Twenty-one subjects completed two sets of (8) laps around a 7.5-mile test track during two 4-hour sessions. They drove an instrumented vehicle while performing a combination of car following, peripheral target detection, and secondary (in-vehicle) tasks of varying complexity. Subjects performed one set of laps with each of two interfaces, voice-based and visual/manual. Secondary tasks comprised three categories including baseline tasks (radio tuning, phone dialing), simple tasks (message retrieval plus voice memo creation), and complex tasks (simple task components plus phone dialing and information retrieval from automated phone systems). Measures of driving performance, target-detection, secondary task performance and eye movements were recorded. Analyses were conducted to determine whether the voicebased interface reduced the relative distraction potential for secondary tasks of varying complexity. Generally, differences between tasks were stronger than differences between interface conditions. Measures of car-following performance, target detection, and secondary task performance revealed differences attributable to task complexity. Differences between the two interfaces were observed on peripheral target detection measures and on several driving performance measures. Overall, the benefits of using the voice-based interface were not large enough to appreciably reduce the distraction potential associated with performing the secondary tasks in the car-following scenario.