Cell Phone Use on the Roads in 2002

You don't have the latest version of Adobe Flash Player.

Please update your flash player.

Please login or register to make a comment!

...Description...... more. less.

by Westat, Inc under the direction of NHTSA 9s National Center for Statistics and Analysis under federal contract number DTNH22-00-07001. Also we thank Mike Goodman and Julie Barker in NHTSA 9s Office of Advanced Safety Research, and Paul Tremont in the Office of Research and Technology for helpful comments. Abstract With the increasing popularity of cellular phones and public concern about the safety of using phones while driving, there has been increased interest in tracking the incidence of driver cell phone use. This report presents the National Highway Traffic Safety Administration 9s (NHTSA) most recent results on this topic, which come from NHTSA 9s National Occupant Protection Use Survey (NOPUS). The survey estimated that during daylight hours, drivers of cars, trucks, vans, and sport utility vehicles used hand-held phones during 4% of their driving time in 2002, up from 3% in 2000. These results were obtained by observing actual traffic. In fact the NOPUS is currently the only source of probability-based observed national data on driver cell phon e use. While the NOPUS percentages are small, they indicate a substantial and growing phenomenon, translating into approximately 600,000 drivers on the road at any given time between 8 AM and 6 PM using hand-held phones in 2002, up from 400,000 in 2000. When combined with data from other surveys on hands-free use, NOPUS that nearly finds that 900,000 drivers on the road at any given daylight time are using cellular phones in some manner, whether by holding the phone or using a hands-free device. However, while NOPUS finds an increase in hand-held use behind the wheel, the survey data indicate a shift toward the use of hands-free devices. The growth in hands-free use while driving that occurred between 2000 and 2002 outpaced the growth in hand-held phones (a 33% growth for hand-held phones, compared to 100% for hands-free). NOPUS first observed cell phone use in 2000, and so the data in this report represent some of the first direct measurements of this form of driver distraction on the road. 17. Key Words cellular phones, wireless phones, driver distraction, observational surveys, hand-held phones, hands-free devices 18. Distribution Statement Document is available to the public through the National Technical Information Service, Springfield, VA 22161 http//:www.ntis.gov 19. Security Classif. (of this report) Unclassified 20. Security Classif. (of this page) Unclassified 21. No. of Pages 46 22. Price Form DOT F 1700.7 (8-72) Reproduction of completed page authorized i National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 TABLE of CONTENTS Tables and Figures.....................................................................iii 1. Executive Summary............................................................ 4 1.1 More Drivers Are Holding Phones .....................................................................................4 1.2 Over Half a Million Drivers Are Holding Phones At Any Given Time.............................5 1.3 The Shift Toward Hands-Free Use .....................................................................................6 1.4 Other Major Findings of the 2002 Survey..........................................................................7 2. Background: A Primer on Wireless Phone Technology..... 9 3. Driver Cell Phone Use Nationwide................................... 12 3.1 The NOPUS National Estimate........................................................................................12 3.2 Estimates of the Use of All Wireless Phones....................................................................13 3.3 Comparing the NOPUS Findings to Results of Other Surveys ........................................13 4. Subnational Estimates....................................................... 15 4.1 Changes in Age Patterns ................................................................................................... 15 4.2 Males Use Phones Hands-Free Slightly More than Women............................................15 4.3 Increased Use in Urban Areas...........................................................................................18 4.4 Increases in Pickup Trucks, in the West, and Among Minorities.....................................18 4.5 Additional Hand-Held Estimates......................................................................................19 5. NOPUS Survey Methodology........................................... 25 5.1 The Basic Survey Methodology........................................................................................25 5.2 Improvements in 2002 ...................................................................................................... 26 5.3 What Does NOPUS Consider cUsing a Hand-Held Phone d?...........................................26 5.3 Design Aspects that Impact Data......................................................................................27 5.4 Assessing Significance..................................................................................................... .27 6. Derived Estimates............................................................. 28 6.1 Data Sources .............................................................................................................. .......29 6.2 The Extent of Hand-Held Versus Hands-Free Use...........................................................30 6.3 Trip Characteristics....................................................................................................... ....34 ii National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 6.4 Applying these Estimates to NOPUS ...............................................................................37 6.4.1 Extrapolating NOPUS to Estimates of All Cell Phones ..................................................37 6.4.2 The Numbers of Vehicles with Drivers on Phones..........................................................39 6.4.3 Relating Phone Use to Trip Length and Duration...........................................................40 7. References......................................................................... 42 iii National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 Tables and Figures Figure 1: Percent of Drivers Holding Phones.................................................................................4 Figure 2: The Number of Drivers Holding Phones At Any Given Time........................................5 Figure 3: Percent of Drivers Using Phones, by Means of Use .......................................................5 Figure 4: Number of Drivers on Phones At Any Given Time, By Means of Use.........................6 Figure 5: Hand-Held Phone Use By Select Characteristics............................................................7 Table 1: Cell Phone Use Nationwide............................................................................................1 2 Figure 6: Cell Phone Use Nationwide..........................................................................................12 Figure 7: The Growth in Cellular Phone Subscribers...................................................................13 Figure 8: Growth in the Use of Wireless Devices ........................................................................14 