Radio jet and core data for a complete sample of 98 FRII sources with z < 1 are analysed with a Markov-Chain Monte Carlo (MCMC) model fitting method to obtain constraints on bulk-flow speeds in the beam. The Bayesian parameter-inference method is described and demonstrated to be capable of providing meaningful constraints on the Lorentz factor at both kiloparsec and parsec scales. For both jets and cores we show that models in which some intrinsic dispersion is present in the features' intrinsic prominence, bulk-flow speeds or both provide the best fit to the data. The constraints on the Lorentz factor on parsec scales are found to be consistent with the expected values given VLBI observations and other evidence, with mean gamma ~ 10-14. On kiloparsec scales, the Lorentz factor is found to be ~ 1.18 - 1.49, in agreement with the results of previous analyses of radio jet data. These values are clearly not consistent with the gamma ~ 10 speeds required by beamed inverse-Compton models of X-ray emission from quasar jets; our results therefore support models that require velocity structure in powerful jets.

In a long-term observing project we have imaged a complete sample of FRII quasars and radio galaxies with z < 1.0 at high resolution and high sensitivity with the VLA and MERLIN. This sample of 98 sources includes 15 quasars, 11 broad line radio galaxies and 57 narrow line radio galaxies, allowing unification to be considered in terms of source morphological properties. Radio maps of all the targets have been presented in earlier papers. Here we carry out a systematic analysis of the properties of the jets, cores, lobes and hotspots of objects in the sample. The majority of the tests that we perform show that the data are consistent with a model in which quasars and broad-line radio galaxies are unified with narrow-line objects. Relativistic beaming is the main effect that determines the properties of kiloparsec-scale jets, and it may also have some effect on hotspots. However, some properties of the sample are difficult to account for in simple unified models.