A confounding factor is a factor that influences the relationship between variables that are investigated. In dietary research, for example, a lack of physical exercise may potentially confound the association between the consumption of soda beverages and overweight: when those who drink the most soda beverages are generally more overweight, but also exercise less, the association between the consumption of soda beverages and overweight may be caused partly or entirely by a lack of exercise. Other common confounding factors include age, gender, whether people smoke or drink alcohol, occupation, educational attainment, or income. Confounding factors are a potential bias in statistical research and may lead to over- or under-estimating the relationship between variables. While theoretically their number can be infinite, statistical research often accounts for a set of known potential confounding factors.