Why I’m disappointed in Ebsco’s new visual search, part 1 of ?

So the other day I got an email from a colleague pointing out that Ebsco had changed its visual search interface.
I haven’t seen much discussion about this, and in the past year or so I
haven’t found many people who routinely use the visual search in
library instruction (or if they do, they don’t really talk about it).
This is a little surprising to me, because most of my colleagues at OSU
use it regularly, and find that students like it.  I’m really
disappointed with the new interface – I think that everything my
students liked about the old interface has been lost, and I’m really
not sure how to introduce the new one to them. 

This probably won’t be the first post I make on this topic, and I
expect to get pretty far afield from what we can expect from Ebsco
visual search, so bear with me…

In the last year or so the visual search has become one of those
must-see things to show undergrads — the standard trio for our
beginning composition class is citation generation, Lexis-Nexis and
Ebsco’s Visual Search (they’ve already had a little experience with
standard Ebsco searching before the class starts). 

Last term, I kept an eye on the comp students when they moved to the
hands-on computers for the second half of the class to see where they
decided to start looking for resources.  Admittedly, my methods weren’t
that rigorous, but I’d say that about a third of the students in any
given class would head to Lexis-Nexis first, drawn to the new thing
they hadn’t tried before.  One or two students would go to the library
catalog.  The rest of the class went to Ebsco, and the vast majority –
maybe four out of five of them – went straight to the visual search.

This is what the visual search looked like then.
Results were grouped by subject (the circles) and students could click
on a big circle, to find further subsets or refinements of their
topics.  Individual articles were represented as squares.

(the link above shows the 2 new result displays in the new interface)

Now like Caleb here, I don’t really like visual search for myself.
When I’m searching, I like to feel like I haven’t missed anything and
when the articles were sorted for me by a computer I don’t really
trust, I never felt like I am getting everything.  A long list of
citations I can check off as I go works a lot better for me.  But –
also like Caleb, I had decided a while ago that just because it didn’t
work for me, didn’t mean that it didn’t work for them.

One thing that I did think the students were getting out of the old
search was the ability to browse related articles on a topic easily –
they could zoom deeper into a topic, and if things weren’t working out
for them, they could back off and go in another direction.  The Playful Librarian
points out that the visual search is really mis-named; it’s not so much
a visual search but a visual way to browse and refine results.  And I
don’t disagree with that – in fact, I think that very thing is what
made it so useful for my students.

When I look at a long list of text results, I’m usually conversant
enough with the topic or the discipline I’m exploring that I can
mentally sort things into relevant-not relevant-potentially
relevant-not at all relevant but still interesting types of
categories.  When the students in my classes are starting off their
research, especially in a course like beginning composition where their
topics aren’t grounded in the rest of the class material, they can’t
necessarily do that.  When we start them off exploring on their topics
in that course we send them to Wikipedia so they can learn a little bit
about the topic – at least enough to be able to start placing their own
ideas within the larger discourse. 

The visual search interface gave them another way to do that
exploring, right within the search tool most of them used to get their
articles.  Even though we could complain about the indexing in the
databases, and criticize the topic groupings that Ebsco would come up
with, for a lot of our students it was a way to start.   They didn’t
need a list of 10,000, or 5,000 or even 1,000 results.  They needed
help making sense of those results.  They needed a way to say "I want
to look at this subset of my results, and that subset — not so much."

For a course like first year composition, it’s not like they need to
be comprehensive on their topics.  They have a requirement to use eight
sources, and of those eight, some must be books and websites.  So
finding seven or eight relevant articles of sufficient quality on a
topic is more than enough research to get them started writing these
papers.  And a browsing interface gives them a way to start thinking
about the results in an analytical way – it helps them start to think
about classifying their results, thinking about the different
dimensions of their topics.  So yeah, it is a visual browsing/ refining
tool – and that’s just exactly what a lot of them need.

Increasingly, I’m thinking that we focus too much on the search in
library instruction, and in libraries more generally as we develop
tools to find library resources.  A couple of years ago at the Access
conference in Ottawa, when some of the presentations focused on
different ways of visualizing information, I started thinking how many
of the students I encounter would benefit from rich, powerful browsing
interfaces.  I hear frequently, from colleagues around the library
world, that our students don’t care about the metadata and that normal
people just want the fastest access they can get to the article they
need.  But I don’t think that’s entirely true.  I mean, I do think they
do want the fastest access possible to the information they need – who
doesn’t want that?   But they do care about the metadata.  They just
want it used in a way that gives them results they can manipulate,
browse, move through in a way that makes thinking about them fun,
productive, useful — that can actually help them make sense of the
results they find.

Most undergraduates go into their research process not only knowing
that they need to find information, but knowing that they need to find
specific types of information.  They are constantly given arbitrary
requirements like "five newspaper articles" or "a peer-reviewed journal
article" as part of the their assignments.  Those students care about
more than just finding an article – they want to find the right kind of
article.  They care a lot about that.  And they’re not always sure how
to do it — while these requirements might seem simple, they require
the student to know some pretty sophisticated things about how
information is produced, indexed, and retrieved. 

