Aims: To analyse the relative importance of different biological and technical factors on the analytical sensitivity of conventional culture methods for detection of Salmonella Dublin in cattle faeces. Methods and Results: Faeces samples collected from six adult bovines from different salmonella-negative herds were split into subpools and spiked with three strains of S. Dublin at a concentration level of c. 10 CFU g(-1) faeces. Each of the 18 strain-pools was divided into two sets of triplicates of four volumes of faecal matter (1, 5, 10 and 25 g). The two sets were pre-enriched with and without novobiocin, followed by combinations of culture media (three types) and selective media (two types). The sensitivity of each combination and sources of variation in detection were determined by a generalized linear mixed model using a split-plot design. Conclusions: Biological factors, such as faecal origin and S. Dublin strain influenced the sensitivity more than technical factors. Overall, the modified semisolid Rappaport Vassiliadis (MSRV)-culture medium had the most reliable detection capability, whereas detection with selenite cystine broth and Mueller Kauffman tetrathionate broth combinations varied more in sensitivity and rarely reached the same level of detection as MSRV in this experiment. Significance and Impact of the Study: The study showed that for MSRV-culture medium and xylose lysine decarboxylase agar as the indicative medium, the sensitivity of the faecal culture method may be improved by focusing on the strain variations and the ecology of the faecal sample. Detailed investigation of the faecal flora (pathogens and normal flora) and the interaction with chemical factors may result in developing an improved method for detection of S. Dublin.