In my review of Modern Romance, I shared a graph from the Economist that depicts how couples met from 1940 to 2010:

A quick glance tells us that the online category ascended sharply throughout the 1990’s and into the 2000’s. Online seemingly stole market share from every other method of meeting potential romantic partners, except the bar/restaurant category. What gives?

A prima facie observation would say that bars and restaurants experienced a resurgence at the turn of the century, causing the online category (in the heterosexual market) to level off. Instead, I suggested:

I suspect the uptick in the “Bar/restaurant” category can be explained by couples too embarrassed to admit that they met online.

How did I arrive at that conclusion? Through curiosity, imagination, and domain-specific experience.


When analyzing data, the first question should be why. Why was meeting online virtually nonexistent until the 1990’s? Why did it become so popular? Why did other categories decline in popularity? Why was the bar/restaurant category an exception?

Begin with the mindset of a child. Children are naturally curious because everything is new. Their minds haven’t calcified with preconceptions yet.

Ask why even when the answer appears obvious. Yes, the obvious answer is usually correct. Meeting online was virtually nonexistent before the 90’s because “online” didn’t exist before then.1Perhaps the 1980’s blip was DoD employees hooking up through ARPANET. It also makes sense that less couples met through other channels if more couples met online.

However, the obvious answer is not always correct. That’s where imagination comes into play.


Wikipedia defines imagination as the “ability to produce and simulate novel objects, peoples and ideas in the mind without any immediate input of the senses”.

We use imagination to run through various “what if” exercises. The first “what if” you should ask is regarding data quality. What if the data is inaccurate? What if there’s no resurgence in the bar dating scene?

Imagination helps you “think outside the box”. A coin that comes up heads a thousand times in a row didn’t invalidate probability theory; it’s likely a loaded or two-headed coin. Don’t be Vizzini:

Westley seemingly poisons one of two wine goblets with iocane, and challenges Vizzini to figure out which one is poisoned. Vizzini dies after they both drink, and Westley reveals to Buttercup that Vizzini would have died no matter which goblet he drank from: Westley had successfully built up an immunity to iocane over the last 5 years, so he poisoned both goblets.

Princess Bride

Your imagination grows as a function of your general life experience. After all, popular conceptions of alien lifeforms are merely adjustments to known forms, such as additional eyeballs or adding tentacles. The more varied your experiences, the more powerful your imagination.

Your imagination can be exercised. You can become an idea machine by coming up with ten new ideas per day. Similarly, you can improve your imagination by using it more often.

Keep improving your imagination and next time you’re an industry analyst pondering why regional sales of Belle & Sebastian’s “The Boy With The Arab Strap” have skyrocketed, you’ll correctly guess that a cute girl used one of the lyrics as her yearbook quote.2Okay, that’s unreasonable.

Domain-Specific Experience

When it comes to understanding data, nothing beats domain-specific experience. In fact, it’s hard to understand data without it.

But it’s not possible to have sufficient experience in every domain. You would need many lifetimes for that. If you lack the necessary experience in the relevant domain, enlist the help of someone who has it.

In the case of how couples meet, firsthand experience tells me that there hasn’t been a revival in the heterosexual bar dating scene. Someone who’s been married for 20 years might not know that though. They’d have to ask people who are still actively dating. And I’d have to ask gay friends whether such a revival took place in the homosexual dating market.

Knowing that there was no bar scene revival, I thought about why the data might suggest otherwise. It didn’t take long to reach a hypothesis. The data resulted from a survey. People lie on surveys, especially when an incentive exists. Until very recently, online dating carried a social stigma. Thus, the bar/restaurant “uptick” was likely due to embarrassed couples creatively interpreting the definition of “meet”.

On Compensation

Let’s look at another chart:

Seems like a positively-skewed distribution with a glaring data error. Why else would there be a large spike at 160K a year? Perhaps a lot of respondents lied, like they might have lied about how they met their significant other. Note the word “reported”. If Family Feud asked “Name something people commonly exaggerate”, salary would be a good guess.

But you might (astutely) observe that more than 18% of respondents could not have independently arrived at 160K a year if they were exaggerating. You’d see variance in the exaggerations.

What if I then told you that the salaries were reported by newly-minted lawyers in 2010? Anyone with domain-specific experience would immediately have a light bulb moment. White shoe law firms all started their entry-level employees at the same 160K salary.3I’m told that the number is 190K in 2020.

Sometimes the data is inaccurate, but not always. Stay curious, imaginative, and acquire (a wide network of people with) domain-specific experience.

Featured photo by Isaac Smith.