Analysis of Agreement Between Two Long Ranked Lists
Srinath Sampath · OhioLink ETD Center (Ohio Library and Information Network) · 2013
An alternative approach to the problem recently posed by Hall and Schimek (2012) is proposed: determining at what point the agreement between two rankings of a long list of objects degenerates into noise.To this end the method of estimation in Fligner and Verducci (1988)'s multistage model for rankings is modified from maximum likelihood of conditional agreement over a sample of rankings to a locally smoothed estimator of stage-wise agreement.An extension to the case of overlapping but different sets of objects in the two lists is also provided.Simulations show that this approach performs very well under several conditions.The technique is next applied as a stopping rule to augment the tau-path algorithm, developed by Yu, Verducci and Blower (2011), in an analysis of associations between gene expression and compound potency in cancer data, and the detection of the endpoint of agreement.The methodology is also applied to a database of popular names for newborns in the United States and insights into trends as well as differences in naming conventions between the two sexes are uncovered.