• methods: data analysis;
  • methods: statistical;
  • galaxies: kinematics and dynamics;
  • cosmology: observations;
  • large-scale structure of Universe


We constructed a Bayesian hyper-parameter statistical method to quantify the difference between predicted velocities derived from the observed galaxy distribution in the IRAS-PSCz (Point Source Catalogue Redshift) survey and peculiar velocities measured using different distance indicators. In our analysis, we find that the model–data comparison becomes unreliable beyond 70 h−1 Mpc because of the inadequate sampling by the IRAS survey of prominent, distant superclusters, like the Shapley Concentration. On the other hand, the analysis of the velocity residuals shows that the PSCz gravity field provides an adequate model to the local, ≤70 h−1 Mpc, peculiar velocity field. The hyper-parameter combination of ENEAR, SN, A1SN and SFI++ catalogues in the Bayesian framework constrains the amplitude of the linear flow to be β = 0.53 ± 0.014. For an rms density fluctuation in the PSCz galaxy number density inline image, we obtain an estimate of the growth rate of density fluctuations fσ8(z ∼ 0) = 0.42 ± 0.033, which is in excellent agreement with independent estimates based on different techniques.