I agree online dating represents just a part of society. On the other hand, though:

- Nowadays online dating is so widespread that lots of people - I would say most - use it. Hence it's representative enough.

- Asking people about their dating life produces answers that are often biased or untrue (people are afraid of judgment, and women lie especially about their sex life). Analyzing online data instead reveals what they actually DO, thus it's more reliable.

- Valid social analysis requires high numbers (asking your 12 friends is statistically meaningless). OK Cupid's sample was over one million users: a sample way larger than most research, thus statistically more meaningful.

You sounds like one who dislikes the results, so chooses to hide her head under the sand - i.e. denial.

Lastly, it was the article's author who first cited that research. I just corrected her mistaken citation.