A new linear method for the determination of the rate coefficients of complex first-order (or pseudo-first-order) mechanisms is presented and applied to simulated data. The errors associated with parameter recovery are compared with those of the traditional nonlinear least-squares method. Nonlinear methods based on convolution kinetics are also developed, and general convolution equations are obtained. Special attention is paid in both cases to excited state kinetics, where concentrations are usually known only up to a constant factor. The monomer-excimer kinetics is discussed in detail, explicit relations for parameter correlation being obtained. The influence of transient effects is also quantitatively discussed.