older users took the time to
carefully read the te't on a web page before continuing the tasks, which tended to slow them
down. "his finding is supported by the research of Houngs (5777) who found that older users
tended to carefully read information before e'ploring, which further supports the idea of
placing a site inde' at the homepage le!el.
Klder adults also prefer te't to ha!e !ery clear headings (see #artley, 5772). Clear and large
headings help orientate users as well as help those who ha!e difficulty reading because of
degraded or impaired !ision by gi!ing them ob!ious signposts.
sually the greatest factor affecting older web users is their decline in !ision. @or e'ample,
*eale (57<5) stated that there is a /:1 reduction in the amount of percei!ed light at age /:
compared to age 4:, and this reduction increases to <<1 at age <:. %t is therefore
recommended that all sites, but particularly those directed towards older users, ha!e high
contrast between the te't and the background, as well as between darker and lighter colors
in general. %n addition, as lenses thicken, there is a reduction in the transmission of blue light
through it. @or this reason, older people ha!e more trouble sorting or matching colors, and
make more errors in the blue9green and red regions than in the other color
regions. @ortunately, there are many resources a!ailable to help designers create websites
that are accessible to degraded or impaired !ision users. Kne of the better resources is the
*eb (ccessibility %nitiati!e (see in reference section).
Klder adults also tend to ha!e a reduced field of !iew. Consequently, more important
information and links should be placed closer to the center of the screen in order for them to
be more easily detected (#awthorn, 4:::). "his is particularly true if there are many ob)ects
on the screen. "hat is, older adults generally ha!e trouble locating a specific ob)ect if there
are many other ob)ects on the screen (Blude 6 #oyer, 5705), especially if the targeted ob)ect
is located on the screen&s periphery.
Ctudied e'amining the reading performance of
older adults ha!e found that colored te't on a
colored background typically reduces their
reading performance compared to reading black
te't on a white background (Charness 6 Bosman, 577:). Moreo!er, research has shown that
dark te't on light backgrounds is generally superior to light te't on dark backgrounds ("obas,
5703). #owe!er, older adults do ha!e an increased sensiti!ity to glare. "hus, it is
recommended that backgrounds not be pure white, but some form of off9white color (such as
background presented here).
Figure 1. 2ean reading time in seconds
Because of a general decline in acuity, the si+e of te't within a site should also be larger (529
points or more). @or e'ample, older adults (mean of 3:) ha!e also been found to be more
accurate, read faster, and prefer 529pont font si+es (see @igure 5).
"hey also tend to read faster with serif fonts (Bernard, Liao, 6 Mills, 4::5)
-ote;; an e'amination of font preference for children and older adults is discussed in the How
should text be presented within a website link
Klder adults usually ha!e a reduction in motor coordination. "his can greatly affect their
interaction with a website. @or e'ample, #awthorn (4:::) suggest that older adults may ha!e
difficulty with certain actions such as mo!ing and clicking a mouse or using a scroll bar.
%ndeed, a study by Fllis and Jurniawan (4:::) found that older adults indicated that it was
easier to click a link to a number of web pages than to scroll longer pages because the scroll
bar was considered too difficult to manage due to its small si+e. (lso, the scroll bar did not
gi!e them an adequate cue that more information was pro!ided below the screen. %n addition,
the older adults had difficulty clicking the standard9si+ed links. (s a consequence, the si+e of
the link targets were increased by making them into graphic buttons (50: ' 44 pi'els). "o
reduce errors further (caused by tremors, etc), dead space around each link button was
added to clearly separate one button from another. Moreo!er, older adults indicated that the
standard feedback signifying that a link is acti!e (an arrow changing into a hand) did not
pro!ide enough !isual feedback. "hus, they made the link buttons turn into a lighter shade
and appear to rise up when moused9o!er, gi!ing them further indication that the button was
clickable.
Klder adults ha!e reduced attention spans. (ccording to ?ercruyssen (577<), older adults
ha!e difficulties in attention o!er long periods of time. @or instance, #awthorn (4:::)
suggests that older adults may be especially distracted by e'traneous !isual noise, such as
background graphics. Kther distractions may include such things that obscure the screen9
e!en if it is only temporarily !isible. Consequently, items such as pop9out ads and multiple
windows may tend to distract and disorient them to a greater e'tent than younger adults.
#ow can % make sure my site follows general *eb con!entions$
sers o!erwhelmingly prefer sites which employ common *eb con!entions (-ielsen,
5773). #owe!er, common standards are !ery rare in web design. Come of the few that e'ist
are;
@ollow the standard hyperlink colors; blue for non9!isited hyperlinks, purple for !isited
hyperlinks, and red for acti!e hyperlinks (but if you ha!e a site with a blue background 9or
any dark color9 please choose a color that can clearly be contrasted from the background
color. "hat is, if the choice is to either !iolate the standard or not be !isible, choose
!isibilityU).
Blace a na!igational link to the homepage at the upper left corner and bottom of each
page.
Blace the site&s internal hyperlinks at the bottom of each page (in addition to other
places).
%mages and te't space should not cause hori+ontal scrolling on lower resolution screens.
(t the bottom of each page, place the date that the page was updated and the AL
address.
"he te't should facilitate the scanning of information.
se (L"9"ags for graphics, especially for graphics that search as hyperlinks.
@or an e'cellent and more detailed discussion of general *eb con!entions, see the Hale Ctyle
Manual website; http;==info.med.yale.edu=caim=manual=contents.html
Kther fine guidelines include;
%BM; http;==www98.ibm.com=ibm=easy=eouNe't.nsf=FasyBrint=/34
*8C (ccessibility >uidlines 5.:; http;==www.w8.org="A=*(%9*FBCK-"F-"=
?orburger, M.; http;==www.!orburger.ch=kissfp=styleguide=inde'.html
#ow can my website promote customer sales and loyalty$
Knly a small percentage of people actually purchase merchandise online. %n fact, the
percentage of people that actually buy online is appro'imately 8 to / percent (Clatalla, 4::5).
"here are se!eral reasons for this;
@irst, shoppers often do not feel secure in purchasing products online. (ccording to a study by
"ilson, et al. (5770), participants indicated from a list of 2: items that credit card security is
the most important deterrent for them to buy online. ("he other top deterrents were easy
return=e'change methods, detailed descriptions of items, price, secure personal information,
pictures of merchandise, and simple to search, respecti!ely). Kb!iously, using a secure web
ser!er to collect customer data (and communicating that it is secure) promotes the feeling of
security. %t is also the standard for conducting business online. Moreo!er, "ilson et al. further
suggests pro!iding a pri!acy policy on e!ery page and highlight this policy before customers
are e'pected to gi!e personal information. Kther suggestions are to e'plain the benefits from
sharing personal information, pro!ide mechanisms for controlling how this information is
used, and pro!ide a background on the company.
Cecond, is a lack of specific information related to the product and store. Cince online
shoppers cannot physically interact with the product itself, potential customers generally
require more e'tensi!e information about the product 9 in the form of links to detailed
pictures or descriptions of the product. #owe!er, according to Cpiller and Lohse (5770), o!er
/:1 of the e9commerce sites that they studied had less than three lines of te't describing the
products. %t is argued by Lohse and Cpiller (5770) that e9commerce sites should take
ad!antage of the interacti!ity of the web and .offer hyperlinks to more e'tensi!e product
information such as product testimonials (book re!iews at www.ama+on.com) and product
demonstrations (software downloads). (p. 04).
@or e'ample, according to a sur!ey by BricewaterhouseCooper (4::5), pro!iding more
detailed information about the products can significantly increase the likelihood of the
purchase being made (see @igure 5 below).
@eature
1 of shoppers indicating
@eature increases purchase
likelihood
Close9up product images 221
Broduct a!ailability 871
Broduct comparison guides 821
Cearch function 8:1
590:: -umber to contact customer ser!ice representati!e 4/1
Broduct re!iews=e!aluations by online shoppers 421
Catalog quick order 421
Figure 1. Features most likel- to increase the likelihood of online purchase
Moreo!er, many online stores also do not pro!ide information related to the site itself. %n fact,
Cpiller and Lohse found from the sur!eyed e9commerce sites that 7/1 did not ha!e links to
related products on the same site and only 4/1 had a help link to product9si+e selection. %n
addition, only 71 of the sites had a @(V section, e!en though it has been shown that online
stores with a @(V section generally ha!e more !isits that those without them (Cpiller 6
Lohse, 5770). "he sur!ey also found that 231 of the sites did not ha!e an email link. "his is a
!ital (perhaps only) way for customers and potential customer to interact with the store. %t is
essential to pro!ide ways to promote customer=store interactions in order to reduce the
physical and psychological separation between the customer and the store.
@urthermore, in a re!iew of past studies Gar!enpaa and "odd (5773) ga!e recommendations
for presenting online product information;
Kffer discounts, e'cept for unique products.

