A novel methodology used taxi global position system data and high-resolution transit schedule information to compare travel times and travel fares of the two main nondriving travel modes for airport ground access: taxi and transit. Five origin-destination pairs between Pennsylvania Station in New York City and three airports in the New York region were used as an example to demonstrate these methods. An analysis of total trip cost considered hoth travel time and expenditures on fare. A binary logit model was used to model the mode choice of travelers. The results indicate that transit is the more likely choice during most of the day except the midnight period when transit service has longer headways. A sensitivity analysis shows the relationship between the value of time and total trip cost per passenger for different numbers of passengers traveling together and at different times of day. The higher the value of time and the number of passengers in a group, the more likely it is that a taxi is chosen for airport trips. The attractiveness of one mode relative to the other varies spatially and temporally according to the travel time and price. This paper focuses on understanding temporal variation of total cost of each mode and the effect that this variation is likely to have on mode share.