Can Big Data Predict Marriage Success Rates?

While no couple enters into a marriage thinking it’s going to end in divorce, the statistics show that nearly half of all marriages don’t last until “death do us part.” And as clinical as it may sound, big data could prove to be a much-needed solution.

The Analytical Side of Marriage

Research shows that the three most common causes of divorce are basic incompatibility (43 percent), infidelity (28 percent), and money issues (22 percent). Other causes lag far behind.

Keeping this in mind, you and your spouse can decrease your likelihood of divorce rather substantially by paying attention to these three areas. This means developing greater computability, setting up boundaries in the marriage, and regularly talking through finances.

What we just did right there was an exercise rooted in data. We took some available data points and used them to (hopefully) increase the success and longevity of your marriage or relationship. And while this was a very rudimentary and basic exercise, it’s the same sort of thing data scientists, relationship experts, and dating websites are doing with their clients on a regular basis.

eHarmony is a great example of this. As one of the most trusted names in online dating, they’re always looking for ways to improve user experience and help clients find the perfect partner. Until just a few years ago, the matching process used to require a lot of moving parts and manual involvement. Then they incorporated IBM analytics into the system and everything changed.

“The company wants to effectively ensure its customers find love using its service and also aims to reduce the divorce rate. So far, they’re finding success,” data expert Kevin Small writes. “Since eHarmony employed predictive analytics, it’s able to make 3.5 million matches a day for its users as well as track those relationships’ success rates, according to IBM Analytics.”

IBM analytics reports that, since implementing the predictive analytics element into the matching process, the divorce rate for matched couples is just 3.8 percent (compared to 7 percent for other dating websites and 8 percent for couples meeting offline).

While most people are pretty comfortable with the idea of a predictive analytics engine matching them with a partner, researchers at the University of Southern California are attempting to take this to another level. They’ve reportedly developed a computer algorithm that can predict whether a married couple’s relationship has improved or declined over the years – based entirely on the tone of voice they use when speaking with one another.

“The algorithm broke the conversations down, using speech processing technology, into its acoustic characteristics – volume, pitch, intensity, jitter and shimmer, which can measure whether someone’s voice is shaking or warbling, perhaps due to emotion,” Madhumita Murgia writes for The Telegraph.

The results? The algorithm was correct a whopping 79 percent of the time, which was a higher rate of accuracy than session notes provided by therapists.

Divorce in the Age of Data

There’s no way to predict whether or not a marriage will work out with 100 percent accuracy, but there are certainly some powerful methods for predicting a likely outcome. And in an age where divorce rates are sky-high, big data certainly has a place in the dating and relationships industry.

The question is, how do we use data in a way that respects the emotional aspect of human relationships while helping people objectively evaluate tendencies and probabilities? In other words, will people actually be able to take these insights and apply them to their own lives, or will they be so blinded by love that they continue to make decisions that defy statistics? Only time will tell.