Power Line Communications (PLC) emerged as a competitive technology for indoor broadband communications. It exploits the existing power line infrastructure for data transmissions. Transmission delays may essentially degrade the performance of applications relying on two-way communications such as videoconferencing, remote control systems, online games, and so on. Reported PLC channel models indirectly characterize signal propagation delays through channel transfer functions or impulse responses, but they do not address more important delays occurring because of multiple retransmissions of lost data packets in noisy PLC channels. This paper proposes a cost-efficient data delay measurement and a high-level statistical modeling approach for applications that are sensitive to data delays. Hybrid statistical distributions are explored for accurate delay characterization. © 2014 IEEE.