See this Skeptical Cardiologist post for more discussion of the topic covered in my last post—how silly observation studies mine data for spurious or meaningless associations, and even sometimes suggest association equals causation.