Content-Type: text/html

Readers  Response to Digital News Stories Presented in Layers and Links
 
Print news, a traditionally change-resistant profession, has surprisingly
embraced the coming of the Internet and World Wide Web. But as newspapers race
into electronic publishing in record numbers (2,297 online worldwide at the
latest Editor & Publisher count), they seem to be forgetting the most important
thing: their readers.
Perhaps motivated by dreams of profit, or perhaps by a fear of getting left
behind, many newspapers are entering the information superhighway without
knowing exactly where they re going. Pegie Stark Adam of the Poynter Institute
for Media Studies noted in a 1995 conference on interactive media that
newspapers are rushing  to dump anything they can online, as fast as they can,
with little customization  (Reason, Oct. 29, 1995, par. 9). Even industry
executives admit this is the case. Mercury Center managing editor Bruce Koon
told attendees at the same conference that writing for an online newspaper  is
like a revolutionary war. You go into action not when you are ready, but when
the opportunity presents itself  (Reason, Oct. 30, 1995, par. 3).
The result, according to media futurist Roger Fidler, who is quoted in a 1995
Editor & Publisher story:  Online publications are still uncompelling,
frustrating and time-consuming to use  (Webb, p. 36).
Although good things can come out of the profession s open-mindedness and
willingness to experiment, things seem to be moving too fast. Currently,  so
much experimentation is going on so fast you can see the changes daily  (Pogash,
1996, par. 9). Indeed, the industry would do well to keep its enthusiasm for
electronic publishing but back away from this haphazard, trial-and-error
approach   which has included everything from a love-letter service offered by
the Raleigh News and Observer s NandO.net to a virtual museum exhibit of Russian
history offered by the St. Petersburg Times   and instead try to find out what
really works. In other words, newspapers need to work with readers to find out
what they really want in an electronic paper. Only then will they know how to
make the best use of this new medium.
Among the many questions that need to be addressed is how to best make use of
the hyperlinked environment, or the ability to build stories in layers of
information so related material or additional details are just a click away.
Currently, according to a 1995 speech by Nora Paul of the Poynter Institute,
 Too many electronic news products are simply shovelware   scooping up the old
flat text used in the ink-on-paper product and throwing it on the screen  (par.
20). Hypertext, however, gives us the opportunity to  rethink reporting as a
layering of news   (Paul, 1995, par. 20)   the opportunity to create content
with depth by providing links to other relevant documents.
This study addresses how to best present stories in such layers and links to
make them of maximum usefulness and appeal to readers. Specifically, this study
attempts to determine the amount of information most desirable in the top layer
of any story and the best way to incorporate links from a story to sidebars or
other related material. For purposes of this study, the top layer refers to the
initial screen that lets readers know what stories are available to them. The
term link, which generally refers to the physical connector, or the place a
reader would click to move from one story to another, also includes the
description of the related material to which the connection leads.
As Slatin (1990) points out:
The more cryptic the link or node identifiers are, the harder it is for the
reader to predict the results of activating a particular link. The harder it is
to make such predictions, the greater the likelihood that the reader will simply
opt out of the process in frustration (p. 880).
 
 
 
