This paper derives the individual and the total change in welfare because of changes in frequency and reliability for a scheduled service. These values are derived from scheduling preferences and it is assumed that travelers plan their trip. The model takes into account anticipating behavior of travelers meaning that travelers may shift to an earlier connection if arrival times are more variable. First it is shown that the marginal social benefits of improved frequency and arrival time reliability do depend on the distribution of preferred arrival times that is assumed. Under the assumption that lateness is more severe than earliness, increasing the frequency is more beneficial for the early segment of the peak period. Second, demand elasticity is included in the model. It is shown that the benefits of improvements in frequencies and reliability are overestimated if the elasticity is ignored. The analysis is done for planning travelers and therefore the results are useful when analyzing the individual and social cost of unreliable train or air connections.