2009A&A...503..909C

Statistical methods for quantifying clustering have not yet been tested against anti-clustered data. We investigate the performance of the normalised correlation length, s{bar}, and Q when applied to anti-clustered data. We tested the operation of s{bar} on simulated data and recently published observations. s{bar} successfully distinguishes anticlustered data in simulations and indicates anticlustering in an area where it is speculated that triggered star formation has occurred. The existence of radial anticlustering can cause the Q clustering analysis method to malfunction. Users of Q should therefore use s{bar} to rule out anticlustering before applying Q.