Gas shale swelling during hydraulic stimulation is one of the major challenges in unconventional gas development. It is hypothesized that the large volumetric strain upon water imbibition is a consequence of dramatic changes in capillary pressure and disjoining pressure. However, how they are combined to induce bulk deformation in these partially saturated nanoporous materials is unknown. This doctoral thesis set out to develop an effective stress framework for partially saturated gas shales. Laboratory testing was performed to fill the methodological and evidence gaps, under two overarching goals: to establish water retention behavior and to quantitatively study gas shale elasticity in partially saturated conditions. Controlled suction experiments were carried out on intact samples to construct water retention curves in terms of degree of saturation over a wide range of total suction. The obtained curves were divided into three distinctive saturation domains that represent adsorbed water, capillary water, and hydrocarbon-wet pore volume. The high capacity axis translation (HCAT) triaxial apparatus was developed and verified for its performance. Equipped with the new filter, Shale Disc, the apparatus is now capable of reproducing reservoir conditions by controlling capillary pressure up to 11 MPa under triaxial stress. A series of hydrostatic compression tests were performed at different saturation states to understand the impact of water saturation on gas shale elasticity. It was found that the drained bulk modulus and its stress dependency vary significantly with water saturation. The collected experimental evidence was then compiled to develop an effective stress framework. The effective stress expression incorporates the contributions of suctions by two hydraulic stress variables: capillary pressure and a macroscopic average of disjoining pressure. As a preliminary development, the proposed framework was validated against unconfined swelling and shrinkage data and showed good agreement. The key contribution of this thesis is the experimentally determined water retention behavior and bulk modulus combined in an effective stress framework. It is anticipated that the work described here will serve as an empirical basis to improve the prediction of gas shale swelling and shrinkage.