LITERATURE REVIEW
 
Because of the size of the online world, it s hard to believe it is just getting
past infancy. But the truth is, electronic publishing   at least as defined by
today s technology   is indeed a new concept. So while many articles in the
popular press and trade journals discuss current trends and future possibilities
of electronic media, little academic literature is available. In terms of
academic research, studies that address reader preferences when dealing with
electronic media are especially rare.
Until the last three years, most studies related to electronic publishing
primarily were concerned with whether computer presentation of news may be
inferior to traditional newspaper presentation in terms of readers  processing
information. However, the studies showed there were no differences in reading
speed, comprehension or recall of facts between news stories presented on paper
and the same stories presented on a computer screen. Specifically, Oborne and
Holton (1988) reported no significant differences in either reading speed or
comprehension between screen and paper. Later, Melvin DeFleur, Davenport, Cronin
and Margaret DeFleur (1992) reported that facts from news stories presented by
newspaper or computer screen were recalled at a significantly higher rate than
facts from the same stories presented by radio or TV. The researchers were
surprised to find that recall for computer screen presentation did not differ
significantly from that of newspaper presentation.
In another comparison among media, Thompson (1995) tested reaction to stories
and pictures presented in three different modes   on paper, on computer screen,
and on computer screen with a 5- to 8-second sound bite available. All
presentations used the same design, but the multimedia condition obviously had a
feature that could not be incorporated into the other two. Thompson then tested
the perceived simplicity, interestingness and pleasingness of the three
different presentation modes. He also asked the 75 study participants to rate
the simplicity, interestingness and pleasingness of the text in each mode and
the image in each mode. Finally, he collected open-ended responses about each
mode. Thompson found no significant differences in any of the simplicity
comparisons. However, he found the multimedia condition consistently ranked
significantly better than the other two conditions in terms of overall
interestingness and pleasingness, text and image interestingness and image
pleasingness.
In analyzing the responses to the open-ended questions, Thompson (1995) found
that negative comments about the paper presentation outnumbered positive
comments 25-15. But in the  computer condition, positive comments outnumbered
negative ones 30-10. For the multimedia condition, positive comments outnumbered
negative ones 60-14. Thompson interpreted the comments as a sign that young
readers would encourage and embrace multimedia presentation and would offer
increased resistance to ink on paper.
One of the first studies that actually tested audience reaction to electronic
delivery of news and information involved teletext. Although today s technology
is far more advanced than that used in the teletext trials of the early 1980s,
we can still learn some lessons from those experiments. Among the highlights of
a teletext study by Elton and Carey (1983): 1) The most frequently used sections
of the teletext newspaper were news, sports, weather, entertainment listings and
business information. Sections offering community and consumer information were
accessed less frequently; 2) Users rarely read long items completely, usually
going only two to three screens into a story before going on to something else;
3) Users complained about the slowness of the service; 4) Sections updated
regularly also were those accessed most frequently; 5) Many users complained
that the technology was not simple enough.
More recently, an experiment by Mueller and Kamerer (1995) asked 62 college
students to rate their satisfaction with electronic newspapers and preferences
for electronic newspapers vs. traditional newspapers after spending 30 minutes
exploring San Jose s Mercury Center. The researchers found that:
Subjects understand the choices available to them most of the time, believe it
is a useful tool, would rather search for a topic in electronic newspapers than
traditional, believe that with more exposure to the new medium they would become
more at-ease using it, and find information more current in electronic
newspapers than traditional (Mueller and Kamerer, 1995, p. 11).
However, Mueller and Kamerer (1995) also  found the new medium to be unappealing
to browse leisurely, inappropriate for all news material, uncomfortable to
travel through, not preferable over traditional newspapers, and more difficult
to read than a traditional newspaper  (p. 11). The researchers also asked
participants about their media habits, and found that, generally, media use
correlated positively with the electronic news satisfaction variables.
Another study reports on the effect of the Internet on time spent with
traditional news media. In this study, Bromley and Bowles (1995) collected
information from 98 users of the Blacksburg (Va.) Electronic Village, a
seven-month-old service that offered online users free access to detailed local
information and to the Internet.  While survey respondents reported spending an
average of 78 minutes a day using online services, the majority said the amount
of time they spent with newspapers, television and radio had not changed since
they began using the Electronic Village. Among those who said their traditional
media use had declined, the time was more likely to be subtracted from
television viewing than from newspaper reading or radio listening. Specifically,
18.4 percent said they spent less time with television, 7.1 percent said they
spent less time with newspapers and 4.1 percent said they spent less time with
radio. An Editor & Publisher article also suggests television is more vulnerable
to online usage than are newspapers since most online activity in U.S.
households occurs between 7 and 10 p.m.,  right in the middle of network TV s
prime time viewing  (Consoli, 1997, p. 36).
Finally, a recent Newspaper Research Journal study provides baseline information
about startup costs, revenue strategies, staffing, content and number of
subscribers to online newspapers. The study by C. Harper (1996), which drew
responses from 40 online newspapers, found that publishers  approaches to the
new medium varied widely. For instance,  budgets ran the gamut from virtually no
startup costs, excluding salaries, to budgets in excess of $1 million per year
(C. Harper, 1996, p. 7). Further,  in the smaller markets, only one person
supervised the electronic newspaper. In larger markets, the staff ranged from
five to 50 people  (C. Harper, 1996, p. 9).  Overall, however, the trend was
toward smaller budgets and leaner staffs. For instance, 15 of the 27 respondents
to the budget question said their first-year budgets were under $5,000, and 19
of the 33  respondents to the staffing question said five or fewer people were
running the electronic product.
C. Harper (1996) also found that revenue strategies varied, with 13 of the
newspapers surveyed charging for subscriptions and seven selling Internet access
accounts to readers. Seven of the publications carried no ads, and three others
allowed free advertisements. Among those with paying  advertisers, some based
their rates on the number of hits to the publication while others charged daily
or monthly rates. Only 1/3 of the papers provided different or additional
content in their electronic products, and only 11 of the 40 included audio or
video presentations in the electronic versions.
None of these studies, however, directly addresses presentation in layers and
links. The computer revolution, and, in particular, the birth of hypertext, has
given us what Slatin calls  a new medium for thought and expression (1990, p.
870). Now, as Slatin explains, instead of creating conventional text within a
marked beginning and a marked end, we can create documents with multiple entry
and exit points, and with different possible pathways in between. However, a
content analysis of a sample of 83 U.S. online newspaper sites (Gubman and
Greer, 1997) found few online newspapers adapting their writing style to use the
opportunities provided by hypertext. Only 13 (15.7 percent) of those newspapers
surveyed used any kind of nontraditional storytelling or linked boxes.
Fredin (1997) addresses this lack of innovation by proposing prototypes for
hypertext stories that call for layers and links. One of his prototypes is the
simple-digression format, much like a main news story linked with sidebars
determined by a journalist. Another is the complex-digression format, for
detailed and continuing stories such as U.S. health care, in which readers can
chart their own pathway by changing the selection of the main story (and thus
the links) to fit their interests. Fredin suggests long headlines on
complex-digression format stories. He says these are necessary for two reasons:
1)  each file has many links, so each headline should contain more than one
point, and 2) the application of what he calls the  first a little, then a lot
(Fredin, 22). These more detailed headlines that provide an abstract of the
story are reminiscent of nut graf summary decks in print newspapers or news
summaries written in display type and placed between the headline and the story.
The rationale for summary decks, as explained by Hilliard (1991) in Newspaper
Research Journal, is much the same as that for layer one summaries of stories in
an electronic publication.
The nut graf summary deck affords the scanning reader a quick, more complete
synopsis of the story than is provided by the headline and with a greater
summary of key information than can be provided by the lead paragraph(s), yet
not attempting to provide complete details (Hilliard, 1991, p. 78).
Hilliard (1991) says the product gives readers something similar to a radio or
TV news summary, with the added benefit of allowing  immediate access to the
story s complete details  (p. 78) for those readers who decide they want more
information. Further, for readers who don t want the full story, the nut graf
summary deck  permits some interpretation of the context or impact of the story,
a journalistic aspect often difficult to achieve in the typical single- or even
multiple-deck headlines  (Hilliard, 1991, pp. 78, 79).
Investigators tested the idea in Washington on both subscribers and
non-subscribers of the Walla Walla Union-Bulletin. The summary decks turned out
to be popular, with participants from both groups saying nut grafs helped them
take in complex information quickly and generally agreeing they were worthwhile
despite the space they took up (Hilliard, 1991).
 
