An extensive questionnaire was developed and used to collect data from 33 herds that were on the New York State Paratuberculosis Control Program, to study farm factors associated with the presence of Mycobacterium paratuberculosis infection in dairy herds. The results of the last whole herd paratuberculosis fecal culture were used to indicate presence of infection in a herd, with herds having one or more animals positive classified as 'infected'. The average prevalence within herds was 5.2%. Fourteen herds were uninfected and 19 herds had prevalences ranging from 0.7%-28.2%. Data on 31 continuous and 67 categorical risk factors were collected by questionnaire. Ten factors were significantly associated with prevalence risk of infection in the univariable logistic regression. These factors were: the type of farm operation (commercial/registered or both); earlier diagnosis of the disease before entering the control program; number of clinical cases in the previous year; whether clinical cases were raised or purchased animals; typical signs in clinical cases; exposure of calves 0-6 weeks of age to feces of adult cows; contact of young stock with adult animal feces from using the same equipment to clean the housing for both groups of animals; spreading feces on fields from which forage is later harvested and fed to animals of any age group; what is done with animals that are suspected of having paratuberculosis or test positive on culture; and frequency of cleaning the cow barn. Stepwise logistic regression was used to determine the significance of each risk factor while controlling simultaneously for the effect of other factors. The significant factors were the type of farm operation, clinical signs, and exposure of calves to feces of adult cows. Commercial herds, presence of clinical signs typical of paratuberculosis in animals, and exposure of calves 0-6 weeks old to feces of adult cows all indicate a higher likelihood that a herd is infected with M. paratuberculosis.