We present a comparison of electron densities derived from optical forbidden line diagnostic ratios for a sample of over a hundred nebulae. We consider four density indicators, the [OII] λ3729/λ3726, [SII] λ6716/λ6731, [ClIII] λ5517/λ5537 and [ArIV] λ4711/λ4740 doublet ratios. Except for a few HII regions for which data from the literature were used, diagnostic line ratios were derived from our own high quality spectra. For the [OII] λ3729/λ3726 doublet ratio, we find that our default atomic data set, consisting of transition probabilities from Zeippen (
1982MNRAS.198..111Z) and collision strengths from Pradhan (
1976MNRAS.177...31P), fit the observations well, although at high electron densities, the [OII] doublet ratio yields densities systematically lower than those given by the [SII] λ6716/λ6731 doublet ratio, suggesting that the ratio of transition probabilities of the [OII] doublet, A(λ3729)/A(λ3726), given by Zeippen (
1982MNRAS.198..111Z) may need to be revised upwards by approximately 6 per cent. Our analysis also shows that the more recent calculations of [OII] transition probabilities by Zeippen (
1987A&A...173..410Z) and collision strengths by McLaughlin & Bell (1998, J. Phys. B, 31, 4317) are inconsistent with the observations at the high and low density limits, respectively, and can therefore be ruled out. We confirm the earlier result of Copetti & Writzl (
2002A&A...382..282C) that the [OII] transition probabilities calculated by Wiese et al. (1996, Atomic transition probabilities of carbon, nitrogen, and oxygen: a critical data compilation, Washington, Am. Chem. Soc. for the N.I.S.T.) yield electron densities systematically lower than those deduced from the [SII] λ6716/λ6731 doublet ratio and that the discrepancy is most likely caused by errors in the transition probabilities calculated by Wiese et al. (1996). Using our default atomic data set for [OII], we find that N
e([OII])≲N
e([SII])≃N
e([ClIII])<N
e([ArIV]).
