The effect of uniaxial and multiaxial mechanical strain on the electrical conductivity of particle filled polymer composites is investigated in the framework of concentration-driven percolation. For composites consisting of low aspect ratio, rigid conductive particles in a compliant polymer matrix, a simple argument leads to the conclusion that the effective volume fraction of conductive particles (the ratio of total particle volume to the total volume of the deformed composite) plays a dominant role, with conductivity remaining isotropic despite the directional bias of the strain state. As such, conductivity is expected to exhibit classical power, law-dependence on concentration, which in this case takes the form of a strain-dependent effective volume fraction. Consideration of deformation effects on particle agglomerates suggest, however, that particle-to-particle network connections are likely to be affected most significantly along directions experiencing the most severe strains, introducing a directional bias in network connectivity at a higher length scale. To assess the importance of this possible directional bias, random resistor network models are used to study the conductivity of uniaxially strained composites. For conservative assumptions on the severity of the bias in bond probabilities, network conductivities exhibit approximately isotropic, concentration-driven behavior for moderate strains, supporting the predictive utility of the simple percolation conduction-effective volume fraction approach. Further corroboration is provided by experiments in the literature on silicone-graphite composites subjected to uniaxial compressive strain, where good agreement is obtained through moderate strains for the theoretically correct value of the conduction exponent in concentration-driven percolation.