An extensive set of experimental data, for zero pressure gradient boundary layers, over a wide range of Reynolds number is re-evaluated with the help of a composite profile fitted to the mean-velocity data. Boundary layer parameters such as Pi, H and the time scale ratio, Omega, are then carefully examined for consistency among the various experiments and their dependence on Reynolds number. Based on the predictions of the classical theory for these parameters, several criteria are established to evaluate whether a data set can be classified as 'well-behaved', i.e. representative of the desired 'canonical' state. We find that, when carefully applied, the different criteria are very consistent between themselves and can effectively be used interchangeably. The analysis can furthermore help identify the causes for deviations from the desired TBL state and hence serve as a guide in the design of future experiments.