Table 2: The Usual Means of Cell Phone Use for Male and Female Drivers..............................16 Table 3: Driver Cell Phone Use in 2002 by Gender and Means of Use.......................................16 Table 4: Daylight Driving Time in 2002, by Gender....................................................................17 Table 5: Time Spent by Men and Women Driving While on the Phone in 2002.........................17 Figure 9: Driver Cell Phone Use, by Urbanization, Region, and Time of Day and Week...........19 Figure 10: Hand-Held Use, by Gender, Race, and Vehicle Type.................................................19 Table 6: Driver Hand-Held Cell Phone Use During Daylight Hours by Select Characteristics...20 Table 7: Driver Hand-Held Cell Phone Use During Daylight Hours by Gender and Vehicle Type ............................................................................................................................... ............21 Table 8: Driver Hand-Held Cell Phone Use During Daylight Hours by Driver Age and Vehicle Type .......................................................................................................................... ........21 Table 9: Driver Hand-Held Cell Phone Use During Daylight Hours by Race and Vehicle Type 22 Table 10: Driver Hand-Held Cell Phone Use During Daylight Hours by Urbanization and Vehicle Type................................................................................................................... ..22 Table 11: Driver Hand-Held Cell Phone Use During Daylight Hours by Region and Vehicle Type .......................................................................................................................... ........23 Table 12: Driver Hand-Held Cell Phone Use During Daylight Hours by Time of Day/Week and Vehicle Type................................................................................................................... ..24 Table 13: Derived Estimates and the Auxiliary Data Sources Used to Derive Them..................28 Table 14: The Means of Cell Phone Usually Used While Driving in 2002 .................................31 Table 15: The Extent to Which Hands-Free Devices Were Used in North Carolina in 2002......31 Table 16: The Extent to Which Hands-Free Devices Were Used Nationwide in 2002................32 Table 17: Calculation of the Extent of Hands-Free Use Among Part-Time Hands-Free Users...33 Table 18: A Typical Snapshot of Drivers on Cell Phones Nationwide in 2002, by Type of User and Means of Use..............................................................................................................3 3 Table 19: A Typical Snapshot of Drivers on Cell Phones Nationwide in 2002, by Phone Used.34 Table 20: Trip Characteristics in 1995 and 2001, by Time of Day..............................................35 Table 21: Trip Characteristics in 2000 and 2002, by Time of Day..............................................36 Table 22: The Means Usually Used While Driving to Use a Phone in 2000 ...............................38 Table 23: Driver Cell Phone Use Nationwide .............................................................................38 Table 24: Drivers on the Road at Any Given Time 1 , By Phone Use...........................................39 Table 25: Driver Cell Phone Use, As It Relates to the Distance and Duration of Trips..............40 4 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 1. Executive Summary With the increasing popularity of cell phones and public concern about the safety of using phones while driving, there has been increased interest in tracking the incidence of driver cell phone use. This report presents the National Highway Traffic Safety Administration 9s (NHTSA) most recent results on this topic, which come from NHTSA 9s National Occupant Protection Use Survey (NOPUS). The survey estimated that during daylight hours, drivers of cars, trucks, vans, and sport utility vehicles used hand-held phones during 4% of their driving time in 2002, up from 3% in 2000. These results were obtained by observing use in actual traffic. In fact the NOPUS is currently the only source of probability-based observed national data on driver cell phone use. While the NOPUS percentages are small, they indicate a substantial and growing phenomenon, translating into approximately 600,000 drivers on the road at any given time between 8 AM and 6 PM using hand-held phones in 2002, up from 400,000 in 2000. When combined with data from other surveys on hands-free use, NOPUS that nearly finds that 900,000 drivers on the road at any given daylight time are using cellular phones in some manner, whether by holding the phone or using a hands-free device. (See Section 6 for the derivation of these statistics.) However, while NOPUS finds an increase in hand-held use behind the wheel, the data also indicate a shift toward the use of hands-free devices while driving. The growth in hands-free use while driving that occurred between 2000 and 2002 outpaced the growth in hand-held phones (a 33% growth for hand-held phones, compared to 100% for hands-free). (Again see Section 6 for derivations.) NOPUS found increases in hand-held use among drivers in the 16-24 and 25-69 age ranges and on urban roads, and continued to find similar use by males and females. NOPUS first observed cell phone use in 2000, and so the data in this report represent some of the first direct measurements of this form of driver distraction on the road. 1.1 More Drivers Are Holding Phones Figure 1: Percent of Drivers Holding Phones 3% 4% 0% 1% 2% 3% 4% 5% 2000 2002 Percent of Drivers Holding Phones Source: National Occupant Protection Use Survey, National Center for Statistics and An l i HT A *Drivers observed holding phones between the hours of 8 AM and 6 PM. 5 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 In 2002, 4% of drivers observed at intersections controlled by stop signs or stoplights during daylight hours were observed holding phones, up from 3% in 2000. These results are from the National Occupant Protection Use Survey (NOPUS), which is conducted by the National Center for Statistics and Analysis in the National Highway Traffic Safety Administration (NHTSA) and provides the only probability-based observed data on driver cell phone use in the United States. While the increasing prevalence of drivers holding phones is apparent to many who drive on the nation 9s roads, what NOPUS offers is a scientifically-based quantification of the prevalence of this phenomenon and how much it has increased. It may surprise many that the NOPUS percentages are so low (4% in 2002), compared to their own personal observations, perhaps because drivers holding phones stand out in their minds when they observe them. (Keep in mind that the above estimates concern only hand-held phones. Hands-free estimates will follow below.) Also, we will see next that NOPUS 9s small estimates translate into large numbers of drivers using phones. 