An investigation of factors which differentiate between university lecturers in relation to publication output is reported. The study drew on data from lecturers working full-time at two large Australian universities. Measures of research publication output were used to select two groups of lecturers (N[subscript 1] = 119; N[subscript 2] = 119); that is, those without or with very few publications and those with impressive publication records. Because of the differing metrics of the selected factors, both discriminant and logistic regression analysis were employed to test the discriminative capacity of the various factors and models incorporating some of the factors. A range of factors, including qualifications held, proportion of work time devoted to research tasks and the confidence assigned by lecturers to the execution of certain research tasks, were found to differentiate between the two groups. A logistic regression model incorporating these factors showed good fit of the data, as indicated by the omnibus test result (p less than 0.001), a Cox and Snell R[superscript 2] value of 0.57, and 86.1 percent of the grouped cases being correctly classified. The implications of these results are examined from the perspectives of university managers, who may wish to create an environment that fosters greater research and publication, and those who engage in these activities. (Contains 5 tables.)