AbstractThe present article describes an attempt to improve the prediction of dental attendance, results by using an additional number of variables and performing discriminant analyses, clone separately for men and women, It appears that regular and irregular dental attenders might be discriminated on the basis of mathematically combined variables and interactions. The (number of) variables selected by the analysis to differentiate regular from irregular male attenders differ(s) from those selected for female attenders. About 80% of all respondents can be classified correctly, but this is largely due to the assignment of actual regulars to classified regulars. The use of differing sets of prior probabilities affects the classification results, namely, either the classification of the regulars or that of the irregulars improves. In the discussion attention is given to the seemingly contradictory results of the Mann‐ Whitney tests per variable on the one hand and the discriminant analyses on the other. In this connection, the finding that ‘education’ plays a different role for men than for women is discussed. The validity of the variable ‘last visit’ is dubious. It is concluded that when differentiating regulars from irregulars, a division of the respondents according to sex makes sense because of the (number of) variables associated with then‐ regular dental attendance. Furthermore, it seems warranted to say that the Classification of the irregulars fails because he reversed scores on the variables with which the regulars are classified don't contain all information needed for the prediction of irregular dental attendance. Lastly, notwithstanding the use of more variables in the present study, the results are not better than those in the previous one, in which just three factor