Travel time reliability is a key metric of interest to practitioners and researchers. There are several metrics available for travel time reliability - buffer index, planning index, probability of travel time being higher than a specific value, etc. Travel time reliability metrics have been defined for transit and automobile, and plays as important as a role as average congestion in affecting travel choices. Most of these metrics have forms such as the ratio of percentile to sample mean, the ratio of percentiles to the median, etc. Traditionally, transportation engineers and planners have used point estimates to compare travel time reliability for various modes. Ready-made procedures to perform statistical inference - confidence intervals and hypothesis tests are not available for these metrics. This project will evaluate and develop methods to determine confidence intervals and hypothesis tests for select travel time reliability parameters. Another complicating factor is the travel time distribution, which is not normal and often asymmetric and varies depending on the location and time. Travel time distributions vary depending on facility type, traffic conditions, signal settings, and composition. We will focus on distribution-free confidence intervals and hypothesis test procedures which work for a wide range of shapes of distributions. We will also study the applicability of existing travel time reliability metrics for class one vehicles (bicycles and motorbikes) and the feasibility of defining an overall travel time reliability of an arterial segment that considers all modes.