Consulting Best Practices

Consultants Need to Understand Type I and Type II Errors

Type I errors and Type II errors, otherwise known as false positives and false negatives, are two types of error that everyone should understand. In a post for the Institute of Management Consultants USA, Mark Haas breaks down how they work. Read this, and you will never make a mistake again for the rest of your life, maybe.

Debug Mode

In dealing with a false positive, it means you have rejected a null hypothesis when it should have been accepted. In other words, you looked at the data and decided a connection existed between two things when that was actually not the case. For instance, you won $10 on the lottery, so you decided to sell your house to spend on more lottery tickets because you were on an unbreakable hot streak baby. A more practical example would be that a consultant, eager to please a client, “discovers” a problem that does not really exist. He or she can “fix” the nonexistent problem and then take credit for improvements that also may or may not exist.

A false negative is the opposite—missing a connection that really exists. These can occur when, for instance, trying to get a pulse on company culture and attitudes. An employee survey that is not comprehensive can perhaps make it look like management is doing a great job, when really management is a bunch of dinguses. Businesses are sunk when the roots of bad situations are left unfound.

Haas ultimately recommends you consider Type I and II errors in medical testing terms:

The worst outcome when looking for a serious disease is to conclude it is not present when it is (Type II). To accommodate that, we use screening procedures that are relatively fast, cheap and for which we can tolerate a Type I (false positive) error. As a consultant, you may want to develop protocols that let you quickly tease out potential problem areas and for which you recognize there may be Type I errors… Then you can proceed with more focused and rigorous protocols to look more closely at an issue, recognizing that what you want to avoid is a Type II error (false negative).

Approaching risk and errors with this attitude will be a value-add to your services all by itself. You can view the original post here:

Show More