@ocus on products that ha!e low deli!ery costs.

Benchmark the e9store against traditional retail stores and catalog stores.

Fmphasi+e brand name products and product quality.

Aeduce shopping effort by pro!iding search techniques.

Aespond promptly to questions.

Bro!ide no9cost=no9hassle return policy.

Bro!ide rich product descriptions including images and words.

Fmphasi+e security measures.

Bro!ide customer testimonial.
Lastly, many online stores do not pro!ide adequate ser!ice information. "hat is, in the e9
commerce websites studied by Lohse and Cpiller (5770), close to one9third did not pro!ide
company policies, background, or history. Moreo!er, most of the sites that did pro!ide
information pro!ided less than 5: lines of te't.
"hird, shoppers must feel comfortable in !iewing the shopping items before buying. (s with
any store, shoppers should feel free to e'amine the merchandise without ha!ing to gi!e out
personal information before committing to buy. %n other words, placing items into a shopping
cart should be effortless and anonymous. "his will allow them to feel at ease and safe while
shopping online. -ot doing this can ha!e dire consequences. @or e'ample, in a study of e9
commerce sites, Chaparro, Childs, Braheswari, 6 Aappard (*inter=4:::) found that sites
requiring the users to register with the site before adding items to a shopping cart deterred
them from adding items to the cart. %t also had the effect of producing a low amount of
satisfaction with the site.
Con!ersely, a well designed shopping interface can promote the shopping e'perience (see
Dack.com (5777) for a nice o!er!iew of shopping carts and checkout interfaces. Kther studies
such as by Aohn (5770) pro!ide further information pertaining to e9commerce usability dos
and don&ts).
@ourth, websites should be &credible& to the shoppers. @ogg, et al. (4::5), conducted a study
that sur!eyed o!er 5,2:: participants and found two elements that affected the credibility of
a site; commercial implications and amateurism. "he items that most negati!ely affected the
commercial implication of sites were ones that;
makes it hard to distinguish ads for content
automatically pops up new window with ads
has one or more ads on each page
sites that require a paid subscription to gain access.
%mportantly, the item site makes it hard to distinguish ads from content recei!ed the
lowest score in the entire study, and thus it is recommended to make that distinction
within a site.
@or the amateurism scale, sites that;
are rarely updated with new content
has a link to a site that is percei!ed as .not credible.
has a link that doesn&t work,. is .sometimes une'pectedly una!ailable.
has has a .typographical error.
were considered the most amateurish, respecti!ely. Cites that ha!e been;
updated since your last !isit
offers information in more than one language
were considered to be more professional. Kther items that had a positi!e effect on
website credibility were sites that;
pro!ide a quick response to your customer ser!ice questions
lets you search past content
looks professionally designed
send emails confirming transactions you make
@rom their results @ogg et al. (4::5) made se!eral credibility promoting recommendations;
%nclude elements that highlight the brick9and9mortar nature of the organi+ation it
represents, such as listing a physical address.