 
 
RESEARCH PROBLEM
 
This study differst from those mentioned above because of its focus on layers
and links. The general question being addressed was: How can newspapers best
present electronic news in layers and links to make it of maximum usefulness and
appeal to potential readers? Although various areas may be addressed under this
rather broad heading, the focus of this study was to try to find out how readers
would react to different lengths and placements of summaries and links.
Research questions included:
1) How do initial summary and link length affect subjects  decisions to read a
story?
To answer that rather broad question, the following areas were addressed:
          How do initial summary and link length affect subjects  reports of whether
they have enough information to make a decision about reading the story?
          How does link length affect subjects  ability to predict story content?
          How do initial summary and link length affect subjects  motivation to read
the stories?
          How do initial summary and link length affect subjects  reports of feeling
informed without reading the story?
2) How does link presentation affect subjects  ability to find the link?
3) Do subjects think they would like reading news presented in layers and links?
In particular, subjects were asked whether they would like routinely having the
option of seeing a brief summary of an article without having to see the whole
story and whether they would like routinely having the chance to get more depth
on a story if they wanted it.
4) Which summary and link length and which link presentation do subjects say
they prefer?
Specifically, subjects were asked:
          Whether they preferred a link to a sidebar to be a separate element or words
within the text of the main story;
          Whether they preferred the link length to be a few words, a headline or a
headline and story abstract;
          Whether they preferred the initial summary length to be a headline only, a
headline and deck or a headline and story abstract.
Questions were designed to measure reader attitudes, not behaviors.
 
