Where attitudes toward money are concerned, opposites really do seem to attract. Many couples have settled into roles where one spouse is, at least in relative terms, the tightwad and the other is the spendthrift.

Now a trio of researchers, from the University of Pennsylvania and Northwestern University, have set out to explore those relationship dynamics. Using surveys to rank married people on something called the Tightwad-Spendthrift Scale (isn’t academic terminology wonderful!), they first tested the hypothesis that opposites attract.

Answer: Yes, they do. The research also offers a clue about why this happens:

we find that the extent to which people are attracted to mates with opposing emotional reactions toward spending is significantly correlated with the extent to which they are dissatisfied with their own emotional reactions toward spending.

In other words, if you wish you weren’t such a tightwad, you’ll seek out a mate who loves to go on shopping sprees.

Interestingly, though, people don’t make such decisions consciously. The researchers also surveyed unmarried people, asking about the qualities they would find ideal in a mate. Here, most respondents wanted someone with money habits similar to their own:

Unmarried people tend to think they would be happiest with mates with similar emotional reactions toward spending. Thus, if anything, the observed complementary attraction occurs in spite of people’s tendency to seek mates with similar emotional reactions toward spending. As in many other domains (e.g., women who say they prefer nice guys as long-term partners, but actually desire bad boys for short-term encounters …), people’s accurate forecasts of what will make them happy in the long-run fail to predict what they will initially find attractive when actually encountering potential mates.

The sad truth is that the single folks are probably accurate about the kind of mate who’s best for them: In spendthrift-tightwad marriages, the authors find, money becomes a big source of conflict and unhappiness.