Notes. Top section: fits results for all samples considered in this work. For reference, we give the means ⟨z⟩ and ⟨L⟩/L₀ for each sample in the second and third column. The fourth to sixth columns contain the best-fit intrinsic alignment amplitude, and the corresponding reduced χ² and p-values of the fit for 4 degrees of freedom. To facilitate comparisons with preceding work, we also fit the amplitude of the version of the NLA model based on Hirata & Seljak (2004), shown in the last column. Note that these fits produce the same χ² as the foregoing ones. We also show the fit results for MegaZ-LRG samples with a colour cut as introduced in Sect. 5.4. The numbers in parentheses indicate Π_max in units of    h^-1 Mpc for the MegaZ-LRG samples. Bottom section: Amplitude fits to MegaZ-LRG samples neglecting the contributions by galaxy-galaxy lensing (gG) and magnification-shear cross-correlations (mG). The rightmost column contains best-fit amplitudes when including the gG but not the mG signal. Due to the very similar r_p-dependence of the gI, gG, and mG signals, the χ²-values of these fits are almost identical.