Driver Selection Tests and Measurement (2012)

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.

38 CHAPTER FOUR SURVEY METHODS AND RESULTS Chapters two and three reviewed research and trade lit- erature on driver individual differences and selection tests and measurements, respectively. An additional vehicle for obtaining information for this study was project surveys. Two similar survey forms were used for two different respondent groups. Most important was a survey of current CMV fleet safety managers. The safety manager survey asked respon- dents their opinions on driver risk factors, what selection practices they used, and their ratings of the effectiveness of these practices. Survey opinions are not taken as empirical facts, but rather as indicators of industry thinking on safety management questions. Of secondary importance, but still of interest, was a survey of other experts in motor carrier safety. This survey form addressed the same general topics, but was limited to opinions because the respondents were not current practitioners. The two survey forms are provided in Appendix A. This chapter describes the survey approach and specific methods, and provides principal results for each respondent group. Results for the two respondent groups are presented separately because of their different perspectives on the problem and because the two forms differed some- what in their questioning approaches and content. A general caveat is that most of the survey responses represent subjective responses to subjective questions. A few questions were objective (e.g., asking safety managers whether they use a particular safety management practice), but most called for subjective judgments by respondents. Another caveat is that both samples must be regarded as convenience samples of interested, knowledgeable indi- viduals, not as representative samples of larger populations. Conceptually, both the safety manager and “other expert” populations are amorphous and not captured by any list. In addition, the safety manager population is extremely large (in the hundreds of thousands in the United States), diverse, and problematic from the sampling perspective. OVERVIEW OF SURVEY APPROACH, ANALYSIS, AND INTERPRETATION Sampling Concept The conceptual population for the safety manager survey was North American motor carrier (truck and bus) carrier safety managers. This population is somewhat amorphous, as there is no consistent definition or criterion for “carrier safety manager.” Also, there is no central potential respon- dent list on which to base systematic sampling. The safety manager sample consisted of individuals par- ticipating in trade associations or national meetings relat- ing to motor carrier safety. The e-mail addresses of these individuals were known to the project team, or paper survey forms were distributed directly to them in trade association meetings. The sample is presumed to be strongly biased toward organizations and individuals with more experience, past success, safety sophistication, and safety conscientious- ness than the overall population. Those returning the survey (whose responses are pre- sented here) are the respondents. Just as the sample space was likely a biased slice of the population, the sample was likely a biased slice of the sample space, because those responding tend to be more committed and interested in the topic than those not responding. Moreover, they tend to be more educated and verbal (Walonick 2010). Both sources of bias operated strongly in the present safety manager survey and to a lesser extent in the “other expert” survey. A larger study focusing on the survey per se might be better able to capture the larger population, increasing the size and representativeness of the sample space and obtain- ing a higher survey response rate. Study resources did not permit a more extensive, rigorous, and layered subject sam- pling approach. The sample obtained, even if it represents a skewed sample of the most knowledgeable and safety-con- scious respondents, still provides valuable information and accomplished the following objectives: • It tapped the views and practices of industry leaders. • It provided information on subjects’ relative opinions on the various traits and practices presented (e.g., which personal traits are rated most safety-relevant by respondents). • It provided contacts for follow-up interviews with safety managers on the practices of safety-active companies. Data Analysis and Interpretation There were three general types of questions on the sur- veys: questions about respondent opinions, questions

