Thursday, March 1, 2018

Tuning Your Gaydar

After Vancouver Men’s Chorus rehearsal last night, I was explaining to a friend what today’s blog essay would be about. This made me realize it’s actually just an elaborate pretext for gossiping about the percentage of singers in various choirs who are gay.

Seattle Men’s Chorus isn’t included in the actual comparison, so I wanted to give a shout out to SMC as a longtime beacon of trans, bi, and straight inclusion. I also neglected to make a guess about gaydar statistics for the Mormon Tabernacle Choir….

Yesterday I wrote about how everyone is born with a primitive form of gaydar. As with so many other things, the gays are merely more refined. Nevertheless, even expert gaydar requires delicate calibration.

Years after leaving Brigham Young University, I accompanied my first gay roommate to Utah for a colleague’s wedding. When we toured BYU's campus, Phillip was startled to find virtually every guy set off his gaydar. I had to explain that clean-cut Mormon men aren’t all gay. They simply give off a wholesome, caffeine-alcohol-and-testosterone-free vibe, as if they'd just walked out of a milk commercial or a Mitt Romney home video. My roommate needed to adjust his instinctive gaydar to fit the unique Utah environment.

Statisticians have a formula for integrating new information into our existing probability assumptions. It’s called “Bayes’s Theorem,” after the eighteenth century English clergyman who introduced the concept. In statistics guru Nate Silver’s book The Signal and the Noise, he introduces Bayes's Theorem with the example of “living with a partner and coming home from a business trip to discover a strange pair of underwear in your dresser drawer. You will probably ask yourself: what is the probability that your partner is cheating on you?”

Silver uses the following propositions to show how Bayesian probability works:
  • Most importantly, “you need what Bayseians call a prior probability (or simply a prior). What is the probability you would have assigned to him cheating on you before you found the underwear?” Studies show about 4 percent of married partners cheat on their spouses in any given year. 
  • Separately, “you need to estimate the probability of the underwear’s appearing as a condition of the hypothesis being true – that is, you are being cheated upon.” As Silver observes, “If he’s cheating on you, it’s certainly easy enough to imagine how the panties got there. Then again, even (and perhaps especially) if he is cheating on you, you might expect him to be more careful.” Silver therefore assigns a 50 percent probability that a cheating partner would leave behind such damning evidence. 
  • Finally, “you need to estimate the probability of the underwear’s appearing as a condition of the hypothesis being false. If he isn’t cheating, are there some innocent explanations for how they got there?” Silver runs through a few colorable theories – a secret gift for you? A new cross-dressing hobby for him? – and assigns them a collective probability of 5 percent.

Running these numbers through the algebraic formula described in Bayes’s Theorem, you come up with 29 percent as the posterior possibility your partner is cheating on you. That figure (instead of the original 4 percent) becomes the new prior the next time you find a strange pair of underwear.1

1Running the revised numbers back through Bayes Theorem results in an 80 percent probability of cheating the next time. A third pair of suspicious underwear takes you to 98 percent….

I’m an English major. I have no clue how to run the numbers. [Ed. note: actually, he managed to come up with the figures in that last footnote all by himself. Plus a calculator.]

Instead, I’m most interested in Bayes’s Theorem as an example of the kinds of tools available to us for making choices under conditions of uncertainty, including the choice of what to believe. Bayesian probability is also consistent with pragmatism, in the old-fashioned William James sense that is the bedrock of American (and my) philosophy.

One important aspect of pragmatism is that everything is negotiable in the search for what works. As Kahneman argues, both our subconscious and conscious minds are constantly adjusting our mental model of reality, like an internal kaleidoscope that brings coherence to the world around us. Ordinarily that involves tweaking the model around the edges, but sometimes we’re due for a paradigm shift.

Fortunately, as both a practical and philosophical matter, everything is not on the table all at once. Instead, everything simply has the potential for revision. Like our assessment of who might be gay, and who is merely fashionable, sensitive, hot, and/or European.

So what does Bayesian probability have to do with gaydar, other than the fact Nate Silver is gay?

As one New York Times contributor observed, it's “notoriously difficult” to estimate “what percent of American men are gay.” The best approximations, from self-reporting in large anonymous Gallup surveys, put the number between 3 and 4 percent.2

2A much larger male population – over 7 percent – identifies as gay, bisexual, or transgender. This can be attributed to two main factors: bisexuals surprisingly outnumber gays, and millennials are increasingly likely to identify as LGBT.   
Women (and lesbidar) are beyond my expertise. However, I’m intrigued by one divergent statistic: if the option of being “mostly attracted to men” is offered in surveys, the proportion of women exclusively attracted to men declines to 79%. In contrast, 93% of men would still describe themselves as “exclusively attracted to women.”

