Various approaches have been proposed for the conformational behaviour of macromolecules with mixed statistics, i.e. chain molecules which are described by different statistics on different length scales. An example is the concentration blob model where the chain experiences excluded volume effects on a small length scale whereas these effects are screened on large length scales. In the present work, I propose a generic model for the description of chain statistics which includes the blob model as a special case. The chain is treated as a succession of n segments. These segments are steps of an uncorrelated random walk. The conformational behaviour of each segment is determined by the exponent nu relating the average extension of the segment to its contour length. Explicit expressions are given for the mean squared end-to-end distance R-2 and the radius of gyration R-g of the chain for arbitrary n and nu. For nu = 1 the transition from rigid rod (n = 1) to broken rod (n > 1) behaviour is described. The case 1/2 < nu < 1 yields a general description of chain statistics in the blob model. The models proposed comprise two cases : model I, where all subsegments have the same length; and model II, where the breaking points between two segments are distributed randomly along the chain. It is shown under which conditions the widely used relation R-g(2) = R-2/6 loses its validity.