¶ … fishing: A public health example.

Imagine this scenario. You are newly self-employed as a freelance sports journalist. You apply for a health insurance plan. You are dumbfounded to find yourself rejected, given that you are trim, fit, and exercise regularly. Little do you know the real reason behind the insurance company's rationale -- it has surmised (correctly) that you are an avid cyclist, given that your last job was as an editor for a biking magazine. Given that cycling is a risky sport, the insurance company decides that you are a bad risk. "Data dredging or data fishing is the inappropriate (sometimes deliberately so) search for 'statistically significant' relationships in large quantities of data" (Data dredging, data fishing, 2010, B2B). Health insurance companies already use a wide variety of risk-management techniques, such as using the patient's gender and age to calculate premiums, as well as screen individuals for pre-existing conditions. Using data dredging could increase such company's ability to carefully screen their clients and deny people coverage exponentially.

Employers could also misuse data to their advantage. For example, when considering a new employment prospect, if a company discovered, upon consulting with the individual's previous employer, that he or she had contracted cancer several years ago, they might decide to forgo hiring the individual. Also, by virtue of a prospective employee's age, location, and demographic profile, they might conclude that she was statistically likely to have a child within three to five years, and thus decide not to hire her because of the potential expense she might incur due to maternity leave. The surmises derived from data mining are not always accurate, but many organizations would jump at the chance to avoid any possible financial or legal risk, even if this would be unfair to an individual, ethically and morally.

Reference

Data dredging, data fishing. (2010). B2B. Retrieved February 13, 2010 at http://www.businesspme.com/uk/articles/technologies/13/Data-dredging,-data-fishing.html