Changeable Message Sign Displays During Non-Incident, Non-Roadwork Periods (2008)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

15 STATUS OF CHANGEABLE MESSAGE SIGNS Background Once a CMS system is installed, a question always arises concerning when messages should be displayed during non-incident/non-roadwork periods. There are the following two schools of thought on this topic: 1. Display messages only when unusual conditions exist on the freeway. 2. Always display messages regardless of whether or not unusual conditions exist on the freeway; or, at a mini- mum, always display a message during the peak periods and only when unusual conditions exist during the off- peak periods. The authors of early CMS message guidelines recom- mended the first of the two approaches (display messages only when unusual conditions exist) unless travel-time messages could be displayed (2–4). The recommendations were based on human factors principles and came about because of dif- ficulties in designing and displaying other types of messages when incidents or roadwork were present, particularly during peak periods. The second approach (always display a mes- sage) leads to violation of the following two important human factors principles for a CMS: • Don’t tell drivers something they already know; and • For more effective systems, use the CMSs only when some response by drivers is required (i.e., change in speed, path, or route). In the absence of incidents during the peak periods, more often than not, bottleneck locations and the subsequent loca- tions and durations of congestion can be predicted by mo- torists. Consequently, the same congestion information will most likely be displayed almost daily. There is concern by some that display of repetitive information will result in many drivers failing to read the CMS, even when important information is given, owing to the psychological visual change detection phenomenon referred to as change blind- ness (22). Change blindness is the phenomenon in which a person viewing a visual scene apparently fails to detect large changes in the scene. With respect to the issues discussed herein, a driver may fail to recognize that the current message is different than the message the driver previously saw on the CMS; thus, the driver fails to read the message. Associated with the change blindness phenomenon is a potential credi- bility problem. The issue of change blindness has caused some agencies to consider the use of flashing beacons on CMSs to attract the attention of motorists when incident, roadwork, or similar messages are displayed, even though CMSs, particularly light-emitting signs, are designed to at- tract the attention of motorists. At this time, it is not known if the practice of using flashing beacons improves motorist attention to the CMS message. To date, no research has been conducted to evaluate the effects of change blindness with re- spect to CMS messages. Another consideration, with respect to displaying messages for recurring congestion, is that one simply runs out of descrip- tors for the various possible levels of congestion. For example, if descriptors such as HEAVY CONGESTION or MAJOR DELAY are used to describe recurrent congestion during the peak period, then descriptors are not available for the more severe congestion when incidents occur. An example of negative public reaction to displaying messages during non-incident/non-roadwork periods was the experience of the TMC in Los Angeles. Initially, Caltrans personnel in the Los Angeles TMC displayed public service messages on freeway CMSs to avoid having blank signs. Al- though these messages were transportation-oriented in nature (e.g., NEXT TIME TRY AMTRAK TO LAS VEGAS, RELIEVE CONGESTION–RIDESHARE, etc.) they did not relate to the operation of the freeway system. Public reaction to the use of the CMSs in this manner was quite negative. There was a belief among traffic operations professionals that such use led to a public disregard of messages on the CMSs, thus making the signs less effective when traffic operational messages were displayed. The practice was discontinued, and the CMSs were then only used for messages pertaining to unusual real-time traffic flow conditions (D. Roper, former Deputy District Di- rector, Operations, Caltrans, personal communication, Mar. 21, 2002). As will be discussed later, the TMCs in Los Angeles and other locations in California now display travel time. In contrast, in 1992, Smith reported that CMSs were in- stalled for the INFORM (INformation FOR Motorists) Project on Long Island more than 18 months before the sys- tem became operational because of delayed construction schedules (23). Adverse public reaction to having expensive CHAPTER THREE NON-INCIDENT/NON-ROADWORK MESSAGES

CMSs sitting idle for several months prompted New York DOT to adopt a policy of displaying some type of message on the freeway CMSs at all times. Survey Results General A summary of the status of the CMSs during non-incident/ non-roadwork periods reported by the agencies that responded to the questionnaire is shown in Figure 2. The results indicated that during the peak period, 55% of the TMCs do not display messages, whereas 45% regularly or periodically display a mes- sage. During the off-peak period, 59% do not display messages, whereas 41% regularly or periodically display a message. Over the years, a decreasing number and percentage of transportation agencies have left CMSs blank during non-incident/non-roadwork periods. Conversely, the number that regularly or periodically displays messages has increased. In 1997, 77% of the state DOTs that responded to a survey had a policy of displaying messages only when unusual con- ditions were present on the facility and left the CMSs blank during other times (24). 16 Basis for Decision to Leave the CMS Blank or to Display a Message The TMCs were asked for the basis for the decision to leave the CMSs blank or to regularly or periodically display messages during non-incident/non-roadwork periods. The results are summarized in Table 4. Interestingly, the decision by 78% of the TMCs as to whether to leave the CMS blank or to display a message dur- ing non-incident/non-roadwork periods was based solely on agency administrative/upper-management preference and/or TMC manager/supervisor preference. An additional 7% included feedback from telephone calls, newspapers, radio, and/or television. Only 15% of the TMCs based their decision partly on information from focus groups or research. Blank CMSs (No Message) Status of CMSs When No Message Is Displayed Not all of the 59 TMCs that do not display messages during non-incident/ non-roadwork periods leave the CMSs totally blank. Nine percent display a single pixel, and 5% display a set of pixels. A summary of the status of the CMSs for those TMCs that do not display messages during non-incident/non-roadwork periods is shown in Figure 3. Public Response to Blank CMSs A summary of the re- sponses to the question, “What has been the public reaction to leaving the CMSs blank?” is shown in Figure 4. Twenty-six percent of the 59 TMCs that do not display messages indi- cated that they have received very favorable (13%) or favor- able (13%) responses. Thirty-seven percent indicated a somewhat favorable (13%) or neutral (24%) response. Thirty-seven percent indicated that they did not have suffi- cient information to comment on the public’s response. Experiences and Lessons Learned about Blank CMSs The experiences of the TMCs that do not display messages 55 59 45 41 0 10 20 30 40 50 60 70 Peak Off-Peak Period of Day Pe rc en t o f T M Cs No Message Regularly or Periodically Display Message FIGURE 2 Status of CMSs during non-incident/non roadwork periods. Basis Percent of TMCs Percent of TMCs Agency administrative/upper-management preference only 27 TMC manager/supervisor preference only 17 Preference Only Agency administrative/upper-management preference, and TMC manager/supervisor preference 34 78 Preference and Feedback Agency administrative/upper-management preference, and TMC manager/supervisor preference, and Feedback from telephone calls, newspapers, radio, and/or television 7 7 Agency administrative/upper-management preference, and TMC manager/supervisor preference, and Research conducted by agency and/or research conducted by others 3 Research conducted by agency only 2 Research conducted by others only 7 Preference and/or Research or Focus Group Focus group studies only 3 15 100 100 TABLE 4 BASIS FOR DECISION TO DISPLAY OR NOT DISPLAY MESSAGES DURING NON-INCIDENT, NON-ROADWORK PERIODS

17 • Test the CMSs, so the public knows they are working. Do not test the signs with a time and temperature dis- play feature (1 TMC). • CMSs should be reserved for traffic impacting incidents unless used for a safety campaign during off-peak hours (1 TMC). • Drivers do not appreciate being “lectured to” with safety-oriented messages, even when there is not any- thing else important to say (1 TMC). • Less maintenance is required on CMSs (1 TMC). • We get complaints if we over-use taxpayer money being spent for signs that are not needed (1 TMC). • We get complaints if we overuse them for construction lane closures off the system (adjacent freeway) (1 TMC). • Displaying messages that may be unrelated to the motorist’s travel could actually increase motorist disre- gard for the CMS when there are relevant messages dis- played (1 TMC). • A lot of questions and confusion are eliminated by leav- ing the CMSs blank unless posting travel times or other advisory or incident messages improves the focus of the project and does not create questions for the public. Back and forth flashing symbols to indicate the sign is working result in confusion and questions (1 TMC). • We should develop procedures to better utilize the CMS (1 TMC). Experiences and Lessons Learned about Displaying One or More Pixels As noted earlier, some of the TMCs display one or more pixels on the CMSs rather than leaving the signs totally blank. Some experiences and lessons learned by TMCs are as follows: • Some drivers believe that a single pixel displayed on a CMS is a malfunctioning sign (1 TMC). • We do get some negative feedback when a flashing dot is used. We still think leaving them blank is better (1 TMC). • At one point, we tried displaying safety-related messages, and the feedback was very negative, so we went back to just a flashing pixel in the off-peak hours (1 TMC). Concerns and Challenges The primary concerns and chal- lenges regarding leaving the CMSs blank centered on (1) ways to convince motorists that the signs are functional and (2) pres- sures from administrators/upper-level management and others to display messages. Plans for Changing Policy for Blank CMSs The agencies that currently leave the CMSs blank during peak, off-peak, or both, were asked whether they plan to change their policies. As shown in Figure 5, 90% of the 59 TMCs that leave their CMSs blank did not intend to change, whereas 10% planned to change their policy. About one-half of the TMCs that indicated that they intend to change their policy would like to display travel-time information whenever resources become available. The results indicate that the vast majority of the TMCs that leave the CMSs blank are satisfied with their current policy. Blank CMS 86% Single pixel 9% Set of pixels 5% FIGURE 3 Status of CMSs for TMCs that do not display messages during non-incident, non-roadwork periods (n = 59). 13 13 13 24 0 0 0 37 0 10 20 30 40 Ve ry fav or ab le Fa vo ra ble So m ew ha t fa vo ra ble Ne utr al So m ew ha t u nfa vo ra ble Un fav or ab le Ve ry un fav or ab le No in for ma tio n av aila ble Public Response Pe rc en t o f T M Cs FIGURE 4 Reported public response to blank CMSs during non-incident/non-roadwork periods (n = 59). appeared to be good, and several advantages were noted. Some of the experiences and lessons learned reported by the TMCs regarding CMSs left blank are as follows: • No adverse comments from the public (18 TMCs). • Driver acceptance/satisfactory experience (13 TMCs). • Drivers pay more attention to the message when one is displayed. Messages are more effective when they are displayed. Frequent display of non-essential messages will result in drivers ignoring more important messages (15 TMCs). • The CMSs are seen by some as underutilized (2 TMCs). • CMS conspicuity and message urgency is preserved (1 TMC). • Relevant, timely information enhances driver respect for the CMSs (1 TMC). • Credibility is the key to success (1 TMC). • Reduces the potential for displaying erroneous infor- mation (1 TMC). • A lot of questions and confusion are eliminated by leav- ing the CMSs blank unless posting travel times or other advisory or incident messages (1 TMC). • Time and temperature displayed is not the same as the time and temperature displayed in everyone’s vehicles and a lot of people call to let you know “your time is off” or “your temp is wrong” (1 TMC).

