This is a very interesting study. Anyone who has worked in the consumer credit industry will know that credit scores are correlated with a whole bunch of things -- your standard credit default to insurance claims to employment. Anytime you need to identify 'responsible' individuals, credit scores can act as a pretty good proxy although using credit scores is legally kosher only in certain situations.

I think their trustworthiness hypothesis is probably weaker than other parts of their analysis. I have my own hypothesis based on my experience working in this industry: Staying in a committed relationship may not be really about 'responsible' behavior but could be due to similarity of financial management styles. Most marriage counselors talk about 'sorting out your finances' before getting married; credit scores are essentially capturing that issue -- people who have very good credit scores tend to be savers (although credit scores do not use income or wealth into account) and ones who have poor scores tend to be financially poor and for the large part spend more than they can afford. Also people with good scores tend to be manage finances in a particular way -- pay bills early, monitor balances regularly, etc. -- while the ones with poor credit scores do not usually pay much attention to finances and miss payments frequently.

When you have two people with two different styles of managing money, it can lead to a strained relationship and subsequently lead to separation.