 
 
METHODOLOGY
 
The researchers asked permission of professors in various university departments
to recruit students from their classes. One researcher then visited those
classes, explained to students the requirements for participation  and asked for
volunteers.
Test sessions were conducted with up to 15 students at a time in a Macintosh
computer lab. Each student had his own computer and was able to proceed at his
or her own pace.
Each test session began with a brief training period to show subjects how to
move from a story summary to a main story, how to follow a link to a sidebar and
how to return to a previous page. The actual test began with subjects seeing
three story summaries stacked in a column on an otherwise empty page. Some
groups saw short summaries of the stories (a simple headline), some saw
medium-length summaries (a headline plus deck), and some saw long summaries (a
headline plus a 3-sentence abstract of the story). Headlines for each group of
stories were the same. Stories were old Akron Beacon Journal articles.
The full text of each story was available at the click of a mouse. All stories
were laid out the same, and each had one link to a sidebar, with the only
difference coming in how those links to the sidebars were incorporated.
For one story, the link was simply several boldface words underlined within the
text of the main story. For the second story, the link was a 6- to 10-word
headline set as a separate element in the shoulder of the second column of type.
In the third design, the link again was a separate element but included a
headline and 3-sentence abstract of the sidebar.  All subjects saw each type of
link. Link types were rotated among the different stories to minimize potential
bias from story content (see Appendix A for example).
Subjects were instructed to read each summary, the full text of the main stories
and the related sidebars. They filled out a questionnaire as they went through
the stories. The question survey asked subjects to respond to 104 Likert scale
statements.
In the first section of the survey, subjects were asked if the summary gave them
enough information so they could make a decision whether to continue reading,
and what that decision would be. The purpose of these questions was to determine
if varying length summaries affected subjects  ability to make a decision about
reading further or affected their actual decision.
Next, the survey asked how interested subjects were in the story topic, based
only on what they knew of the topic from the summary.
The survey also asked subjects if they felt informed enough about the story
subject from the summary, without reading the full story. This question was
important in determining how effective the different summary lengths were in
conveying the major points of the story and in determining if subjects who said
they felt informed enough were any less likely to want to continue reading.
In the second section of the survey, subjects followed the link to the full text
of story and reevaluated whether they would read it if they came across it in a
newspaper.  Specifically, comparing the responses with those from earlier
questions told whether the number of people who said they wanted to read the
story increased or decreased once they actually saw it.
Next, subjects evaluated the link to the story sidebar, telling if they found
it, if it was difficult to find and if it was distracting (see Appendix A for
example).
Subjects also were asked if they thought they had enough information to know
what the  sidebar to which the link connected would be about, a question
designed to address how link length affects readers  ability to predict content.
As with the initial summaries, subjects also were asked if the link gave enough
information for them to decide whether to go ahead and read the sidebar, what
that decision would be and whether they felt informed about the subject without
reading the full story.
In the third section of the survey, subjects followed the link to the sidebar
and, after reading the story, reevaluated whether the link gave enough
information for them to know what the story would be about. Finally, subjects
reevaluated whether they would read the story if they came across it in a
newspaper.
This process was then repeated for the second and third stories and their
sidebars.
After going through all the stories and seeing the three different types of
links, subjects were asked 1) which link placement they preferred (a few words
within the text of the story or a separate box of information), and why;  and 2)
which link length they preferred (a few words, a headline, a headline and
3-sentence abstract), and why.
Next, subjects viewed all three initial summary lengths and told which they
preferred and why.  In addition, subjects were asked whether they would like the
option of seeing a story summary without having to see the whole story and
whether they would like the option of getting more information about a story if
they wanted it.
Finally, a demographics section asked routine questions about gender, age and
race, and also asked participants how often they used computers and online
services and how often they read a traditional or online newspaper.
 
 
 
