Upper panel: Pseudo-C_ℓ measured from GLASS mock samples. The data points of each series are the average PCL from 40 realisations; the error bars are the square root of the covariance diagonal given by NAMASTER. The χ_d² values, which describe the difference between each dataset and the No selection PCL, are calculated similarly to Eq. (18). Lower panel: Relative difference of each case with respect to the no selection case. The shaded regions are the ℓ modes corresponding to a physical scale smaller than 8 h⁻¹Mpc estimated at the mean redshift.