39 about specific carrier practices (safety managers only), and questions about respondents themselves and their organizations. Opinion questions were subjective and called for subjective, judgmental responses, mostly in the form of Likert scale ratings. It is important that these responses not be misinterpreted as objective facts. Ques- tions about specific carrier practices used (yes-no) were on the safety manager survey forms only. Questions about the respondents themselves (e.g., years of experience) were also objective. All of the caveats on sample repre- sentativeness apply to all questions on both forms. Thus, none of the survey results on either form can be general- ized to larger respondent groups or populations such as “North American carrier safety managers” or “experts in motor carrier safety.” The value of the survey results is not based on representativeness to larger populations, but rather on respondents’ answers to specific questions relative to other, similar questions (e.g., Which personal traits were rated most relevant to crash risk? Which safety practices were rated as most effective?). Nonuse of Response Percentages Per CTBSSP policy, the survey results tables in this chap- ter, and survey results cited elsewhere in this report, do not include results percentages. Instead, raw numbers are cited. This practice reduces the likelihood that survey results will be misinterpreted or incorrectly cited as representing larger respondent populations. Readers may generate their own percentages, but it is important that they be stated as being representative of larger groups. Likert Scale Means Likert scales are numeric rating scales, often with five choices numbered from 1 to 5. Likert scales usually have word descriptors for each choice, or “anchor” choices at the ends and perhaps the middle. Two different Likert scales were used in project surveys: • A five-point scale relating driver personal/psychologi- cal traits to crash risk. Choices ranged from “little or no association (1)” to “very strong association (5).” • A five-point scale rating the effectiveness of driver hir- ing practices. Choices ranged from “highly ineffective (1)” to “highly effective (5).” Results are provided in the form of respondent counts for each choice along with the weighted arithmetic mean of all choices. TRB’s online survey service provided these sta- tistics automatically in survey reports. For paper surveys, the survey statistics were obtained from Excel spreadsheets used to enter and reduce the data. MOTOR CARRIER SAFETY MANAGER SURVEY METHODS This section describes methods specific to the safety man- ager surveys. Safety managers were the respondent group of greatest interest for the study. These individuals have com- pany titles such as Safety Manager, Safety Director, Director of Compliance, and Vice President for Safety (and/or Com- pliance). A few have titles relating to HR management or operations. The respondent pool (sample space) consisted of individuals participating in national industry groups sup- porting safety, or who had attended safety meetings and whose contact information was available to the project team. As discussed in the previous section, this pool may be char- acterized as representing safety-conscious carrier safety and HR managers who are willing to participate in such research. All of the sampling and data analysis issues discussed in the previous section apply to the safety manager survey. Questionnaire Design and Content The safety manager survey questionnaire is provided in Appendix A. It consisted of the following general sections: • A brief statement of the study and survey purpose, with a confidentiality assurance • Two related five-choice questions on general factors affecting safety and crash risk (used on paper form only) • Two related five-choice questions on carrier practices affecting driver crash risk • Twelve driver personal/psychological traits, each rated on a one- to five-point Likert scale for association with crash risk • Thirteen carrier driver hiring practices, with a two- part answer for each: – Yes-no for whether the practice was used by the manager’s fleet – If yes, a one- to five-point Likert scale to rate the practice’s safety effectiveness • A single question on the carrier’s planned use of the Pre-Employment Screening Program (PSP) • An open “comments” space • Four questions on respondent’s professional experi- ence and fleet characteristics • A space for an optional e-mail address to which to send the project report PDF file • A space to volunteer for a paid interview on innovative carrier practices for the project case studies. The paper version of the questionnaire (administered principally at a truck and a bus trade association meeting) contained all of the previously mentioned sections. The online version was streamlined slightly by omitting the ini- tial two questions on crash risk factors.

40 Questionnaire Distribution and Analysis Two CMV trade associations, the Truckload Carriers Asso- ciation (TCA) and the Bus Industry Safety Council (BISC), assisted the study by distributing paper survey forms for this project and MC-22 (Safety Management in Small Motor Carriers) at national meetings. The National Private Truck Council (NPTC) assisted the effort by e-mailing the online survey solicitation to its Safety Council members with its endorsement. Paper surveys were formatted on a single front-and-back sheet where answer choices were circled or penciled in. At the TCA meeting, approximately 100 survey forms (for each of the two projects) were distributed, and 20 were returned. Two other truck forms were obtained through personal con- tacts. At the BISC meeting, approximately 50 forms were distributed, and 