1.2 Over Half a Million Drivers Are Holding Phones At Any Given Time While the use rates found by NOPUS are small numbers, they indicate a sizable phenomenon because of the sheer volume of vehicles on the road. At any given time between the hours of 8 AM and 6 PM in 2002, there were about 14 million privately owned vehicles on the road. The Figure 2: The Number of Drivers Holding Phones At Any Given Time Figure 3: Percent of Drivers Using Phones, by Means of Use 385,434 566,788 0 200,000 400,000 600,000 800,000 2000 2002 Avg. No. of Drivers Holding Phones During Each Moment Data derived from the following sources: The 2000 and 2002 National Occupant Protection Use Surveys, National Center for Statistics and Analysis, NHTSA, ; the 2001 National Household Travel Survey, DOT; and the 1995 National Personal Travel Survey, DOT *Based on drivers observed holding phones between the hours of 8 AM and 6 PM. 3% 1% 4% 4% 2% 6% 0% 2% 4% 6% 8% Drivers holding phones Drivers using phones hands-free Drivers using phones Percent of Drivers Using Phones 2000 2002 Data derived from the following sources: The 2000 and 2002 National Occupant Protection Use Surveys, National Center for Statistics and Analysis, NHTSA; the 2001 National Household Travel Survey, DOT; the 1995 National Personal Travel Survey, DOT; and the CTIA's Semi-Annual WirelessIndustry Indices , April 2003 Edition *Drivers observed holding phones between the hours of 8 AM and 6 PM. 6 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 NOPUS use rate of 4% means that nearly 600,000 of these vehicles were driven by someone holding a phone. (See Section 6 for the derivation of these statistics.) 1.3 The Shift Toward Hands- Free Use Although the 2002 survey only observed hand-held phones, by combining its estimates with other data sources, we can estimate the extent to which drivers used phones hands-free. Using data from interviews of cell phone users, we find dramatic growth in hands-free use while driving 3 a 100% growth in the percent of drivers using phones hands-free, over the level seen in 2002, and a nearly a 150% growth in the number of drivers on the road at any given time who are using phones in the hands-free mode. (See Figures 3 and 4. Estimates in Figure 4 do not necessarily sum to totals due to rounding.) Note that during the same period of time, the percent of drivers on hand-held phones only increased by 33%, indicating a shift towards using hands-free phones when people are behind the wheel. (See Figure 3.) Nearly 900,000 vehicles on the road at any given daylight time are being driven by someone on a phone, with about 600,000 drivers holding the phone and the remaining 300,000 using phones in some type of hands-free mode. (Again see Section 6 for derivations.) Figure 4: Number of Drivers on Phones At Any Given Time, By Means of Use 385,434 116,158 501,593 12,346,222 566,788 283,965 850,753 13,068,396 0 3,000,000 6,000,000 9,000,000 12,000,000 Drivers holding phones Drivers using phones hands- free Drivers using phones Drivers not on phones Avg. No. of Drivers On the Road Each Moment* 2000 2002 Data derived from the following sources: The 2000 and 2002 National Occupant Protection Use Surveys, National Center for Statistics and Analysis, NHTSA; The 2001 National Household Travel Survey, DOT; The 1995 National Personal Travel Survey, DOT; (Stutts et al., 2003); and (Royal, 2003) *Between the hours of 8 AM and 6 PM. 7 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 1.4 Other Major Findings of the 2002 Survey Other major findings of the 2002 NOPUS include the following. Each of the findings concerns the use of hand-held phones by drivers during daylight hours: " Use increased among young adults and adults, ages 16-24 and 25-69. Use is now statistically lower among older drivers than among those under 70 years of age. " Males and females continue to exhibit similar use, with each talking on hand-held phones for 4% of their daylight driving time in 2002. " Use increased on urban roads, with a statistically significant increase from 2% in 2000 to 5% in 2002. Report Organization This report is organized as follows. Section 2 presents basic information concerning wireless phones that the lay reader may find helpful. Section 3 presents and analyzes the NOPUS estimates for national patterns. Section 4 examines subnational estimates, such as use by age and gender. Section 5 describes the NOPUS design and data collection methods. The estimates that are obtained from NOPUS data in conjunction with data sources, including many estimates in this Executive Summary, will be derived in Section 6. Along the way, we will derive the independently interesting statistics in Section 6.2 that that a snapshot of drivers on phones on the road (during daytime, in 2002) would consist of an estimated 67% using hand-held phones and 33% using hands-free phone devices. Figure 5: Hand-Held Phone Use By Select Characteristics 3% 3% 3% 3% 3% 1% 2% 3% 3% 4% 4% 4% 5% 4% 1% 5% 5% 2% 0% 1% 2% 3% 4% 5% Overall Male Female Age 16-24 Age 25-69 Age 70+ Urban Suburban Rural Percent of Drivers Holding Phones 2000 2002 *Drivers observed holding phones between the hours of 8 AM and 6 PM. Source: National Occupant Protection Use Survey, National Center for Statistics and An l i HT A 8 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 Terminology In this report we will use the words cwireless phone d and ccellular phone d (or ccell phone d) interchangeably. We shall also sometimes simply use the term cphone d to describe either. The term csubscriber d will be used to include prepaid subscriptions as well as subscribers to service plans. See Section 2 for a description of various types of wireless phones. We will refer to the period between 8 AM and 6 PM as cdaylight d hours. Entries of tables might not sum to totals due to rounding. NOPUS Is the Official Measure of Use Most estimates concerning driver cell phone use come from interviews of drivers, such as those in (Boyle et al., to appear), (Royal, 2003), and (Stutts, 2003). While interviews provide crucial information on aspects such as the situations in which people will use phones while driving, the NOPUS estimates are obtained from observations of actual traffic, and so better reflect actual use on the road. In fact, NOPUS provides the only probability-based observed data on driver cell phone use nationwide. In this way NOPUS provides a unique role in the study of driver cell phone use, and, in NHTSA 9s view, provides the best national measure of the incidence of driver cell phone use. NOPUS first observed cell phone use in 2000, and so the data in this report represent some of the first measurements of the actual occurrence of this form of driver distraction on the road. Wireless subscriptions have risen dramatically in recent years and are expected to increase substantially in the future. (Roche et al., 2003; Yankee Group, 2002) Consequently driver cell phone use has the potential for large increases as well. As research provides a greater understanding of the risks of this form of driver distraction and several jurisdictions consider restricting cell phone use, it will be important to continue to track actual cell phone use on the roads. Data collection, estimation, and variance estimation for the NOPUS were conducted by Westat, Inc under the direction of NHTSA 9s National Center for Statistics and Analysis under federal contract number DTNH22-00-07001. 