RESULTS/DISCUSSION
 
 
The Sample
The sample consisted of 162  students recruited from a variety of classes and
from 25 different majors. The group was split nearly evenly between under- and
upperclassmen and also included eight post-baccalaureate or graduate students.
Ages of  subjects ranged from 18 to 47, with about 80 percent falling between 18
and 23. About two-thirds of the subjects were female. The sample included 124
Caucasians, 19 African Americans, 2 Hispanics and 2 of mixed race. Fifteen left
the race question blank. Overall, 72 percent said they use a computer at school
daily or weekly and 55 percent said they use a computer at home daily or weekly.
Fifty-seven percent said they go online from school daily or weekly and 40
percent said they go online from home daily or weekly. Asked how often they read
a newspaper, 54 percent said daily, 36 percent said once or twice a week and 10
percent said once or twice a month or less. About 20 percent said they read a
newspaper online on a daily or weekly basis.
    1   How do initial summary and link length affect subjects  decisions to
read a story?
Initial summary length significantly affected whether subjects thought they had
enough information to decide about reading the full story on only one story (see
Table 1). An index comparing the abstract group against the other two groups
combined shows the differences are statistically significant when totaled across
all three stories (see Table 2).
[  INSERT TABLES 1 AND 2 ABOUT HERE  ]
It is interesting that agreement for even the shorter summaries is 70 percent or
above. Although one might expect larger differences between groups, these
numbers are not unusual when considered along with studies that show most
readers of traditional newspapers make their decision whether to continue
reading within the headline or first few sentences of a story (Schramm, 1947;
Swanson, 1948). At the same time, however, it follows that, as Slatin pointed
out, the more information a reader has, the more likely he will be to feel
comfortable making that call. That would explain why the abstract groups
consistently are most likely to say they have the information to make that
decision. But for the same reason, the story in which the head-plus-deck group
reported feeling least able to decide whether to read the full story stands out.
The reason for this deviation remains unclear (see Appendix A for example).
The same overall trend was apparent with the links to the sidebars (see Tables
3-4). Here, the differences in whether the various links to the sidebars gave
enough information for subjects to decide about reading the sidebar were
statistically significant twice.
[ INSERT TABLES 3 AND 4 ABOUT HERE ]
These figures also are interesting because of how they line up next to the
numbers from the same question about the initial summaries. Specifically, it is
notable how similar the statistics from the headline and abstract conditions
are. This seems to indicate that even though the links to the sidebars are
worked into the layout of the main stories, those stories and the context they
provide are not necessarily giving readers any extra advantage in making a
decision about  reading the sidebars.
The variation in the percentage of subjects who said they had enough information
with the words-in-text links to make a decision also is worth noting.  Even
though the percentages were lower than they were for the other types of links,
more than two-thirds of  subjects in these groups still said they had enough
information to decide about two of the stories. But for the third story, the
percentage who said they had enough information dropped to 33 percent. The
difference is undoubtedly because the first two links, although only a few
words, are fairly explicit. But the third link is more vague and ambiguous. This
is important, however, because the connection or tie from a story to a sidebar
often is somewhat indirect.  And to try to express what that sidebar is about in
a few words in the middle of a sentence of the main story   without interrupting
the flow of that story   is very likely to result in such a vague link. This
could be an important argument against the use of such links in stories.
Adding to the credibility of that argument is subjects  report of how well they
could predict, based on the links, what the sidebars were about. As expected,
the more information provided in the links, the more likely subjects were to
report they could predict what the sidebars were about.  The differences were
statistically significant for two stories and approached significance for the
third (see Tables 5-7). Again, the discrepancy was especially evident in the
third sidebar, the one in which the words-in-text link was more vague and
indirect than the other words-in-text links.  As mentioned, it is unusual that
with some stories and types of links, more subjects said they had enough
information to decide about reading the stories than said they had enough
information to predict what those stories were going to be about.
[  INSERT TABLES 5, 6 AND 7 ABOUT HERE ]
It also is interesting that, in general, the percentage of subjects who thought
there was enough information in the links for them to predict what the sidebars
were going to be about decreased after they actually saw the sidebars. In other
words, upon seeing the sidebars that the links described and getting a chance to
reevaluate the effectiveness of the links, subjects decided the links did not
describe the stories as well or as clearly as they had expected.
Again, the results are significant by type for two of the stories, and they
follow the same trend for the third story. In all cases, subjects in the
abstract link groups were most likely to agree, after having read the actual
sidebars, that the links they followed there gave enough information for them to
know what the stories were going to be about. Subjects in the words-in-text link
groups were least likely to agree. Taken together, these numbers indicate that
the shorter links may at best be counted on only to communicate a general story
topic, while the longer links may be able to provide at least some insight into
angle and direction but still may not contain enough information for all readers
to accurately predict story content.
Even though, evaluations of the links generally decreased once subjects got a
chance to see the actual stories, there were no corresponding decreases in the
number of people who said they wanted to read the stories.  Again, this may
boost the idea that subjects are making their decision to read   or not read   a
story from the topic rather than from a firm understanding of what the story is
about. This would explain why they still wanted to keep reading even though the
stories were not exactly what they had expected.
One of the most crucial aspects of the predictability and decision-making issues
is how they affect readers  desire to actually read a given story.
The amount of information in the summaries made no significant differences in
the  percentage of subjects who said they wanted to read the stories. Nor did
the amount of information in the links to the sidebars significantly affect
whether subjects reported wanting to read the sidebars.
Overall, 74 percent to 82 percent of respondents said they would read the main
stories because they were interested in the various story topics, and 57 percent
to 75 percent said they would read the sidebars based on their interest in the
individual story topics. One area of further research will be to see whether
these results are repeated with other stories of varying interest levels.
Likewise, subjects  ability to predict what the sidebars were about based on the
link information did not affect whether they said they wanted to read those
stories.
In all groups, across all three stories, more subjects said they would read the
story after they saw it than said they wanted to read the story after just
seeing the initial summary. This was surprising since it would seem more likely
for readers to be hooked by a good headline or summary, then get to the full
story and decide they had seen enough and didn t need to read about all the
details.  As with the main stories, more subjects said they would read the
sidebars after they actually saw them than said they wanted to read them based
on the information in the links.
The amount of information in the summaries and links did affect whether subjects
reported feeling informed enough without reading the full stories. There were
significant differences between groups for two of the three main stories, and
the same trend was evident for the third (see Tables 8-9). As expected, subjects
were more likely to agree they felt informed in the abstract condition than in
the headline-alone or head-plus-deck conditions.
[ INSERT TABLES 8 AND 9 HERE ]
Length also made a difference in whether subjects reported being informed enough
by the links to the sidebars (see Tables 10-11). The difference was significant
for two of the stories, with the abstract links about twice as likely to elicit
agreement as the headline or words-in-text links.  Still, only about 50 percent
of the subjects in the abstract conditions reported being informed enough by the
links.  Oddly, one story did not follow this trend, with subjects in the
headline condition most likely to say they were informed enough.
 [ INSERT TABLES 10-11 HERE ]
Interestingly, even though the abstract groups were more likely to say they felt
informed, they still reported wanting to read the full stories and sidebars at
the same level as subjects in other groups. This was a surprise since it would
seem logical based on knowledge of current readership habits, that many subjects
would not want to wade through the full story once they felt they knew the
basics.  Once more, it could be that the overall appeal of the story topics made
subjects want to continue even after they thought they had the basic facts.
Again, further research might clarify the reasons for this finding.
 
