Statistics and the Immigrant Ban — where’s the problem?

I thought it might be helpful to evaluate the positives and negatives of the Immigration Ban using scientific principles. I always try to be optimistic that clarity in our principles will lead to better decisions.

In many areas of research, we talk about Type 1 and Type 2 errors — or false positives and false negatives. Generally, we do our best to minimize both types of errors, but, in some cases, one type of error is worse than the other.

An example of how we use these concepts in making policy is the death penalty. One of the most compelling arguments against the death penalty is the desire not to kill innocent people, a type 1 error/false positive. Americans would rather not kill a few guilty people (a Type 2/false negative error) in order to avoid the death of an innocent person.

Lynching in the South was an example of massive Type 1 errors. A person was guilty of whatever because of the color of their skin. It was a decision based on hate, emotion, and the desire to eliminate people who were “different”.

Using stereotypes to judge people typically results in massive Type 1 errors.

For terrorism, Americans want to maximize security, so we need to assess our Type 2 errors, because we want these (missing a terrorist) to be as small as possible, without harming innocent people (Type 1 errors).

As White House spokespeople said, Obama cited the countries involved in the ban as potential sites of terrorism and, in this process, strengthened the vetting process to avoid type 2 (missing terrorists) errors. We have not had an act of terrorism on US soil from anyone from these countries since 9/11 (and most of the 9/11 terrorists were from Saudi Arabia). So, it appears that the Obama rules were accurate and sufficient.

However, there was a great deal of fear-mongering in the election, so that many people overestimate the Type 2 errors and feel safer with the ban. Fear pushes us into a mode that increases the use of stereotypes, so people are more willing to make decisions that result in Type 1 errors.

In great part because our Type 2 errors have been so small, the ban has unintended consequences.

The Immigrant Ban, because is broad and groups people by state and/or religion, it leads to large Type 1 errors. Many people in the US and worldwide know that they and their friends and family are not terrorists, so see this broad ban as unnecessary and based on fear and, possibly, hate and ignorance. It presumes guilt — even when, for many immigrants, there is already vetting in place that has been successful, so it harms more innocent people without finding more terrorists among them.

Our country loses credibility by basing decisions on the apparent belief that everyone from some group, whether country, political party, or race, is bad.

The US government has never made important decisions based on tweets or memes or fear, but we have always, before, spent time discussing issues and decisions at a deep level — and we need to continue this.