Messages Displayed Public Response to Messages Displayed during Non-Incident/ Non-Roadwork Periods Twenty-three of the 45 TMCs (51%) that regularly or periodically display messages during non-incident/non-roadwork periods as an alternative to leav- ing the signs blank reported a very favorable (22%) or favor- able (29%) public response. Several of these TMCs currently display travel-time information. Twenty percent reported a somewhat favorable (7%) or neutral (13%) response. Comparable to TMCs with blank CMSs, 27% reported that they had insufficient information to comment on the public’s reaction. The public response is summarized in Figure 6. Experiences and Lessons Learned about Displaying Messages Some of the experiences and lessons learned reported by the TMCs regarding messages displayed on CMSs are as follows: • Positive feedback from motorists about travel time (10 TMCs). • If the messages are unfamiliar or too long, motorists sometimes slow down to read messages (3 TMCs). • Public likes that the CMSs are being used and not blank (3 TMCs). • Some motorists like messages; some do not (2 TMCs). • Keep messages simple (2 TMCs). 18 • Display messages that only are related to a motorist’s commute (1 TMC). • We get complaints if we override the trip times with other information, such as “big event this weekend” (1 TMC). • Being able to display travel times has helped us to stay away from safety messages since we do not leave the CMSs blank (1 TMC). • In addition to travel times, we have also displayed public service campaign messages (MOVE OVER FOR STOPPED EMERGENCY VEHICLES, BOOZE AND LOSE IT, CLICK IT OR TICKET) on a very limited basis. These have only been over weekends and only in con- junction with statewide media blitz campaigns. These types of media campaigns are limited to roughly one weekend once a quarter (1 TMC). • We prefer to post positive messages such as TRAFFIC MOVING WELL TO EXIT X to reassure drivers that the system is working. If a sign is blank it could mean that nothing is going on or the sign is broken (1 TMC). • Well received by other agencies (1 TMC). • Leaving the CMSs blank is more effective since it re- moves the tendency for motorists to become complacent. More use is leading to less attentiveness (1 TMC). • Motorists appreciate information disseminated for con- gestion and other motor-related laws (1 TMC). • We have received numerous calls from the public in support of our traffic safety messages (1 TMC). • We get to “exercise” the CMSs and make sure that we keep good preventive maintenance on them (1 TMC). As noted, many TMCs display travel time during non-incident/non-roadwork periods. Several positive com- ments reported by the TMCs will be discussed further in the upcoming section on Travel-Time Messages. Concerns and Challenges The primary concerns expressed and challenges noted by TMCs that display messages other than travel time were (1) uncertainty as to the best types of messages that should be displayed and (2) that when dis- played, the messages are not changed often enough. Plans for Changing Policy of Displaying Messages Plans for the 45 TMCs that regularly or periodically display No 90% Yes 10% FIGURE 5 Percent of TMCs that plan to change policy regarding blank CMSs (n = 59). 22 29 7 13 0 2 0 27 0 10 20 30 40 Ve ry fav or ab le Fa vo ra ble So m ew ha t fa vo ra ble Ne utr al So m ew ha t u nfa vo ra ble Un fav or ab le Ve ry un fav or ab le No in for m ati on av ail ab le Public Response Pe rc en t o f T M Cs FIGURE 6 Public response to displaying messages during non-incident/non-roadwork periods (n = 45). No 93% Yes 7% FIGURE 7 Percent of TMCs that plan to change policy regarding displaying messages during non-incident/ non roadwork periods (n = 45).

19 messages during the peak, off-peak, or both, when incidents or roadwork are not present are summarized in Figure 7. Ninety-three percent stated that they do not have plans to change their policy, whereas 7% plan to change. TRAVEL-TIME MESSAGES Background Types of Time-Related Information As a prelude to discussion of the survey results dealing with travel-time messages, it is important to review the different types of time-related information that could be displayed on CMSs. There are five ways in which time-related information may be displayed on CMSs when a freeway is the driver’s primary facility. They are as follows: 1. Travel time on the freeway, 2. Comparative travel times on the freeway and alternate route, 3. Time saved by taking an alternate route, 4. Delay on the freeway, and 5. Delay avoided by taking the alternate route. Travel time is simply the time in minutes required to traverse from one specified location to another. It does not necessarily imply that a problem of some sort exists on the freeway. The other four types of time-related information are classified as “effects” of some specific problem occurring on the freeway. Travel time and delay refer to the freeway only. Conversely, comparative travel times, time saved, delay, and delay avoided, presume that an alternative route exists and refer to a difference in travel times between the freeway and the alternative route. Travel time and comparable travel-time messages are briefly discussed in the paragraphs that follow. Travel time displayed on a CMS (e.g., 20 MIN TO [desti- nation]) may be easily checked by the driver who scans a watch or dashboard clock. If the actual driving time differs significantly from the promised 20 min, there is concern that credibility of the CMS may be weakened. Some TMCs dis- play a travel time range (e.g., 20–25 MIN TO [destination]) to reduce the possibility of error. It should be understood that many factors could affect the actual travel time of a given vehicle after the driver reads the message on the CMS. For example, it has been observed that travel time between two points on a freeway can significantly change during the transitional period between the off-peak to peak periods or when a crash occurs on the freeway. Thus, there is the concern that the travel time displayed could be significantly incorrect (underestimated). Another considera- tion of displaying travel-time information is that drivers are left to their own resources to decide whether to use the free- way or to divert to an alternative route. There is no hint of whether the alternative route being considered is actually “faster” than the freeway. Travel times can also change significantly between two points on a freeway during the transition between peak and off- peak periods. In these cases, the calculated and posted travel time may be significantly higher than that experienced by drivers. Although erring on the high side (overestimating) is not as critical as erring on the low side for the driver’s trip, the erroneous messages compromise credibility. Messages with comparative travel-time information dis- play the travel times along both the primary and alternative routes to a specific destination point. In contrast to time saved, delay, or delay avoided, comparative travel times would normally be displayed on a separate CMS and not in a message in combination with other freeway-related infor- mation (e.g., ACCIDENT, 2 LANES CLOSED, etc.). Note that comparable travel-time information leaves the driver the task of subtracting one value from the other to deter- mine time savings. It takes longer to read than simply giving the time saved. It requires mental arithmetic and checking which route has which time. The driver must also be careful not to accidentally invert the two routes in performing the arithmetic operation. Furthermore, some analysts believe that most drivers expect CMS messages to advise them as to which route to take. Comparative travel-time information does not do this. Displaying Travel Times In 1991, Dudek and Hutchingson (25) recommended that transportation agencies display travel time on CMSs as an alternative to leaving them blank as soon as the necessary hardware, software, and funding became available. Travel time is very useful to motorists because it gives them some in- dication of the potential arrival time at their destination. Also, travel times can be displayed during the peak and off-peak pe- riods and provides the added advantage that a message will be displayed on the CMS more frequently rather than having the sign blank in the absence of an incident or roadwork. Because of rapidly changing traffic conditions, it is difficult to post travel time information manually. It is more efficient to dis- play travel times automatically using system software. Travel time is generally calculated from speed measure- ments taken at loop detector stations or measured directly with automated vehicle identification or toll tag sensors. A com- puter algorithm calculates the estimated travel time between two points on the highway. It is important to recognize that the data available from these sources are estimated travel times of current conditions (loop detectors) or the travel times of the vehicles that recently traveled between two automated vehicle identification sensor stations. In essence, it is historical travel time. The process of, and algorithms for, accurately predicting the travel times of drivers are not currently available. Although display of travel times is advantageous, the fol- lowing possible credibility issues have created concerns for some CMS operators:

• Display of historic travel times; and • Daily repetition of the same travel times displayed to commuters. First, as previously noted, current technology does not allow TMCs to accurately predict travel times; therefore, recent his- torical travel times are displayed. Motorists can easily measure their own travel times and dispute incorrectly posted travel times. If the posted travel times are not accurate, credibility may be weakened. To circumvent this concern, Houston’s TransStar displays the time of day of the most recent calcula- tion of travel times. Another approach used by TransGuide in San Antonio is to display a range of the estimated travel time. In 2000, Dudek et al. (26) evaluated two alternative travel- time message formats (shown in Exhibit 2) that were similar to those displayed by TransStar and TransGuide. The results in- dicated that displaying recent historical travel times may not pose a credibility issue provided that the differences in ex- pected and actual travel times are not significantly different. The second concern with displaying travel time on a reg- ular basis is the possibility that commuter drivers may see the same travel times posted daily if traffic conditions do not change from day to day and may begin to ignore the CMS at later dates and not read the sign when important incident in- formation is presented—the so-called change blindness effect described previously. To date, no research has been conducted to validate or disprove this concern. In 2004, PBS&J (27) conducted a scan of practice by means of interviews with representatives from 12 state DOTs and 2 FHWA division offices and made the following observations: • In new deployments, seek feedback from, and educate, the public before travel-time messages are instituted; • Travel times must be dynamic; • Travel-time messages can be structured to benefit more than the local traveler; • Messages for travel time should be considered differ- ently from emergency messages; and • Travel times should not be simultaneously provided for both high-occupancy vehicle and general-purpose lanes on the same sign because it is too much information for drivers to absorb. 20 TRAVEL TIME TO I-610 20 MIN AT 8:20 TRAVEL TIME TO US-281 8-12 MINUTES Alternative 1 TranStar (Houston) Alternative 2 TransGuide (San Antonio) Exhibit 2 Does not Display 70% Peak & Off- Peak 24% Peak Period Only 6% 70 77 24 21 6 2 0 10 20 30 40 50 60 70 80 90 Peak Off-Peak Time of Day Pe rc en t o f T M Cs Does Not Display Travel Time & Other Types of Messages Travel Time Only FIGURE 8 Percent of TMCs that display travel time. FIGURE 9 Percent of TMCs that display travel time and other messages during non-incident/non-roadwork periods. Survey Results Display of Travel Time by TMCs The percentage of TMCs that display travel-time messages during the peak and off-peak periods is shown in Figure 8. Thirty percent of the TMCs that responded to the survey dis- play travel-time messages. Twenty-four percent display travel time during both the peak and off-peak periods and 6% dis- play travel time only during the peak period. During the off-peak period, 21% display other types of messages in addition to travel time, and 2% only display travel time. The percentage of TMCs that only display travel time and those that display other messages in addition to travel time are shown in Figure 9. The specific TMCs that display travel time are given in Table 5. The frequency of display of travel-time messages during the peak and off-peak periods is shown in Figure 10. The results show that 26% of the TMCs regularly display travel time during the peak periods and 18% regularly display travel time during the off-peak period. A summary of the relative number of CMSs on which TMCs display travel-time information during the peak and off-peak periods is shown in Figure 11. The results show that, as a rule, most TMCs that display travel time do not dis- play these messages on all of their CMSs. During the peak period, only 3% of the TMCs display travel time on all CMSs and only 5% display travel time on most CMSs. Likewise, only 3% display travel time on all the CMSs during the off- peak period, whereas 4% display travel time on most CMSs.

21 Agency/Traffic Management Center Location Peak Off-Peak California Department of Transportation • Irvine • Los Angeles • Oakland • Sacramento • San Bernardino • San Diego X X X X X X X X X X X X X X X X X Colorado DOT • Golden (statewide) X Delaware Department of Transportation (statewide) X Florida DOT • Orlando X Georgia DOT • Atlanta X Harris County Toll Road Authority X X Illinois DOT • Chicago X Illinois Tollway X X Kansas DOT • Leeís Summit (bi-state) X Kentucky Transportation Cabinet • Louisville X Louisiana DOTD • Baton Rouge • Shreveport X X X X Minnesota DOT • Minneapolis–St. Paul X North Carolina DOT • Raleigh X Oklahoma • Oklahoma City X Oregon DOT • Portland X Tennessee DOT • Knoxville • Nashville X X X X Texas DOT • Fort Worth • Houston • San Antonio X X X X X X Utah DOT • Salt Lake City X WSDOT–NW Region • Bellingham • Seattle X X X X Wisconsin (statewide) X X *The TMC in Kansas City, Missouri, began displaying travel time after the survey was completed and is not shown in the table. TABLE 5 TMCs THAT DISPLAY TRAVEL TIME*

Reasons for Not Displaying Travel Time The primary reasons cited as to why travel-time information was not displayed were that (1) infrastructure and/or software was not available and (2) congestion was not a problem. A summary of the specific reasons given follows: • Lack of funding, lack of adequate infrastructure and/or software to support display of travel times (18 TMCs). • Congestion is not a problem where CMSs are located (11 TMCs). • Detectors lose their accuracy too quickly and need to be recalibrated. This is a time-consuming process (1 TMC). • Currently displaying messages manually; travel time needs to be displayed automatically (1 TMC). • Problems with confidence in data from detectors (1 TMC). • Freeway system is too small to display travel times (1 TMC). • CMSs are not in appropriate locations where travel-time messages would be useful (1 TMC). Public Response Sixty percent of the TMCs reported that they received a very favorable response (43%) or favorable response (17%) to travel-time messages. Thirty-seven percent reported that public responses have been somewhat favorable (27%) or neutral (10%). Only 3% reported that they had insufficient information to judge the reaction from the public. A sum- mary of the reported responses from the public about travel- time messages is shown in Figure 12. 22 Examples of Typical Travel-Time Messages Exhibit 3 provides examples of typical messages that were reported by the 30 of 100 TMCs that display travel times. The examples in Exhibit 3 indicate that there is little uni- formity with respect to message format among the TMCs. In addition, many of the messages far exceed the maximum number of units of information that should be displayed to allow unfamiliar drivers to read and comprehend the messages while traveling at typical freeway speeds, and will result in information overload. However, if the same travel-time messages were displayed every day on a given CMS, familiar drivers would quickly learn that the only in- formation that changes is the actual time to the destinations, and they would learn to ignore the other elements in the message. Experiences and Lessons Learned Some of the TMCs that display travel-time information com- mented on the need to make sure that the travel times are accurate. Also, experiences indicated that it is difficult to dis- play accurate travel times during rapidly deteriorating traffic operating conditions (e.g., transition between off-peak and peak periods and occurrence of incidents). A summary of the experiences and lessons learned with displaying travel time follows: • When displayed, travel times need to be accurate; CMS system credibility is at risk and drivers may not react positively to other travel-time messages (6 TMCs). • It is very difficult to accurately calculate travel times during rapidly deteriorating conditions. To address this problem, travel-time messages should be immediately deleted during situations that will likely result in rapidly deteriorating conditions (4 TMCs). • Messages must be short (2 TMCs). • Some drivers are requesting that travel-time messages be displayed during construction (2 TMCs). 70 77 26 18 3 21 30 0 10 20 30 40 50 60 70 80 90 Peak Off-Peak Time of Day Pe rc en t o f T M Cs Does Not Display Regularly Often Occasionally Rarely FIGURE 10 Travel time display frequency. 70 77 3 35 42 1 18 12 2 3 0 10 20 30 40 50 60 70 80 90 Peak Off-Peak Time of Day Pe rc en t o f T M Cs Does Not Display All CMSs Most CMSs About 50% of CMSs Some CMSs Very few CMSs FIGURE 11 Relative number of CMSs on which travel-time messages are displayed. 43 17 27 10 0 0 0 3 0 10 20 30 40 50 Ve ry fav or ab le Fa vo ra ble So m ew ha t fa vo ra ble Ne utr al So m ew ha t u nfa vo ra ble Un fav or ab le Ve ry un fav or ab le No in for ma tio n a va ilab le Public Response Pe rc en t o f T M Cs FIGURE 12 Public responses to travel-time messages (n = 30).

