The paper discusses methods for estimating heterogeneous effects in labor economics, focusing on settings where researchers compare various units to infer differences in effects. It presents a unified framework based on a linear model with normal random coefficients and errors. The model allows for the recovery of mean and dispersion of effects, as well as the construction of predictors of these effects. The paper highlights the economic content of key assumptions and discusses practical estimation strategies. It provides moment conditions on the model's parameters and outlines various estimation approaches. The normal random coefficients (RC) model is central to the analysis, allowing for the estimation of moments, higher-order moments, and distributions of effects. The paper also addresses the challenges of estimating these effects in complex data settings with high-dimensional models and large numbers of covariates. It discusses the implications of normality and linearity assumptions and explores ways to relax these assumptions. The paper reviews related literature and discusses specification choices for the model, including the variance-covariance structure of errors and the mean and variance of unit-specific effects. The paper concludes with a discussion of estimation strategies for the parameters of the normal RC model and the quantities of interest.The paper discusses methods for estimating heterogeneous effects in labor economics, focusing on settings where researchers compare various units to infer differences in effects. It presents a unified framework based on a linear model with normal random coefficients and errors. The model allows for the recovery of mean and dispersion of effects, as well as the construction of predictors of these effects. The paper highlights the economic content of key assumptions and discusses practical estimation strategies. It provides moment conditions on the model's parameters and outlines various estimation approaches. The normal random coefficients (RC) model is central to the analysis, allowing for the estimation of moments, higher-order moments, and distributions of effects. The paper also addresses the challenges of estimating these effects in complex data settings with high-dimensional models and large numbers of covariates. It discusses the implications of normality and linearity assumptions and explores ways to relax these assumptions. The paper reviews related literature and discusses specification choices for the model, including the variance-covariance structure of errors and the mean and variance of unit-specific effects. The paper concludes with a discussion of estimation strategies for the parameters of the normal RC model and the quantities of interest.