Relationship between smoking and calculus deposition.

An extensive study of the relationship between smoking and calculus deposition in nearly 6,000 men was undertaken by Pindborg,1 in 1949. Chilton,2 in 1967, used these data to illustrate the familiar test for association in a contingency table but noted that the analysis was limited because, although it established the existence of a significant association between smoking and calculus deposition, it did not answer the important question of which particular groups (or combinations of groups) differed significantly. He concluded that "what is needed here are analogues of the tests* suggested by Tukey3 and by Scheff6,"4 but apparently he failed to realize that such analogues are, in fact, available and may easily be used to extract additional valuable information from contingency table analyses. Two procedures that accomplish this extended analysis and illustrate the power of the method in the analysis of Pindborg's' data are outlined in this study.