Recent simulation work has successfully captured the formation of the star clusters that have been observed in merging galaxies. These studies, however, tend to focus on studying extreme starbursts, such as the Antennae galaxies. We aim to establish whether there is something special occurring in these extreme systems or whether the mechanism for cluster formation is present in all mergers to a greater or lesser degree. We undertake a general study of merger-induced star formation in a sample of 5 pc resolution adaptive mesh refinement simulations of low-redshift equal-mass mergers with randomly chosen orbital parameters. We find that there is an enhanced mass fraction of very dense gas that appears as the gas density probability density function evolves during the merger. This finding has implications for the interpretation of some observations; a larger mass fraction of dense gas could account for the enhanced HCN/CO ratios seen in ultraluminous infrared galaxies and predict that αCO is lower in mergers, as for a given mass of H2, CO emission will increase in a denser environment. We also find that as the star formation rate increases, there is a correlated peak in the velocity dispersion of the gas, which we attribute to increasing turbulence driven by the interaction itself. Star formation tends to be clumpy: in some cases there is extended clumpy star formation, but even when star formation is concentrated within the inner kpc (i.e. what may be considered a nuclear starburst) it still often has a clumpy, rather than a smooth, distribution. We find no strong evidence for a clear bimodality in the Kennicutt-Schmidt relation for the average mergers simulated here. Instead, they are typically somewhat offset above the predicted quiescent relation during their starbursts.