26 were returned. At the latter, meeting attendees included a significant proportion of non-safety managers (e.g., government officials, trade association offi- cials, vendors, consultants) for whom the survey was not intended. The exact number of carrier safety managers in the room is not known. An additional effort to obtain safety manager respondents was made using TRB’s online survey service. The online survey had the same content as the paper survey, except for the omission of the first two questions relating to general crash risk factors. These two questions were “thought ques- tions” that required more time for response than others on the survey. They were omitted from the online version to streamline the survey and perhaps increase response rates. E-mail requests were sent to 105 respondents believed to be current motor carrier safety managers based on their busi- ness cards and contact information gathered at various recent motor carrier safety conferences. An additional solicitation was sent from an NPTC official to NPTC Safety Council members. Twenty-one people took the online survey, which brought the total safety manager survey sample to 69. Paper survey answers were entered into an Excel spread- sheet for analysis. Online survey tabulations were generated and added to the Excel sheet totals. This experience suggests that both methods are viable. Handing out paper surveys at trade association meetings with the support of the organizers likely yields a higher return than sending e-mail solicitations. Carrier officials are often the targets of product marketing and other promo- tions, and thus may tend to be wary of responding to external e-mails in general. They may have confidentiality concerns, even if confidentiality statements are prominent in survey materials. Walonick (2010) provides a more extensive dis- cussion of the difficulties of obtaining survey data from vari- ous respondent groups. In spite of the challenges of obtaining a robust survey sample and the acknowledged unrepresentativeness of the sample in relation to all safety managers, the 69 responses provided sufficient data for analysis as well as many useful comments. In addition, a number of respondents volunteered for follow-up structured interviews. Follow-Up Structured Interviews The last question of the safety manager survey form asked respondents if they would be interested in participating in a paid follow-up interview to discuss innovative fleet prac- tices. The question included the assurance, “Responses will be confidential; no interviewees or carriers will be identified unless desired.” The key purpose of the interviews was to gather information and opinions for project case study write- ups. If respondents did volunteer, and had a relatively large number of “yes” responses under carrier practices (indicative of more developed driver hiring systems), they were con- tacted to schedule an interview. These interviews covered both this project and MC-22. A total of 20 respondents were contacted, usually both by e-mail and by phone, of whom 10 agreed to participate. These 10 provided substantial informa- tion on innovative carrier approaches and practices for hiring better drivers. This information is presented in chapter five. MOTOR CARRIER SAFETY MANAGER SURVEY RESULTS Factors Affecting Safety and Crash Risk Questions 1 and 2 addressed factors affecting safety and crash risk. These were also the first two questions of the MC-22 survey, as the two questions were pertinent to both studies. The same five choices were presented in each. Ques- tion 1 asked for the respondent’s choice of up to two factors having the greatest effect, whereas Question 2 asked for the one factor with the least effect. Table 3 presents responses. Note that Questions 1 and 2 were omitted from the online version of the survey in order to reduce survey length and increase response rates. As expected, choices for the two opposite questions (greatest and least) were more or less inversely related. Driver-related choices (a) and (b) were regarded as having the greatest effect on crash risk. The other three choices (vehicle, roadway, and weather) were all regarded as having much smaller effects. Choice (a) has the greatest relevance to driver selection, because driver selection procedures attempt to discern persistent personal traits predictive of crash risk. Both (a) and (c) are fundamentally driver assessment activi- ties, whereas the other three choices are primarily behavior change interventions. Figure 11 presents a histogram of the safety manager Question 1 “Most” votes for the five crash risk factor categories.