Make the site easy to use.

Con!ey e'pertise, such as including credentials, citations, and references.

Con!ey trustworthiness by stating a policy on content, etc.

"ailor the site to the !isitor and acknowledging their presence.

(!oid o!erly commercial elements on a website (see abo!e).

(!oid amateurism (see abo!e).
"he use of pictures may increase user trust for e9commerce websites. (ccording to research
by Cteinbruck, Cchaumburg, Duda, and Jruger (4::4), the inclusion of a photograph for an e9
bank website significantly increased the percei!ed trustworthiness of that website. %n this
study, the condition that had the highest trustworthiness ratings was for a website with a
welcome page that contained a picture of a person that was identified as a customer ser!ice
agent. "he second highest rating was for the picture only, and the lowest rating was for a
welcome page without any picture. Moreo!er, a study by Aiegelsberger and Casse (4::4) did
find that photographs increased the relationship with the brand for some indi!iduals, but as a
whole, reactions !aried greatly to the use of photographs.
#ow can % make my site more appealing to international users$
"oday, a ma)ority of the world&s *eb traffic is generated outside of the nited Ctates.
#owe!er, many websites in western countries99particularly in the .C.99do not take the idea
of a global audience into consideration, e!en though a substantial portion of their re!enue
may be foreign. *hat is needed is an understanding of how different cultures respond to
!arious website designs.
Color has psychological effects on users that are different across cultures. (ccording to Ausso
and Boor (5778), color can present opposite meanings, such as yellow for cowardice in the
nited Ctates, and >race and -obility in Gapan. Cee "able 5 for some cultural associations of
color.
Culture Aed Blue >reen Hellow *hite
nited Ctates Danger Masculinity Cafety Cowardice Burity
@rance (ristocracy @reedom=Beace Criminality "emporary -eutrality
Fgypt Death
?irtue=@aith=
"ruth
@ertility=
Ctrength
#appiness=
Brosperity
Goy
%ndia Life=Creati!ity
Brosperity=
@ertility
Cuccess Death=Burity
Gapan (nger=Danger ?illainy
@uture=Houth=
Fnergy
>race=
-obility
Death
China #appiness #ea!ens=Clouds
Ming Dynasty=
#ea!ens=Clouds
Birth=*ealth=
Bower
Death=Burity
*able 1. 9xamples of cultural associations of color (From +usso : 3oor7 1%%3).
?arious cultures respond to layout designs differently. Aesearch e'amining multicultural
differences and their interaction with website design is sparse99mostly coming from Marcus
and >ould&s (4:::) application of #ofstede&s (5775) research on world culture to web design.
#ofstede stated that world culture consists of fi!e dimensions, four of which are rele!ant to
the *eb.
"he first dimension, power9distance, refers to the degree in which indi!iduals with less power
e'pect and accept unequal distributions of power within a culture. Cultures with a high
amount of power9distance (BD) tend to ha!e centrali+ed political power and deep hierarchies.
(ccording to #ofstede, these types of societies emphasi+e hierarchical relationships,
authority, e'perts, certifications and official logos, leaders, security, and an acceptance for
restrictions to information access. Cultures with low BD emphasi+e flatter hierarchies and
greater equality in relationships. "hus cultures with a high BD may typically feel comfortable
with a greater emphasis on highlighting the accomplishments of high9ranking indi!iduals
within a company, as well as pro!iding a greater hierarchical relationship between different
di!isions and positions within the site. Low BD cultures, such as Denmark, would tend to be
more comfortable with sites that showcase &common& indi!iduals or both genders. "hey would
also tend to de9emphasi+e hierarchical differences between indi!iduals within the same
company (see "able 4 for scores of countries that were studied).
"he second dimension, indi!idualism, refers to the degree to which a culture emphasis the
self and immediate family o!er the society at large. Cultures with large amounts of
indi!idualism (%-D), such as the nited Ctates, !alue personal freedom and rewards, pri!acy,
and di!ersity of opinion. #ere freedom of the press and self9actuali+ation are pri+ed.
(ccording to Marcus and >ould, cultures that are high in %-D will emphasi+e images of
