Tests of gravity on large-scales in the universe can be made using both imaging and spectroscopic surveys. The former allow for measurements of weak lensing, galaxy clustering and cross-correlations such as the ISW effect. The latter probe galaxy dynamics through redshift space distortions. We use a set of basic observables, namely lensing power spectra, galaxy-lensing and galaxy-velocity cross-spectra in multiple redshift bins (including their covariances), to estimate the ability of upcoming surveys to test gravity theories. We use a two-parameter description of gravity that allows for the Poisson equation and the ratio of metric potentials to depart from general relativity. We find that the combination of imaging and spectroscopic observables is essential in making robust tests of gravity theories. The range of scales and redshifts best probed by upcoming surveys is discussed. We also compare our parametrization to others used in the literature, in particular the gamma parameter modification of the growth factor.