We use the difference (delta) between the position angles of the nuclear radio emission and the host galaxy major axis to investigate the distribution of the angle (beta) between the axes of the nuclear accretion disk and the host galaxy disk in Seyfert galaxies. We provide a critical appraisal of the quality of all measurements, and find that the data are limited by observational uncertainties and biases, such as the well known deficiency of Seyfert galaxies of high inclination. There is weak evidence that the distribution of delta for Seyfert 2 galaxies may be different (at the 90% confidence level) from a uniform distribution, while the Seyfert 1 delta distribution is not significantly different from a uniform distribution or from the Seyfert 2 delta distribution. The cause of the possible nonuniformity in the distribution of delta for Seyfert 2 galaxies is discussed. Seyfert nuclei in late-type spiral galaxies may favor large values of delta (at the similar to 96% confidence level), while those in early-type galaxies show a more or less random distribution of delta. This may imply that the nuclear accretion disk in noninteracting late-type spirals tends to align with the stellar disk, while that in early-type galaxies is more randomly oriented, perhaps as a result of accretion following a galaxy merger. We point out that biases in the distribution of inclination translate to biased estimates of beta in the context of the unified scheme. When this effect is taken into account, the distributions of beta for all Seyferts together, and of Seyfert 1s and 2s separately, agree with the hypothesis that the radio jets are randomly oriented with respect to the galaxy disk. The data are consistent with the expectations of the unified scheme, but do not demand it.