Drexel University Home Pagewww.drexel.edu DREXEL UNIVERSITY LIBRARIES HOMEPAGE >>

iDEA: Drexel E-repository and Archives > Drexel Academic Community > College of Information Science and Technology > Research Day Posters (IST) > Overcoming groupthink in distributed heterogeneous data analysis

Please use this identifier to cite or link to this item: http://hdl.handle.net/1860/3833

Title: Overcoming groupthink in distributed heterogeneous data analysis
Authors: Pellegrino, Don
Chen, Chaomei
Keywords: Decision making;Information retrieval
Issue Date: 23-Apr-2009
Series/Report no.: IST Research Day 2009 posters
Abstract: The aim of this project is to decrease researchers' susceptibility to groupthink and associated deficiencies in the decision making and discovery processes. Many fields have ongoing initiatives to increases digitization of resources which are already growing at an exponential rate. Examples include digital archives of journal articles, e-science, open-notebook science and national intelligence data. Many research initiatives attempt to address the issues of data overload although recent research in citation analysis has shown that increasing the online availability of articles may accelerate the development of consensus viewpoints. This may have the effect of stifling novel or surprising views in favor of the popular view. It is hypothesized that information retrieval (IR) algorithms such as Google's PageRank, which weigh the results by popularity and recency factors, contribute to the acceleration of consensus building. As information retrieval results are often a significant part of the input to a decision making process they may be systematically introducing a bias into the overall discovery processes. It is further hypothesized that interfaces which stress simplicity over transparency such as keyword based search and list based result views compound this issue as users are generally unaware of the existence of such bias and are therefore unable to properly account for it. Measures of interest such as those that have been successfully applied in bioinformatics might be applied in IR to provide results that complement the consensus hits with meaningful alternative viewpoints. Information visualizations may be used to increase the transparency of results and expose more sophisticated dimensions of relevance. The research question is how such measures and visualizations can best be applied to the design of IR systems to help researchers overcome the limitations of groupthink.
URI: http://hdl.handle.net/1860/3833
Appears in Collections:Research Day Posters (IST)

Files in This Item:

File Description SizeFormat
Pellegrino_2009.pdf20.31 MBAdobe PDFView/Open
View Statistics

Items in iDEA are protected by copyright, with all rights reserved, unless otherwise indicated.


Valid XHTML 1.0! iDEA Software Copyright © 2002-2010  Duraspace - Feedback