personal success, youth, change. "he con!erse of this, collecti!ism, emphasi+es group
harmony, e'perience of older and wise leaders, and tradition and history. "hus, indi!iduals in
countries such as in the nited Ctates should feel more comfortable !isiting sites that
promote a !ery modern9looking, youthful, and indi!idually successful looking design.
Collecti!e cultures, such as in >uatemala, should feel more comfortable with sites that
promote the history of the company and how it helps the >uatemalan society at9 large.
%mages could reflect this !iew by prominently showing company leaders working together
with the >uatemalan people.
"he third dimension, masculinity, refers to the degree to which traditional masculine roles of
asserti!eness and competition are emphasi+ed. Cultures with high masculinity (M(C) stress
and !alue these !alues, whereas cultures that de9emphasi+e them tend to stress mutual
cooperation and family support. "hus Gapan, which has the highest M(C inde', may as a
society be more comfortable with sites that promote traditional male and female roles, as
well as ha!e certain sections of a site specifically dedicated to each gender. Countries such as
Cweden, on the other hand, should probably not ha!e sites that emphasi+e gender roles
because they ha!e a !ery low M(C score.
"he fourth dimension, uncertainty a!oidance, refers to the degree to which indi!iduals ha!e
an'iety about uncertain e!ents. #ofstede points out that cultures with a high amount of
uncertainty a!oidance (() tend to be e'pressi!e, ha!e more formal and simple rules, and
desire structure in organi+ations. Low ( cultures tend to desire more informal business
arrangements and are more rela'ed. "hus, indi!iduals within countries with a high (, such
as >reece, would tend to prefer sites with limited, simple, and redundant na!igational
de!ices, whereas indi!iduals in #ong Jong would tend to prefer greater comple'ity and less
control o!er na!igation.
BD %D? M(C (
rank score rank score rank score rank score
Arab Countries 7 80 26/27 38 23 3 27 68
Argentina 3/36 !" 22/23 !6 20/2# 6 #0/# 86
Australia !# 36 2 "0 #6 6# 37 #
Austria 3 ## #8 2 7" 2!/2 70
$elgium 20 6 8 7 22 ! /6 "!
$ra%il #! 6" 26/27 38 27 !" 2#/22 76
Canada 3" 3" !/ 80 2! 2 !#/!2 !8
Chile 2!/2 63 38 23 !6 28 #0/# 86
Columbia #7 67 !" #3 ##/#2 6! 20 80
Costa &i'a !2/!! 3 !6 # !8/!" 2# #0/# 86
(enmar) # #8 " 7! 0 #6 # 23
*ast Afri'a 2#/23 6! 33/3 27 3" !# 36 2
*'uador 8/" 78 2 8 #3/#! 63 28 67
+inland !6 33 #7 63 !7 26 3#/32 "
+ran'e #/#6 68 #0/## 7# 3/36 !3 #0/# 86
,erman- !2/!! 3 # 67 "/#0 66 2" 6
,reat $ritain !2/!! 3 3 8" "/#0 66 !7/!8 3
,ree'e 27/28 60 30 3 #8/#" 7 # ##2
,uatemala 2/3 " 3 6 !3 37 3 #0#
Hong .ong #/#6 68 37 2 #8/#" 7 !"/0 2"
/ndia #0/## 77 2# !8 20/2# 6 ! !0
/ndonesia 8/" 78 !7/!8 #! 30/3# !6 !#/!2 !8
/ran 2"/30 8 2! !# 3/36 !3 3#/32 "
/reland !" 28 #2 70 7/8 68 !7/!8 3
/srael 2 #3 #" ! 2" !7 #" 8#
/tal- 3! 0 7 76 !/ 70 23 7
0amai'a 37 ! 2 3" 7/8 68 2 #3
0apan 33 ! 22/23 !6 # " 7 "2
1ala-sia # #0! 36 26 2/26 0 !6 36
1exi'o /6 8# 32 30 6 6" #8 82
2etherlands !0 38 !/ 80 # #! 3 3
2ew 3ealand 0 22 6 7" #7 8 3"/!0 !"
2orwa- !7/!8 3# #3 6" 2 8 38 0
4a)istan 32 !7/!8 #! 2/26 0 2!/2 70
4anama 2/3 " # ## 3! !! #0/# 86
4eru 2#/23 6! ! #6 37/38 !2 " 87
4hilippines ! "! 3# 32 ##/#2 6! !! !!
4ortugal 2!/2 63 33/3 27 ! 3# 2 #0!
5alvador #8/#" 66 !2 #" !0 !0 /6 "!
5ingapore #3 7! 3"/!# 20 28 !8 3 8
5outh Afri'a 3/36 !" #6 6 #3/#! 63 3"/!0 !"
5outh .orea 27/28 60 !3 #8 !# 3" #6/#7 8
5pain 3# 7 20 # 37/38 !2 #0/# 86
5weden !7/!8 3# #0/## 7# 3 !"/0 2"
5wit%erland ! 3! #! 68 !/ 70 33 8
6aiwan 2"/30 8 !! #7 32/33 ! 26 6"
6hailand 2#/23 6! 3"/!# 20 !! 3! 30 6!
6ur)e- #8/#" 66 28 37 32/3 ! #6/#7 8
7rugua- 26 6# 2" 36 !2 38 ! #00
75A 38 !0 # "# # 62 !3 !6
8ene%uela /6 8# 0 #2 3 73 2#/22 76
9est Afri'a #0/## 77 3"/!# 20 30/3# !6 3! !
:ugoslavia #2 76 33/3 27 !8/!" 2# 8 88
*able 2. *he scores for the four dimensions of world culture. (From ;ofstede7 1%%1).
-ote; %f you see an &international& focused website that !iolates the con!entions of your
culture, please email me the type of !iolations so % can report them. mberWacm.org
Aeferences
Bauer, D., 6 Ca!onius, C., A. (570:). %mpro!ing the legibility of !isual display units through contrast re!ersal. %n F. >rand)ean, F. ?igliani (Fds.),
