Information Technology Reference
In-Depth Information
p-message and send it to a provenance system (outside of Transana), describing this
coding event in some detail. Later, the same or a person different from the coder
would retrieve a bunch of coded events from the Transana database and aggregate
them into various counts, for instance using a spreadsheet program or a statistical
package such as SPSS. The spreadsheet program, if provenance enabled, would cre-
ate a p-assertion describing this operation. Finally, a word processor would be used
to write a research report, using the frequency figures so support a specific claim
about a finding. That part of the report would be linked back to the frequency table,
and the word processor would issue a p-assertion to the provenance store describing
this link. The graph shows a similar chain for the case of processing answers to a
questionnaire or interview.
Having documented such chains of events in a provenance store, anybody inter-
ested in finding out how claims in the report are related to data could use the
provenance system query interface to do so. To the extent that the outputs of the var-
ious steps are stored in some form digitally and are accessible through the Internet,
the intermediate outputs can be displayed. To the extent that the applications are
available as services, the process yielding the evidence chain can even be replayed.
A number of provenance systems have been developed in science e-research
projects, for instance,
CombeChem
(Taylor et al., 2006) and
MyGrid
(Stevens,
Robinson, & Goble, 2003). These tools track the experiment and analysis process
and automatically create provenance record as well as allow researchers to anno-
tate key research steps using controlled vocabulary or narrative descriptions. While
in the social sciences the automatic tracking of data-producing processes is not as
easy as in the scientist's laboratory, provenance approaches have been developed
there as well, for instance, in the
PolicyGrid
project (Philip, Chorley, Farrington, &
Edwards, 2007a).
In learning science research the provenance architectures must enable the cap-
ture of not only how data were created, but also the context in which data were
originated. For design-based research, for instance, two key aspects of the prove-
nance are important: (a) pedagogical or learning design provenance, related to the
design of the pedagogical intervention; and (b) methodological or study design
provenance, related to the accompanying study, data acquisition (observations,
tests, etc.), and analysis. While structures and ontologies for representing sci-
ence experiments can be adapted for methodology provenance (e.g., Taylor et al.,
2006), new structures and ontologies need to be created to represent pedagogical
designs.
Conclusions
In this chapter we introduced some existing and some emerging technologies and
discussed the potential to enhance learning research. To repeat: technologies are
evocative and open for many interpretations. How technology-enhanced research
will be taken up, what kinds of research practices they will support, and for what
kinds of tasks they will be utilized for is by large dependent on their use: Neither can
