Warnings and Warning Labels

Marc Green

Human factors designers attempt to make products as safe as possible. Some products, however, carry inherent hazard: consumer appliances that require electricity, drugs, swimming pools, machinery, etc. When hazards cannot be completely removed by design, the next step is to provide a "guard," a physical means of reducing the danger. This may be a barrier, such as a fence, or a "lock-out," such as a plug with asymmetrical pins that can only fit an outlet in the proper way. Warnings are a last resort to promote safety if the hazard is not obvious. (No one needs a warning to avoid putting his hand into a fire.)

The warning should alert the user of the danger, describe its nature, and explain what will happen if the user fails to comply. For example, a label might say "Warning: Keep out of the reach of children. May cause stomach irritation and blindness." Like most warnings, the statement starts with a signal word, "Warning." It then says what to do ("keep out of reach of children") and what the consequences will be for failure to comply ("May cause stomach irritation and blindness.") Some warnings may also include statements about what to do if a failure to comply occurs: "See doctor immediately" or "Induce vomiting immediately."

Warning Effectiveness

The problem with warnings is that they often fail to work. In order to be effective, the user must process the warning in a series of mental operations. Below, I explain these operations and factors that influence warning effectiveness.

1. Notice the warning

The first stage is simply becoming aware that the warning exists. Sometimes the issue is conspicuity - the ability of something to draw attention even when the user is not looking for it. (See Error, Accident and "Inattentional Blindness" for more information.) This is a huge topic, but there are a few general points to make. First, people typically focus attention in the same direction as the eyes point. This means that warnings not located close to the line of sight will not be readily seen. Physical properties such as size and high contrast will likely further encourage conspicuity. However, much recent research suggests that many factors often thought to promote conspicuity, such as motion and color, do not always work. Moreover, other recent research on "inattentional blindness" has shown that people often fail to notice objects which are right in front of their eyes.

Second, warnings should be placed where they are likely to be seen. This rule may sound obvious, but it is often ignored. Everyone has heard of the woman who was burned by hot coffee at McDonald's and won a large damage claim¹. McDonald's, as part of their defense, claimed that the cup had a warning about the hot coffee. This may be true, but how many people actually read their coffee cups? The point is that warning designers must understand how the user will interact with the product and position warnings accordingly.

In some cases, however, users may actively seek warnings. If the user believes that the product is dangerous, s/he might look for a warning. On the other hand, if the user has experience with the same or similar products, there will be less likelihood to seek a warning. A user may well assume that the warning is irrelevant when s/he has a history with a particular type of product and has not had negative interaction. A key to determining cause in any accident analysis is to examine, not only the environment at the time of the accident, but also the actor's previous experience in that environment.

2. Perceive the warning

Once the user notices a warning, it must be perceived. By this, I mean that the words or graphics must be read - the sensory input is given meaning by drawing on memory. In order to read a warning, it must be high contrast and, in the case of text, be large enough and rendered in a legible type font and family. Exact size is partly a function of the user, e .g., larger print for older users. It's ironic that warnings on prescription bottles are often small even though the elderly, who have reduced vision, use prescription drugs with high frequency.

If people believe that there is little hazard, they are likely to refrain from bothering to "read" the warning, even if they are aware that there is one present. Moreover, people often perceive the wrong part of the warning. There are many cases, for example, of a nurse looking at a vial and then injecting the patient with the wrong drug. In most cases, the nurse noticed the label but failed to perceive it. People have a tendency to generalize the cues that they use to make decisions. If it is the case that different labels have different colors, the nurse might unconsciously depend on color (or even more subtle cues like position on the shelf) to signal the warning and then cease reading the label text.

3. Understand the warning

Once the warning is perceived, the user must properly understand its meaning. The message should be clear and easily understood, but it is difficult to ensure clarity for several reasons. Sentences may be poorly constructed or contain words that are unfamiliar (Flange?). Sometimes the message is simply too vague. A warning that says "ensure adequate ventilation" is problematic because it is up to the user to define "adequate."