Surveys also show Americans are terrible at numbers. This phenomenon isn't limited to English majors. According to Gallup, Americans lavishly overestimate the percentage of gays and lesbians in the general population, putting the number at a whopping 23 percent. But it’s not just a gay thing: “Americans estimate that a third of the U.S. population is black, and believe almost three in 10 are Hispanic, more than twice what the actual percentages were as measured by the census.”

The self-reported percentage of gay men is comparable to the numbers of Mormons or Jews in the general population – a fact the LDS church should be more sensitive to the next time it embarks on another anti-gay campaign

Of course, you wouldn’t assume only three percent of BYU students are Mormon. Or that only three percent of Yale Law School students are Jewish. Context matters.

FYI, I understand Vancouver Men’s Chorus currently has two self-identified straight men singing with us, out of something like 120 men on the risers. You’ll have to guess who they are at our concert in June.

So how do your brain’s Thing 1 and Thing 2 combine statistical information with instinctive gaydar to make assessments about the men we encounter?

As I recently reported, last month I attended a performance in Seattle by the Whiffenpoofs of 2018. The Whiffenpoofs are the Ivy League’s oldest and best a capella singing group. For many decades, fourteen male singers from each Yale graduating class have been tapped as Whiffs. Based on their group portrait, how many do you think identify as gay?

Going into the Whiffenpoof concert in Seattle, I could have relied on Gallup’s national survey data for my Bayesian prior – resulting in the assumption approximately one out of fourteen current Whiffenpoofs is gay. However, the singing group’s very name makes that estimate seem low.

Fortunately, I have several additional pieces of data. First, in addition to singing with Vancouver Men’s Chorus, Seattle Men’s Chorus, and Windy City Gay Chorus, I’ve also sung with several ostensibly non-gay church choirs. The ratio of gay singers was definitely more than one in twenty. Particularly tenors.

Second, last year Vancouver Men’s Chorus had the privilege of singing in a concert with Chanticleer, the world’s premiere a cappella singing group. We also had a chance to sit down with the twelve professional singers and ask about their lives and experience. Chanticleer's concert set was so flawlessly glorious I decided singing is pointless for the rest of us. Fortunately, after our last number (by Kylie Minogue, with dancing boys and 100 coloured fans) the four gay boys from Chanticleer leapt to their feet in an enthusiastic standing ovation.

Finally, after the last time the Whiffenpoofs toured Seattle, I happened to run into one of their roommates at a gay bar. He confirmed my gaydar estimate: four of that year’s vintage were indeed gay.

So I went into last month’s concert with a rough estimation that around four of this year’s Whiffs were likely to be gay. Let’s call that a 30 percent likelihood for any individual singer.

First, however, I wanted to listen to the Thing 1 part of my brain's intuitions before exercising my Thing 2’s conscious judgment. Could I tell, based on specific visual or auditory evidence, whether any particular Whiffenpoof was gay?

Unfortunately, twenty-something Ivy Leaguers who sing Cole Porter songs while wearing white ties, white gloves, and tails all seem pretty damn gay. During the performance nothing was conclusive – gestures, solos, glove work, facial expressions, nada. Sure, my intuition told me a couple of the guys should have their odds bumped up a few percentage points. But it would still be a matter of guesswork in any individual case.

Gaydar shouldn’t be mere astrology for homosexuals. I felt like my bewildered roommate Phillip visiting BYU years ago – my gaydar failed me, so I had to go with the odds. Or ask someone for help.


After the Whiffenpoof concert last month, I stayed at a friend’s apartment in Seattle rather than drive home to Bellingham. As I was lounging on the futon with my iPhone, I recognized one of the singers logged onto a gay dating app. It wasn’t difficult to spot him – he was wearing a tux in his profile picture, and said he was currently visiting Seattle while on tour with a collegiate singing group.

So I adjusted the likelihood of this particular Whiffenpoof being gay from 30% to a virtually 100% probability.

Now look back at the picture of fourteen Yalies wearing white tie and tails. Guess who?

Applying Bayes's Theorem to a single piece of evidence
offered in support of a dubious hypothesis:
the Moon is made of cheese

No comments:

Post a Comment