When we focus on the search at the expense of that context, then I’m
not sure how much we help them.  When I sit at the reference desk, I
often feel like the students that ask for help are already frustrated
because Google, or other powerful search tools, have failed them.  They
either can’t find what they want, or they can’t make sense of what they
find. When the trouble is that they’re not finding anything – the
trouble tends to be with the initial query — with choosing their
keywords — which again comes back to their thinking about their
topics.  Just as often, they have found results, they’re just not sure
how to classify or prioritize those results; they don’t know what
they’ve found.  I think we can spend hours talking to these about how
to drive the different interfaces, or about how to connect or
manipulate keywords, and we’re not going to help them at all.

A lot of this really came together for me when I read Gary
Marchionini’s 2006 article Exploratory search: From finding to
understanding.*  Marchionini argues that there are three different
kinds of search activities: looking stuff up; learning; and
investigating.  Of these, only the first is improved by focusing
attention to the query —

In general,
lookup tasks are suited to analytical search strategies that
begin with carefully specified queries and yield precise results
with minimal need for result set examination and item comparison.

Learning
searches won’t be improved by better queries, because by definition
they require the user to go back and forth in an iterative process,
changing the query as they learn more.  Similarly, they won’t be
supported well by a search tool that focuses on bringing back the
perfect result, fast and on the first try.  When searching to learn, as
scholars must do, the first try just isn’t ever going to be the only
try —

Much of the search time in learning search
tasks is devoted to examining and comparing results and
reformulating queries to discover the boundaries of meaning for
key concepts. Learning search tasks are best suited to
combinations of browsing and analytical strategies, with lookup
searches embedded to get one into the correct neighborhood for
exploratory browsing.

More
and more, I think undergraduate research instruction needs to focus on
analyzing and contextualizing – the higher-order thinking about their
topics — which comes down to what they know about them and if they
have the vocabulary to articulate what they know. 

And more and more I think that the tools themselves also need to
help them do that.  And that’s where I think visual displays, and
browseable results, can be powerfully useful for new scholars.  Lorcan
Dempsey referred a while ago to the concept of glanceability,
which is a term usually used to describe interfaces or displays where
users can pick up the information they need at a glance.  Dempsey used
a slightly different definition, which I find extremely useful in
describing what I think visual browseability can do for undergraduate
users: 

Glanceability is about enabling users to understand information with low cognitive effort.

The first time I read that definition, I thought of the Presidential speeches tag cloud —

As a former historian, I love the way this simple interface takes
familiar texts, and by redisplaying them sparks new ideas and
connections, as well as a sense of change over time, in the user’s
brain.  It’s quick, it definitely takes low cognitive effort, and it’s
even fun if you’re a politics geek like me — but it’s not cheap nor is
it superficial.  One couldn’t learn everything worth knowing about
these speeches with this tool, but it can spark actual higher order
thinking in a low-effort way.

The Neoformix blog
presents a lot of different ways to manipulate and present texts,
visually.  One of my favorites is the Document Arc diagram, which shows
the connections between authors and arguments within a paper – here’s
the document arc of a paper I co-wrote a few years ago on using blogs
in the classroom.

One of my favorite things about this is that it shows at a glance
where the key in-text citations repeat – showing that the argument in
the paper is connected to the larger discourse.

The American Studies tagline project
takes the Presidential speeches idea in a slightly different direction
— focusing on the discourse within a discipline.  Taking a collection
of standard "significant documents" from the American Studies
discipline, this tagline presents the texts visually, letting the user
see at a glance the evolution of discourse within the discipline.  It’s
not hard to see how a student could use a tool like this to locate
their own ideas within a discipline or discourse.

But where I really think the power of visualizations as a way to
help students understand scholarship, and discourse, can be seen is in
tools that show how the different voices within a discourse are
connected.  It’s probably not surprising that Citegraph,
a legal research tool, illustrates this well, given how overtly
connected legal texts are.  This shows how the connections between
ideas and citations within related texts can be displayed and
manipulated visually.  And I’ve long been intrigued at the thought of
using Many Eyes‘ social network visualization tool to visually examine the connections and networks between scholars.

For too many examples of scholarly visualizations to capture here, look at these possibilities using CiteWiz.  These are the figures and tables from a 2007 article in Information Visualization,
and they include timelines, concept maps, tag clusters and more.  The
whole article (cited below)** is worth reading – but the figures alone
are plenty to pique my interest.

So that’s part 1 of why I’m disappointed in the new Ebsco visual
search.  It may be more of a search than a browse, though I’m not sure
about that, but that’s a loss I think for my students.  And I think
that browse, especially visual, exploratory browsing, is something to
think about further.

______________________

*Marchionini, Gary (2006). Exploratory search: From finding to understanding. Communications of the ACM, 49: 4, 41-46.

**Elmqvist, Niklas & Tsigas, Philippas (2007).  CiteWiz: a tool for the visualization of scientific citations networks.  Information Visualization, 6, 215–232