We present detailed clustering measurements from the 2dF QSO Redshift Survey ( 2QZ ) in the redshift range 0.8 < z < 2.1 .
Using a flux limited sample of \sim 14 , 000 objects with effective redshift z _ { eff } = 1.47 , we estimate the quasar projected correlation function for separations 1 < r / { h ^ { -1 } { Mpc } } < 20 .
We find that the 2-point correlation function in real space is well approximated by a power law with slope \gamma = 1.5 \pm 0.2 and comoving correlation length r _ { 0 } = 4.8 ^ { +0.9 } _ { -1.5 } h ^ { -1 } { Mpc } .
Splitting the sample into three subsets based on redshift , we find evidence for an increase of the clustering amplitude with lookback time .
For a fixed \gamma , evolution of r _ { 0 } is detected at the 3.6 \sigma confidence level .
The ratio between the quasar correlation function and the mass autocorrelation function ( derived adopting the concordance cosmological model ) is found to be scale independent .
For a linear mass-clustering amplitude \sigma _ { 8 } = 0.8 , the “ bias parameter ” decreases from b \simeq 3.9 at z _ { eff } = 1.89 to b \simeq 1.8 at z _ { eff } = 1.06 .
From the observed abundance and clustering , we infer how quasars populate dark-matter haloes of different masses .
We find that 2QZ quasars sit in haloes with M > 10 ^ { 12 } M _ { \odot } and that the characteristic mass of their host haloes is of the order of 10 ^ { 13 } M _ { \odot } .
The observed clustering is consistent with assuming that the locally observed correlation between black-hole mass and host-galaxy circular velocity is still valid at z > 1 .
From the fraction of haloes which contain active quasars , we infer that the characteristic quasar lifetime is t _ { Q } \sim { a few } \times 10 ^ { 7 } { yr } at z \sim 1 and approaches 10 ^ { 8 } yr at higher redshifts .