This study investigates the use of polarimetric variables for estimating hail size. Archived data from the KOUN polarimetric radar in Norman, Oklahoma are obtained for three hail-producing events: 24 May 2004, 29 May 2004, and 2 June 2004. A total of 45 hail reports are used for analysis, including hail sizes ranging from 0.75" to 4.25". Horizontal and vertical structures of reflectivity (ZH), differential reflectivity (ZDR), and correlation coefficient (ρHV) associated with these reports are examined using Interactive Data Language (IDL) programs. Comparison of these images allow for the hail to be categorized into two groups based on ZDR and ρHV signatures: category 1: hail less than 1.75" in diameter and hail 1.75" or higher in diameter. RHI images reveal differences in the vertical structures of these categories; extended columns of low ZDR and ρHV are observed for the larger hail. Also, vertical profiles, produced for each report, show a more substantial decrease in ρHV below the melting layer for hail 1.75" or larger. Box-and-whisker plots and discriminant analysis are then used to determine the ability of ZDR and ρHV to distinguish between the two categories. Although ZDR appeared promising for estimating hail size, the discriminant analysis revealed that ρHV is the best variable to discriminate between the categories.