Abstract A broad class of solid phase catalytic chemical reaction kinetics can be modelled by the Hougen—Watson approach based upon the Langmuir adsorption isotherm. An appropriate statistical procedure for determining the parameters of the resulting models is nonlinear least squares. However, nonlinear least squares estimators are only asymptotically efficient, the departure from the asymptotic properties for small sample sizes being strongly dependent upon the form of the model. It is shown that there is a general form of parameterization for the Hougen—Watson approach that corresponds to a “close-to-linear” model, that is, one whose statistical properties approach that of a linear model even for small samples. The general formulation has all its parameters in the denominator of the expression for the reaction rate.