Rejoinder: Testing for Independence in a Two-Way Table: New Interpretations of the Chi-Square Statistic