In this thesis I examine some extensions to the concordance model of cosmology and their observational
consequences. In particular, I look at how vector perturbations sourced by cosmic strings
could generate a B-mode signal in weak lensing surveys. For a string network close to the current
observational upper bound, this signal could be detected within the next 5-10 years. I also
examine model independent modified gravity and forecast the constraints that could be obtained
with the combination CMB+weak lensing+galaxy cluster counts. For the specific example of Planck
(CMB+cluster catalogue) and DES(weak lensing), deviations from General Relativity could be constrained
to within 5% in the next 3-4 years. The forecast for cluster counts relies on mass functions
calibrated for ∧CDM+GR cosmologies, so I then investigate how reliable the predictions are in the
case of modified gravity. At first glance, it appears that the mass function formalism is reliable and,
with a little cheating, can even work in the case where modified gravity has a screening scale.