In an effort "to assess public understanding of selected traffic control devices (TCDs)" and, "if appropriate, to develop specific recommendations to improve comprehension of these devices," Robert W. Stokes, Margaret J. Rys, Eugene R. Russell, and Jeff Kerbs of Kansas State University's Department of Civil Engineering surveyed over 500 Kansas motorists. The Kansas Department of Transportation published the results of their study in a March 1995 report entitled "Motorist Understanding of Traffic Control Devices in Kansas." This is a summary of their methods and findings.

Previous Research

Motorists must understand the messages traffic signs and pavements markings are designed to convey in order to make appropriate, safe driving decisions. As the number of signs and markings used on our highways increases and designs change, traffic safety professionals are concerned about the continued effectiveness of some TCDs. The authors of this report reviewed previous research in this area to guide them in designing and administering their survey.

A study conducted in 1987 classified 17 standards in the Manual on Uniform Devices (MUTCD) for Streets and Highways "as having significant need for additional research." Many of these were signs. The authors concluded that this need may result in part from signs being "the oldest traffic control devices currently in use." Moreover, newer signs have sometimes been included in the MUTCD "without any research concerning their effectiveness." Stokes, et al. theorized "the design and use of many signs may be based on inadequate, nonexistent or outdated research."

Studies conducted for the American Automobile Association in 1979 and 1980 sampled thousands of drivers from across the United States. These surveys evaluated driver understanding of some traffic sign symbols, traffic signals, and pavement markings. Researchers concluded that understanding was "generally poor" and that older drivers had a poorer understanding of the TCDs evaluated than did younger drivers. A Federal Highway Administration study in 1986 yielded similar results and found certain groups of signs to be more confusing than others. Especially confusing were curve versus turn signs and pedestrian versus school crosswalk signs.

In 1978 and 1981, the Texas Transportation Institute (TTI) evaluated motorists' understanding of TCDs in Texas. Their 1978 research found 11 of 27 TCDs were "seriously misunderstood," and the 1981 study found 19 of 63 TCDs in need of improvement. Recent TTI research has identified more specifically which signs and roadway markings were least well understood.

In selecting traffic control devices to include in their survey, Stokes, et al. referred to findings from previous research efforts. They discovered that the signs and pavement markings TTI researchers and others had consistently identified as misunderstood were:

TURN vs CURVE signs, School/Pedestrian Zone vs School/Pedestrian Crossing signs, Speed Zone signs, Double Turn signs, Two-Way Left Turn Lane signs and markings, and Divided Highway signs.

In summarizing their research review, Stokes, et al. reported that few studies had gathered information to help highway professionals develop programs to improve motorist comprehension of TCDs. They also found a lack of research that would lead to suggestions for changing the TCDs themselves to improve their comprehension.

Design of Present Study

Using information from previous research and input from the Kansas Department of Transportation advisory committee for this project, the researchers selected 43 signs and pavement markings to include in a survey of Kansas motorists. They chose 25 warning signs, 13 regulatory signs, and 5 longitudinal pavement markings. The latter included:

1) the single broken yellow center line, 2) the solid single yellow center line, 3) the single broken white lane line, 4) the solid white edge line, and 5) the two-way left turn lane lines and pavement marking arrows.

In their study, Stokes, et al. used self-administered questionnaires composed of sign faces and plan views. At the end of the survey were 17 demographic and driving experience questions.

The primary survey form showed colored illustrations of the 43 TCDs and elicited multiple-choice responses as to the meaning of each and the appropriate motorist response to each. Respondents chose from a correct response, two incorrect responses, and a "not sure" response. Motorists were instructed to answer with "not sure" rather than guess. Space for comments was included.

An alternate survey form asked respondents to write their own descriptions of the meanings of 10 signs. Results from this open-ended questionnaire allowed researchers to compare responses and learn to what extent the multiple-choice nature of the primary survey gave respondents clues to correct answers. If fewer respondents were able to correctly identify a sign's meaning when they had to come up with the answer themselves rather than select the answer from a list of three, the researchers would assume respondents' guesses were helped by the multiple-choice format.

Fifty Kansas Department of Transportation employees and 50 students from Kansas State University pilot tested the survey instruments. Proposed surveys were found effective, and the researchers made only minor changes.

Using a quota sampling plan, researchers selected a representative sample of 500 motorists. This group included drivers from seven Kansas counties and allowed "for meaningful analysis of various population subgroups (i.e., males, females, different age and ethnic groups) with varying levels of driving experiences." Surveys were administered at civic and social functions, court houses, city halls, and places of employment.

Study Limitations

Stokes, et al. intended this survey to identify problems with motorist understanding of the safety-related messages conveyed by certain signs and pavement markings. The study did not explain why motorists might misunderstand these messages, nor did it find out whether such misunderstanding might lead to inappropriate or unsafe driving. Moreover, the researchers did not necessarily design the study to produce recommendations for correcting the problems they found.

Use of a written survey assumed respondents were sufficiently literate to read and understand the survey itself. Moreover, the researchers cautioned that signs and markings were presented in an isolated, out-of-context format. Motorists completing the survey did not have the benefit of environmental clues that would normally help them interpret TCDs in a real-life driving situation.

Finally, although "a reasonable effort" was made to collect surveys from a representative sample of Kansas drivers, in the interest of time the researchers collected information from survey sites where they expected to have ready access to many potential respondents. Therefore, the sampling technique was not truly random.

Analysis of the Data

Approximately 500 motorists completed the multiple-choice survey, and an additional 25 responded to the open-ended questionnaire. While this group was largely representative of the driving population of Kansas, urban male respondents from 35 to 54 years of age were over-represented and motorists from 16 to 24 were under-represented. The researchers stated, "In terms of basic demographics, the typical respondent can be characterized as a middle-age male with some college education who resides in a non-rural area of the state." These individuals tended to be experienced, mature drivers who had been licensed in Kansas for more than ten years. Most had taken a driver education class; however, most took it over ten years ago.