2   How does link presentation affect subjects  ability to find the link?
Overall, 90 to 100 percent of subjects were able to locate the links to the
various sidebars. There were no significant differences by type in the number of
subjects who found the link to the sidebar for each story.  For one story, there
were significant differences in the number of subjects who reported difficulty
in finding the link.  For this story only, it appears subjects had trouble
recognizing the words-in-text link. One possible explanation for this is that
the subjects in this group saw the abstract and headline link options first and
may have had trouble adjusting to a different presentation because of a learning
effect.
An important difference in demographics also surfaced in this area.
Specifically, in all three stories, whites were significantly more likely to
find the links to the sidebars. And in two cases, blacks were significantly more
likely to say finding the link was difficult. The third case approached
significance. These differences showed up even though correlations of overall
computer and online use with questions about finding the links revealed no
significant differences.  However, an analysis of the computer use figures by
race shows that whites are significantly more likely than blacks to use a
computer or to go online from home.  Although the numbers are much closer in
terms of computer and online use at school or work, the differences in finding
links may be attributable to different levels of experience with computers.
Regardless of the reason, this is an important finding that underlines the need
of newspapers and all business and industry to consider the best way to present
digital information to the computer literate and illiterate, and to avoid the
assumption that almost everyone is familiar with and has the technology to
access digital information.
Overall, only 10 to 20 percent of subjects found any of the links distracting.
The differences among link types were significant only once, with the abstract
link condition causing the most distraction in that instance. It is interesting,
however, that the type of link subjects saw first seems to have affected their
judgments of the links they saw later. For instance, those who saw the abstract
link first were more likely to say the words-in-text link was distracting when
they did see it than those who saw the words-in-text link first. Again this may
be the result of a learning effect.
The reasons subjects gave for being distracted by the various types of links
help explain these results. The most common cause for distraction with the
words-in-text and headline links was placement, while placement and length were
both a problem for the abstract link.  In fact, given a chance to comment about
link placement, several said they liked having the additional information in the
abstract links, but suggested putting the link box at the end of the main story
so it would not be so likely to interrupt or distract them. The problem with
this of, course, is that readers who chose not to read entirely through the main
story would almost certainly overlook the link. Second, moving the link box to
the end of the story would take away page designers  ability to use it as a
graphic element to help break up text and could leave readers with a very gray
and unattractive page if there were no other pictures or graphics to accompany a
story.
 
