What do medicine and statistical models have in common?

We have all seen the winged staff with two snakes more than once. These represent the therapeutic dose and the toxic dose, the poison. The winged staff represents medicine, or whoever represents it professionally, because it rises above the parts as an expert of both.

The same applies to statistical models, where the therapeutic dose turns out to be what explains the model of the phenomenon of business interest, such as turnover, probability of potential customer conversion, etc., the toxic part turns out to be the model’s error term, which is almost always present in statistical models. The cost of approximating reality to an equation, mathematicians would say.

The statistician, ideally as well as for many physicians, has the knowledge to balance error and utility of the model. Not necessarily models that explain 40 percent should be thrown out, just as not necessarily models that explain 87 percent become a security: the poison, the error, should be monitored just as the dentist makes a few follow-up visits after wisdom tooth extraction, the surgeon visits after surgery. Error left to its own devices, such as mishandled wounds or “forgotten gauze,” can become breeding grounds for pathogens. From bacteria to business processes disconnected from reality.

Just as surgeons, who see patients only as numbers, increase the risk of medical malpractice, statisticians, or data scientists, who do not interact with customers can also engage in bad statistics.

If you want to rely on a statistics firm that cares about you, let’s check the conditions of this harmony in a free call.

 

Privacy Policy