The article reported on study results in five categories: warning signs, regulatory signs, pavement markings, open-ended questionnaires, and effects of demographic and driving characteristics.

The overall comprehension level for all warning signs was about 75 percent. The authors noted, "As an aside, a 'percent correct' score of 80% is required to pass the written portion of the Kansas Driver's License Exam." Using the 75 percent figure as a breaking point, the following warning signs fell below the average comprehension level and may be considered "poorly understood."

• School crossing
• Railroad Advance Warning
• Roadway Alignment
• Reverse Turn
• Turn
• Meaning of Shape and Color of Yellow Sign
• School Zone
• Pavement Ends
• Reverse Curve
• Lane Reduction Transition

Judging from the number of "not sure" responses to the following warning signs, Stokes, et al. classified them as possibly confusing:

• Meaning of Shape and Color of Work Zone Signs
• Truck Crossing
• Narrow Bridge

Stokes, et al. cautioned that while these results highlighted a lack of understanding of the precise meaning of some warning signs, that lack of understanding would not necessarily result in unsafe driving behavior. The researchers elaborated on misunderstanding the meaning of the School Crossing sign. They wrote:

For example, 82.4% of the survey respondents confused the School Crossing sign with the School Zone sign. The safety- related question associated with this perception is 'are drivers sufficiently alert to the potential presence of pedestrians to be prepared to stop for pedestrians, as required by the School Crossing sign?

The overall comprehension level for all regulatory signs was 78 percent. Using the 78 percent figure as a breaking point, the following regulatory signs fell below the average comprehension level and may be considered "poorly understood."

• Reduced Speed 30
• Divided Highway
• Keep Right sign

A Reduced Speed 30 sign informs motorists that the speed limit ahead will decrease to 30 miles per hour. The researchers noted that more than two-thirds of the respondents thought that a Reduced Speed 30 sign meant they must immediately reduce their speed to 30. The authors observed, "It could be argued that this interpretation would not result in unsafe driving behavior."

The Divided Highway sign shows motorists that they will have to cross the median of a divided highway before turning left from their present location. Sixty percent of survey respondents chose an incorrect meaning for this sign or responded that they were not sure of the sign's meaning. The researchers pointed out, "This confusion could result in potential safety problems."

The overall comprehension level for all pavement markings was about 70 percent. Using the 70 percent figure as a breaking point, the following pavement markings fell below the average comprehension level and may be considered "poorly understood."

• Solid White Edge Line
• Single Broken White Lane Line
• Two-Way Left Turn Lane Marking

While more than one-third of the respondents felt they were not to cross a solid white edge line for any reason, the researchers again noted that this perception arguably would not result in unsafe driving. The correct choice for the solid white edge line was "to let you know where the edge of your driving lane is."

A single, broken white lane line in the middle of the road tells motorists they are on a one-way road where lane changes are permitted. While 38 percent of respondents erroneously indicated this marking identified "two-way roads where drivers are allowed to change lanes," the authors felt this incorrect interpretation would not necessarily cause unsafe driving.

The drawing used in the survey to show the two-way left-turn lane did not show right-turn and left-turn arrows within the turn lane. The researchers theorized this lack of arrows may have accounted for some of the confusion about these lane markings. Over 22 percent of respondents said they were "not sure" of the meaning of the two-way left-turn lane markings shown and almost 15 percent chose incorrect responses.

The open-ended questionnaire, which asked respondents to write their own explanations of the meanings of ten TCDs, produced results that suggested the multiple-choice format of the primary survey provided clues that helped respondents select correct answers. The percentage of correct responses to signs on the open-ended survey (except for the divided-highway sign) was significantly lower than the percentage of correct responses to the same signs on the multiple-choice survey. The researchers recognized, "The implication of this conclusion is that the results from the multiple choice survey instrument may over-state the general population's understanding of some TCDs."

Researchers evaluated differences in correct response rates for the various demographic groups and for the groups identified as having distinct driving characteristics. No significant differences were found based on gender, age, ethnicity, or education. In addition, no differences were found based on years licensed or driving type (urban/rural/both); however, drivers who never had a driver's education class were found to have a significantly lower correct response rate than those who had completed such a class.

Conclusions and Recommendations

Comparing these figures to results from previous research, Stokes, et al. found them to be "in general agreement." They reiterated that the findings might suffer from bias due to the over-representation of mature, experienced drivers and the use of a multiple-choice survey format that increased the incidence of correct guessing. These factors may have resulted in study data that overstated the level of understanding of TCDs in the general driving population.

According to the researchers, the findings suggested that "there is a need to either redesign or modify certain TCDs currently in use in Kansas, or to initiate programs to educate the motoring public on the appropriate response to these devices." Considering this conclusion, Stokes, et al. made four recommendations:

1. Send the results to the Federal Highway Administration to guide future research that will help determine to what extent the misunderstandings revealed by the study result in unsafe or inappropriate driving behavior.
2. Send the results to the Kansas Motor Vehicle Department and to Kansas education officials to guide design of driver education courses that emphasize misunderstood TCDs and to inform the training program for driver education instructors. In addition, the Motor Vehicle Department might continue this research by determining whether incorrect responses to questions on driver licensing and driver education tests support the results of this study.
3. Incorporate, or continue to incorporate, verbal messages to supplement the symbolic messages on certain TCDs. As an example, they recommended adding a "PED XING" panel to the Pedestrian Crossing sign.
4. Send the results to the Kansas Department of Transportation Public Information Office to be used in creating an awareness campaign that will inform motorists about the problems of misunderstood TCDs.