41 TABLE 3 SAFETY MANAGER RESPONSES RELATING TO FACTORS AFFECTING SAFETY AND CRASH RISK (1) Factors Affecting Safety and Crash Risk: Consider the entire fleet of North American commercial vehicles (trucks and buses). Across all these drivers and vehicles, which factors have the greatest association with crash risk? Pick up to two of the factors below which, in your opinion, have the greatest association with crash risk. (2) In your opinion, which one factor has the least asso- ciation with crash risk? (1) Most (2) Least (a) Enduring/long-term driver traits (e.g., age, physical abilities, medical conditions, personality, behavioral history) 29 5 (b) Temporary driver states (e.g., moods, daily circadian rhythms, effects of recent sleep, effects of recent food and fluids, effects of environmental conditions in cab) 29 4 (c) Vehicle characteristics (e.g., configuration, safety equipment, load) and mechanical condition (e.g., brakes, tires) 7 11 (d) Roadway characteristics and traffic conditions (e.g., undivided vs. divided highways, construction zones, traffic density, speed limits, lane restrictions) 9 15 (e) Weather and roadway surface conditions (e.g., wet vs. dry, road surface friction, visibility, wind) 10 9 Total Responses: 84 44 FIGURE 11 Safety manager “most” votes for the five crash risk factor categories. Most Important Carrier Practices Questions 3 and 4, and all subsequent questions, were included on both the paper and online versions of the survey. Questions 3 and 4 addressed the importance of five different areas of carrier safety management. The same five choices were presented in each. Question 3 asked for the respondent’s choice of up to two practices having the greatest importance, whereas Question 2 asked for the one practice with the least importance. Table 4 presents responses. Again, choices for the two opposite questions (greatest and least) were more or less inversely related. Choice (d) driver evaluation was rated overall as most important, fol- lowed by choice (b) driver selection. Selection and evalua- tion are related in that both are forms of driver assessment aimed at identifying good and bad drivers. Choice (a) driver preparation and (e) rewards/discipline were rated as having relatively low importance. Figure 12 presents a histogram of the safety manager Question 3 “Most” votes for the five types of carrier practices. TABLE 4 SAFETY MANAGER RESPONSES RELATING TO GENERAL CARRIER PRACTICES (3) Most Important Carrier Practices: All elements of driver training and companies’ safety management practices are important, but some may be more impor- tant than others. Pick up to two of the carrier practices below which, in your opinion, have the greatest effect on drivers’ safety behaviors and safety records. (4) In your opinion, which one practice has the least effect on driver safety outcomes? (3) Most (4) Least (a) Driver preparation; pre-hire CMV driving training and testing (e.g., basic school training and CDL testing) 14 22 (b) Driver selection and hiring; company driver recruit- ing, screening, selection, and hiring (include both man- datory and voluntary hiring practices) 34 4 (c) Company communications to drivers; driver orien- tation, finishing, safety meetings, refresher training, policy announcements, safety reminders 23 9 (d) Driver evaluation; company monitoring and evalua- tion of individual drivers (e.g., violation and incident tracking, ride-alongs, covert observations of driving, onboard computer monitoring) 38 2 (e) Company rewards and discipline (e.g., incentives, feedback, recognition, letters (both commendations and reprimands), bonuses, pay increases/decreases, other consequences imposed by management) 11 24 Total Responses: 120 61 FIGURE 12 Safety manager “most” votes for the five types of carrier practices. The tabulations indicate that respondents regarded driver traits as having paramount importance in relation to risk, and carrier practices to assess driver traits and behaviors to be the most important carrier safety practices. These findings are a testimony to the importance of the topics addressed in this study. Driver Personal Characteristics Questions 5–16 presented 12 driver personal traits or other characteristics and asked respondents to rate the association of each with crash risk on a five-point Likert scale, with the following instructions:

42 Driver Personal/Psychological Traits What driver characteristics are most associated with risk? In general and across all drivers, HOW STRONG IS THE ASSOCIATION of each of these per- sonal characteristics with DRIVER CRASH RISK? 1 = Little or no association. 5 = Very high association. Choose one number for each. If you are unsure or have no opinion, leave it blank. The five Likert scale choices were as follows: 1. Little or No Association 2. Some Association 3. Moderate Association 4. Strong Association 5. Very Strong Association. Table 5 provides the number of responses for each choice, the total number of responses (N), and the weighted arith- metic average or mean of responses (Avg.). Averages are rounded to the nearest tenth. Three personal traits received average Likert scale rat- ings of more than 4.0: aggressive personality, risk-taking personality, and poor vehicle handling. Two received ratings of less than 3.0: introverted/unsociable and poor English lan- guage skills. Safety manager ratings for personality traits were generally consistent with research findings relating to individual differences discussed in chapter two. For exam- ple, aggressiveness/hostility has a strong relation to crash and other accident risk. At the other extreme, introversion as a personality trait is generally beneficial to safety, as these