We revisit the relation between H_2_O maser detection rate and nuclear obscuration for a sample of 114 Seyfert galaxies, drawn from the CfA, 12um and IRAS F25/F60 catalogs. These sources have mid-infrared spectra from the Spitzer Space Telescope and they are searched for X-ray and [OIII] (5007{AA}) fluxes from the literature. We use the strength of the [OIV] (25.9um) emission line as tracer for the intrinsic AGN strength. After normalization by [OIV] the observed X-ray flux provides information about X-ray absorption. The distribution of X-ray/[OIV] flux ratios is significantly different for masers and non-masers: The maser detected Seyfert-2s (Sy 1.8-2.0) populate a distinct X-ray/[OIV] range which is, on average, about a factor four lower than the range of Seyfert-2 non-masers and about a factor of ten lower than the range of Seyfert-1s (Sy 1.0-1.5). Non-masers are almost equally distributed over the entire X-ray/[OIV] range. This provides evidence that high nuclear obscuration plays a crucial role for the probability of maser detection. Furthermore, after normalization with [OIV], we find a similar but weaker trend for the distribution of the maser detection rate with the absorption of the 7um dust continuum. This suggests that the obscuration of the 7um continuum occurs on larger spatial scales than that of the X-rays. Hence, in the AGN unified model, at moderate deviation from edge-on, the 7um dust absorption may occur without proportionate X-ray absorption. The absorption of [OIII] appears unrelated to maser detections. The failure to detect masers in obscured AGN is most likely due to insufficient observational sensitivity.