Feature: Type ‘L’ for Love

Jordan Ramsey reveals how computers are being used to simulate love and investigate our choice of life partners.

'

From the first kiss to that last vow, each person’s relationship trajectory is unique. At least, that’s what we tend to think after swearing off men, women, or both for the sixth or seventh time. Demographic age-at-first-marriage studies, however, show a trend in relationships with a similar overall pattern across many countries and recent years. This trend shows that your chance of getting married for the first time increases from a minimum age to a peak—somewhere in your twenties or thirties for most developed countries—then drops off as you age. What kind of behaviour do we exhibit in our relationships on an individual level to produce such consistent widespread phenomena at the population level? And if we can find this typical behaviour of individuals in a population, can we somehow exploit this knowledge to put ourselves in a better position to find ‘the one’ (and to potentially optimize who ‘the one’ is)?

Researchers Peter Todd, Francesco Billari and Jorge Simão used a unique method to model population marriage trends with plausible psychological phenomena. Todd and his team used computer simulations to model a group of males and females assigned a rank from 0 to 100 rating his or her desirability as a mate. Each mate had incomplete knowledge of the pool of possible candidates, meaning he or she had to select a partner with limited rationality. In addition, each individual searched sequentially through potential partners, without knowing the quality to expect in the next mate and without being able to return to previous partners—a concept that is all too familiar to many of us. Time constraints on biological reproductive potential limited the number of partners and time spent with each. Can you feel that biological clock ticking now?

Under these circumstances, research into the psychology of decision making has shown that people generally stick to simple heuristics. This means that instead of coming up with a complex set of rules and probabilities to determine whether to stay in a relationship or move on and find someone better, we stick to simple rules based on experience. Todd’s team chose to simulate a ‘satisficing’ rule (a combination of satisfy and suffice), in which a person chooses to settle down with anyone ranked above his or her threshold of acceptability, or aspiration level.

Computers can tell us a lot about our love lives.

In the simulation, the aspiration level was set during a learning phase. For this, the team used previous studies in which researchers found that people of similar attractiveness tend to pair up. This could be extended to similarity in general, for example couples often form through shared hobbies and interests. The trick is to learn roughly how you rank in a population and choose a partner based on this. The researchers ‘taught’ their population how they ranked during the learning phase by simulating a series of opposite sex encounters in which each person was either accepted or rejected by a potential mate. With every ‘date’, a simulated person made an adjustment to his or her own ranking in the population and set an aspiration level for a future partner.

The researchers found that with fewer than 20 encounters, simulated males and females were able to roughly assess their own value in a population, set an appropriate aspiration level and find a mate with relative ease. More than 20 encounters in the learning phase, however, significantly reduced the number of mating pairs. In essence, the population became too picky and was unable to find someone to match his or her standards. Todd and his colleagues found that a reasonable number of encounters before entering the mating phase was 12—this has been dubbed the ‘12 bonk rule’. The researchers were able to reproduce the characteristic age-at-first marriage trends after adjusting for individual variation in the number of ‘bonks’ in the learning phase.

As an alternative to this highly entertaining model, Todd and his team formulated a more realistic model in which the learning phase was replaced with a courtship period. During the courtship period the couple continued to meet new people, at which point they decided either to stay in the current relationship or switch to a better partner. Very romantic indeed. Individual aspiration levels began low but were raised depending on how their partner ranked and lowered depending on the length of time the person waited for a better partner to come along. After a pre-determined courtship period, the couple mated permanently and were removed from the pool of ‘singles’. Entirely too realistic. Varying the courtship period also produced realistic age-at-first-marriage graphs.

These models give us interesting—and potentially alarming—insights into the dynamics of individual relationships producing the patterns seen in populations. But in these models each individual uses the same rule of ‘satisficing’ to choose a partner. What happens in a real population where individuals use competing strategies to find a mate? Researchers in Indiana and Berlin simulated a population using three different mating strategies based on speed, quality or harmony. Each strategy had its own strengths and weaknesses. In the speed strategy, individuals proposed to each person he or she ‘dated’, regardless of rank. In the quality search, individuals only made offers to potential partners with a ranking above a certain threshold. Finally, in the harmony strategy individuals proposed to those only within a specific range of their own ranking. The strategies were then evaluated based on their abilities to achieve these goals of speed, quality and harmony in a competitive environment.

S imulations show us how previous relationships help us set standards for the future.

Simulations showed that, though the harmony strategy won out when these competing groups were less choosy (that is, the aspiration level was set low in the quality strategy and the range was wide in the harmony strategy), speed quickly became the best strategy in a more discriminating population. In this case, an individual who proposes on every ‘date’ wins in terms of the speed necessary to find a mate and the quality and harmony of that chosen mate.  The picky population ensures the best possible outcome for someone looking to settle down with the next person who comes along. In other words, it would seem that when everyone else is worrying about the rank of his or her potential mate, the individual employing speed as a strategy does not need to—the rest of the population takes care of that for him. On the other hand, quality was rarely the best strategy to employ. Take note, gentlemen. The hot girl in the club is sick of your unwelcome advances.

Simulating the romantic relationships of a population based on psychological models can yield surprising insights into our own behaviour. It can be comforting to know that the process of finding a mate—an ordeal that can feel isolating and painful on an individual basis—is one that everyone goes through in a similar way. So next time your most recent boy or girlfriend breaks up with you, try not to be heartbroken. Instead, try to think of it as setting an aspiration level for your next, more suitable, mate. Now there’s a silver lining.

Jordan Ramsey is a 1st year PhD student in the Department of Chemical Engineering and Biotechnology

  • http://thenatsci.wordpress.com/2012/05/01/how-to-win-the-dating-game/ How to win the Dating Game « thenatsci

    [...] http://www.bluesci.org/?p=6998#more-6998 tells us that after a given number of social interactions (the learning phase) with members of the [...]