On the robustness of reciprocal associations between personality and religiosity in a German sample.

The very small size of these reciprocal effects, along with their sensitivity to model specifications, suggest that conclusions about causal effects of personality and religiosity should be drawn very cautiously.