Introduction

While distracted driving has been around since the advent of automobile travel, recently it has become a focus of increasing interest.  Some of this interest is attributable to the increased use of cell phones and to a recent surge in state legislation to curb cell phone use while driving, possibly because cell phones are among the newer and more visible array of driver distractions. Drowsiness is another condition that takes drivers’ attention away from the road. 

Despite the growing concern over these behaviors, little information is available on the specific conditions under which drivers engage in various distracted and drowsy driving behaviors.  In addition, little information exists on the characteristics of the drivers who exhibit these behaviors.

To help answer these and other questions surrounding distracted and drowsy driving, the National Highway Traffic Safety Administration (NHTSA) has undertaken a national survey of the driving public's attitudes and experiences related to distracted and fatigued driving behaviors.  NHTSA contracted with The Gallup Organization to conduct two surveys of the driving public’s attitudes and behaviors regarding speeding, and a second dealing with unsafe driving.  Both surveys covered aggressive driving, driver distraction and fatigued driving. 

Similar methods were used to field the two surveys and many of the questions were asked on both surveys to provide more robust estimates.  The data collected in the two surveys resulted in two separate reports: this report on Distracted and Drowsy Driving and a second report on Speeding and Aggressive Behaviors, which is published under a separate cover.  

Methods

Sampling Objective

The sampling requirement of the two studies was the same: acquire a representative national sample of drivers age 16 and older in the 50 U.S. states and Washington, D.C. 

In order to accommodate the need to acquire data on topics of speeding, aggressive and unsafe driving, distracted and fatigued driving, two separate surveys were undertaken. The first survey focused on speeding behaviors, while the second survey focused on aggressive driving and other unsafe driving behaviors.  Both versions measured distracted driving and drowsy driving.  In addition, split-sample procedures were used within each version to extend the number of questions that could be asked within the 18-minute telephone survey.

Each survey instrument was fielded as an independent national sample and was constructed in an identical manner.  Gallup used a three-stage procedure to meet the sampling objective:

1.                   Gallup first identified the universe of residential telephone listings within each of the eight U.S. Census Regions.  

2.                   Second, Gallup drew a systematic sample of telephone 100-number blocks within each region. Gallup then randomly generated the last two numbers for a full 10-digit phone number within each valid block selected in the previous stage.  This procedure provides for an equal probability of selection for each working residential telephone number in the United States (both listed and unlisted residential telephone households).

3.                   Next, a single driver age 16 or older was randomly selected (using the “most recent birthday” method described in the Methods report) for inclusion from all eligible members of the driving public residing in that household.

Up to 14 attempts were made to reach each randomly selected respondent.  Seven attempts were made to reach the household, and once a respondent in the household was identified, Gallup made up to seven additional attempts to reach that person.

Using the two surveys, Gallup completed a total of 4,010 telephone interviews with vehicle drivers age 16 and older between February 4, 2002 and April 14, 2002.  Interviews were completed in both English and Spanish, using a computer-assisted telephone interviewing  (CATI) system.

Sample Weighting

While the two samples were weighted separately, similar sample weighting was carried out for each sample.  The final telephone samples of drivers age 16 and older were weighted to equalize selection probabilities (at both the household and the individual levels ¾ particularly since we excluded non-drivers), and to adjust for non-response bias by demographics.  In the last stage of the weighting process, the adjusted results were projected to the number of drivers age 16 or older in the United States  A detailed description of the weighting procedures can be found in      Volume II: Methods.

The final number of weighted and unweighted interviews by age and gender appear below:

Precision of Sample Estimates

All sample surveys are subject to sampling error in that results may differ from what would be obtained if the whole population had been interviewed.  The size of such sampling error depends largely on the number of interviews. For the main sample of 4,010 telephone interviews, the expected maximum sampling error range is approximately +/- 1.5% at the 95% level of confidence. The table above shows the sampling error ranges by age and gender at the 95% level of confidence.  Due to the stratification and other complexities of the sample design, in some cases (particularly among smaller sub-groups of the population) the error ranges will be slightly larger than those shown in the table.  This information is provided to offer the reader a general sense of the range of the true estimates.  The report Volume II: Methods, presents a table showing the expected sampling error ranges for sub-group sizes in the sample.  

Data Presented

It should be noted that this is a top-line report on survey data and includes responses from more than 4,000 persons of driving age on more than 200 survey questions.  The report is not intended to provide in-depth analyses of any one topic, but rather to give the reader a general overview of the data. Additional analyses may be done at the reader's discretion.

The data in this report are based on driver responses from two separate surveys conducted concurrently.  Some of the questions were shared between the two surveys, while others were unique to one of the two surveys. The two surveys were referred to as “Speed” and “Unsafe” to identify their primary topic differences. Figures in the report identify from which of the two surveys (Speed or Unsafe) the data are based.    In addition, within each survey version some questions were asked of a random half-sample of drivers, rather than the entire survey base.  These items are noted with an “(A)” or “(B)” marker in the figure.