9 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 2. Background: A Primer on Wireless Phone Technology. For the lay reader 9s information we include some basic information on wireless technology, at least as it existed in 2000 and 2002, the primary data years for this report. First we note that different people might consider different technologies to be a cwireless phone d, a ccell phone d, both, or neither. For technical definitions of cellular phones, wireless phones, satellite phones, and other wireless communications, the reader should consult documents by the Federal Communications Commission (FCC), e.g. on their website www.fcc.gov . By far the most common type of wireless phone is the cellular phone 3 a portable phone that transmits speech and other information through a network of ground-based cell towers. Cellular phones may be held by hand or may be used with a hands-free attachment such as an earpiece with an attached microphone or a headset. In addition, some cell phones come with a cpush-to- talk d feature, in which the user can use the phone e.g. in his/her lap by pushing a button when s/he wishes to speak. The user can listen to the other party on the phone by releasing the button. Another communications technology that can be used in a vehicle is a car phone, which is a phone installed and integrated into the car. Car phones either attach via cords to the dashboard, or utilize speakers and microphones in the dashboard, such as the Onstar system. In particular, car phones might be utilized in either a hand-held or hands-free mode. A satellite phone is a hand-held phone that transmits speech and other information via satellite. These phones are generally larger than cellular phones, and may have hands-free attachments. However because they are expensive and are used mostly by the military, it would be rare to observe a driver using a satellite phone on U.S. roads. Other wireless phones are incorporated into certain Personal Data Assistants (PDAs), such as Blackberries. These transmit information via cell phone towers, and may have hands-free attachments such as earbuds or headsets. However this technology was relatively rare during the time the 2000 and 2002 NOPUS were conducted. What Does cHand-Held Use d Mean? The notion of cusing a hand-held phone d might be interpreted by different people as comprising different sets technologies (e.g. is cpush-to-talk d a hand-held technology?) or different sets of behaviors (does checking email on a PDA constitute cusing a phone d?). Some people might consider cpush-to-talk d hands-free use in the context of driving a vehicle, since one is only required to touch the phone when one is speaking. Some people might consider checking email on a PDA cusing a phone d since they are using their PDA and the PDA has phone capabilities. 10 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 Some people might even limit their notion of cusing a phone d to the activities of conversing, i.e. speaking and listening to a party on the other end of the line, and so not consider dialing to be cusing a phone d. Since this report presents a variety of hand-held use estimates from different sources, it is worth summarizing what chand-held use d means in each context. Additional detail are supplied in the methodology sections of this report. The NOPUS Estimates The majority of the estimates in this report on hand-held use, and indeed all of the hand-held estimates presented by this report as new, are produced by the NOPUS survey. The survey data were obtained by data collectors observing vehicles stopped in actual traffic on the roads. Data collectors classified a driver as cusing a hand-held phone d if the driver was holding what appeared to be a phone to his/her ear . Observers were not trained in the various types of wireless phones, and so may or may not have counted technologies such as PDAs or corded car phones (or, in the event they saw one, a satellite phone) as phones, depending on their interpretation of what constitutes a cphone d. This definition of chand-held use d was implemented in NOPUS because it is a simple observable definition that captures much hand-held activity. Note however that this definition gives rise to some odd characterizations. Activities such as dialing with one 9s fingers, which could be difficult from the roadside, would not be counted as cusing a hand-held phone d in NOPUS (unless the driver were somehow manually dialing with the phone to his/her ear), while voice- activated dialing would constitute use. In summary, the NOPUS hand-held estimates reflect the types of phones subjectively viewed by the data collectors as cphones d, and drivers are characterized as cusing a hand-held phone d if they are holding the cphone d to their ear. Additional detail are provided in Section 5. Other Data Sources This report also cites estimates on hand-held use from other data sources, namely from (Boyle et al., 2001), (Royal, 2003), and (Stutts et al., 2003). These data are used for the sole purpose of constructing hands-free estimates from the NOPUS hand-held estimates. These additional surveys were conducted via telephone interviews of respondents, and so relied to some extent on the respondents 9 notions of what constitutes cusing a hand-held phone d. In some cases, survey questions might have mentioned particular types of phones, telecommunications service (such as GSM), or communications protocols (such as PCS), and this information might have influenced the respondents 9 answers. The reader is encouraged to consult these surveys for the specific wording of questions. In particular, the estimates of hand-held use from these surveys might reflect various types of wireless phones, depending on the survey questions and respondents 9 notions of hand-held use. 11 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 What Constitutes cHands-Free Use d? As with hand-held use, the notion of cusing a phone hands-free d can mean different things to different people. Is cpush-to-talk d hands-free? Is dialing considered cuse d? The estimates in this report concerning hands-free phone use derive from the NOPUS estimates of hand-held use and telephone surveys (Boyle et al., 2001), (Royal, 2003), and (Stutts et al., 2003). Again the telephone surveys necessarily relied in part on respondents 9 notions of wireless phones, what they consider use, and what use behaviors they consider to be hands-free. The reader should consult the individual surveys for additional detail. 12 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 3. Driver Cell Phone Use Nationwide This section provides information that further illuminates the NOPUS nationwide use rates. Section 3.1 discusses the national estimates of hand-held use from NOPUS, and Section 3.2 presents derived estimates on the use of cell phones through any means, whether by holding the phone or using a hands-free device. Section 3.3 compares the NOPUS estimates with those from other surveys. 