individuals tend to be non-sensation-seeking and generally conservative in their behaviors. Hiring Practices and Tools Questions 17–29 presented 13 carrier practices and first asked respondents to state whether or not they regularly used the practice (yes or no). Respondents answering “yes” on a question were then asked to rate the effectiveness of the practice on a five-point Likert scale, with the following instructions: Which Driver Hiring Practices and Tools Do You Regularly Use to Select Safe Drivers? For each of the hiring practices below, please circle yes or no as to whether your organization uses the practice. If yes, rate the effectiveness of the method using the 1–5 scale provided. If no, leave the ratings blank. The five Likert scale choices were as follows: 1. Highly Ineffective 2. Ineffective 3. Not Sure/Neutral 4. Effective 5. Highly Effective. Table 6 provides the number of respondents reporting using each practice. Table 7 shows the effectiveness ratings given by users of the practice. Nonusers were instructed to leave these TABLE 5 SAFETY MANAGER LIKERT SCALE RATINGS FOR ASSOCIATIONS BETWEEN PERSONAL CHARACTERISTICS AND CRASH RISK Rating or Statistic: Personal Trait 1 2 3 4 5 N Avg. (5) Aggressive personality 2 4 5 25 33 69 4.2 (6) Risk-taking personality 0 2 2 28 37 69 4.4 (7) Dishonest/untrustworthy 3 4 12 33 16 68 3.8 (8) Introverted/unsociable 19 23 17 7 1 67 2.2 (9) Low intelligence/mental abilities 5 13 28 17 5 68 3.1 (10) Poor English language skills 14 16 19 14 3 66 2.6 (11) Unhappy/personal problems 0 5 14 39 11 69 3.8 (12) Financial problems/in debt 0 12 16 27 12 67 3.6 (13) Dissatisfied with driver job/profession 0 6 20 28 14 68 3.7 (14) Poor general physical health 2 4 23 29 11 69 3.6 (15) Overweight/obese 6 13 22 18 9 68 3.2 (16) Poor vehicle handling (e.g., backing, parking) 0 1 7 23 38 69 4.4

43 items blank or, in the online version, were not presented with the questions. Statistics provided include the number for each Likert scale choice, the total number of responses (N), and the weighted arithmetic average or mean of responses (Avg.). Averages are rounded to the nearest tenth. TABLE 6 SAFETY MANAGER RESPONDENT USE OF HIRING PRACTICE/TOOL Rating or Statistic: Driver Hiring Practice/Tool Yes No N (17) Give on-road driving test 67 1 68 (18) Range/yard maneuvering test (e.g., backing, parking) 59 9 68 (19) Standardized interview (set list of questions) 44 23 67 (20) Check criminal record 63 4 67 (21) Check credit history and rating 20 46 66 (22) Determine likely safety belt use (by observa- tion, interview, questionnaire, etc.) 41 24 65 (23) General medical history questionnaire 43 22 65 (24) Mental ability test (e.g., math, reasoning) 10 55 65 (25) English language test 12 54 66 (26) Any computer-based dynamic performance test (e.g., hand-eye coordination, tracking) 7 59 66 (27) Job satisfaction or job choice questionnaire 16 49 65 (28) Personality questionnaire (e.g., aggressive- ness, risk-taking, attitudes) 21 43 64 (29) Questionnaire about driving behaviors (e.g., following distances, turn signal use) 19 45 64 Respondents used an average of 6.1 of the 13 practices listed. The most frequently used were road tests, range tests, and checking criminal records. The least frequently used were computer-based dynamic tests, mental abil- ity tests, and English language tests. Hiring practices receiving the most favorable ratings included the road and range tests, computer-based dynamic tests (though used and rated by only seven respondents), personality ques- tionnaires, and questionnaires about driving behaviors. Checking credit history and rating received the lowest average rating. On paper forms, Question 30 in this section was a “write- in” item where respondents could write in another hiring practice and rate it. Few respondents answered this question. Two respondents mentioned physical tests given to applicants relating to job requirements like loading and unloading. Additional Questions Question 31 asked respondents if they planned to use the new FMCSA PSP. Answers were as follows: Yes (45), No (5), and Not Sure (15). Thus, the PSP will become a standard procedure for most carriers participating in the survey. Question 32 asked respondents if they had any comments on the previous questions (or any related issue). Few com- ments were made. One respondent believed that a variety of driver personality characteristics “comes with the territory” and that, therefore, the primary focus should be on driving skills and behaviors. TABLE 7 SAFETY MANAGER LIKERT SCALE RATINGS OF EFFECTIVENESS OF HIRING PRACTICE/TOOL Rating or Statistic: Driver Hiring Practice/Tool 1 2 3 4 5 N Avg. (17) Give on-road driving test 0 0 5 44 18 67 4.2 (18) Range/yard maneuvering test (e.g., backing, parking) 0 0 2 39 18 59 4.3 (19) Standardized interview (set list of questions) 0 4 13 24 4 45 3.6 (20) Check criminal record 2 2 15 34 10 63 3.8 (21) Check credit history and rating 3 2 7 6 2 20 3.1 (22) Determine likely safety belt use (by observa- tion, interview, questionnaire, etc.) 2 1 14 19 6 42 3.6 (23) General medical history questionnaire 0 0 15 18 9 42 3.9 (24) Mental ability test (e.g., math, reasoning) 0 2 3 5 1 11 3.5 (25) English language test 0 1 3 7 2 13 3.8 (26) Any computer-based dynamic performance test (e.g., hand-eye coordination, tracking) 0 0 1 3 3 7 4.3 (27) Job satisfaction or job choice questionnaire 0 0 8 7 1 16 3.6 (28) Personality questionnaire (e.g., aggressiveness, risk-taking, attitudes) 0 0 2 11 7 20 4.3 (29) Questionnaire about driving behaviors (e.g., following distances, turn signal use) 0 1 0 13 6 20 4.2

