Adjustment to the aggregate association index to minimize the impact of large samples

Title: Adjustment to the aggregate association index to minimize the impact of large samples
Creator: Beh, Eric J.; Cheema, Salman A.; Tran, Duy; Hudson, Irene L.
Relation: Advances in Latent Variables: Methods, Models and Applications p. 241-251
Relation: Studies in Theoretical and Applied Statistics
Publisher Link: http://dx.doi.org/10.1007/10104_2014_24
Publisher: Springer
Resource Type: book chapter
Date: 2015
Description: The past few decades have seen a great deal of attention given to the development of techniques to analysis the association between aggregated categorical data. One of the most recent additions to this analysis has been the development of the aggregate association index (AAI). One feature of the AAI is that its magnitude is affected by the sample size; as the sample size increases so too does the AAI, even when the marginal proportions remain unchanged. In this article, we propose adjustments to the AAI to overcome the effect of increasing sample size. The Adjusted AAI is shown to be more stable than the original AAI in response to any increase in the sample size. Fisher's criminal twin data (Fisher, J. R. Stat. Assoc. Ser. A 98, 39-82, 1935) is used to demonstrate the adjustments.
Subject: aggregate association index; 2 x 2 tables; Pearson's chi-squared statistics; large samples
Identifier: http://hdl.handle.net/1959.13/1324592
Identifier: uon:25071
Identifier: ISBN:9783319029665
Language: eng

		Thumbnail	File	Description	Size	Format