The large-scale distribution of matter in the Universe encodes a wealth of information about the most fundamental ingredients of our Universe. The most popular way to describe this information is via the power spectrum (the 2-point correlation function), but there is additional important information contained in higher-order correlation
functions. In this talk, I will describe a formalism that allows to easily evaluate these higher-order statistics using so-called “response functions”, which describe how the power spectrum “responds” to large-scale perturbations. This is done by combining perturbation theory methods with a simulation technique called “Separate Universe
Simulation”. As a concrete application, I will show how responses have proved extremely useful in evaluating (as well as understanding the physics behind) the covariance matrix of the weak-lensing power spectrum, which is a 4-point function that plays a central role in the data analyses of current and future-surveys.