44 Information About Respondents and Their Fleets Safety managers were also asked two questions about their professional experience and two questions about their fleet’s characteristics. Question 33 asked their years of experience as a safety manager or human resource manager, and Ques- tion 34 asked their total years of experience in commercial truck/bus operations. Table 8 provides summary statistics of their answers (Note: SD = standard deviation). TABLE 8 SUMMARY STATISTICS ON PROFESSIONAL EXPERIENCE OF SAFETY MANAGER RESPONDENTS Statistic: Question Range Median Mean SD (33) Number of years experi- ence as carrier safety man- ager or human resources manager 2 to 36 10 11.8 8.1 (34) Total years experience in commercial truck/bus operations 5 to 55 22.5 22.8 11.9 The 69 safety manager respondents claimed a combined 793 years experience as safety managers and 1,551 years experience in CMV transport. As a group, they are highly experienced. Question 35 asked respondents to state the approximate number of power units (i.e., tractors or trucks) currently in their fleets. Table 9 provides summary statistics of their answers. TABLE 9 SUMMARY STATISTICS ON SAFETY MANAGER RESPONDENT FLEET SIZE Statistic: Question Range Median Mean SD (35) Approximate number of power units currently in fleet 8 to 15,000 143 827 2,016 There are no definitive population statistics to compare with the respondent individual and fleet statistics. Never- theless, it is clear that survey respondents were generally more experienced than most individuals with motor car- rier safety management responsibilities, and that their fleets were generally much larger than average. This reflects that larger fleets are overrepresented at virtually all national and regional safety conferences and as active members of national and state truck and bus transport organizations. Question 36 asked respondents to select the truck or bus operation type that best characterized their fleet. The number of responses in each category is listed in Table 10. Although the question asked for “the” best characterization, many bus safety managers selected two choices (g and h). Therefore, that dual selection is listed as a separate choice. TABLE 10 SAFETY MANAGER RESPONDENTS’ FLEET OPERATION TYPES Operation Type No. Safety Managers (a) For hire: long haul/truckload 24 (b) For hire: long haul/less-than-truckload (LTL) 0 (c) For hire: local/short haul (most trips < 100 miles) 2 (d) Private industry: long haul 8 (e) Private industry: local/short haul (< 100 miles) 8 (f) Passenger carrier: scheduled service 5 (g) Passenger carrier: charter 12 (g+h) Passenger carrier: both scheduled service and charter 8 (h) “Other” 1 Total (N): 68 “OTHER EXPERT” SURVEY METHODS The secondary project survey was of other experts in motor carrier safety. These individuals were primarily professional associates of the principal project investigators. They were either known personally or selected on the basis of their job positions or other professional activities. They included pro- fessionals in government, industry trade associations, other industry roles (e.g., safety consulting), and research. Many of these individuals are actively involved in other TRB truck and bus safety activities. Even though these individuals are highly knowledgeable, they are regarded as secondary respondents because by definition they are not currently car- rier practitioners. As a result, their survey forms included opinion items but not items on their practices related to driver hiring and selection. The data from this “other expert” survey were of interest, though, to (a) gauge expert opinion on questions, (b) indicate areas that may deserve more con- sideration, and (c) identify ongoing related research. Questionnaire Design and Content The “other expert” survey questionnaire was similar to that for safety managers. The form is provided in Appendix A. It consisted of the following general sections: • A brief statement of the study and survey purpose, with a confidentiality assurance • Two related five-choice questions on general factors affecting safety and crash risk (used on paper form only) • Two related five-choice questions on carrier practices affecting driver crash risk • Nineteen driver personal/psychological traits rated on a one- to five-point Likert scale for association with crash risk