3.1 The NOPUS National Estimate NOPUS found that hand-held use increased from 3% in 2000 to 4% in 2002. That is, in 2002, 4% of drivers stopped at a stop sign or stoplight between the hours of 8 AM and 6 PM were holding a phone. This estimate reflects a csnapshot d of use, but one can also infer from it that drivers were using hand-held phones for 4% of their driving time in 2002. Table 1: Cell Phone Use Nationwide In 2002 In 2000 2000-2002 Change Estimate Standard Error Estimate Standard Error Estimate Standard Error The Percent of Drivers Holding Cell Phones 1 4% 0.7% 3% 0.5% 1% 0.8% 1 Drivers of passenger vehicles with no commercial markings observed between 8 AM and 6 PM at intersections controlled by a stop sign or stoplight. Source: National Occupant Protection Use Survey, National Center for Statistics and Analysis, NHTSA The increase in hand-held use, from 3% in 2000 to 4% in 2002 is statistically significant with 80% confidence. That is, we are fairly confident (80% confident) that use increased. In addition the measured increases that NOPUS saw in virtually all of the subcategories in which it observed use, such as vehicle type, age of drivers, and urbanization, are also consistent with greater use nationwide. (See Section 4 for the NOPUS subnational estimates.) We note that during the period 2000-2002, one state (New York) Figure 6: Cell Phone Use Nationwide 0% 1% 2% 3% 4% 5% 6% 2000 2002 All Cell Phones Source: National Center for Statistics and Analysis, NHTSA, 2002 NOPUS; [G]; [S] 4% 3% Hand-Held Phones 4% 6% 13 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 passed a law banning the use of hand-held phones while driving. It is possible that the nationwide use of hand-held phones would have increased more without the NY ban. If more jurisdictions restrict driver cell phone use, this could impact future use rates. In fact when the NOPUS observations were made in June 2002, the only state that substantially restricted driver cell phone use was New York, which in November 2001 banned the use of hand-held phones while driving. Other jurisdictions such as New Jersey and the District of Columbia have since enacted hand-held bans (New Jersey in May 2004 and D.C. in July 2004), but these newer laws would not have affected the NOPUS data, which was last collected in 2002. A few other states, such as Massachusetts, allow cell phone use as long as it does not interfere with the operation of the vehicle. Some major cities, such as Santa Fe, have hand-held bans similar to that in New York. (Governor 9s Highway Safety Association website) Many jurisdictions, however, have no or relatively weak restrictions on cell phone use, and this might explain to some extent why use is as prevalent as it is. 3.2 Estimates of the Use of All Wireless Phones Using auxiliary data sources, one can extrapolate the NOPUS estimates, which only concern hand-held phones, to all wireless phones. Doing this, we estimate that 6% of drivers stopped at a stop sign or stoplight between the hours of 8 AM and 6 PM in 2002 were using some type of wireless phone, compared to 4% in 2000. These estimates are derived in Section 6. 3.3 Comparing the NOPUS Findings to Results of Other Surveys Although NOPUS 9s estimates might initially seem low, they actually indicate use to be quite substantial. NOPUS found that drivers are using cell phones on average 6% of the time during daylight hours. The percentage of trips in which a driver used a phone, and the percentage of drivers who use phones during some substantial fraction of their trips are bound to be higher, assuming most phone conversations are short compared to the duration of the trip. Indeed NOPUS 9s results are consistent with higher use rates measuring different aspects of use in other surveys. A survey conducted by the Gallup Organization in 2002 for NHTSA found that drivers who own cell phones made outgoing calls on 18% of their trips on average, and took incoming calls on 17% of their trips. (Royal, 2003) Since Figure 7: The Growth in Wireless Subscriptions 0 50,000,000 100,000,000 150,000,000 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Number of Wireless Subscribers Source: The CTIA Semi-Annual Wireless Industry Indices, April 2003 E i i n 14 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 many trips during which a driver makes or receives a call also involve periods of time in which no call is being made, it is quite conceivable that calls were made 6% of the time that drivers were on the road. In addition, since the data in (Royal, 2003) were obtained from telephone interviews, rather than from observation, the estimates from (Royal, 2003) are susceptible to errors in recollection. It is possible that respondents perceive their use to be greater or less than it is. The survey in (Royal, 2003) also found that 13% of drivers made outgoing calls on at least ¾ of their trips, with 13% taking incoming calls. The average length of a call, including both incoming and outgoing calls, was found to be 4.5 minutes. The NOPUS findings also mirror growth in wireless use in general, as reported by the wireless industry. The Cellular Telecommunications and Internet Association reports dramatic increases in the number of wireless subscribers and in the amount of time that wireless devices are used. (We note that wireless devices include devices such as pagers and PDAs, as well as cellular phones.) Wireless subscriptions increased by 40% from 2000 to 2002, from 97 million in June 2000 to 135 million subscriptions in June 2002. The total time that wireless devices were used also increased dramatically during the same time period, from 4 billion hours in 2000 to 10 billion in 2002. (Source: The CTIA 9s Semi-Annual Wireless Industry Indices, April 2003 Edition) Figure 8: Growth in the Use of Wireless Devices - 2,000,000,000 4,000,000,000 6,000,000,000 8,000,000,000 10,000,000,000 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Hours of Use Source: The CTIA Semi-Annual Wireless Industry Indices, April 2003 Edition 15 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 4. Subnational Estimates This section presents use rates for hand-held phones, broken out in various ways, such as by gender, age, and geographic region. Among these, we only have data on the hand-held / hands- free distribution for gender. Since the distribution in a subcategory might differ substantially from the ratios used to derive the 2000 and 2002 national estimates of the use of hands-free cell phone devices in Section 6, we only derive subcategory estimates of hands-free use for gender. For instance, we estimate in Table 17 that drivers in the West hold phones during 5% of their daylight driving time, but we do not estimate how often cell phones are used hands-free in the West. 4.1 Changes in Age Patterns Young drivers, ages 16-24, increased their daytime use of hand-held phones from 3% in 2000 to 5% in 2002, and drivers in the 25-69 age range increased from 3% use to 4% use. Although these increases are not statistically significant, they were large enough to make hand-held use among older drivers (ages 70 and higher), which remained constant at 1%, significantly lower than use among 16-69 year olds. See Table 6 for these estimates. In 2000, no differences in use were detected by age, but the 2000 survey might have observed at too few sites to detect a difference. Note that age is assessed subjectively by the NOPUS data collectors, and this may affect the NOPUS age estimates to an unmeasurable degree. 