The participants were 30 drivers who all have their own cell phones and have used the phone while driving. The drivers were sampled from three age groups:

The rationale for selecting these particular age groups is that the novice drivers are very experienced in using the phone but not in driving, and their psychomotor abilities to integrate dual tasks is the greatest. The experienced drivers are experienced in both tasks and in combining them as well. The older drivers typically do not time-share phone conversation with driving, and are the least capable of learning to combine these tasks.

Equipment and Measurement Devices:

The participants drove a STISIM PC-computer-based simulator, installed in a cab of a passenger car. The driver’s field of view was projected on an 8x6-foot screen located in front of the car, subtending a 40-degree horizontal visual field with a 1:1 magnification ratio. The STISIM is a fixed-base simulator with interactive steering wheel, brake pedal, and accelerator. The simulator has a scene definition language that allows flexibility in the design of the route and traffic. For this study, the roadway consisted of a relatively straight two-lane highway with few turns and little traffic. There were cars that occasionally came towards the participants’ vehicles in the opposite lanes, and other cars that passed them. However, except in the car-following task, on no occasion did the driver come up to a car proceeding at a slower speed.

Built in with the driving task was an “off-road” secondary target detection task. Its purpose was to add to the driver’s information processing load in an otherwise easy drive. It consisted of two diamond shapes positioned on the right-top and left-top corners of the screen. In a random manner, one of them would change shape from a diamond to a triangle, and would remain in that shape for 5 seconds. The participant’s task was to move the signal handle in the direction of the diamond that changed shape. A total of 12 signals were given on each session, with 0-3 signals on each of nine 3-minute segments. Because in some conditions no signal appeared at all, performance data on this task were not analyzed as a function of the distractions.

To simulate the cellular phone demands, two hands-free dual tasks consisting of listening and responding to verbal questions were used. All of the information was given through a dedicated speaker installed on the dashboard to the left of the steering wheel. Two different kinds of distracting “phone tasks” were given:

  1. Math operations task – the drivers were presented with a sequence of numbers and operations and asked to provide the final answer. For example: [(3 + 6 + 9 – 2 + 4) X 2 + 6] / 2 =? Answer: 23. This task was identical to the one used by McKnight and McKnight (1993). The numbers were presented at the rate of 1 per 2 seconds, and the participants were allowed five seconds for their final response.

  2. Emotionally-involving conversation – Prior to the driving, as part of the background interview, the participant was asked about his/her school, work, social habits, hobbies and interests. Based on these, a series of questions were developed to generate conversations that would be emotionally challenging. The topics differed for the different age groups and different participants within each age group. For example, if the driver was an avid sports fan of a specific team then one conversation topic would be about how that team is doing, challenging the participant that his/her team has not been doing that great lately. As in a real phone conversation, the experimenter could not see the driver or the driving scene during the conversation, and thus its pace (as far as the experimenter was concerned) was uninfluenced by the driving task.


Each participant participated in 5 sessions, extending over 14 days with no less than one day between sessions and no more than 4 days between sessions. The first session lasted close to an hour and the rest of the sessions lasted slightly over 30 minutes.

In session 1 the participant was given a brief questionnaire, interviewed about his/her hobbies and interests, and then asked to drive the simulator without any distracting tasks for a 4-minute practice session. Next the participant received three minutes of practice in each of the two distracting “phone” tasks. The length of this practice session was based on previous research experience with this simulator. The practice period was followed by a short break after which the experimental session began.

Each experimental session consisted of three 9-minute blocks of driving – each divided further into three 3-minute segments. The three segments consisted of a random order of the three levels of the distracting phone tasks: no distraction, math computations, and an emotionally-involving conversation. There were no breaks between the segments within each 9-minute driving block, but there was a one-minute break between driving blocks. The three blocks consisted of three “speed conditions”: driving while attempting to maintain a constant speed of 50 mph, driving while attempting to maintain a constant speed of 65 mph, and driving while following a car whose speed varied between 50 and 65 mph. The order of the phone conditions (within the blocks) and the order of the speed conditions (between the blocks) were counterbalanced within and between subjects.

All drivers were asked to drive as close as possible to the center of the lane. In the constant-speed conditions, whenever the driver’s speed deviated from the required speed by 10 mph the driver was reminded of the required speed and asked to resume it. This actually happened very rarely, and only on the first session.

At the end of each session the participants were asked to fill out a questionnaire relating to their perceptions of their task that day. There were seven questions, and for each question the participants had to rate their feelings on a scale of 1 to 9, with 1 indicating “very little” and 9 indicating “very much.” The questions were:


The study design was a within subject repeated measures paradigm. Performance was evaluated on the basis of driving-related measures and phone-related measures.

The independent variables whose effects on driving-related measures were evaluated included:

The driving-related dependent measures included the following:

The distraction-related dependent measure was the number of correct responses in the math operations task. It was evaluated relative to the three independent measures (driving condition/speed, practice, and driver age/experience).

The subjective evaluation scores – provided at the end of each session – could only be evaluated relative to the practice effect.