Speakers were told to share a "close call" story that they had not shared with the driver before.
There were two distinct sections of the experiment: dual-task and single-task. During the dual-task scenario dyads were conversing either in person or over a simulated phone call. During the single-task there was only driving and no conversation: the single-task scenario serves as the control. The order of the assignments was randomized. In all driving conditions, drivers were given certain instructions on what exit ramp to use. The whole experiment took 60 minutes, but individual driving sequences lasted ten minutes for both single-task and dual-task sections.
What types of measures do they use?
Measures include those involving driving performance and conversation analysis. Driving performance measures include operational, tactical, and strategic measures. Operational measures include "how well participants stayed in the center of the lane without lateral moving or drifting," (Drews, Pasupathi & Strayer 2008, p. 395). Tactical measures included average speed and average following distance (distance between the driver and next car ahead). Strategic performance measures included navigational issues such as whether they took the correct exit.
What is the hypothesis?
The hypothesis is stated on page 393: "the different contexts affect the ability to allocate attention to a task differently…the allocation of attention is not independent of contextual variables, even if the task at the onset seems identical."
The research explores the difference between drivers conversing with passengers and drivers conversing with a person over the phone. The researchers hypothesize that a conversation with a passenger is qualitatively different than a conversation on a cellular phone. The reason for the difference is that attention is allocated differently in phone vs. in-person conversations. The researchers also suggest that driving while speaking with a passenger is relatively safe, whereas driving while talking on a cell phone can lead to an accident.
Conversations were transcribed and analyzed. Although the close-call story was required, the conversations drifted and were analyzed in terms of references to situational variables like traffic. Turn-takes in the conversation were also analyzed. Word complexity, measured by number of syllables per word, was analyzed. The impact of the driving environment, or how...
397). In tactical performance, no changes in driving speed were noted for each conversational condition. However, following distance was greater during the cell phone conversation. Drivers using cellular phone were four times more likely to fail the navigational challenge.
Fewer references to traffic conditions occurred during the cellular phone conversation. Number of turns was greater for the passenger vs. cellular phone conversations. Production rate of conversation was greater in the cellular phone condition, but complexity was not significantly different.
What are some critiques? Ideas for future directions? (say something beyond what is presented in the article.)
Future research might attempt to analyze a real driving scenario, which is bound to have a different cognitive affect on both the driver and the conversation partner.
Is the sample sufficiently large and representative of the population of concern?
The sample is sufficiently large, but not representative of the population of concern because each state and each region will have different styles of driver. Also, familiarity with the road and region might affect driver performance.
Are there any groups (based on gender, age etc.) left out that results might not generalize to?
The authors do not indicate any demographic variables besides age and gender. Ethnicity is not taken into account, and nor is length of driving experience.
Are there any weaknesses in the study design?
Other than the aforementioned critiques, the study design is not problematic.
Are there outside variables that are not being controlled for?
Outside variables include the drivers' level of experience, emotional state during the experiment, and relationship with the friend/partner.
Is the study ecologically valid? will findings generalize for the real world, or only work in an experimental setting?
The experiment uses a simulation and so is not ecologically valid. The conversations were "naturalistic" but not entirely so because of the simulation.