3   Do subjects think they would like reading news presented in layers and
links?
As expected, almost all subjects (93 percent) thought they would like seeing a
summary without having to see the full text of a story.
Subjects also said they liked the idea of routinely having the chance to get
more information on a story if they wanted it. In fact, 99 percent   all but one
person   said they would like the option of getting more depth on a story if
they wanted it. Of course these numbers are no surprise since the extra
information would not be a burden for anyone who was not interested in it. At
the same time, however, a bit of caution may be in order when considering these
results since the numbers may also simply reflect a normative or socially
desirable response from subjects.
 
4   Which summary and link length and which link presentation do subjects say
they prefer?
 
Presentation of link to sidebar
Overall, subjects said they preferred the link to the sidebar to be a separate
element (64.6 percent) as opposed to words highlighted in the text (35.4
percent).  The separate box  gives room for the reader to find out more,  one
subject said. It also allows the reader  to get a lot more information quickly,
another said.
Another common reason for this choice was that subjects found the words-in-text
link distracting. With the separate element, on the other hand, they said they
could just skip past it, then come back to it when they felt ready instead of
being interrupted by a link in the middle of the main story.
 
Length of link to sidebar
More than twice as many subjects preferred the link length to be a headline and
abstract (54.3 percent) as preferred the other two options (headline, 22.2
percent; a few words, 23.5  percent). The most common reason for preferring the
abstract was because the extra information allowed subjects to make a better
decision about whether they would want to follow the link. With an abstract,
 you know what to expect,  one said. A few words or a headline, on the other
hand,  is simply not enough to let you grasp the content  of the story.  It s
short, but there s still some meat to it,  another said.
However, nearly all those who chose the headline link as preferable said they
thought the abstract condition was too long. The headline condition was better,
they said, because it gave just enough information about the topic for them to
know if they wanted to read it.   It gives an idea without telling the whole
story; it s detailed, but not too much,  one said. Further, the  context of the
main story should make it clear what the sidebar is going to be about, some
said.
Similarly, those who preferred the words-in-text link thought a few words were
enough to let them make a decision.  A few words are enough   if they are the
right words,  one said.
 
Length of initial summary
        Overall, subjects preferred the summary length to be a headline and deck (56.8
percent), followed by head and abstract (32.1 percent) and headline only (11.1
percent). The  initial summary length was the only one of the overall
preferences statistically significant by type (see Table 12). Interestingly, the
abstract group was almost evenly split between the abstract option and the
head-plus-deck alternative. The head-only group and the head-plus-deck group
both chose the head-plus-deck option as preferable.
[ INSERT TABLE 12 ABOUT HERE ]
Most of those who chose this option said they thought the headline alone was too
short and vague, and the abstract too much to read while the head-plus-deck
option  gave all the basic information but wasn t too long.
Many also said this length was best at capturing their interest and motivating
them to continue reading, whereas the abstract would probably keep them from
reading the full story. The abstract  defeats the purpose  of following the
link, several said, while the head-plus-deck  leaves something unknown, gives
you a reason to click.   Others said the abstract was so long, they stopped
reading after the first sentence.
Those who chose the abstract condition as preferable, however, were just as
convinced that the other two options did not provide them with enough
information. Like those who preferred the abstract links to the sidebars, many
said several sentences were necessary to let them find out if it really was
something they were interested in so they could judge better whether to follow
through to the full text of the story. With an abstract,  you re not misled in
any way,  one said.
Some also said the longer abstracts made them more interested in reading the
story, and again, most realized that it allowed them to be better informed about
the story subject if they didn t have time or simply did not care to read any
farther.
Of the minority who chose the headline option, many said the headline was the
only option that left them wanting to know more.   If you re interested, you
want to go right to it instead of reading the same information twice,  one said.
 The story seems repetitive if you ve already read a summary,  another said.
 
 
 
