•  
  •  
 

Abstract

Researchers from most academic disciplines use statistical tools to analyze experimental and observational data. Unfortunately there is evidence that a disturbing number of peer-reviewed articles that use statistical methods use them incorrectly. Prior research has found this to be a problem in journals dedicated to a variety of fields including medicine, psychology, business and education. Currently, the identification of common statistical errors in journal articles involves a manual and thus very labor-intensive process. As a result, only a small number of articles in a few disciplines from a limited set of journals have been assessed. In this paper we propose developing text mining models to partially automate the identification of common statistical errors in journal articles. We use the term literature mining to refer to the application of text mining tools to academic journals. Once developed, such models can be used to audit the existing journal-based body of knowledge and rapidly compile lists of articles that likely contain a particular statistical error. Looking forward, the models can also be used by editors and reviewers to screen papers prior to publication. The identification of new research opportunities is also discussed.

Included in

Business Commons

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.