45 • An open “comments” space • Two questions on respondent’s years of motor carrier safety-related experience and on specific types of posi- tions held. Questionnaire Distribution and Analysis The “other expert” survey was administered only online, through TRB’s online survey service. The survey solicita- tion was sent by e-mail to 128 individuals, with a second e-mail reminder sent several weeks later. A total of 34 online surveys were completed (27%). Survey results were tabu- lated by the reports program. “OTHER EXPERT” SURVEY RESULTS Factors Affecting Safety and Crash Risk Questions 1 and 2 addressed factors affecting safety and crash risk. The same five choices were presented in each. Question 1 asked for the respondent’s choice of up to two factors having the greatest effect whereas Question 2 asked for the one factor with the least effect. Table 11 presents responses. TABLE 11 OTHER EXPERT RESPONSES RELATING TO FACTORS AFFECTING SAFETY AND CRASH RISK (1) Factors Affecting Safety and Crash Risk: Consider the entire fleet of North American commercial vehicles (trucks and buses). Across all these drivers and vehicles, which factors have the greatest association with crash risk? Pick up to two of the factors below which, in your opinion, have the greatest association with crash risk. (2) In your opinion, which one factor has the least associa- tion with crash risk? (1) Most (2) Least (a) Enduring/long-term driver traits (e.g., age, physical abil- ities, medical conditions, personality, behavioral history) 19 2 (b) Temporary driver states (e.g., moods, daily circadian rhythms, effects of recent sleep, effects of recent food and fluids, effects of environmental conditions in cab) 24 3 (c) Vehicle characteristics (e.g., configuration, safety equipment, load) and mechanical condition (e.g., brakes, tires) 3 13 (d) Roadway characteristics and traffic conditions (e.g., undivided vs. divided highways, construction zones, traffic density, speed limits, lane restrictions) 11 6 (e) Weather and roadway surface conditions (e.g., wet vs. dry, road surface friction, visibility, wind) 6 10 Total Responses: 63 34 As expected, choices for the two opposite questions (greatest and least) were more or less inversely related. Driver-related choices (a) and (b) were regarded as having the greatest effect on crash risk, whereas vehicle-related choice (c) was regarded as having the least effect. This was followed by weather and roadway surface condition (e). Questions 3 and 4 addressed the importance of five dif- ferent areas of carrier safety management. The same five choices were presented in each. Question 3 asked for the respondent’s choice of up to two practices having the great- est importance, whereas Question 2 asked for the one prac- tice with the least importance. Table 12 presents responses. TABLE 12 OTHER EXPERT RESPONSES RELATING TO GENERAL CARRIER PRACTICES (3) Most Important Carrier Practices: All elements of driver training and companies’ safety management practices are important, but some may be more impor- tant than others. Pick up to two of the carrier practices below which, in your opinion, have the greatest effect on drivers’ safety behaviors and safety records. (4) In your opinion, which one practice has the least effect on driver safety outcomes? (3) Most (4) Least (a) Driver preparation; pre-hire CMV driving training and testing (e.g., basic school training and CDL testing) 8 13 (b) Driver selection and hiring; company driver recruit- ing, screening, selection, and hiring (include both man- datory and voluntary hiring practices) 19 3 (c) Company communications to drivers; driver orien- tation, finishing, safety meetings, refresher training, policy announcements, safety reminders 10 10 (d) Driver evaluation; company monitoring and evalua- tion of individual drivers (e.g., violation and incident tracking, ride-alongs, covert observations of driving, onboard computer monitoring) 24 1 (e) Company rewards and discipline [e.g., incentives, feedback, recognition, letters (both commendations and reprimands), bonuses, pay increases/decreases, other consequences imposed by management] 6 7 Total Responses: 67 34 Again, choices for the two opposite questions (greatest and least) were more or less inversely related. Choice (d) driver evaluation was rated overall as most important, fol- lowed by choice (b) driver selection. The other three choices (driver preparation, company communications, company rewards and discipline) were all rated as having relatively low importance. Driver Personal Characteristics and Potential Value of Testing Questions 5–23 presented 19 driver personal traits or condi- tions and asked respondents to rate the association of each with crash risk on a five-point Likert scale. The specific instructions were as follows: Driver Personal/Psychological Traits What driver characteristics are most associated with risk? In general and across all drivers, HOW STRONG IS THE ASSOCIATION of each of these per-