Frgonomic (spects of ?isual Display "erminals (pp. 5839524).London; "aylor 6 @rancis.
Benway, G.B. (5770). Banner blindness; "he irony of attention grabbing on the world wide web. Broceedings of the #uman @actors and Frgonomics
Cociety 24nd (nnual Meeting, C(, 5, 2<892<3.
Benway, G. B., 6 Lang, D. M. (Dec=5770). Banner blindness; *eb searchers often miss .ob!ious. links, %nternetworking, 5.8. Aetrie!ed 0=4:=:4;
http;==www.internettg.org=newsletter=dec70=bannerNblindness.html
Bernard. M. (4::4). F'amining the Fffects of #yperte't Chape on ser Berformance. sability -ews, 2, 4.
http;==psychology.wichita.edu=surl=usabilitynews=24=hyperte't.htm
Bernard. M. (4::4). ser F'pectations for the Location of Common F9Commerce *eb Kb)ects sability -ews, 2. 5.
http;==psychology.wichita.edu=surl=usabilitynews=25=web;ob<e't=e'om>htm
Bernard, M. (4:::). @inding information on the *eb; Does the amount of whitespace really matter$ sability -ews 4.5.
http;==psychology.wichita.edu=surl=usabilitynews=4*=whitespace.htm
Bernard, M (4::5). De!eloping schemas for the location of common web ob)ects. sability -ews 8.5.
http;==psychology.wichita.edu=surl=usabilitynews=8*=*ebNob)ect.htm
Bernard, M. (4:::). Constructing user9centered websites; "he early design phases of small to medium sites. sability -ews 4.5.
http;==psychology.wichita.edu=surl=usabilitynews=4*=webdesign.htm
Bernard. M. (4:::). F'amining user e'pectations of the location of web ob)ects. %"> %nternetworking 8.8 OKnlineQ.
http;==www.internettg.org=newsletter=dec::=articleNbernard.html
Bernard, M. (5777). Breliminary findings on the use of sitemaps . sability -ews 5.5.
http;==psychology.wichita.edu=surl=usabilitynews=5*=Citemaps.htm
Bernard, M. (5777). Citemap design; (lphabetical or categorical$ sability -ews 5.4.
http;==psychology.wichita.edu=surl=usabilitynews=5C=sitemap.htm
Bernard, M., Baker, A., 6 @ernande+, M. (4::4). Baging !s. Ccrolling; Looking for the Best *ay to Bresent Cearch Aesults. sability -ews, 2. 5.
http;==psychology.wichita.edu=surl=usabilitynews=25=paging.htm
Bernard. M., @ernande+, M., 6 #ull, C. (4::4). "he Fffects of Line Length on Children and (dults& Knline Aeading Berformance. sability -ews, 2, 4.
http;==psychology.wichita.edu=surl=usabilitynews=24=te'tNlength.htm
Bernard. M., #amblin, C., 6 Ccofield. B. (4::4). Determining Cogniti!e Bredictors of ser Berformance within Comple' %nterfaces. sability -ews, 2,
4. http;==psychology.wichita.edu=surl=usabilitynews=24=cogniti!eNfactors.htm
Bernard, M., 6 C. #ull, C. (4::4). *here Chould Hou But the Links$ Comparing Fmbedded and @ramed=-on @ramed Links sability -ews, 2. 5.
http;==psychology.wichita.edu=surl=usabilitynews=25=LinkNposition.htm
Bernard, M., Lida, B., Ailey, C., #ackler, "., 6 Gan+en, J. (4::4). ( Comparison of Bopular Knline @onts; *hich Ci+e and "ype is Best$ sability -ews, 2.
5. http;==psychology.wichita.edu=surl=usabilitynews=25=onlinete't.htm
Bernard, M., #ull, C., 6 Drake, D. (4::5). *here should you put the links$ ( comparison of four locations. sability -ews 8.4.
http;==psychology.wichita.edu=surl=usabilit-news/35/lin)s>htm
Bernard, M., Larsen, L. (4::5). *hat is the best layout for multiple9column web pages$ sability -ews 8.4.
http?//ps-'holog->wi'hita>edu/surl/usabilitynews=8C=layout.htm
Bernard, M., Liao, C., 6 Mills, M. (4::5). Determining the best online font for older adults. sability -ews 8.5.
http;==psychology.wichita.edu=surl=usabilitynews=8*=fontCA.htm
Bernard, M.6 Mills, M. (4:::). Co what is the most legible type, si+e, and format of common font to use for websites$ sability -ews 4.4.
http;==psychology.wichita.edu=surl=usabilitynews=4C=font.htm
Bernard, M.6 Mills, M. (4::5). *hich font to children prefer to read online$ sability -ews 8.5.
http;==psychology.wichita.edu=surl=usabilitynews=8*=fontGA.htm
Bernard, M., Mills, Beterson, M., 6 Ctorrer, J. (4::5). ( comparison of popular online fonts; *hich are best and when$ sability -ews 8.4.
http;==psychology.wichita.edu=surl=usabilitynews=8C=font.htm
Bouch, (., Juchnisky, (., 6 Bhatti, -. (4:::). Vuality is in the eye of the beholder; Meeting users& requirements for %nternet quality of ser!ice.
Broceedings of C#%& ::, 47398:2.
Byrne, M. D. (5778). sing icons to find documents; Cimplicity is critical. %nterchi& 78, 22<92/4.
Briley, G. 6 Ctolt+, C. (March 3, 5777). Comething annoying in the air; (irline web sites. *ashington Bost, F:5.