4.2 Males Use Phones Hands-Free Slightly More than Women Both male and female drivers used hand held phones for 4% of their daylight driving time in 2002, up from 3% for each group in 2000. This is fairly strong evidence that males and females engage in this behavior with the same frequency. We will see later in this section, however, that men spend more daylight time driving than women (see Table 4), and so the NOPUS results indicate that the total daylight driving time that men spend on hand-held phones substantially exceeds that spent by women. (See Table 5.) Because we have a hand-held / hands-free distribution for men and women, we can calculate estimates of the use of all cell phones by gender. By conducting a survey that interviewed respondents, (Royal, 2003) finds hands-free use to be more popular with men. 16 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 Table 2: The Usual Means of Cell Phone Use for Male and Female Drivers 1 Among drivers who use phones on at least 25% of their trips. Unknown responses have been distributed. Source: (Royal, 2003) In Section 6, we will use auxiliary data, namely from (Stutts et al., 2003), on the amount of time that part-time hands-free users use hands-free devices to obtain an extrapolation of the NOPUS national hand-held estimate to an estimate of hands-free use. (In this paper, we refer to people who use phones hands-free at least part of the time cpart-time hands-free users d.) If we had such auxiliary data broken out by gender, we could use the same technique to derive estimates of hands-free use for males and females. Unfortunately (Stutts et al., 2003) does not break out their data by gender, so this approach was not possible. However we can obtain cruder estimates of male and female hands-free use by adjusting the NOPUS estimates by the data in the previous table. Doing this, we would estimate that approximately 3% of male drivers on the road during the daytime are using phones via hands-free devices, since 3% = 4% x 39/61. The corresponding calculation for females estimates that females drivers use hands-free devices for 2% of their driving time. Table 3: Driver Cell Phone Use in 2002 by Gender and Means of Use Driver Gender Means of Use Male Female Holding the Phone 4% 4% Using a Hands-Free Device 3% 2% Total 7% 6% Data derived from the following sources: " The 2002 National Occupant Protection Use Survey, National Center for Statistics and Analysis, NHTSA " (Royal, 2003) That is, hands-free units are sufficiently more popular with men to make our (crude) gender estimates of hands-free use while driving different. We do not have sufficient statistical data to determine whether this difference is statistically significant. We do not have data on the hand-held / hands-free distribution by gender for the 2000 data year, so we cannot make the corresponding calculation from the previous table for the 2000 NOPUS data. Percent of Drivers Who Usually Use the Cell Phone via the Means of Use (When They Use a Phone While Driving) 1 Means of Use Men Women Holding the Phone 61% 69% Using a Hands-Free Device 39% 31% 17 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 We can also calculate the total daylight time spent on the phone while driving, by gender. Using the online survey calculator on the website of the National Household Travel Survey, we calculated that men drove for a total of 26 billion daylight hours during the course of 2002, compared to 21 billion hours for women. Table 4: Daylight Driving Time in 2002, by Gender Total Time Spent Driving Between the Hours of 8 AM and 6 PM Each Day, in Millions of Hours Driver Gender In 1995 In 2001 Average Annualized Increase In 2002 Male 22,394 25,398 2% 25,936 Female 16,367 19,924 3% 20,588 Data derived from the following sources: " The 2001 National Household Travel Survey, Dept. of Transportation " The 1995 National Personal Travel Survey, Dept. of Transportation Applying our cell phone use rates, we find that the number of hours spent by men on phones while driving in daylight exceeds that for women. The estimates that include hands-free devices in the following table are somewhat crude since they are based on a somewhat crude estimate of hands-free use by gender. Table 5: Time Spent by Men and Women Driving While on the Phone in 2002 1 Total duration of all trips in privately owned vehicles that start between the hours of 8 AM and 6 PM. 2 Based on observations of passenger vehicles with no commercial or government markings at intersections governed by a stop sign or stoplight. Data derived from the following sources: " The 2002 National Occupant Protection Use Survey, National Center for Statistics and Analysis, NHTSA " The 2001 National Household Travel Survey, Dept. of Transportation " The 1995 National Personal Travel Survey, Dept. of Transportation Since men are involved in fatal crashes at a higher rate than women, their use of potentially distracting devices could result in greater numbers of fatalities and serious injuries. In 2001, 63% of occupants of vehicles in fatal crashes were male. (This statistic was calculated from NHTSA 9s Fatality Analysis Reporting System, which contains data on all fatal motor vehicle crashes on public roadways in the U.S.) (Boyle et al., 2001) found that among cell phone users in 2000, men spent more time on their phones than women, talking for an average of 589 minutes per month, compared to 394 minutes for women. Consequently men spent on average 49% more time on wireless phones. Daylight Driving Time Each Day, in Millions of Hours 1 Men Women Total driving time 1 , in millions of hours 25,936 20,588 while holding a phone 2 1,044 879 while using a phone hands-free 2 667 395 while using a phone 2 1,711 1,274 while not using a phone 2 22,514 18,040 18 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 4.3 Increased Use in Urban Areas Driver hand-held cell phone use in urban areas increased from 2% of daylight driving time in 2000 to 5% in 2002, a statistically significant increase. (See Figure 9.) Use is now significantly lower in rural areas than in urban and suburban areas. It should be kept in mind that urbanization is assessed by the data collectors on scene, as opposed to from independent sources, such as data on population density, and so these estimates may reflect the subjectivity of the data collectors. Also note that NOPUS estimates use by drivers observed in urban areas, not the use by drivers who live in urban areas. 4.4 Increases in Pickup Trucks, in the West, and Among Minorities Although we cannot assert an increase with 95% confidence, we are fairly confident that daytime use of hand-held phones increased in pickup trucks (93% confidence), the West (90% confidence), blacks (91% confidence), and other races (89% confidence). (See Figure 10.) 19 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 4.5 Additional Hand-Held Estimates Figure 9: Driver Cell Phone Use, by Urbanization, Region, and Time of Day and Week Figure 10: Hand-Held Use, by Gender, Race, and Vehicle Type 1% 2% 3% 4% 5% Urban Suburban Rural Northeast Midwest South West Weekday Weekend Rush Non-Rush 2000 2002 H *Use between 8 AM and 6 PM. H,L: Significantly high or low in category. Source: National Occupant Protection Use Survey, National Center for Statistics and Anal sis NHTSA Significant increase 1% 2% 3% 4% 5% Overall Male Female White Black Other Pass Cars Vans & SUVs Pickups 2000 2002 Use between 8 AM and 6 PM. H,L: Significantly high or low in category. Source: National Occupant Protectino Use Survey, National Center for Statistics and Analysis, NHTSA H 20 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 Table 6: Driver Hand-Held Cell Phone Use During Daylight Hours by Select Characteristics Use in 2002 Use in 2000 2000-2002 Change Characteristic Estimate 1 Standard Error Estimate 1 Standard Error Estimate 2 Standard Error Overall 4% 0.7% 3% 0.5% 1% 0.8% Male 4% 0.7% 3% 0.5% 1% 0.9% Female 4% 0.8% 3% 0.6% 1% 1.0% Age 16-24 5% 1.0% 3% 0.8% 2% 1.2% Age 25-69 4% 0.8% 3% 0.5% 1% 0.9% Age 70+ 1% (L) 0.4% 1% 0.4% 0% 0.5% White 4% 0.6% 4% 0.8% 0% 1.0% Black 5% 1.0% 2% 0.9% 3% 1.3% Other 5% 2.0% 2% 0.7% 3% 2.1% Passenger Cars 4% 0.7% 3% 0.5% 1% 0.9% Vans & SUVs 5% 0.7% 5% 1.0% 0% 1.2% Pickup Trucks 3% 0.6% 2% 0.6% 1% 0.9% Urban 5% 0.7% 2% 0.7% 3% (S) 1.0% Suburban 5% 0.9% 3% 0.8% 2% 1.2% Rural 2% (L) 0.6% 3% 0.8% -1% 1.0% Northeast 3% 0.5% 3% 1.2% 0% 1.3% Midwest 5% 1.7% 4% 1.1% 1% 2.0% South 3% 0.5% 4% 0.9% -1% 1.0% West 5% 1.8% 2% 0.7% 3% 1.9% Weekday 4% 0.7% 3% 0.6% 1% 0.9% Weekend 4% 1.5% 2% 0.4% 2% 1.5% Weekday Rush Hour 3% 0.4% 2% 0.5% 1% 0.6% Weekday Non-Rush Hour 5% 0.9% 5% 0.9% 0% 1.3% 1 Estimated cell phone use among drivers of passenger vehicles with no commercial markings observed between 8 AM and 6 PM at intersections controlled by a stop sign or stoplight. Estimates marked with an cH d or cL d are significantly high or low in their category. 