The Constrained Variables Method (CVM) was one of the first and most common approaches in demographic and social research to identify the separate effects of age, time period and birth cohort. The method, which became widely known through an article by Mason and colleagues in 1973, is applied by constraining at least two categories of the age, period and/or cohort variables to be equal. A lively and still ongoing debate has emerged concerning the appropriateness of the CVM to estimate age, period and cohort (APC) models. The most important objection to this method is the potential arbitrariness of the chosen constraint(s), which may lead to biased if not invalid conclusions. As a consequence, APC scholars have turned towards other, often purely statistical, approaches to circumvent the identification problem. But has the CVM rightly been abandoned? In this chapter we describe the essentials of the CVM, give an overview of the main criticisms and pitfalls, and provide some counter-examples and recommendations to overcome these pitfalls. We conclude that the CVM is by no means useless when constraints are made explicit and have a strong theoretical and empirical foundation.