Chaparro, B., Childs, C., Braheswari, H., 6 Aappard, (. (4:::). Knline shopping; "?&s, toasters, and toysU oh myU sability -ews.
http;==psychology.wichita.edu=surl=usabilitynews=8*=webNusability.htm
Charness, -., Bosman, F. (577:). #uman factors in design, in G. F. Birren, J. *. Cchaie (Fds.), #andbook of the Bsychology of (ging, 8, (cademic
Bress, Can Diego, C(, pp. 22<92<8.
Dack.org (5777). Frnst 6 Houng&s sur!ey; Best practices for designing shopping cart and checkout interfaces. Aetrie!ed 5:=5<=:5. (defunct page).
http;==www.dack.com=web=shoppingNcartNprint.html
Delleart, B. >. C., 6 Jahn, B. F. (5770). #ow tolerable is delay$; Consumers& e!aluations of internet web sites after waiting. Aetrie!ed 0=4:=:4;
http;==ideas.uqam.ca=ideas=data=Bapers=dgrkubcen5770<2.html
Duchnicky, G. L., 6 Jolers, B. (. (5708). Aeadability of te't scrolled on !isual display terminals as a function of window si+e. #uman @actors, 4/, <089
<74.
Fllis, A. D., 6 Jurniawan, C. #. (4:::). %ncreasing the usability of online information for older users; ( case study in participatory design.
%nternational Gournal of #uman9Computer %nteraction, 54 (4), 4<8943<.
@araday, B. (4::5). (ttending to web pages. C#% :5& %nteracti!e Boster, 5/795<:.
@ishler, J. D. (Cummer=5707). Jids talk back. (dobe Maga+ine. Aetrie!ed 0=4:=:4 (BD@);
http;==www.adobe.com=products=adobemag=archi!e=pdfs=
"8smfe)f>pdf
@itts, B. M. (57/2). "he information capacity of the human motor system in controlling the amplitude of mo!ement. Gournal of F'perimental
Bsychology, 23, 8059875.
@ogg, B., Marchall, G., Laraki, K., Ksipo!ich, (., ?arma, C., @ang, -., Baul, G., Aangnekar, (., Chon, G., Ceani, B., 6 "reinen, M. (4::5). *hat makes web
sites credible$ ( report on a large quantitati!e study. Broceedings of C#% :5, <59<0.
@orsythe, C., Aing, L., >rose, F., Bederson, B., #ollan, G., Berlin, J., 6 Meyer, G. (577<). #uman @actors Aesearch and De!elopment for the %nternal *eb
at Candia -ational Laboratories; ( Ae!iew and pdate (conference proceedings).
@oss, C. L. (5707). "ools for reading and browsing hyperte't. %nformation Brocessing Management, 4/, 2:39250.
@uccilla, G. 6 Bi++olato, G. (Gune=5770). Creating web site designs based on user e'pectations and feedback. %nternetworking. Aetrie!ed 0=4:=:4;
http;==www.internettg.org=newsletter=)une70=webNdesign.html
>eorgia "ech Aesearch Corporation. >?&s *** ser Cur!ey. Aetrie!ed 0=4:=:4; http;==www.g!u.gatech.edu=userNsur!eys=
>ilbert, J., 6 Cchleuder, G. (577:). Fffects of color and comple'ity in still photographs on mental effort and memory. Gournalism Vuarterly, <3, 3279
3/<.
>regory, M., 6 Boulton, F. C. (573:). F!en !ersus une!en right9hand margins and the rate of comprehension reading. Frgonomics, 58, 2439282.
>roff, L., Liao, C., Chaparro, B., 6 Chaparro, (. (5777). F'ploring how the elderly use the web. sability -ews.
http;==wsupsy.psy.tcwsu.edu=surl=usabilitynews=5C=elderly.htm
#anna, L., Aisdan, J., C+erwinski, M., (le'ander, J. (5770). "he role of usability research in designing children&s computer products. Aetrie!ed
0=4:=:4; http;==www.research.microsoft.com=users=maryc+=druin70.htm
#artley, G. (5772) Designing instructional te't for older readers; ( literature re!iew. Aetrie!ed 0=4:=:4;
http;==cogprints.ecs.soton.ac.uk=archi!e=:::::/74=::=5770:4::4.html
#awthorn, D. (4:::). Bossible implication of aging for interface designers. %nteracting with Computers, 54, /:39/40.
#orton, *. (5707). Designing and writing online documentation; #elp files to hyperte't. Gohn *iley 6 Cons; -ew Hork.
#orton, *. (5773). Designing icons and !isual symbols. C#% &73. Aetrie!ed 0=4:=:4; http;==www.acm.org=sigchi=chi73=proceedings=tutorial=wh.htm
#ofstede, >. (5775). Cultures and Krgani+ations; Coftware of the Mind; %ntercultural Cooperation and its %mportance for Cur!i!al, Mc>raw #ill, -ew
Hork.
#uey, F. B. (57<0). "he psychology and pedagogy of reading. Cambridge, M(; M%" Bress.
Gacko, G., (., 6 Cal!endy, >. (577<). #ierarchical menu design; breadth, depth, and task comple'ity. Berceptual and Motor Ckills, 04, 5503954:5.
Gar!enpaa, C., "odd, B. (., (5773). Consumer reactions to electronic shopping on the world wide web. %nternational Gournal of Flectronic Commerce, 5
(4) /7900.
Gupiter Communications (4:::). "argeting teens is a gender game; @ocus on marketing to girls and programming to boys. Aetrie!ed 0=4:=:4 (BD@);