Users may also fail to understand the warning if they have little familiarity with the product and its operation. The words may be comprehended but the relationship to product use unclear. Some users, such as children (lack of language skills), older people (declining cognitive abilities) and non-native speakers (insufficient English knowledge) may also have difficulty in understating warnings.

In order to address users with lower language skills, there has been great interest in rendering warnings with pictographs, small pictures, such as a person recoiling from an electric shock, or symbols, abstract graphics, such as the one used to note radioactive material.

It is very difficult, however, to ensure that pictographs and symbols will be properly interpreted. One famous example is the traditional symbol for poison, a skull-and-cross-bones. Many children were poisoned because they interpreted the graphic as meaning "pirate food," and therefore believed that the substance was not only safe but might be fun to eat. (The skull-and-cross-bones has been replaced by "Mr. Yuck," a schematic, frowning face.)

To show the difficulty of creating understandable symbols, I have run classes where students privately created symbols for a variety of concrete and abstract concepts. Each student then presented his symbols to class to test comprehension. Success rate was very low, usually 20% at best. Many research studies have recorded similarly low rates of graphic interpretation. The American National Standards Institute (ANSI) suggests that graphic warnings be tested and have a recognition rate of 85%, but there is no requirement that manufacturers test graphic warnings. (Performing a test in order to support litigation is often a good idea.)

Another reason that warnings are often ineffective is that they are designed by people who already know about the hazard and who are highly familiar with the product. These designers are unable to put themselves in the place of a naive user who approaches the product for the first time. Warnings should be tested on typical users before a product is released.

4. Comply with the warning

Even if the user understands the warning, s/he may not comply. One common reason is the cost of compliance. If a warning says not drive for two hours after taking a drug, then the likelihood of compliance is a function of the time and expense required to find other transportation. If a warning says, "do not enter," then compliance likelihood lies in the availability of an alternative route. One study showed that the users are more likely to comply if the user's cost is directly lowered: users were more likely to comply with a warning to use safety gloves when the gloves were attached to the product. In sum, users perform a mental cost-benefit analysis where perceived likelihood and severity of the hazard are weighed against cost of compliance. Any factors which increase cost or reduce perceived risk (such as product familiarity) will hurt compliance.

Another example comes from the large number of diving accidents occurring with use of in-ground pools. Conventional pools usually have both a shallow and deep end. If someone sees a "no diving" or "shallow" notice at one end of the pool, s/he need only go to the other end to dive safely. In-ground pools have only one height, so there the person seeing the warning must give up diving altogether, making compliance more difficult. (Of course, there are other reasons that in-ground pools are more dangerous: conventional pools have a diving board to act as a safe diving cue, so swimmers are more likely to believe signs.)

Other factors lie inside the user. Some people are risk takers and more willing to risk the consequences. Other people are control-oriented and do not like to have their behavior limited by "orders" from the manufacturer.

Finally, people who have used a product many times with no negative consequences (and possibly know of other people who have had similar experiences) will be less likely to comply. People frequently ignore swimming pool signs concerning possible dangers. This occurs, at least partly, because most people haven't broken their necks by jumping off a diving board and don't know anyone who has. Since younger people have had less experience in life, they are less likely to perceive significant risk. Conversely, people who know of bad experiences of failure to comply will be more likely to heed a warning. I only started wearing safety glasses around power tools after meeting someone who had lost an eye from kickback on a table saw. The hazard was no longer an abstract concept.

Conclusion

I have only attempted to provide a broad overview of the issues involved in warning labels. Each situation has its own variables and therefore requires detailed analysis. Lastly, warning labels are only a last resort when hazards cannot be eliminated by careful design. They do not replace good design or excuse poor design.

Endnotes

¹ The woman in question was elderly and received third degree burns, so the injury was not trivial. She sued only to recover medical costs. The jury awarded the large punitive amount because of McDonald's failure to act on several previous complaints that the coffee was too hot and was dangerous.

---

A division of 2057949 Ontario, Inc.
Copyright © 2024 Marc Green, Ph. D.
Home Page
Contact Us