23 • Drivers were slowing down to read the signs (2 TMCs). • Initial deployment of travel-time messages caused addi- tional congestion (1 TMC). • Check the times regularly to make sure they are accu- rate (1 TMC). • If the travel times are thought to be inaccurate by the TMC staff, the messages should be deleted while investigat- ing the inaccuracy (1 TMC). • We hear complaints when times are not being displayed as a result of special events or ozone alerts. We no longer post ozone alerts (1 TMC). Exhibit 3 For Single Destination EST. TIME TO WADE AVE. 15 TO 17 MIN MAIN ST/EXIT 12 8 MILES AHEAD TRAVEL TIME 10–12 MIN DOWNTOWN LOOP 12 MILES TRAVEL TIME 11–13 MIN Example 1 Example 2 Example 3 I-40 JCT TRAVEL TIME 8–10 MIN 8 MILES TRAVEL TIME TO US-90 6 MIN AT 11:12 Example 4 Example 5 For Two Destinations TRAVEL TIME TO I-294 10 MIN I-215 10 MIN DOWNTOWN 40 MIN FREEWAY TIME TO DOWNTOWN 15 MIN HAWLEY 8 MIN TRAVEL TIME TO LOOP 410 10–12 MIN IH-35 20–22 MIN Example 1 Example 2 Example 3 TRAVEL TIME GREEN ST EXIT 15 MIN LYNNWOOD 15 MIN EVERETT 32 MIN Example 4 Example 5 For Single Destination with Two Alternative Routes TIME TO WEST HILLS ON 40 W ON 640W 15–17 MIN 16–18 MIN SEATTLE VIA SR-520 VIA I-90 22 MIN 17 MIN Example 1 Example 2

• It is not necessary to display travel times on all CMSs. Some CMS locations do not lend themselves to travel- time messages (1 TMC). • Traffic- and incident-related messages take priority over travel-time messages (1 TMC). • For some locations, we installed additional detection equipment to enhance the accuracy of the travel-time calculations (1 TMC). • Any time we employ a new CMS for travel times, a few motorists complain that it’s slowing traffic down. The complaining lasts only a day or two (1 TMC). • Displaying travel times on CMSs manually is very difficult and can cause considerable delay in posting. It is best to have an automated system to display travel time (1 TMC). • About 20% of comments from the public were asking for travel times on alternate routes so they can make route choices (1 TMC). Concerns and Challenges The primary concerns and challenges dealt with ensuring that the travel times displayed are accurate because of loss of in- frastructure and problems with rapidly deteriorating traffic conditions during certain times of the day and when incidents occur. A list of the comments received follows: • Accuracy is a concern owing to poor reliability of detec- tors, loss of communications, or inadequate spacing of detectors. Ensuring that detectors are calibrated regu- larly and maintained is a major concern (8 TMCs). • Travel-time information is not very accurate during rapidly deteriorating traffic conditions such as transition from off-peak to peak traffic flows or when incidents occur. The lag in recognition time during these condi- tions causes the displayed times to be too short during the window of time characterized by rapid deterioration (3 TMCs). • Guidance is needed on the best message content and format (1 TMC). • The travel time experienced by drivers can be longer than that shown on the CMS. Need information on the best ways to educate the public that the travel times displayed are not yet predictive. Also, how will we ever get the public to understand that if an incident happens after they pass under the sign that the infor- mation was correct at the time they saw the message (1 TMC)? • Current system does not allow us to automatically update travel times. Displaying travel times manually is very difficult and often results in inaccurate travel times dis- played on CMSs (1 TMC). • Guidance is needed on the frequency that travel times should be updated during congested periods (1 TMC). • There are concerns about driver complacency after see- ing the same travel time information every day because traffic conditions are the same each day (1 TMC). 24 • We have an internal struggle over the format for travel- time messages and whether the travel-time values should be rounded to the nearest 5-min interval (1 TMC). • It is not clear whether a two-phase message should be used when travel times are displayed (1 TMC). Update Frequency of Travel Times The agencies and TMCs were asked to comment on the fre- quency with which they update the travel-time messages dis- played on CMSs. The results of the 30 TMCs that responded are shown in Figure 13. A total of 77% of the TMCs update the travel time dis- played on CMSs within 5 min. Seventeen percent update every minute, 20% every 2 min, 27% every 3 min, and 13% every 5 min. Thirteen percent of the TMCs change the travel times when traffic conditions change. Interestingly, 3% update every 10 min, 3% every 15 min, and 3% update every 30 min. Cost to Implement Travel-Time Messages The cost to develop and add software to existing systems to dis- play travel time messages varied widely and ranged between $30,000 and $250,000. The difference in cost was due pri- marily to each TMC’s current infrastructure and software capabilities. Reasons for Not Displaying Travel Time The agencies and TMCs that do not currently display travel- time messages were asked to cite the reasons for their deci- sions. The results are shown in Figure 14. Forty-four percent of the 70 respondents reported that they did not have the infrastructure to accurately calculate travel times, 29% stated that they did not have the necessary software, 16% were systems in rural areas and did not have the need and facilities to display travel time, and 10% re- ported that they did not have sufficient recurrent congestion to justify displaying travel time. One TMC (1%) reported 17 20 27 13 3 3 3 13 0 50 10 15 20 25 30 Time Interval Up da te In te rv al 1 Minute 2 Minutes 3 Minutes 5 Minutes 10 Minutes 15 Minutes 30 Minutes As Conditions Dictate FIGURE 13 Update frequency of travel-time information displayed on CMSs (n = 30).

25 that travel time was not displayed owing to the lack of fund- ing to implement the necessary hardware and software. CONGESTION MESSAGES Background Sometimes CMSs are used to present information about traf- fic conditions when the freeway becomes congested. The dif- ficulty in doing so involves the large continuum of possible traffic operational conditions that are very difficult to describe on CMSs. Additionally, if a multitude of descriptors are used during the peak period in attempts to describe congested traf- fic conditions, then it becomes very difficult for the CMS message designer and CMS operator to think of words to describe congested conditions when incidents occur. For example, if the term HEAVY CONGESTION is displayed during the peak period, then what descriptor should be used to describe an even more congested state when an incident occurs? Human factors laboratory studies were conducted as early as the mid-1970s to evaluate descriptors for levels of conges- tion. Letter grades (A, B, C, D, and F), number coding (1 to 10), and word descriptors (e.g., CONGESTED, MODERATE CON- GESTION, HEAVY CONGESTION, and JAMMED TRAFFIC) were studied in laboratory studies conducted at different locations in the United States (2). The studies revealed that the subjects could not associate letter grades or number codes with the degree of congestion. In addition, there were incon- sistencies between small and large city motorists as to the rel- ative level of congestion associated with the word descriptors. Interpretation problems with the letter grades were also found in actual field studies conducted in Houston and reported by Stockton et al. (28). In 2001, FHWA’s Johnson (29) noted that care must be exercised in displaying messages that provide motorists with information about the state of traffic. Messages such as CONGESTION AHEAD, NORMAL TRAFFIC, or EXPECT DELAYS do not offer meaningful or useful information to travelers and contribute to the erosion of the public’s trust in these systems and of the information they provide. Inaccurate, incomprehensible, or inappropriate information displayed on a 44 16 10 1 29 0 10 20 30 40 50 Reason Pe rc en t o f T M Cs Lack of Infrastructure Lack of Software Rural Area Little or No Recurrent Congestion Lack of Funding FIGURE 14 Reasons for not displaying travel time (n = 70). Does Not Display 65% Peak & Off- Peak 31% Peak Period Only 4% FIGURE 15 Display of congestion messages during non-incident/non-roadwork periods as an alternative to leaving the CMSs blank. 65 69 12 98 5 12 12 3 5 0 10 20 30 40 50 60 70 80 Peak Off-Peak Time of Day Pe rc en t o f T M Cs Does Not Display Regularly Often Occasionally Rarely FIGURE 16 Display frequency of congestion messages during non-incident, non-roadwork periods as an alternative to leaving the CMSs blank. CMS can cause motorists to question the credibility and ignore all CMS messages. CMSs are also used to display the limits of congestion— the locations of the beginning and end of congestion. Dis- playing the limits of congestion is useful because it helps the familiar motorist understand the extent of the problem and assess where to return to the freeway if the motorist decides to avoid the congestion by routing around the problem (7). Survey Results Display of Congestion Messages The percentage of TMCs that regularly or periodically display congestion messages during non-incident/non-roadwork periods as an alternative to leaving the CMSs blank is shown in Figures 15 through 17. Figure 15 shows that 65% of the TMCs do not display congestion messages, 31% display con- gestion messages during both peak and off-peak periods, and 4% display congestion messages only during the peak period. As shown in Figure 16, 65% of the 100 TMCs reported that they do not regularly or periodically display congestion mes- sages in peak periods during non-incident/non-roadwork peri- ods, whereas 35% of the TMCs regularly or periodically display congestion messages. Only 12% of the TMCs indicated that they regularly display congestion messages and only 8% often display congestion messages as an alternative to leaving the CMSs blank during non-incident/non-roadwork periods.

