Recent advances in cosmological simulations have made it possible to produce high resolution simulations of galaxies with accurate gas physics and stellar feedback. One advantage of these simulations is they allow for realistic mock observing of galaxies at all scales. We use dwarf galaxies simulated with the GIZMO hydrodynamics code, and the FIRE-2 feedback model to investigate biases in CMD-derived star formation histories, and velocity dispersions from RGB stars. For star formation histories, strong age gradients mean that biases from incomplete coverage of the galaxy can be quite large, about 10 to 20% even for modest coverage fractions. For velocity dispersions, the error from incomplete coverage is small, and converges rapidly with increasing coverage. However the RGB stars which are used as kinematic tracers potentially trace a biased stellar population which can lead to errors of about 10% in the velocity dispersions of dwarf galaxies. We use these results to make predictions of how to avoid biases in star formation histories and velocity dispersions with future observations.