2 Change estimates marked with an cS d are statistically significant with 95% confidence. Source: National Occupant Protection Use Survey, National Center for Statistics and Analysis, NHTSA 21 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 Table 7: Driver Hand-Held Cell Phone Use During Daylight Hours by Gender and Vehicle Type Use in 2002 Use in 2000 2000-2002 Change Characteristic Estimate 1 Standard Error Estimate 1 Estimate 1 Standard Error Males 4% 0.7% 3% 0.5% 1% 0.9% Passenger Cars 4% 0.9% 3% 1.0% 1% 1.4% Vans & SUVs 4% 0.7% 3% 0.8% 1% 1.1% Pickup Truck 3% 0.5% 3% 0.8% 0% 0.9% Females 4% 0.8% 3% 0.6% 1% 1.0% Passenger Cars 4% 0.7% 3% 0.6% 1% 0.9% Vans & SUVs 6% 0.9% 6% 1.7% 0% 1.9% Pickup Truck 4% 1.7% 1% (L) 0.4% 3% 1.7% 1 Estimated cell phone use among drivers of passenger vehicles with no commercial markings observed between 8 AM and 6 PM at intersections controlled by a stop sign or stoplight. Estimates marked with an cH d or cL d are significantly high or low in their category. 2 Change estimates marked with an cS d are statistically significant with 95% confidence. Source: National Occupant Protection Use Survey, National Center for Statistics and Analysis, NHTSA Table 8: Driver Hand-Held Cell Phone Use During Daylight Hours by Driver Age and Vehicle Type Use in 2002 Use in 2000 2000-2002 Change Characteristic Estimate 1 Standard Error Estimate 1 Estimate 1 Standard Error Age 16-24 5% 1.0% 3% 0.8% 2% 1.2% Passenger Cars 5% 1.1% 3% 0.8% 2% 1.3% Vans & SUVs 4% 1.2% 6% 2.6% -2% 2.9% Pickup Truck 5% 1.6% 1% 0.7% 4% (S) 1.7% Age 25-69 4% 0.8% 3% 0.5% 1% 0.9% Passenger Cars 4% 0.9% 3% 0.6% 1% 1.1% Vans & SUVs 5% 0.6% 5% 1.2% 0% 1.4% Pickup Truck 4% 0.9% 3% 1.1% 1% 1.4% Age 70+ 1% 0.4% 1% 0.4% 0% 0.5% Passenger Cars 1% 0.3% 1% 0.5% 0% 0.6% Vans & SUVs 2% 1.1% 5% 3.2% -3% 3.4% Pickup Truck 1% 0.5% 1% 0.5% 0% 0.7% 1 Estimated cell phone use among drivers of passenger vehicles with no commercial markings observed between 8 AM and 6 PM at intersections controlled by a stop sign or stoplight. Estimates marked with an cH d or cL d are significantly high or low in their category. 2 Change estimates marked with an cS d are statistically significant with 95% confidence. Source: National Occupant Protection Use Survey, National Center for Statistics and Analysis, NHTSA 22 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 Table 9: Driver Hand-Held Cell Phone Use During Daylight Hours by Race and Vehicle Type Use in 2002 Use in 2000 2000-2002 Change Characteristic Estimate 1 Standard Error Estimate 1 Estimate 1 Standard Error White 4% 0.6% 4% 0.8% 0% 1.0% Passenger Cars 4% 0.7% 4% 0.9% 0% 1.1% Vans & SUVs 5% 0.6% 5% 1.1% 0% 1.2% Pickup Truck 4% 0.7% 2% 0.6% 2% (S) 0.9% Black 5% 1.0% 2% 0.9% 3% 1.3% Passenger Cars 4% 1.0% 1% 0.7% 3% (S) 1.2% Vans & SUVs 6% 2.0% 4% 2.2% 2% 3.0% Pickup Truck 5% 2.5% 1% 0.6% 4% 2.6% Other 5% 2.0% 2% 0.7% 3% 2.1% Passenger Cars 5% 2.6% 1% 0.7% 4% 2.7% Vans & SUVs 6% 1.6% 3% 2.9% 3% 3.3% Pickup Truck 3% 0.9% 10% 6.7% -7% 6.8% 1 Estimated cell phone use among drivers of passenger vehicles with no commercial markings observed between 8 AM and 6 PM at intersections controlled by a stop sign or stoplight. Estimates marked with an cH d or cL d are significantly high or low in their category. 2 Change estimates marked with an cS d are statistically significant with 95% confidence. Source: National Occupant Protection Use Survey, National Center for Statistics and Analysis, NHTSA Table 10: Driver Hand-Held Cell Phone Use During Daylight Hours by Urbanization and Vehicle Type Use in 2002 Use in 2000 2000-2002 Change Characteristic Estimate 1 Standard Error Estimate 1 Estimate 1 Standard Error Urban 5% 0.7% 2% 0.7% 3% (S) 1.0% Passenger Cars 4% 0.6% 3% 0.8% 1% 1.0% Vans & SUVs 5% 1.2% 3% 1.3% 2% 1.8% Pickup Truck 6% 1.6% 2% 1.4% 4% (S) 2.1% Suburban 5% 0.9% 3% 0.8% 2% 1.2% Passenger Cars 4% 1.0% 3% 0.7% 1% 1.2% Vans & SUVs 6% 1.0% 6% 1.7% 0% 2.0% Pickup Truck 4% 0.8% 1% (L) 0.4% 3% (S) 0.9% 23 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 Use in 2002 Use in 2000 2000-2002 Change Characteristic Estimate 1 Standard Error Estimate 1 Estimate 1 Standard Error Rural 2% 0.6% 3% 0.8% -1% 1.0% Passenger Cars 2% 0.6% 2% 0.9% 0% 1.1% Vans & SUVs 3% 0.8% 7% 2.4% -4% 2.5% Pickup Truck 2% 0.6% 3% 1.5% -1% 1.6% 1 Estimated cell phone use among drivers of passenger vehicles with no commercial markings observed between 8 AM and 6 PM at intersections controlled by a stop sign or stoplight. Estimates marked with an cH d or cL d are significantly high or low in their category. 2 Change estimates marked with an cS d are statistically significant with 95% confidence. Source: National Occupant Protection Use Survey, National Center for Statistics and Analysis, NHTSA Table 11: Driver Hand-Held Cell Phone Use During Daylight Hours by Region and Vehicle Type Use in 2002 Use in 2000 2000-2002 Change Characteristic Estimate 1 Standard Error Estimate 1 Estimate 1 Standard Error Northeast 3% 0.5% 3% 1.2% 0% 1.3% Passenger Cars 2% 0.5% 3% 1.0% -1% 1.1% Vans & SUVs 3% 0.9% 3% 2.2% 0% 2.4% Pickup Truck 2% 0.3% 1% 0.6% 1% (S) 0.7% Midwest 5% 1.7% 4% 1.1% 1% 2.0% Passenger Cars 5% 1.7% 2% 0.8% 3% 1.9% Vans & SUVs 5% 2.3% 6% 2.5% -1% 3.4% Pickup Truck 3% 1.1% 4% 2.7% -1% 2.9% South 3% 0.5% 4% 0.9% -1% 1.0% Passenger Cars 3% 0.4% 3% 1.0% 0% 1.1% Vans & SUVs 5% 0.8% 7% 2.1% -2% 2.3% Pickup Truck 3% 0.7% 1% 0.6% 2% (S) 0.9% West 5% 1.8% 2% 0.7% 3% 1.9% Passenger Cars 5% 2.1% 2% 0.9% 3% 2.3% Vans & SUVs 6% 1.3% 3% 1.4% 3% 1.9% Pickup Truck 4% 1.6% 2% 1.2% 2% 2.0% 1 Estimated cell phone use among drivers of passenger vehicles with no commercial markings observed between 8 AM and 6 PM at intersections controlled by a stop sign or stoplight. Estimates marked with an cH d or cL d are significantly high or low in their category. 2 Change estimates marked with an cS d are statistically significant with 95% confidence. Source: National Occupant Protection Use Survey, National Center for Statistics and Analysis, NHTSA 24 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 Table 12: Driver Hand-Held Cell Phone Use During Daylight Hours by Time of Day/Week and Vehicle Type Use in 2002 Use in 2000 2000-2002 Change Characteristic Estimate 1 Standard Error Estimate 1 Estimate 1 Standard Error Weekday 4% 0.7% 3% 0.6% 1% 0.9% Passenger Cars 4% 0.8% 3% 0.6% 1% 1.0% Vans & SUVs 5% 0.7% 6% 1.5% -1% 1.6% Pickup Truck 4% 0.7% 2% 0.6% 2% (S) 0.9% Weekend 4% 1.5% 2% 0.4% 2% 1.5% Passenger Cars 4% 1.7% 2% 0.8% 2% 1.9% Vans & SUVs 6% 2.0% 1% 0.7% 5% (S) 2.1% Pickup Truck 2% 0.7% 3% 1.9% -1% 2.0% Weekday Rush Hour 3% 0.4% 2% 0.5% 1% 0.6% Passenger Cars 3% 0.4% 2% 0.5% 1% 0.7% Vans & SUVs 4% 0.6% 5% 1.5% -1% 1.6% Pickup Truck 3% 0.5% 1% 0.5% 2% (S) 0.7% Weekday Non-Rush Hour 5% 0.9% 5% 0.9% 0% 1.3% Passenger Cars 4% 1.0% 4% 0.9% 0% 1.4% Vans & SUVs 6% 1.0% 8% 2.5% -2% 2.7% Pickup Truck 3% 0.8% 3% 1.6% 0% 1.8% 1 Estimated cell phone use among drivers of passenger vehicles with no commercial markings observed between 8 AM and 6 PM at intersections controlled by a stop sign or stoplight. Estimates marked with an cH d or cL d are significantly high or low in their category. 