During the off-peak, 69% of the TMCs reported that con- gestion messages are not displayed during non-incident/ non-roadwork periods. Thirty-one percent stated that con- gestion messages are regularly or periodically displayed. Of these, 14% display congestion messages regularly (9%) or often (5%), and 17% display congestion messages occasionally (12%) or rarely (5%). The results in Figure 17 show that during peak period congestion messages are displayed on all of the CMSs by only 2% of the TMCs. Nine percent of the TMCs display congestion messages on most CMSs and 5% on about 50% of CMSs. Similarly, during off-peak periods only 2% of the TMCs display congestion messages on all of CMSs, 10% on most of CMSs, and 1% on about 50% of CMSs. Basis for Decision to Display Congestion Messages The basis for the decision of the 35 TMCs to display conges- tion messages during non-incident/non-roadwork periods is summarized in Table 6. As the results show, the decision of 26 Basis Percent of TMCs Percent of TMCs Agency administrative/upper-management preference only 14 TMC manager/supervisor preference only 20 Preference Only Agency administrative/upper-management preference, and TMC manager/supervisor preference 29 63 Preference and Feedback Agency administrative/upper-management preference, and TMC manager/supervisor preference, and Feedback from telephone calls, newspapers, radio, and/or television 25 25 TMC manager/supervisor preference, and Research conducted by agency and/or research conducted by others 9 Research conducted by agency only 0 Research conducted by others only 0 Preference and/or Research or Focus Group Focus group studies only 3 12 100 100 TABLE 6 BASIS FOR DECISION TO DISPLAY CONGESTION MESSAGES DURING NON-INCIDENT, NON-ROADWORK PERIODS AS AN ALTERNATIVE TO LEAVING THE CMSs BLANK (n = 35) 23 26 20 0 3 00 29 Public Response 0 10 20 30 40 Ve ry fav or ab le Fa vo ra ble So m ew ha t fa vo ra ble Ne utr al So m ew ha t u nfa vo ra ble Un fav or ab le Ve ry un fav or ab le No in for m ati on av aila ble Pe rc en t o f T M Cs FIGURE 18 Public response to congestion messages during non-incident, non-roadwork periods (n = 35). 65 69 2 2 9 105 1 15 10 4 8 No CMSs All CMSs Most CMSs About 50% of CMSs Some CMSs Very Few CMSs 0 10 20 30 40 50 60 70 80 Peak Off-Peak Time of Day Pe rc en t o f T M Cs FIGURE 17 Number of CMSs on which congestion messages are displayed as an alternative to leaving the CMSs blank. 63% of the 35 TMCs was based only on agency administrative/ upper-management preference, TMC manager/supervisor preference, or both. An additional 25% was influenced by feedback from telephone calls, newspapers, radio, and/or television. Therefore, a total of 88% of the 35 TMCs did not base this decision on objective results from focus group or other research studies. As shown in Table 6, only 12% made decisions based on information from focus group or other research results. Public Response A summary of the reported public response to congestion messages during non-incident/non-roadwork periods is shown in Figure 18. Forty-nine percent of the 35 TMCs indicated that they received very favorable (23%) or favorable (26%) comments from the public. Somewhat favorable responses were reported by 20% of the TMCs. Twenty-nine percent of the TMCs had no information regarding public response to congestion messages.

27 Examples of Typical Congestion Messages Exhibit 4 shows examples of typical messages that were re- ported by the 35 of 100 TMCs that regularly or periodically display congestion messages rather than leaving CMSs blank. Experiences and Lessons Learned The experiences and lessons learned most noted by the TMCs that display congestion messages was that the messages must be accurate and up to date. A summary of comments from TMCs regarding congestion messages follows: • Messages must be accurate and kept up to date as the queue changes (11 TMCs). • Display the beginning and end points of congestion to make it valuable to motorists (2 TMCs). • Turn message off when there is not a need for a con- gestion message (2 TMCs). • Motorists do not like to see messages saying simply CONGESTION AHEAD (1 TMC). • Keep the message brief (1 TMC). • Do not display these messages during routine commute times, where congestion occurs on a regular day-to-day basis, but only when congestion occurs where it is not anticipated (1 TMC). • If messages are displayed at locations where recurring congestion occurs on a daily basis, commuters may ne- glect the messages as it shows on CMS every day at same time (1 TMC). • Although the local daily commuters understand the traffic flow scenarios associated with interstate travel in the area, through-traffic commuters do not have the same prevailing knowledge of the roadway topol- Exhibit 4 CONGESTION EXIT 10–EXIT 20 CONGESTION I-55 TO STATE LINE EXPECT DELAYS HEAVY CONGESTION I-64E TO I-264 USE ALT RT I-71N Example 1 Example 2 Example 3 CONGESTION AHEAD FROM WURZBACH TO CROSSROADS STOP AND GO TRAFFIC 1 MILE AHEAD EXPECT DELAYS CONGESTION NEXT 1 1/2 MILES EXPECT DELAYS Example 4 Example 5 Example 6 CONGESTION AT 11TH ST USE CAUTION CONGESTION AT WEBER PREPARE TO STOP CONGESTION AHEAD PLEASE USE CAUTION Example 7 Example 8 Example 9 CONGESTION AHEAD 5 MILES HEAVY TRAFFIC AHEAD CONSIDER ALT RTES CONGESTION 2 MILES AHEAD TO BUTLER BLVD Example 10 Example 11 Example 12

ogy and issues created by congestion in this area. Pro- viding congestion information to motorists allows the local commuter to consider an alternate route or arterial and provides additional information to unfamiliar mo- torists so that they may drive more safely on roadways with which they are not familiar (1 TMC). • Traffic center technicians have to have regional and geo- graphical knowledge. They must have all available geo- graphical information (1 TMC). • We have learned to vary congestion messages to re- move motorist complacency (1 TMC). • Motorists need to know what is ahead of them. My theory is that the signs can have a “calming effect” on the motorist. No one likes surprises (1 TMC). • Limited number of CMS locations limits amount of in- formation to motorists (1 TMC). • Motorists may overlook CMSs when messages are reg- ularly displayed (1 TMC). • Monitor congestion even after operation hours (1 TMC). • Any message creates somewhat of a hazard because mo- torists instinctively brake to read the entire message. That is why we try to stay with one-phase messages (1 TMC). • Public feedback indicates that motorists favor queue length messages rather than delay times (1 TMC). • Congestion messages are good for unusual circum- stances when motorists would not expect heavy con- gestion at that time of day or place (1 TMC). • There is perception that on certain sections of the high- way, motorists tend to slow down prior to a CMS to read the message on the sign. The congestion message on the sign can thereby contribute to the existing delay (1 TMC). • Ample distance between incident and sign is very im- portant (1 TMC). • Trucking firms and truckers like travel times rather than congestion messages (1 TMC). • Congestion messages are favored by trucking, taxi ser- vices, and professional drivers (1 TMC). Concerns and Challenges Several concerns and challenges were noted by the TMCs. A summary of the concerns and challenges follows: • One challenge is ensuring that accurate information is displayed at all times (4 TMCs). • A major challenge is an understanding of the differences between the various levels of congestion (e.g., heavy, moderate, and light) (1 TMC). • Giving drivers up-to-date location and duration/length of congestion is difficult even with cameras and sensors. Traffic conditions may change quickly while the opera- tor is responding to a crash or other problem (1 TMC). • Often numerous small incidents happen that will clear too quickly for specific messages but will cause residual delays; therefore, generic messages needed but still have to be different enough in hot spots to get motorists’ atten- tion from day to day (1 TMC). 28 • The nature of urban congestion is such that when you try to report on the location of congestion (e.g., “Delays to Exit 290”) by the time you enter the message, the end of congestion has moved—like a caterpillar (1 TMC). • In areas where there are no cameras getting updates, it is a challenge to ensure the messages are up to date (1 TMC). • A concern and challenge is how to measure congestion to display accurate information (e.g., delay time and queue length) (1 TMC). • Displaying congestion messages consistently during all hours (1 TMC). SPEED MESSAGES Background Travel speed does not rank high on the list of information needs of motorists when incidents occur on the freeway or roadwork is affecting traffic flow. In addition, unless the speed information is displayed in a correct format, it is likely that some motorists will misinterpret the information and some may react in adverse ways. Therefore, travel speed should not be displayed as part of an incident or roadwork message. In an early study in 1971, Dudek et al. (30) found that travel speed between various reference points ahead was not consid- ered as important to motorists as the location and length of congestion, degree of congestion, and the reason for the con- gestion. In another study conducted in Houston and Dallas in 1978 (31), traffic speed was considered less important by com- muters than the type of incident, lane blockage, level of con- gestion, amount of delay, and the location of the nearest exit. In 1971, based on their studies in the Los Angeles area, Hulbert and Beers (32) concluded that referencing to specific miles per hour could elicit dangerous overreaction (deceler- ation) from motorists. In a 1998 study in Toronto, Smiley and Dewar (33) found that in response to the message phrase SPEED 30–40 KM/H NEXT 3 KM, only a slight majority of test subjects in Toronto thought that the speed message re- ferred to the average speed, which is what this message was intended to convey. However, almost as many thought the message referred to a recommended speed, and almost one- fifth thought it referred to the legal speed limit. Results from a study by Vercruyssen in 1997 (34) in Minneapolis on CMS messages for application in highway work zones showed that the most common (67% of the subjects) interpretation of the message CURRENT SPEED/30 MPH was that the speed limit through the construction zone was 48 km/h (30 mph), an incorrect interpretation. Thus, if downstream travel speeds are to be displayed on CMSs to indicate the degree of con- gestion, it is important that the information be specific so that motorists realize that the speeds posted are not speed limits. Another factor to consider is that speed information is not the best choice of alternative traffic descriptors to elicit diversion. In the 1998 Toronto study by Smiley and Dewar (33), the message SPEED 20–40 KM/H (12–25 MPH)

