a b s t r a c tEscherichia coli O:157:H7 adhesion and potential to form biofilm on three different surfaces commonly used in the food industry was evaluated using probabilistic models; the surfaces tested were stainless steel 304 (SS304), poly(vinyl chloride) film covered with thick cloth (PVC1) and poly(vinyl chloride) film covered with thin cloth (PVC2). Using a Central Composite Rotational Design (CCRD), the effect of contact time (0 h, 7 h, 24 h, 41 h and 48 h) and temperature (12 C, 17 C, 28 C, 39 C and 44 C) on the probability of achieving a particular adherent cell count (Log 10 CFU cm À2 ) was determined. By analyzing response surface plots and their corresponding contour plots and by determining quadratic equations for each surface, experimental values were shown to be significant in accordance with predicted values in all cases. The adjusted determination coefficient (R adj 2 ) was 90.5%, 97.2% and 98.9% for SS304, PVC1 and PVC2, respectively, and the level of significance was P 0.001. The bias factor (B f ) and accuracy factor (A f ) both approached 1.0 for the three surfaces evaluated. The model equations for predicting optimum response values were verified effectively by a validation data set for all surfaces evaluated. Therefore, an RSM provides a useful and accurate method for predicting E. coli O157:H7 adhesion and potential to form biofilm on SS304, PVC1 and PVC2 and could be considered to be a standard way to ensure food safety with respect to E. coli O157:H7 contamination through adhesion and biofilm formation.