Electric vehicles (EVs) have a great potential as controllable flexible loads for an electricity market operator. Hence, it is important for an electricity market operator to assess the extent of demand flexibility (i.e., demand shifting and demand curtailment) available from the EVs. In this paper, an electric vehicle aggregator (EVA) mathematically models the demand flexibility based on the data pertaining to its EV consumers and random 24-hour day-ahead price scenarios. The demand flexibility offered by an EVA will be represented using a price elasticity matrix which is calculated with respect to a flat reference price scenario. Convex programming is used to generate the EVA-level data required for constructing the price elasticity matrix. This paper separately analyzes the continuous charging case and the discrete charging case based on the proposed modeling approach. The results also demonstrate that a feasible demand schedule can be obtained for an EVA under different 24-hour day-ahead price scenarios.