CONCLUSIONS
Although the results are not completely clear-cut, this study does provide
important insight into how initial summary and link presentation might affect an
electronic publication s usefulness and appeal.
It is obvious that the initial summaries and the links to the sidebars can be
too long as well as too short. And, of course, both may be  dangerous in terms
of affecting potential readers  judgment of a publication s usefulness and
appeal. However, it is difficult to say exactly where the boundaries ought to
be.
Even though a clear majority of subjects chose the headline-plus-deck option as
the preferred summary length, other statistics from the study suggest it may not
be the most useful length.
Specifically, many subjects said the longer abstracts gave too much information
and therefore would keep them from wanting to read the stories. But the
statistics do not support this argument. In fact, the numbers show that the
varying summary lengths had no effect on whether subjects wanted to read the
stories.
Second, it is important to remember there were significant differences in how
length affected subjects  reports of whether they had enough information to make
a decision about reading the story. So even while 70-plus percent of subjects
said they could make that call even with just a headline, it is evident that the
more information they had, the more comfortably they could make that decision.
Finally, the differences in how informed subjects were from the various summary
lengths also needs to be considered. Again, the abstracts are undoubtedly more
effective in this area. This is important since research has shown that readers
read only about 1/5 of all news items in traditional newspapers (Weaver and
Mauro, 1978). With an electronic publication s different format, however, they
may be enticed to read at least the initial summaries to stories. It would seem
sensible to try to inform them as much as possible with the summaries since they
may not read any further.  This study also shows that if a reader is interested,
he or she will proceed to the story even if the summary communicates the major
points.
When considering these benefits of the abstracts, along with the statistics that
show the longer summaries did not prevent subjects from wanting to read the
stories, the abstracts appear more attractive than the overall preferences would
reveal.
However, should further research show that longer summaries do indeed prevent
readers from wanting to go further into stories, the decision will become more
complicated. In that case, there would be a need to weigh the use of shorter
summaries to motivate people to read further against the use of longer summaries
that would inform readers as much as possible should they choose not to follow
the story any farther. The goal would be to find the best balance without
alienating any readers on either side.
For the same reasons, it would seem that the abstract link condition also might
be the most appealing and useful overall. It is interesting that even though the
abstract summaries and the abstract links were the same length, the abstract
links were accepted to a much greater degree than the summaries.
One possible reason for the greater acceptance of the longer links   and one
that supports the idea that longer summaries are not necessarily a disadvantage
  is presentation. It simply may be that the reason so many subjects found the
abstract summaries so long and unappealing was the format in which they were
presented. In particular, a bigger point size and longer line length probably
made the abstracts seem unattractive and appear much longer than they were.
Meanwhile, the abstract links were presented in a small point size and narrow
font to save space, and they were presented in a shaded box at the top of a
column of type. Besides making the abstract links more visually attractive than
the abstract summaries, these features also made the links appear shorter. It
will be important in future studies to see if readers are more accepting of
initial summary abstracts with a different type of presentation. If that is the
case, the argument for the use of abstracts will be further strengthened.
Of course, another possibility for why the abstract links were preferable is
that the headline and the words conditions were simply too short, leaving the
abstracts as the most attractive option. Again there are several hints that this
could have been the case. First, it appears the words link may have been
dismissed fairly quickly by many subjects since both the research statistics and
subjects  comments show that this type of link can be too vague and misleading.
That would have left subjects with just two choices   the headline or abstract.
However, it seems evident from the summary and link preferences that a headline
alone is not enough  information for most people. So the abstract may have been
the only practical choice since there was no intermediate option. It would be
interesting to see if readers would prefer a headline and deck for the link if
that option were made available.
With all these considerations in mind, it appears that of the various lengths
tested in this research, abstracts may be the most useful and appealing for both
the initial summaries and the links to sidebars. Of course, ideal length is
going to vary according to the individual preferences of each reader. But
overall, subjects  stated preferences combine with statistics regarding power to
inform, predictability and decision-making issues to favor the longer abstracts.
 
Future Research
Because research in digital news publication is so new, future research could,
and undoubtedly will, go in many directions in the coming months and years. This
project alone brings up many areas that deserve additional study.
First, it will be important to complete similar research with stories of varying
interest levels to see if these results are repeated, especially those relating
to how a link s length, predictive power and ability to inform affect a reader s
motivation to follow the link and read the story.
Another area of research will be to see how story interest and relevance affect
the amount of information a reader needs before he feels informed enough about
the topic, as well as the amount of information the reader wants before he feels
comfortable making a decision about whether to read the full story.
The impact of presentation on readers  perception of summary length also
deserves attention. Specifically, further research needs done to see if readers
would be more accepting of abstracts that are more visually appealing.
Future studies that focus on length also need to include an intermediate link
option such as a headline plus deck to see if that length is preferred over the
abstract length. Researchers studying length might also consider how summary and
link length affect reader comprehension of stories, an area that was not touched
on in this study.
In addition, this study presented just a few link display possibilities. Many
other design formats should be considered before a final decision on
effectiveness can be made. Future study should also include the role of graphics
in motivating readers to access stories.
And, of course, any study that can record actual behavior instead of how people
say they would react to certain links and formats would be of great value. In
particular, it would be useful to know how many people would actually follow the
summaries and links to the full stories.
 
 
 
 
 
APPENDIX A
Samples of page layouts