http;==www.e9global.es=:5<=:5<N)upiterNfocus.pdf
Jim, G. 6 Hoo, B. (4:::). "oward the optimal link structure of the cyber shopping mall. %nternational Gournal of #uman9Computer Ctudies, /4, /859
//5.
Jing, *. C., Dent, M. M., 6 Miles, F. *. (5775). "he persuasi!e effect of graphics in computer9mediated communication. Computers in #uman
Beha!ior, 3, 4<79437.
Jno!ed, L., 6 Chneiderman, B. (570<). Fmbedded menus; Celecting items in conte't. Communications of the (CM, 47, 8549850.
Larson, J., 6 C+erwinski, M. (5770). *eb page design; %mplications of memory, structure and scent for information retrie!al. Aetrie!ed 0=4:=:4;
http;==www.research.microsoft.com=users=maryc+=chi705.htm
Lohse, >., 6 Cpiller, B. (5770). Flectronic shopping. Communications of the (CM, 25, 05900.
Lynch, B. 6 #orton, C. (5777). *eb style guide; Basic design principles for creating web sites. Aetrie!ed 0=4:=:4;
http;==info.med.yale.edu=caim=manual=contents.html
Marcus, (., >ould, F. (4:::). Cultural dimensions and global web user9interface design; *hat$ Co what$ -ow what$ <th Conference on #uman @actors
6 the *eb. Aetrie!ed 0=4:=:4; http;==www.tri.sbc.com=hfweb=marcus=hfweb::Nmarcus.html
McDonald, C., 6 Cte!enson, A. G. (5770). Fffects of te't structure and prior knowledge of the learner on na!igation in hyperte't. #uman @actors,
2:,50943.
Mead, C. F., Cpaulding, ?. (., Cit, A. (., Meyer, B., 6 *alker (5773). Fffects of age and training on world wide web na!igation strategies. Broceedings
of the #uman @actors and Frgonomics Cociety 25st (nnual Meeting, 5/495/<.
Mills, C. B. 6 *eldon, L., G. (5703). Aeading te't from computer screens. (CM Computing Cur!eys, 2, 84798/0.
Morkes, G. 6 -ielsen, G. (5773). Concise, scannable, and ob)ecti!e; #ow to write for the *eb. (lertbo'. Aetrie!ed 0=4:=:4;
http;==www.useit.com=papers=webwriting=writing.html
Mullet, J., 6 Cano, D. (577/). Designing !isual interfaces; Communication oriented techniques. CunCoft Bress. Mountain ?iew, C(.
-ielsen, G. (577<). *hy frames suck (most of the time). (lertbo'. Aetrie!ed 0=4:=:4; http;==www.useit.com=alertbo'=7<54.html
-ielsen G. (5773). Gacob -ielsen&s (lertbo'. Aetrie!ed 0=4:=:4; http?//www>useit>'om
-ielsen G. (5777). "op ten mistakes re!isited. (lertbo'. Aetrie!ed 0=4:=:4; http;==www.useit.com=alertbo'=77:/:4.html
-orman, J. L., 6 Chin, G. B. (5700). "he effect of tree structure on search in a hierarchical menu selection system. Beha!iour and %nformation
"echnology, 3, /59</.
-orman, D. (. (5700). "he psychology of e!eryday things. Basic Books -ew Hork.
-ygren, F., 6 (llard, (. (577<). Between the clicks skilled users scanning of pages. Designing for the *eb; Fmpirical Ctudies. #uman @actors and the
*eb=#"ML Conference. Candia -ational Laboratories, (lbuquerque, -M.
Bagendarm, M., 6 Cchaumburg, #. (4::5). *hy are users banner9blind$ "he impact of na!igation style on the perception of web banners. Gournal of
Digital %nformation. Aetrie!ed 0=4:=:4; http;==)odi.ecs.soton.ac.uk=(rticles=!:4=i:5=Bagendarm=
BC Maga+ine, SD-et. (4:::). "he e9complaint window. Aetrie!ed 0=4:=:4 (charge); http;==www.+dnet.com=ibi+mag=stories=re!iews=
0@#0!72@260#03#@00>html
Blamondon, A., 6 (limi, (., M. (577<). Cpeed=accuracy in targeted directed mo!ements. Aetrie!ed 0=4:=:4;
http;==www.bbsonline.org=documents=a=
::=::=:2=70=bbs:::::2709::=bbs.plamondon.html
Blude, D. G., 6 #oyer, *. G. (5705). (dult age differences in !isual search as a function of stimulus mapping and processing load. "he Gournal of
>erontology, 8<, /709<:2.
Boynter %nstitute (4:::). Definitely not your father&s newspaper and surpriseU Aetrie!ed 0=4:=:4; http;==www.poynter.org=centerpiece=:/:8::.htm
BricewaterhouseCooper LLB (4::5). F9retail intelligence system; %nternet users consumer panel. Aetrie!ed 0=4:=:4;
http;==www.pwcglobal.com=e'tweb=
ncpressrelease.nsf=Doc%D=/D8(3<75508/0D8:0/4/<(:0::35DB5:
Briestly, M. (5773). Documenting comple' hierarchies with #"ML frames. Broceedings of BC%>DKC 73, 448948/.
Aamsay, G., Barbesi, (., 6 Breece, G. (5770). Bsychological in!estigation of long retrie!al times on the *orld *ide *eb. %nteracting with Computers,
5:.
Aiegelsberger, G., 6 Casse, M. (4::4). @ace it; Bhotos don&t make a web site trustworthy. Broceedings of C#% &:4, 3249328.
