Decomposition methods have been widely used in econometrics and policy analyses when counterfactual and/or social experiments are impossible to reproduce and there is the need to evaluate the effect of a certain policy or covariate on two different groups which differ over a set of characteristics. Decomposition analysis allows researchers to hold constant those specific groups factors that would tend to distort the comparison between them. During the seminar, I will introduce these methods starting from the seminal work by Evelyn M. Kitagawa (1995), who developed a mean decomposition technique later resumed by the more well-known authors Oaxaca and Blinder. I will then introduce the different methods that have been developed to study distributional effects, such as quantile analyses, and I will give an overview of their applications. I will devote the last part of the seminar to introduce one peculiar application of these methods: their use for spatial analyses. Alternative decomposition methods, such as statistical decompositions, are finally discussed.