In a sensor network, the nodes that have a packet loss rate greater than a given threshold are called lossy nodes. Lossy nodes adversely affect network performance and collectively cause breakdown of network service in many cases. In this paper, we propose a new and simple approach to identify the lossy nodes from passive end-to-end measurements. Specifically, we consider the case of identifying lossy nodes during the aggregation of data from a collection of the sensor nodes to a sink node. The proposed method is based on the hamming distance of sequences on receipt/loss of aggregated data between each pair of parent-child node. With prior knowledge of the reverse data aggregation topology, our approach can identify all lossy nodes effectively. Finally, we validate the proposed approach.