Aetail @oward (4::4). Knline shopping satisfaction high but frustrations persist retail forward sur!ey reports. Aetrie!ed 0=4:=:4;
http;==www.retailforward.com=freecontent=pressreleases=press88.asp
Aohn, G. (5770). Creating usable e9commerce sites. Ctandard ?iew, <, 55:955/.
Ausso, B., 6 Boor, C. (5778). #ow fluent is your interface$ Designing for international users. %-"FAC#% &78. 8249823.
Ccharff, L., @., #ill, (., 6 (ustin, C., @. (577<). Color test results. Aetrie!ed 0=4:=:4; http;==www.thecube.com=color=sur!reslts.html
Cel!idge, B. (5777). #ow long is too long to wait for a website to load$ sability -ews.
http;==psychology.wichita.edu=surl=usabilitynews=5s=timeNdelay.htm
Cel!idge, B. (5777). #ow usable are the airline websites$ sability -ews.
http;==psychology.wichita.edu=surl=usabilitynews=5w=sabilityNairline.htm
Clatalla, M. ((pril, 57, 4::5). Knline shopper; *hen familiarity breeds temptation. -ew Hork "imes. Aetrie!ed 0=4:=:4;
http;==www.nytimes.com=4::5=:2=57=technology=57C#KB.html
Cnowberry, J., Barkinson, C., 6 Cisson, -. (5708). Computer display menus. Frgonomics, 4<, <779354.
Corg, G. (. (570/). (n e'ploratory study of type face, type si+e, and color paper preferences among older adults. npublished masters thesis,
Bennsyl!ania Ctate ni!ersity, ni!ersity Bark.
Cpain, J. (5777). *hat&s the best way to wrap links$ sability -ews. http;==psychology.wichita.edu=surl=usabilitynews=5w=Links.htm
Cpiller, B., 6 Lohse, >. (5770). ( classification of internet retail stores. %nternational Gournal of Flectronic Commerce, 4, 4.
Cteinbruck, ., Cchaumburg, #, Duda, C., 6 Jruger, ". (4::4). ( picture says more than a thousand words; Bhotographs as trust builders in e9
commerce websites. Broceedings of C#% &:4, 3209327.
Culli!an, "., -orris, C., Coloway, F., 6 Beet, M. (4:::). *hen kids use the web; ( naturalistic comparison of children&s& na!igation beha!ior and
sub)ecti!e preferences on two www sites. <th Conference on #uman @actor 6 the *eb. Aetrie!ed 0=4:=:4;
http;==www.tri.sbc.com=hfweb=sulli!an=Culli!an.html
"edeschi, B. (5777). >ood web site design can lead to healthy sales. "he -ew Hork "imes on the *eb F9Commerce Aeport. Aetrie!ed 0=4:=:4;
http;==www.nytimes.com=library=tech=77=:0=cyber=commerce=8:commerce.html
"ilson, A., Dong, G., Martin, C., 6 Jieche, F. (5770). @actors and principles affecting the usability of four e9commerce sites. Broceedings of the #uman
@actors and the *eb. Aetrie!ed 0=4:=:4; http;==www.research.microsoft.com=
users=maryc+=hfweb70=tilson=inde'.htm
"obias, C. L. (5703). Computers and the elderly; ( re!iew of the literature and directions for future research. Broceeding in the #uman @actors Cociety
85st annual meeting, 0<<903:.
"ufte, F. A. (577:). Fn!isioning %nformation. Cheshire, C"; >raphics Bress.
"ullis, ". C., Boynton, G. L., 6 #ersh, #. (577/). Aeadability of fonts in the windows en!ironment. Broceedings of C#% &7/. Aetrie!ed 0=4:=:4;
http;==www.acm.org=sigchi=chi7/=proceedings=intpost=tstNbdy.htm
?ercruyssen, M. (577<). Mo!ement control and the speed of beha!ior. %n (. D. @isk, *. (. Aogers (Fds.), #andbook of #uman @actors and the Klder
(dult, (cademic Bress, Can Diego, C(.
?ora, D. Designing sable *eb Cites. C *est Communications, 5770.
*eale, A. (57<5). Aetinal illumination and age. "ransactions of the %lluminating Fngineering Cociety, 4<, 7/.
*eb (ccessibility %nitiati!e (*(%). Aetrie!ed 0=4:=:4; http;==www.w8.org=*(%=
*ickens, C. (570<). "he ob)ect display; Brinciples and a re!iew of e'perimental findings ("ech. Aept. CBL 0<9<). Champaign,%L; ni!ersity of %llinois
Cogniti!e Bsychophysiology Laboratory, Dept. of Bsychology.
*olfmaier, ". >. (March=5777). Designing for the color9challenged; ( challenge. %nternetworking. Aetrie!ed 0=4:=:4;
http;==internettg.org=newsletter=
mar77=accessibilityNcolorNchallenged.html
*right, B., Milroy, A., 6 Lickorish, (. (5777). Ctatic and animated graphics in learning from interacti!e te'ts. Furopean Gournal of Bsychology of
Fducation
Houngman, M., 6 Ccharff, L. (5770). "e't width and margin width influences on readability of >%s. Couthwest Bsychological (ssociation. Aetrie!ed
0=4:=:4; http;==hubel.sfasu.edu=research=te'tmargin.html
Houngs, F. (Cpring, 5777). Kld and young user interfaces; %s there a .-intendo effect.$ Communications "echnical >roup -ewsletter, #uman @actors
and Frgonomics Cociety, 7954.

Anda mungkin juga menyukai