INFORMATION INTERACTIONS - Human-Computer Interaction and Management Information Systems: Foundations - page 7

Even though defining commonsense terms is something of a thankless task, “information” has
been subject to several attempts at definition over the years. Most noticeably, within the broad
information studies realm (e.g., Williams and Carbo, 1997), a strong theme has been the need to
draw firm lines between raw data and information. Accordingly, data is seen as the base material
that conveys little in and of itself unless ordered, processed, or otherwise made useful to its recip-
ients. Once given meaningful form, data becomes information. Given the strong ties between psy-
chology and both MIS and HCI, it is odd that this distinction has little or no currency in cognitive
psychology, where data and information are typically treated synonymously.
Conceptualizing information does not end with its demarcation from data. Information can itself
be distinguished from knowledge, which in turn can be distinguished from wisdom, and more than
a few topics have been written on these terms and their putative relationships. In their classic text
Working Knowledge , Davenport and Prusak (1997) articulate a view of information as “data that
makes a difference” to the receiver, thereby highlighting the problems associated with extracting
the important elements of a data set or the need to avoid being overwhelmed by too much data.
Information, thus construed, has relevance or purpose, and is extracted or transformed from data
by what these authors refer to as “the five C's”: contextualization, categorization, calculation,
correction, or condensation.
One does not have to accept the “five C's” approach to recognize that technology clearly plays
a role in adding value to data. Thus, one can envisage an MIS- or HCI-style analysis of informa-
tion tools or systems built on this view. Visualization tools can categorize data to be viewed in
new ways, enhancing the emergence of information from complex data sets. Calculation is the
backbone of computing, the defining triumph of what Landauer (1995) described as the first stage
of computing, where technologies performed information tasks at a level impossible for humans
to match. But technological support for Davenport and Prusak's methods is not the full story here,
as these authors themselves acknowledge. More crucially, distinctions between data, information,
and knowledge bring a significant human factor into the discussion, implying that data becomes
information only when a user (a knowing human) makes sense of it and when it makes a difference
for him or her. Making sense is itself multiply determined and context-dependent (Dervin, 1989),
but the crucial distinction rests on the extraction of meaning or the imposition of order through
human processing of some kind.
It is precisely because of this human factor that we cannot simply equate information with
objects such as topics or DVDs, no matter how appealing such a usage is at first. Buckland (1991)
wrote a much-cited article on this in which he referred to the objectification of “information-
as-thing.” My reading of his argument is not that information is an object, but that once informa-
tion is viewed as the potential for intelligent reading of data, there are few objects in our world
that cannot serve as information, under some set of conditions.
Equating information to data that makes a difference for certain people is certainly a start, but
is it a sufficiently strong basis for a field of study? There are alternatives, most noticeably the
mathematical conceptualization of information that defines information in terms of the confi-
dence levels in receivers of the conditions that exist at the source. Put another way, this defines
information as the reduction of receiver uncertainty (e.g., Gharhamani, 2003), which is clearly a
difference that is meaningful. This view of information has had significant impact on the design
of computing systems, and both informed and drew on classic information processing theories of
human cognition. However, I cannot see a way forward for our present concerns by utilizing this
approach since it scales weakly to real-world tasks involving humans and information systems.
Dillon (2004) proposed a view of information as product with purposive process, arguing that
information does not reside in objects or entities, but emerges from the engagement of such objects