29 resulted in a lower percentage of subjects who indicated they would divert than did the message HEAVY CONGESTION. Several studies have been conducted to evaluate advisory speed messages on CMSs in highway work zones. Results of controlled field studies by Richards et al. (35) in Texas indi- cated that advisory speed messages reduced average vehicle speeds by only 7% regardless of the speed posted. Based on these findings, Richards and Dudek (36) recommended that speed messages on CMSs in work zones should be only 8–16 km/h (5–10 mph) lower than the normal operating speed on urban freeways, and 8–24 km/h (5–15 mph) on rural freeways. Survey Results Display of Speed Messages The percentage of TMCs that regularly or periodically display speed messages as an alternative to leaving the CMSs blank is shown in Figure 19. Eighty-five percent of the TMCs do not display speed messages. In contrast, 15% display speed— 14% during both peak and off-peak periods and 1% only dur- ing the peak period. As shown in Figure 20, 12% of the TMCs rarely display speed messages during both the peak and off-peak periods; whereas only 2% regularly display speed messages during the peak, and 1% during the off-peak periods. When used, speed messages are displayed on only a very few CMSs (see Figure 21). Basis for Decision to Display Speed Messages The basis for the decision of each of the 15% of TMCs that regularly or periodically display speed messages as an alter- native to leaving the CMSs blank is summarized in Table 7. Interestingly, the decisions were based solely on administrative/ upper-level management (12%) or TMC manager/supervisor preference (88%). In no case was the decision based on objective data, such as results of focus groups or other types of studies. Public Response The public response reported by the TMCs that display speed messages is shown in Figure 22. Eighty percent of the TMCs had no information from the public to justify the decision to display speed messages and only 7% indicated that they re- ceived a favorable response. Examples of Typical Speed Messages Exhibit 5 shows examples of typical messages that were re- ported by the 15 of 100 TMCs that regularly or periodically display speed messages rather than leave CMSs blank. Experiences and Lessons Learned Very few of the TMCs that display speed provided comments regarding experiences and lessons learned. The two comments that were received follow: • Need regular input from the maintenance crews in the field to keep the speeds appropriate for conditions in rural areas. • It is more feasible to display speeds in 19 km/h (10 mph) increments for our situations. Does Not Display 85% Peak & Off- Peak 14% Peak Period Only 1% 85 85 2 10 01 2 12 12 0 10 20 30 40 50 60 70 80 90 Peak Off-Peak Time of Day Pe rc en t o f T M Cs Does Not Display Regularly Often Occasionally Rarely FIGURE 19 Display of speed messages during non-incident/ non-roadwork periods as an alternative to leaving the CMSs blank. FIGURE 20 Display frequency of speed messages during non-incident/non-roadwork periods as an alternative to leaving the CMSs blank. 85 85 0 1 0 0 1 202 12 12 No CMSs All CMSs Most CMSs About 50% of CMSs Some CMSs Very Few CMSs 0 10 20 30 40 50 60 70 80 90 Peak Off-Peak Time of Day Pe rc en t o f T M Cs FIGURE 21 Number of CMSs on which speed messages are displayed during non-incident/non-roadwork periods as an alternative to leaving the CMSs blank.

Concerns and Challenges Only three comments were received regarding concerns and challenges. Specific comments follow: • A concern and challenge is maintaining credibility. • Challenge is keeping the speed limit in sync with weather conditions and traffic flow. 30 • It is difficult to convey that the speed is an average and the speed may differ along the freeway. PUBLIC SERVICE ANNOUNCEMENTS Background PSAs do not provide drivers with real-time safety or travel efficiency information. PSAs provide motorists with informa- tion that can be given more effectively through other methods such as media campaigns or pamphlets (5–7). These, and other methods, would benefit a greater majority of the motoring public because they would not be limited to only those who travel on freeways with CMSs. One argument in support of not displaying PSAs is the concern that motorists who continually travel a specific route will become accustomed to the PSA and then begin to ignore the CMSs. Subsequent messages indicating lane closures, de- tours, etc., that directly affect the motorists’ travels may then tend to be unnoticed. This is another example of the potential for the change blindness phenomenon. If there is a concern that the CMSs are infrequently used, then it may be desirable to display other information that may affect the motorists’ travel (e.g., existing or planned roadwork on the specific fa- cility or on other intersecting freeways, expressways, or toll roads; or travel-time information) rather than PSAs. 0 0 7 7 6 0 0 80 0 10 20 30 40 50 60 70 80 90 Public Response Ve ry fav or ab le Fa vo ra ble So m ew ha t fa vo ra ble Ne utr al So m ew ha t u nfa vo ra ble Un fav or ab le Ve ry un fav or ab le No in for ma tio n a va ilab le Pe rc en t o f T M Cs FIGURE 22 Public response to speed messages during non-incident/non-roadwork periods as an alternative to leaving the CMSs blank (n = 15). Exhibit 5 DENSE FOG ADVISE 30 MPH SPEED LIMIT 55 DRIVE SAFELY I-285 CONDITIONS WEST EAST 40 MPH 50 MPH Example 1 Example 2 Example 3 Ba si s Percent of TMCs Percent of TMCs Agency administrative/upper-management preference only 12 TMC manager/supervisor preference only 88 Preference Only Agency administrative/upper-management preference, and TMC manager/supervisor preference 0 100 Preference and Feedback Agency administrative/upper management preference, and TMC manager/supervisor preference, and Feedback from telephone calls, newspapers, radio, and/or television 0 0 TMC manager/supervisor preference, and Research conducted by agency and/or research conducted by others 0 Research conducted by agency only 0 Research conducted by others only 0 Preference and/or Research or Focus Group Focus group studies only 0 0 100 100 TABLE 7 BASIS FOR DECISION TO DISPLAY SPEED MESSAGES DURING NON-INCIDENT/ NON-ROADWORK PERIODS AS AN ALTERNATIVE TO LEAVING THE CMSs BLANK (n = 15)

31 Public Response The reported public response is summarized in Figure 26. Only 17% of the TMCs that display PSAs as an alternative to leaving the CMSs blank reported that they received a favorable response from the public. An additional 37% re- ported that the response was somewhat favorable (10%) or neutral (27%). Three percent reported unfavorable response and 43% had no information available to comment on the public’s reaction. Examples of Typical PSA Messages Exhibit 6 shows examples of typical messages that were re- ported by the 30 of 100 TMCs that regularly or periodically display PSAs rather than leaving CMSs blank. Experiences and Lessons Learned The experiences and lessons learned reported by the TMCs that display PSAs centered around (1) the importance of wording, (2) not displaying during the peak traffic flow, (3) limiting the duration of message display, (4) displaying relevant PSAs, and (5) coordination with other agencies. Specific reported experiences and lessons learned are as follows: Does Not Display 70% Peak & Off- Peak 16% Off-Peak Only 14% FIGURE 23 Display of PSAs during non-incident/non-roadwork periods as an alternative to leaving the CMSs blank. 83 70 3 34 58 17 2 5 0 10 20 30 40 50 60 70 80 90 Peak Off-Peak Time of Day Pe rc en t o f T M Cs Does Not Display Regularly Often Occasionally Rarely FIGURE 24 Display frequency of PSAs during non-incident/ non-roadwork periods as an alternative to leaving the CMSs blank. 83 70 4 85 80 1 6 102 3 0 10 20 30 40 50 60 70 80 90 Peak Off-Peak Time of Day Pe rc en t o f T M Cs No CMSs All CMSs Most CMSs About 50% of CMSs Some CMSs Very Few CMSs FIGURE 25 Number of CMSs on which PSAs are displayed during non-incident/non-roadwork periods as an alternative to leaving the CMSs blank. A second argument is the potential negative response by the public when PSAs are displayed because the information is nonessential. As discussed previously, Caltrans personnel in the Los Angeles TMC discontinued displaying PSAs as a result of negative feedback from the public (D. Roper, former Deputy District Director, Operations, Caltrans, personal communication, Mar. 21, 2002). Survey Results Display of PSAs Seventy percent of the TMCs reported that they do not display PSAs as an alternative to leaving CMSs blank, 16% display PSAs during both peak and off-peak periods, 1% during the peak period only, and 14% only during the off-peak period. The percentage of TMCs by the time of day during which PSAs are displayed is summarized in Figure 23. The general frequencies of PSAs displayed during the peak and off-peak period are shown in Figure 24. A higher percentage of TMCs do not display PSAs during the peak period in comparison with off-peak (83% versus 70%). Three percent reported that they regularly display PSAs dur- ing the peak and off-peak periods. The trend for the TMCs that display PSAs is to display the messages occasionally or rarely. As shown in Figure 25, PSAs are displayed on all CMSs by 4% of the TMCs during the peak period and 8% during the off-peak period. Five percent of TMCs display PSAs on most CMSs during the peak and 8% on most CMSs during the off- peak period. Basis for Decision to Display PSAs The basis for each TMC’s decision to display PSAs as an alternative to leaving CMSs blank is shown in Table 8. The decision by 93% of the TMCs that display PSAs was based solely on agency administrative/upper-management or TMC manager/supervisor preference, or both. Only 7% of the TMCs made their decisions based on information received from research.

