Paper Reading #20: The aligned rank transform for nonparametric factorial analyses using only anova procedures

• Title: The aligned rank transform for nonparametric factorial analyses using only anova procedures
• Reference Information:
• Jacob O. Wobbrock, Leah Findlater, Darren Gergle, and James J. Higgins. 2011. The aligned rank transform for nonparametric factorial analyses using only anova procedures.  In <em>Proceedings of the 2011 annual conference on Human factors in computing systems</em> (CHI '11). ACM, New York, NY, USA,  143-146. DOI=10.1145/1978942.1978963 http://doi.acm.org/10.1145/1978942.1978963
• UIST 2010 New York, New York.
• Author Bios:
• Jacob Wobbrock is an Associate Professor in the Information School at the University of Washington.  Wobbrock directs the AIM Research Group which is part of the DUB Group.
• Leah Findlater is currently a professor at the University of Washington but will become an assistant professor at the University of Maryland in January of 2012.  Findlater has developed Personalized GUI's.
• Darren Gergle is an associate professor at the Northwestern University School of Communication.  Gergle is interested in improving understanding of the impact of technological mediation has on communication.
• James Higgins is a professor in the Department of Statistics at Kansas State Unversity. 
• Summary
• Hypothesis:
• The researchers hypothesized that modifying the Aligned Rank Transform to an arbitrary number of factors would be useful for researchers attempting to analyze data.
• Methods
•  The researchers developed the method for the expanded ART and then coded this in both a desktop tool (ARTool) as well as an online, Java-based verson (ARTWeb).  After creating these tools the researchers analyzed three sets of previously published data.  This analysis was to demonstrate its utility and relevance, as opposed to its correctness.
• Results
• Examining old data revealed interactions that had not been seen before.  For example, in a study by Findlater et al. the authors noted that there was a possible interaction that was unexaminable by the Friedman test.  When this data was run using the nonparametric ART method, nonsignificant main effects for Accuracy and Interface were revealed, as well as a significant interaction.
• Contents
• This paper presents a nonparametric ART method, as well as two programs to support the calculation of data using this method.  The system has limitations, such as possibly reducing skew, which may be undesirable.  But, as demonstrated during their tests, the method can help reveal interactions that cannot be discovered through other analyses.
• Discussion
• The researchers were certainly able to prove their hypothesis, as seen by their test cases.  It will be interesting to see whether or not this tool is used for research analysis in the future.  The chart at the beginning of the research paper was somewhat intimidating, listing quite a few already commonly used techniques.  I have a feeling that statisticians will use this so that more intereactions can be observed.  As the saying goes, knowledge is power, so the more the researchers are able to understand the more they can build off of.



Picture Source: "The aligned rank transform for nonparametric factorial analyses using only anova procedures"