2 Change estimates marked with an cS d are statistically significant with 95% confidence. Source: National Occupant Protection Use Survey, National Center for Statistics and Analysis, NHTSA 25 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 5. NOPUS Survey Methodology In this section we present details on the survey 9s methodology. 5.1 The Basic Survey Methodology The NOPUS Sample The results in this note were obtained from the Controlled Intersection Study of NOPUS. This survey uses a multi-stage probability sample of roadways to produce accurate estimates in a cost- efficient manner. The Controlled Intersection sample consists of intersections that are controlled by a stop sign or a stoplight, at which stopped and slowed traffic permit detailed observation. For a complete description of the NOPUS sample design, see (Glassbrenner, 2002). Observation Protocols Data collectors observed cell phone use and demographic characteristics of the drivers of passenger vehicles having no commercial or government markings that were stopped at a stop sign or stoplight during daylight hours between 8 AM and 6 PM. Observations were made both on surface streets and at the ends of expressway exit ramps (when there are controlled intersections). The 2002 survey was conducted between June 3, 2002 and June 22, 2002, and observed 38,000 vehicles. Demographic classifications, as well as urbanization, were made according to the best determination of the data collectors. Rush hour is defined as the period 8:00-9:30 AM together with the period 3:30 3 6:00 PM on weekdays. See (Glassbrenner, 2002) for the States that comprise the four geographic regions. An important piece of information regarding observation protocols is exactly what technologies and behavior were classified in NOPUS as cusing a hand-held phone d. This is described in Section 5.3. Computation of Estimates The field data is entered, edited, and missing values of certain variables (race, gender, and age) imputed. Estimates and sampling errors are computed incorporating the complex sample design. Although we plan to use direct estimation of the variances of differences in future surveys, the variances on the 2000-2002 differences in this note are based on the assumption that the 2000 and 2002 surveys are independent. 26 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 5.2 Improvements in 2002 NOPUS observed 12,000 vehicles at 700 intersections in 2000, compared to 38,000 vehicles at 1,100 intersections in 2002. Note however that many standard errors did not decrease substantially in 2002, primarily because use rates increased. The coefficients of variation were lower on average in 2002 than in 2000. See (Glassbrenner, 2003) for more information on how the new observation sites were chosen. In addition, we calculated a number of new estimates in this report, such as the total numbers of hours and miles driven by people talking on cell phones and the percent of wireless use that occurs behind the wheel, given in the Executive Summary. These calculations were derived in Section 6 and involve extrapolations using auxiliary data sources, such as the National Household Travel Survey. 5.3 What Does NOPUS Consider cUsing a Hand-Held Phone d? Data collectors were instructed to count a driver as cusing a hand-held phone d if the driver was holding what appeared to be a phone to his/her ear . Observers were not trained in the various types of wireless phones, and so the technologies they classified as cphones d likely reflected their own conceptions of what constitutes a cphone d. In particular, individual observers might or might not have classified corded car phones, Personal Data Assistants, or (in the unlikely event they saw any) satellite phones, as cphones d. As noted in Section 2, this definition of chand-held use d was chosen because it is simple (hence leading to fewer observational errors) and can be implemented in the context of observing traffic from the roadside. E.g. the definition purposely excluded activities such as manual dialing since this could be difficult to reliably observe from the roadside, as it can occur below the level of the windows and windshield from the observer 9s point of view. However (as we also noted earlier) this can lead to somewhat odd distinctions as to what is considered cuse d. E.g. manually dialing is not considered to constitute use in NOPUS, but voice-activated dialing is. We note that much of the activity that the NOPUS definition of chand-held use d captures is conversational, i.e. speaking and listening to a person on the other end of the line. However we do not use the term cconversing on a hand-held phone d to describe the NOPUS estimates because this description is also not quite accurate, given that NOPUS counts voice-activated dialing and vocally navigating a customer service menu as cuse d. NOPUS observers did not receive specific training in how to deal with situations in which they initially observed a driver not using a hand-held phone, and later (in the same stopped traffic) observed him/her to be holding a phone. In this situation, observers may have changed their initial assessment from cnot using a hand-held phone d to cusing a hand-held phone d, or not. 27 National Center for Statistics and Analysis, 400 Seventh St., S.W., Washington, DC 20590 The 2002 survey does not observe whether hands-free devices are being used, although the 2004 survey has been expanded to include the observation of headsets into which the driver is speaking. 5.3 Design Aspects that Impact Data Two limitations concerning the NOPUS data are worth particularly noting. First, note that the estimates in this report reflect daytime use only, whereas use patterns may well differ at night. Next note that NOPUS only observes vehicles that are stopped at a stop sign or stoplight. It is entirely conceivable that use rates are higher at such intersections, where e.g. drivers might decide to try to make a short phone call during a red light, than on general roadways. It is also conceivable that drivers might use hand-held phones more frequently on roads that have controlled intersections than on other roadways because they expect to be periodically stopped at lights or stop signs. Consequently, estimates from or based on NOPUS may overestimate use on general roadways, and may over- or underestimate use at general times of day. However because it is the only probability-based observational survey of driver cell phone use nationwide, NOPUS provides the best observed use rates available. Estimates in this report on hands-free use that were derived in part from NOPUS inherit the same data limitations to some degree. For instance, our estimate that drivers used phones hands-free for 2% of their driving time in 2002 was derived based on both NOPUS, which reflects only daytime use, and telephone surveys, which may reflect what the respondents view as their ctypical d behavior, and so may or may not reflect use at other times of day. The net effect of these survey limitations is, of course, unclear. 5.4 Assessing Significance Because NOPUS uses a probability sample, one can calculate the error its estimates incur from observing use for a sample of roadways and times rather than for all roads and times. The actual quantity being estimated by a NOPUS estimate is within twice the standard error of the estimated value with 95% confidence. (Standard errors for sample-based estimates are provided in the tables in this report.) This computation can be used to determine whether

less