32 • A message should be run for a short period of time and/or should not be displayed too often (8 TMCs). • Only display for selected, well-organized and advertised statewide campaigns (3 TMCs). • Messages must provide value to motorists (2 TMCs). • We display the PSAs typically on one CMS per corridor, with travel times being displayed on all others available (1 TMC). • Coordinate any PSA with local government (1 TMC). Concerns and Challenges Concerns and challenges expressed by the TMCs included the lack of clear guidance as to what messages constitute PSAs and how to display consistent messages. In addition, concern was expressed about the requests from other agencies to display 0 17 10 27 0 3 0 43 0 10 20 30 40 50 Public Response Ve ry fav or ab le Fa vo ra ble So m ew ha t fa vo ra ble Ne utr al So m ew ha t u nfa vo ra ble Un fav or ab le Ve ry un fav or ab le No in for m ati on av aila ble Pe rc en t o f T M Cs FIGURE 26 Public response to PSAs during non-incident/ non-roadwork periods as an alternative to leaving the CMSs blank (n = 30). Basis Percent of TMCs Percent of TMCs Agency administrative/upper-management preference only 57 TMC manager/supervisor preference only 13 Preference Only Agency administrative/upper-management preference, and TMC manager/supervisor preference 23 93 Preference and Feedback Agency administrative/upper-management preference, and TMC manager/supervisor preference, and Feedback from telephone calls, newspapers, radio, and/or television 0 0 TMC manager/supervisor preference, and Research conducted by agency and/or research conducted by others 7 Focus group studies only 0 Research conducted by agency only 0 Preference and/or Research or Focus Group Research conducted by others only 0 7 100 100 TABLE 8 BASIS FOR DECISION TO DISPLAY PSAs DURING NON-INCIDENT/NON-ROADWORK PERIODS AS AN ALTERNATIVE TO LEAVING THE CMSs BLANK (n = 30) Exhibit 6 REPORT DWI 1-877-DWI-HALT BLOOD DRIVE HINSDALE OASIS VAN AND CARPOOL CALL 1-800-555-5555 Example 1 Example 2 Example 3 NEW ORLEANS SHUTTLE INFO 1-877- XXX- XXXX AIR QUALITY ALERT TODAY TUNE TO 530 AM BURN BAN IN EFFECT FOR NUECES CTY Example 4 Example 5 Example 6

33 non-traffic-related messages if PSAs are posted. Some TMCs voiced differences of opinion with agency administration/ management about displaying PSAs. Specific comments re- ceived are as follows: • They [PSAs] open the door for other requests that are not transportation related. Denying requests is a problem (2 TMCs). • A concern and challenge is how to display consistent messages (2 TMCs). • Requests for other types of messages often reference the PSAs as the reason (2 TMCs). • Operations staff not wanting to use the signs for this purpose versus administrative requests (2 TMCs). • Lack of clear, concise direction on what qualifies as a PSA (1 TMC). • Wording is critical (1 TMC). SAFETY CAMPAIGN MESSAGES Background Messages that assist the state DOT in improving highway safety are often displayed on CMSs. Normally, these mes- sages are part of statewide safety campaigns. The display of safety messages associated with a safety campaign is allow- able under the current MUTCD (1, §2A07). Safety-related messages should be current, displayed for a limited time, and should relate to a specific safety campaign (17). No reported research studies that objectively address the ef- fects of driver safety campaign messages on driver credibility and change blindness were found in the literature. However, in preference studies conducted by Benson (37) of 517 motorists in the Washington, D.C., area, 67% said that general traffic safety messages should be posted on CMSs. Survey Results Display of Safety Campaign Messages Fifty percent of the TMCs reported that they display safety campaign messages during both the peak and off-peak peri- ods and 33% display the messages only during the off-peak period. Only 17% reported that they do not display cam- paign messages. The distribution of the time of day during which safety campaign messages are displayed is shown in Figure 27. The display frequency of safety campaign messages is summarized in Figure 28. During the peak period, 47% of TMCs do not display safety campaign messages. The mes- sages are displayed regularly by 8% of TMCs, displayed often by 7%, occasionally by 28%, and rarely by 10%. During the off-peak period 17% do not display safety campaign mes- sages, 8% regularly display the messages, 7% display often, 53% occasionally, and 15% rarely. During the peak periods, the majority of TMCs that display safety campaign messages post the messages on all CMSs and most CMSs. Similarly, the majority of TMCs display safety campaign messages on all CMSs and most CMSs during the off-peak periods, as shown in Figure 29. Basis for Decision to Display Safety Campaign Messages Ninety-nine percent of the TMCs reported that the decision to display safety campaign messages was based on agency Does Not Display 17% Peak & Off- Peak 50% Off-Peak Only 33% FIGURE 27 Display of safety campaign messages during non-incident/non-roadwork periods as an alternative to leaving the CMSs blank. 47 17 8 87 7 28 53 10 15 0 10 20 30 40 50 60 Peak Off-Peak Time of Day Pe rc en t o f T M Cs Does Not Display Regularly Often Occasionally Rarely FIGURE 28 Display frequency of safety campaign messages during non-incident/non-roadwork periods as an alternative to leaving the CMSs blank. 47 1717 2524 31 2 87 17 3 2 Does Not Display All CMSs Most CMSs About 50% of CMSs Some CMSs Very Few CMSs 0 10 20 30 40 50 Peak Off-Peak Time of Day Pe rc en t o f T M Cs FIGURE 29 Number of CMSs on which safety campaign messages are displayed during non-incident/non-roadwork periods as an alternative to leaving the CMSs blank.

administrative/upper-management preference, TMC manager/ supervisor preference, both administrative and TMC prefer- ence, or as part of governor safety program, statewide safety program, or law enforcement agency request. Only 1% of the TMCs included research results in their decision. The basis for each TMC’s decision to display safety campaign messages is shown in Table 9. Public Response The reported public response to safety campaign messages reported by the TMCs is shown in Figure 30. Only 12% of the TMCs reported that the response was either very favor- able (2%) or favorable (10%). Thirty-five percent reported that the response was either somewhat favorable (17%) or neutral (18%). The majority of the TMCs (51%) had no in- formation available to assess public opinion. Examples of Typical Safety Campaign Messages Exhibit 7 provides examples of typical messages that were re- ported by the 35 of 100 TMCs that regularly or periodically dis- play safety campaign messages rather than leaving CMSs blank. Experiences and Lessons Learned More comments were received regarding experiences and lessons learned with displaying safety campaign messages as an alternative to leaving CMSs blank than for the other mes- sages discussed in this report. The responses from the TMCs focused on that the safety campaign messages (1) resulted in requests from agencies that other non-traffic-related messages be displayed, (2) uncertainty about the effectiveness of the mes- sages, (3) should be displayed for short periods of time, (4) gen- 34 erated negative feedback (although this was not reflected in the data shown in Figure 29), and (5) should be displayed during certain periods. Comments received from the TMCs follow: • A message should be displayed for a short period of time and/or should not be displayed too often (18 TMCs). • We get negative feedback from the public (8 TMCs). • Only display for selected, well–organized, and adver- tised statewide safety campaigns (7 TMCs). • Public is generally receptive to the messages (6 TMCs). • Keep messages simple and easy to understand (4 TMCs). • They are effective (3 TMCs). Basis Percent of TMCs Percent of TMCs Agency administrative/upper-management preference only 48 TMC manager/supervisor preference only 5 Preference Only Agency administrative/upper-management preference, and TMC manager/supervisor preference 29 Other Other* 17 99 Preference and Feedback Agency administrative/upper-management preference, and TMC manager/supervisor preference, and Feedback from telephone calls, newspapers, radio, and/or television 0 0 TMC manager/supervisor preference, and Research conducted by agency and/or research conducted by others 1 Research conducted by agency only 0 Research conducted by others only 0 Preference and/or Research or Focus Group Focus group studies only 0 1 100 100 *Governor safety program, statewide safety program, or requests from law enforcement agencies. TABLE 9 BASIS FOR DECISION TO DISPLAY SAFETY CAMPAIGN MESSAGES DURING NON-INCIDENT/ NON ROADWORK PERIODS AS AN ALTERNATIVE TO LEAVING THE CMSs BLANK (n = 83) 2 10 17 18 0 2 0 51 0 10 20 30 40 50 60 Public Response Ve ry fav ora ble Fa vo rab le So me wh at fav ora ble Ne utr al So me wh at un fav ora ble Un fav ora ble Ve ry un fav ora ble No in for ma tio n a va ilab le Pe rc en t o f T M Cs FIGURE 30 Public responses to safety campaign messages during non-incident/non-roadwork periods as an alternative to leaving the CMSs blank (n = 83).

