Thursday, September 1, 2016

Knowing Each Other

Interesting bit of psychological trivia:
It’s true that couples are better at knowing each other’s minds than are strangers, but it’s not by very much. Strangers, remember, were able to guess what each other might be thinking and feeling at an accuracy rate of 20 percent. Couples, in comparison, did only a little better, guessing right an underwhelming 35 percent of the time. In his book, Epley describes a similar experiment, one that worked something like a nerdy Newlywed Game: One half of a couple took a series of surveys, while at the same time in a separate room, the other half of that couple was taking the same set of surveys — only they were instructed to predict the way their significant other would likely answer. One question, for example, asked participants questions about their sense of self-worth, ranking how much they agreed on a scale of one to five with statements like “I tend to devalue myself.” (These were not exactly lighthearted questions.) On average, people were pretty good at predicting their partners’ answers, getting them right about 44 percent of the time. But they were not nearly as good as they thought they’d be: When asked to predict their own accuracy, they assumed they were getting these questions right 82 percent of the time. “These couples hit a double,” as Epley explains it, “but they thought they’d hit a home run.”

What’s even more disheartening, though, is that “this overconfidence increased in proportion to how long two people had been together,” he continues. The couples in this study had been dating for up to six years, and the “longer they had been together, the more they thougth they knew about their partner … More time together did not make the couples any more accurate; it just gave them the illusion that they were more accurate.”
Personally I think they didn't have the right people in the study. In terms of knowing each other, dating for six years is trivial compared to having been married for 20.


Shadow said...

Having 5 choices (scale of 1 - 5) for each statement tends to stack the deck against matches, don't you think? Increase the scale to 1 - 10, and you probably have all the subtlety needed to produce the fewest possible matches without running afoul of probabilities. Make the choice binary and watch the percentage soar.

G. Verloren said...

Largely in agreement here - they seem to have gone about studying this is a poor fashion.

That said, even with their own numbers, 20% vs 35% still means a couple is 1.75 times more likely to understand each other than two stangers. Imagine if someone told you they knew a way to grow 75% larger crop yields compared to current methods - you'd be amazed and ecstatic.

Is it even remotely realistic to expect a 82% overall accuracy of prediction, as some of those polled reported they did? Humans are complex creatures, and I think the very nature of our reality largely precludes the possibility of mutually understanding each other that frequently.

These numbers need more context supplied to have any value. We need to know what the realistic ceiling is on these percentages - if you took two people who were as closely in tune with each other as realistically possible, how often would they actually know what the other is thinking? Because if even the very best of the very best cap out at around, say, 60% overall accuracy, then suddenly that increase from 20% to 35% is actually quite impressive.