A method that allows one to determine the relationship between the evolution of copolymer composition and the distribution of glass transition temperatures is presented in this work. The experimental validation of the model is based on histograms of T-g distributions which were computed from DSC thermograms by using a new model-based optimisation technique. This technique allows us to obtain both the average T-g of the overall copolymer and the cumulative distribution of T-g as a function of the degree of advancement of the copolymerisation reaction, even in the case of reaction with composition drift. The modelling strategy was used successfully to analyse the evolution of the T-g with conversion of several free radical copolymers of styrene and butyl acrylate in emulsion. In particular, it has been shown that the value of the glass transition temperature of an ideally alternating copolymer, T-g, used in the Johnston (1973 Macromolecule 6, 453-456) equation, is in fact a function of the copolymer composition, and may vary by as much as 50 degrees C. The model of T-g presented here can be used to control profiles of glass transition temperature distributions (not just averages) during semi-continuous polymerisation processes.