35 • Post only during off-peak or in off-peak direction to minimize unintended congestion (2 TMCs). • Only display safety-related or agency-supported mes- sages (2 TMCs). • Make sure that it is not a distraction to motorists (2 TMCs). • Message must have broad public impact (1 TMC). • Establishing a statewide policy as to what type, and when messages should be displayed, is very important (1 TMC). • Make sure there is value in the message to the public (1 TMC). • Coordinate any safety campaign message with local government (1 TMC). • We receive and deny requests for messages that are advertising (1 TMC). • Generally, these messages do not appear to have any impact on motorists (1 TMC). • Even the radio disc jockeys and traffic reporters make jokes about them (1 TMC). • Public would rather see travel times (1 TMC). • Received positive feedback from law enforcement when conducting routine traffic stops in areas where messages are displayed (1 TMC). • Do not display messages on all CMSs (1 TMC). • Messages that include CALL 911 can result in a lot of phone calls for 911 operators (1 TMC). Concerns and Challenges Concerns and challenges included (1) insufficient research to determine whether the messages are beneficial, (2) appropriate time to display the messages, (3) issues of consistency and credibility, and (4) requests for other types of non-traffic- related messages. Specific comments are as follow: • They open the door for other requests that are not transportation related. Denying requests is a problem (6 TMCs). • Do not know if the messages are beneficial and useful to motorists (3 TMCs). • A concern and challenge is how to display consistent messages (2 TMCs). • Need guidance as to when messages should be displayed and/or how to display consistent messages that provide value to the motoring public (2 TMCs). • There is a lack of clear, concise direction on what qual- ifies as a safety message (1 TMC). Exhibit 7 CLICK IT OR TICKET BUCKLE UP FOR SAFTEY IT'S THE LAW CHILD SAFETY USE SEAT BELTS BOOSTER SEATS Example 1 Example 2 Example 3 DRIVE HAMMERED GET NAILED DRINK DRIVE GO TO JAIL U DRINK U DRIVE U LOSE Example 4 Example 5 Example 6 WORK ZONE SAFETY WEEK SLOW DOWN OR PAY UP Phase 1 Phase 2 Example 7

• Trying to avoid private-interest type requests (1 TMC). • There are differences of opinion between operations staff not wanting to use the signs for this purpose versus administrative requests (1 TMC). • Will support administrative concerns by displaying them (1 TMC). • Some of the public have expressed dislike of unimpor- tant messages being displayed while others like it. We are afraid that people will get complacent seeing these types of messages very much and not pay attention to real emergency messages (1 TMC). TRAFFIC LAW OR ORDINANCE MESSAGES Background No reported research studies that objectively address the effects of traffic law or ordinance messages were found in the litera- ture. However, observations by the author indicated that some state DOTs display traffic law or ordinance messages and, therefore, these types of messages were included in the survey. Survey Results Display of Traffic Law or Ordinance Messages Seventy-four percent of the TMCs do not display traffic law or ordinance messages as an alternative to leaving CMSs blank. Nineteen percent display the messages during both the peak and off-peak periods and 7% display this type of message only during off-peak. The time of day during which traffic law or ordinance messages are displayed is shown in Figure 31. Figure 32 contains a summary of the frequency at which traffic law or ordinance messages are displayed. Only 2% of the TMCs regularly display these types of messages either during the peak or off-peak periods. Most TMCs that display traffic law or ordinance messages display these messages only occasionally or rarely. As shown in Figure 33, only 6% of the TMCs display traf- fic law or ordinance messages on all CMSs. The messages are displayed on most CMSs during the peak period by 8% of the TMCs and during the off-peak by 13% of the TMCs. 36 Basis for Decision to Display Traffic Law or Ordinance Messages Comparable to the previously discussed messages, the primary decision to display traffic law or ordinance messages was made based on agency administrative/upper-management preference, TMC manager/supervisor preference, or both. As shown in Table 10, only 4% of the TMCs that display the messages considered feedback from telephone calls, news- papers, radio, and/or television, and only 4% used support from research results in their decision. Public Response Public response to traffic law or ordinance messages is shown in Figure 34. As indicated, only 12% of the 26 TMCs that display the messages reported very favorable (8%) or favor- able (4%) responses. Fifty-seven percent indicated that the public was somewhat favorable (19%) or neutral (38%). Thirty-one percent of the TMCs had no information available to assess public opinion. Examples of Typical Traffic Law or Ordinance Messages Exhibit 8 shows examples of typical messages that were re- ported by the 26 of 100 TMCs that regularly or periodically Does Not Display 74% Peak & Off- Peak 19% Off-Peak Only 7% FIGURE 31 Display of traffic law or ordinance messages during non-incident/non-roadwork periods as an alternative to leaving the CMSs blank. 81 74 6 68 13 1 14 60 0 0 10 20 30 40 50 60 70 80 90 Peak Off-Peak Time of Day Pe rc en t o f T M Cs Does Not Display All CMSs Most CMSs About 50% of CMSs Some CMSs Few CMSs FIGURE 33 Number of CMSs on which traffic law or ordinance messages are displayed during non-incident/non-roadwork periods as an alternative to leaving the CMSs blank. 81 74 2 25 6 8 94 9 0 10 20 30 40 50 60 70 80 90 Peak Off-Peak Time of Day Pe rc en t o f T M Cs Does Not Display Regularly Often Occasionally Rarely FIGURE 32 Display frequency of traffic law or ordinance messages during non-incident/non-roadwork periods as an alternative to leaving the CMSs blank.

37 display traffic law or ordinance messages rather than leaving CMSs blank. Experiences and Lessons Learned Only a few TMCs reported on their experiences and lessons learned with displaying traffic law or ordinance messages. Those that responded reported positive feedback from law enforcement agencies and cautioned about overuse and problems with displaying the messages at inappropriate times. A summary of the responses from the TMCs follows: • Display of traffic law or ordinance messages MUST be part of an organized campaign, usually kicked off with a media campaign (1 TMC). 8 4 19 38 0 0 0 31 0 10 20 30 40 Public Response Ve ry fav or ab le Fa vo ra ble So m ew ha t fa vo ra ble Ne utr al So m ew ha t u nfa vo ra ble Un fav or ab le Ve ry un fav or ab le No in for ma tio n a va ilab le Pe rc en t o f T M Cs FIGURE 34 Public response to traffic law or ordinance messages during non-incident/non-roadwork periods as an alternative to leaving the CMSs blank (n = 26). Exhibit 8 SLOW DOWN OR MOVE OVER FOR EMERGENCY VEHS MINOR ACCIDENT MOVE IT IT IS THE LAW GEORGIA LAW AFTER AN ACCIDENT MOVE CARS TO SHOULDER Example 1 Example 2 Example 3 DON'T MESS WITH TEXAS GEORGIA LAW HEADLIGHTS ON WHEN RAINING WIPERS ON HEADLIGHTS ON IT'S OUR LAW Example 4 Example 5 Example 6 Basis Percent of TMCs Percent of TMCs Agency administrative/upper-management preference only 46 TMC manager/supervisor preference only 11 Preference Only Agency administrative/upper-management preference, and TMC manager/supervisor preference 35 92 Preference and Feedback Agency administrative/upper-management preference, and TMC manager/supervisor preference, and Feedback from telephone calls, newspapers, radio, and/or television 4 4 TMC manager/supervisor preference, and Research conducted by agency and/or research conducted by others 4 Research conducted by agency only 0 Research conducted by others only 0 Preference and/or Research or Focus Group Focus group studies only 0 4 100 100 TABLE 10 BASIS FOR DECISION TO DISPLAY TRAFFIC LAW OR ORDINANCE MESSAGES DURING NON-INCIDENT/NON-ROADWORK PERIODS AS AN ALTERNATIVE TO LEAVING THE CMSs BLANK (n = 26)

• Received positive feedback from law enforcement when conducting routine traffic stops in areas where messages are displayed. For example, we have a message explaining the Texas law requiring traffic to move over when an emergency vehicle is stopped on the roadway. Troopers have noticed an increase in the number of vehicles observing this law than in the past (1 TMC). • Traffic law or ordinance messages can be useful when they are used in a real-time, logical way such as displaying, HEADLIGHTS ON WHEN RAINING dur- ing a driving rainstorm, or MOVE OVER 1 LANE FOR EMERGENCY VEHS ON SHOULDER when there is an emergency vehicle on the shoulder just ahead. However, displaying traffic law messages when the scenario is not present is not good. For in- stance, displaying HEADLIGHTS ON WHEN RAIN- ING on a sunny day looks stupid to passing motorists (1 TMC). • Can be very effective as part of a public education cam- paign for new traffic laws. They broaden exposure to new laws (1 TMC). • Positive. It should not be displayed for every traffic law. Sign only for those laws that affect freeway travel lanes (1 TMC). • Make sure there is a law or ordinance. Have a program in place to assure the CMS is not abused (1 TMC). Concerns and Challenges The main concern and challenge reported dealt with making sure that the messages were consistently designed and applied. Summary of Message Types That Are Regularly Displayed Of particular interest are the types of messages that are regu- larly displayed on CMSs during non-incident/non-roadwork periods as an alternative to leaving the signs blank. As noted earlier, 43% of the TMCs responding to the survey regularly display messages during non-incident/non-roadwork periods during the peak period. Likewise, 35% regularly display messages during the off-peak period. The types of mes- 38 sages that are regularly displayed by the TMCs are listed in Table 11. From Table 11, the predominant message that is regularly displayed rather than leaving CMSs blank is travel time. Twenty-four percent of the TMCs exclusively display travel time during the peak period, and 15% display travel time during the off-peak period. Congestion messages are exclusively dis- played by 9% and 8% of the TMCs during the peak and off- peak periods, respectively. Speed, safety campaign, PSAs, and traffic law and ordinance messages are regularly displayed by very few TMCs. Percent Of TMCs Message Type Peak Off-Peak Travel Time 24 15 Congestion 9 8 Safety Campaign 3 4 Travel Time Congestion 2 1 Speed 2 1 Travel Time Safety Campaign 0 1 Travel Time PSAs 0 1 Congestion PSAs Safety Campaign 0 1 PSAs Safety Campaign 1 1 Safety Campaign Traffic Law or Ordinance 1 1 PSAs Safety Campaign Traffic Law or Ordinance 1 1 43 35 TABLE 11 TYPES OF MESSAGES REGULARLY DISPLAYED ON CMSs DURING NON-INCIDENT/NON-ROADWORK PERIODS AS AN ALTERNATIVE TO LEAVING THE SIGNS BLANK