Recently it has been suggested that the use of alcohol-containing mouthwashes may increase the risk of oropharyngeal cancer. Heavy alcohol intake and tobacco use are established causes of oropharyngeal cancer. Their use is associated with mouthwash use. In addition, alcohol and tobacco use both tend to be underreported. Here the authors show that, under the hypothesis that mouthwash does not increase the risk of oropharyngeal cancer, confounding due to underascertained exposure to alcohol and tobacco would result in a spuriously elevated odds ratio for mouthwash use. As a general principle, a null association becomes apparently positive if a confounding variable is incompletely ascertained: a spurious association may be produced even in the absence of a difference in the extent of the underascertainment of the confounder among the comparison groups.