First of all, the couples were not married very long. All couples are disproportionately happy at the beginning of their marriages. Second, "happiness surveys," and just surveys in general, rely on self reported data, which has been found to be fairly unreliable. That said, I realize there aren't any other ways to determine people's happiness.
And that's just what I noticed skimming through a few paragraphs. Generally, psych students are pretty terrible about mixing up correlation with causation, and a whole host of other statistical misinterpretations. Then, on top of that, you have a journalist who is completely unequipped to do so interpreting that data. And the actual results of the study, which wasn't very scientific in the first place, being distilled down to a sentence in a retarded game of telephone being that payed by people who don't even speak the language in which the message originated - statistics.