46 sonal characteristics with DRIVER CRASH RISK? 1 = Little or no association. 5 = Very high association. Choose one number for each. If you are unsure or have no opinion, leave it blank. The five Likert scale choices were as follows: 1. Little or No Association 2. Some Association 3. Moderate Association 4. Strong Association 5. Very Strong Association. More trait-rating items were included on the “other expert” form than on the safety manager form because the “other expert” form was otherwise shorter. Also, several of the personal traits on the “other expert” form were worded slightly differently than similar items on the safety manager form, based on assumption that the “other experts” would be more familiar with the traits as personality dimensions (i.e., with high and low values along a scale). Table 13 provides the number of responses for each choice, the total number of responses (N), and the weighted arithmetic average or mean of responses (Avg.). Averages are rounded to the near- est tenth. Personal dimensions rated as having the highest associa- tion with risk included aggressive personality and attitudes, risk perception and attitudes, sleep hygiene habits, and truck road driving as evaluated in a 30-minute ride-along. Those with relatively low-rated associations with risk included debt and credit history/rating, English language skills, cardiac health, and general medical history. Additional Comments Question 24 asked respondents, “Additional comments or suggestions regarding driver personal/psychological dimen- sions and/or tests to assess them?” The following are several responses. A few not provided here promoted commercial products or were unsubstantive. Some responses are edited for brevity. TABLE 13 OTHER EXPERT LIKERT SCALE RATINGS FOR ASSOCIATIONS BETWEEN PERSONAL CHARACTERISTICS AND CRASH RISK Rating or Statistic: Personal Trait: 1 2 3 4 5 N Avg. (5) Aggressive personality and attitudes 0 0 6 15 13 34 4.2 (6) Risk perception and attitudes 0 1 2 17 14 34 4.3 (7) Sensation-seeking 6 6 8 9 3 34 2.9 (8) Conscientiousness/honesty 0 6 11 15 2 34 3.4 (9) Personal/family adjustment and happiness 1 6 8 12 6 33 3.5 (10) Job satisfaction as driver 0 4 16 9 5 34 3.4 (11) Psychological match to the job (e.g., activity preferences, interests) 2 5 17 8 1 33 3.0 (12) Debt and credit history/rating 9 10 10 4 1 34 2.4 (13) Intelligence/mental abilities 1 11 17 3 1 33 2.8 (14) English language skills (e.g., reading, speaking) 7 14 8 3 1 33 2.3 (15) Dynamic sensory-motor performance (e.g., visual tracking, reaction time) 0 5 8 16 5 34 3.6 (16) Body-Mass Index (BMI) 3 12 11 8 0 34 2.7 (17) Sleep apnea (e.g., none, mild, moderate, severe) 0 2 11 10 9 32 3.8 (18) Sleep hygiene habits (e.g., amount and regularity of sleep) 0 2 6 15 10 33 4.0 (19) Cardiac health (e.g., blood pressure, cholesterol) 5 12 12 4 1 34 2.5 (20) General medical history 4 10 13 4 1 32 2.6 (21) Truck driving knowledge (written) 3 14 8 4 3 32 2.7 (22) Truck range maneuvering (e.g., backing, parallel parking) 2 9 8 9 5 33 3.2 (23) Truck road driving (e.g., speed and space man- agement in 30-minute ride-along in traffic) 0 3 8 12 10 33 3.9

47 • The greatest association is the knowledge and skill of the driver instructor. The least association would be CDL test scores. • Attitude and behavior are everything. Regardless of how much training you give someone, it they don’t have a safety attitude and safe behavior habits they are probably never going to be safe. • Drivers must not be preoccupied with work or family or personal hobby problems when they drive. They must be able to put things out of their mind, so they can focus entirely on the driving task. • “Distractedness” is an important driver trait, with dis- tractedness (aka field dependence, impulsivity) at one end and “un-distractedness” or “tunnel-vision” at the other. The least crash prone drivers are in the middle. • It is a good practice to employ new drivers on a proba- tionary basis and require at least 2 weeks of supervised driving where observations can be made about risk behavior, attitudes, knowledge, etc. • Too few drivers are screened for moral and character attributes; yet, there is a direct correlation between increased risk and a driver’s perception of moral and ethical responsibility. Additionally, drivers should be screened for their ability to comprehend their regula- tory and ethical responsibilities. Information About Respondents The years of motor carrier safety experience of the 34 “other expert” respondents, addressed by Question 25, ranged from 7 years to 41 years. The mean was 20.8 years. These respondents were also asked in Question 26 to indicate their professional experience in areas relating to motor carrier safety. The breakdown is shown in Table 14. The percent- ages shown sum to well over 100% because most respon- dents gave multiple responses. TABLE 14 OTHER EXPERT RESPONDENT EXPERIENCE AREAS Operation Type No. Safety Managers (a) Government enforcement 8 (b) Other government (e.g., rulemaking, policy) 9 (c) Industry trade association 12 (d) Commercial driver 5 (e) Carrier safety director/manager 4 (f) Other carrier management position 5 (g) Safety consultant or vendor to fleets 11 (h) Accident investigation/data analysis 14 (i) Motor carrier safety research 24 (j) Journalist 0 (k) Driver trainer/training development 10 (l) Insurance for motor carriers 7 (m) Other 3 Average Number of Experience Areas/Respondent: 3.3 The results show that the experience base of the other experts was both extensive and varied, with heavy repre- sentation of individuals with backgrounds in government, industry trade associations, safety consulting, accident investigation/data analysis, motor carrier safety research, and training.