Yeah, I think the article/study is misleading. The authors of both are making assertions that "it's hard to predict how humans pair-up" based on broad datasets.

Yeah, I think the article/study is misleading. The authors of both are making assertions that "it's hard to predict how humans pair-up" based on broad datasets. The opinion is presented atop the data, instead of approaching it granular-ly like this: Given a single man-woman pairing, the author analyzes the pair and provides an anecdotal/hypothetical assertion of what made the two lovebirds connect. If they are self-reportedly happy, the researcher should also makes a hypothetical assertion of what sustains the pairing. Both of these by collecting a set of observed, qualitative facts. Then categorize the connection, group the connections by type, and then make deductions. Right now, the researcher is seeing that A and B have no similar data. This would look a lot different if approached with common sense "A's relationship makes the most sense because____" and "B's relationship makes the most sense because_______", and then draw connections from these two anecdotal observations.