We propose a coding scheme for the stabilization of continuous linear scalar systems under a communication channel with finite data rate, lossy observations and network-induced delay. Owing to the network-induced delay, uncertain control input is introduced to the corresponding discrete time systems. Furthermore, since the system state is quantized to finite-bit signals and may be randomly lost in the communication, under the proposed coding scheme, the considered scalar system is described by a switched system of more than one dimension with arbitrary switchings. We derive the conditions for first moment stability of the systems, characterizing the relations on quantization cell boundaries, packet loss probabilities, and the network-induced delay. Further, we derive the condition for almost sure stability of the systems. A numerical method is proposed to search the infimum of the data rate for first moment stability under the given condition. It is shown that for the quantizer with the infimum of the data rate for first moment stability, although the quantization cells have different lengths, the matrices corresponding to these quantization cells have the same maximum eigenvalue.