I am not sure how you conduct your experiments. Your results with salts might be  correct considering that the outcomes will be different with and without salts.

Adding of salt sure changes the structure of water and thus the cavities in water.- this has severe efects on solubility of compounds but also (naturally) on the partitioning of compounds. - A prety well known effect in organic chemistry . it has been described as  salting out concerning extractions.

It depends on how much salt you add.  The change in the activity coefficient of water is great if you add salt, but that does not change the mole fraction of water in the water phase all that much (answer to your last questions).  More importantly, you should consider what type of hydrocarbon phase you are using.  All hydrocarbons are not the same, and salts can for ion pairs in some hydrocarbon phases to a larger degree than others.  If the hydrocarbon is a petroleum product (organic mixture), there are often polar groups on some many of the components that salts can interact with.  Regarding Kai's comment, he is correct about the "salting-out effect".  If you have some polarity in the hydrocarbon phase, this will allow water to dissolve extensively in the hydrocarbon (the amount of water in octanol at mutual saturation is about 2.3 Molar).  Adding salt to the water increases the affinity of the water phase for water, and the amount of water in the hydrocarbon phase will be much less.  Hence, adding salt to a liquid-liquid extraction often results in a better extraction of an organic solute because it makes the hydrocarbon phase more hydrophobic, makes the water phase more hydrophilic, and tends to brake any emulsions that form by keeping the water in the water phase.