The bias factor of damped Ly α (DLAs) systems, derived from the cross-correlation with the Ly α forest in absorption spectra of the Baryon Oscillation Spectroscopic Survey of SDSS-III, has been used to infer the characteristic mass of DLA host haloes. So far, no dependence of this bias factor with hydrogen column density NH I or other parameters has been detected. With the aim of measuring the dependence of the bias factor on the strength of metal lines, we define the metal strength parameter S, based on optimally combining equivalent widths of 17 metal lines to obtain the highest possible signal-to-noise ratio of S for individual candidate DLAs (defined as having NH I ≥ 1020 cm−2). We present the distribution of metal strength for these DLAs and the dependence of its mean value on NH I and redshift. We search for systematic effects and variations in the catalogue purity by examining the dependence of the S distribution on the spectral signal-to-noise and the estimated error on S. A catalogue of DLAs with measured equivalent widths for the selected 17 metal lines and the value of S are made publicly available, which are used in a separate paper to measure the dependence of the DLA bias factor on the S parameter. The relation of the metal strength on the gas metal abundances and velocity dispersion is complicated by the saturation of metal lines, and remains to be determined in future work.