Detecting Errors: Separating the noise from the meaningful

How do we determine what’s important? When we drive our vehicles we pay attention to lights, lines, and signs. When we hear the beeping of a truck backing up, we stop, pause and make sure we’re not in the way. These daily beeps, lights, and signs help us cut through the noise of our environment and determine what’s important to pay attention to and what’s not. These principles are important to understand, because they can be applied with great effect to our own productivity and work.

Signal Detection Theory

One way to understand why these tools help us cut through the noise is through Signal Detection Theory (SDT). SDT is a framework for understanding how decisions are made, particularly in circumstances of uncertainty. The framework is widely used in fields like psychology, medical diagnostics, engineering, and military applications. The ideas are also germane to our topic, and thus merits additional fidelity.

For any situation where a system or an environment changes, there exists several elements. Noise is the element of status quo. It is the current state of the environment. You pay no particular attention to it. There is also a signal, which indicates the change in the environment. In some instances, you may detect the signal while other times you may not. The cutoff, or criterion response is the point where the signal is different enough from the noise to allow you to detect changes in the environment. See the figure below:

The diagram shows on the x axis the “internal response” or the awareness an individual has for either the noise or the signal. The y axis shows the probability or likelihood of the signal or noise eliciting that level of awareness.

Four possible outcomes can result from any circumstance.

Correct Rejection: There is no signal, and you did not detect a signal.

Hit: There is a signal and you correctly detect the signal.

False positive (false alarm): There is no signal, but you incorrectly detect a signal.

False negative (miss): There is a signal, but you do not detect it.


Clearly, two of these scenarios are good one is bad, and one is very bad. Correct rejections and hits are the desired outcomes. Meanwhile, false positives are not desirable, but they are disruptive and lead to inefficiencies. We’ve all had the feeling of a phantom cell phone vibrating or are unsure whether or not we heard a knock at the door. They lead to inefficiencies, but the consequences are not horrible. However, false negatives usually have much more dire consequences. A missed phone call may mean missing out on an important opportunity or responding late to somebody in need of your help. This is to say nothing of missing signals from trains and automobiles.

Improving Performance

The diagram above demonstrates that the bad circumstances (false alarms and misses) occur when there is an overlap between signal and noise. This makes intuitive sense; we would only mistake one for the other when they are eliciting the same level of internal response.

The diagrams are also instructive in how to improve our ability to identify changes in the environment, quickly and effectively. If the distance between the noise and the signal is increased the opportunities for misidentification drastically decrease. In short, discrimination between signal and noise becomes easier.

This is exactly what the lights, beeps, sirens, and signs (what I collectively call signal enhancers) seek to do. By utilizing auditory and visual stimuli, signal enhancers increase the distance between signal and noise, decrease the overlap, and aid in quick, effective decision making. The truck beeping as it backed up, the buzzer at the café, the fire alarm, the green walk sign and chirping noise at the intersection are all there to create greater separation from the existing system and environment and the signal.

Strategies

There are many varied strategies that can be used along with the tools of signal enhancers, and in the future, I will dedicate an entire blog post to talking about these strategies and the conditions where they are most successful. For now, consider them at a high level. The three strategies are:

Notification: This strategy utilizes signal enhancers (usually texts like signs and dialog boxes) that the environmental conditions have changed. construction signs with “Workers Present” or email messages telling you how much storage you have left on the cloud are examples of this strategy. They allow the individual to determine if a response needs to be taken and what that response should be.

Deterrence: This strategy creates inconveniences or future inconveniences if an action is performed. A gate in front of a train track is actually quite easy to go around, but it creates an inconvenience of steering the vehicle in a serpentine manner in order to get around the gate, over the tracks, and around the other gate. This maneuvering makes the driver less confident that he or she can go over the tracks before the train arrives, and therefore acts as an effective deterrent. Likewise, fire exit doors that indicate that the alarm will sound when used is an effective deterrent. It discourages people from using the door except in cases of fire because they don’t want the sirens to activate. While there is some constraint put on the individual, the individual is still in control of their decisions and behaviors.

Annoyance: This strategy uses a constant or frequent notification to coax and prod the user into a desired behavior. The notification continues until there is compliance with the preferred course of action. Seatbelt alarms are excellent examples of this. The lights on the dashboard and constant beeping are enough to get most people to buckle their seatbelts, even if it’s simply to get the notifications to turn off.

As each of these strategies indicate, the use of signal enhancers can be effective. By deploying different strategies to the situation signal enhancers can help cut through the noise, focus attention on what’s important and even coax behavior out of individuals. It’s important to remember, that these tools still rely on the individual to take the correct course of action. This limitation can be a blessing or a curse.

Michael Parent

Michael Parent is CEO of the Problem Solving Academy and author of “The Lean Innovation Cycle” a book that explores the intersection of Problem Solving, Lean and Human Centered Design. Throughout his career, Michael has coached executives through strategic problem solving, strategy, and operations management and has led numerous projects in a variety of industries.

Next
Next

Bad Design is like a virus: A look at Latent Errors