[Hermel]

They report correlation, not causation. What you ask for, is that they should put much more effort into researching cause and effect or not publish the data at all. But I for one prefer having data in this form than none at all. Anyone who is reading hackernews every now and then should know about "correlation != causation" by now.

[burgers]

No, I think if they are going to make the generalizations, they need to publish the data they used to make those generalizations. Its the fact that okCupids data is private and proprietary combined with the publishing of correlations that is unethical.