It is much easier to generate real numbers than it is to generate real interest in the truth.

A somewhat interesting conversation in the comments over at A big new NIH survey finds only 1.6% of American adults say they are gay or lesbian by Ann Althouse. Ms. Althouse, a perceptive law professor and blogger, asks, "Why are these numbers so low?"

As many of the commenters point out, it has been known for two or three decades that the percentage of LGBT population is between 1-3%. Sure, there's a lot of parsing of definitions, but there has been a pretty clear consensus among demographers and others such as epidemiologists that the number is about 1-3%.

The comments are fairly predictable.

There are a handful that try and take issue with the empirical results but are refuted by others.

There are many who dismiss it all as irrelevant - don't tell me about your protected class, tell me about how you behave.

There are others who rebuke small minorities for diverting attention from issues affecting the whole population.

Many want to explain that the misestimation of the actual percentage of the population is due to clustering in cities and to overrepresentation in media and Hollywood.

There are some who want to reiterate the point that regardless of percentage, human rights are human rights.

There are some who want to score points off LGBT activists for self-servingly perpetuating false knowledge.

What I find interesting is that Professor Althouse appears not to be aware of the longstanding consensus on the low numbers. She's part of the clerisy - she should know these facts. That she doesn't is surprising. But she clearly isn't alone.

Which really opens up the bigger issue. Just how frequently people badly misjudge empirical, and easily available, reality.

1-3% of the population are LGBT. But what do people think the LGBT population is? 25% according to Gallup. They are overestimating the LGBT population by a factor of 10! Only 4% of the population correctly estimated the LGBT population as less than 5%. 52% of people think that LGBT are 25% or more of the total population.

Why does it matter? For all sorts of economic, legal, epidemiological reasons. If you think that LGBT are 25% of the population and you note that only 10% of authors are LGBT, you might leap to the conclusion that publishers are biased against LGBT rather than recognizing that publishers are either neutral or encouraging of LGBT authors.

Take for example the exercise undertaken by Malinda Lo in I have numbers! Stats on LGBT Young Adult Books Published in the U.S. – Updated 9/15/11. Lo is interested in whether LGBT are adequately represented in published children's and YA books and undertakes the herculean task of quantifying exactly how many books are released each year by major publishing houses with an LGBT protagonist/character/themes. There are all sorts of issues in terms of definitions but ultimately Lo comes up with a reasonably solid number of 0.6%. Her findings:

I often hear people saying that publishers aren’t willing to publish LGBT YA, or that each publisher only publishes one LGBT YA per year. This, statistically, isn’t true. Every one of the big 6 publishers (and plenty of smaller ones) publish LGBT YA titles, and several of them do publish more than one per year.

However, the proportion of LGBT YA to non-LGBT YA is so tiny as to be laughable.

The good news is, the numbers have continued to increase over time, and other than the dip in 2010, the increase has sped up since 2000.

The bad news is, the G in LGBT far outpaces L, B, or T.

The assertion that "the proportion of LGBT YA to non-LGBT YA is so tiny as to be laughable" makes sense if you assume that 25% of the population is LGBT. But if you are comparing 0.6% of books to 1.6% of the population, then it is a different matter. Still underrepresented but in the ballpark. But of course 1.6% isn't the right base line comparison, you actually want to compare LGBT families with non-LGBT families (since it is primarily families buying children's books). Only 16% of LGBT families have children versus 74% of non-LGBT families. Given that 3% of the population are LGBT but only 16% of them have children and 80% of the non-LGBT population (97% of all people) have children, then that means that only 0.6% of all families are LGBT families.

The upshot is that Lo's number refutes the common perception that publisher's shy away from LGBT books. She shows that they do in fact publish LGBT books and they publish them in proportion to the number of LGBT families as a proportion of all families.

Lo indicates a couple of times how depressing she finds the low numbers. But that seems kind of an existential depression in the sense that it might be depressing that there are so few LGBT people. The fact that publishers are publishing LGBT books about a formerly marginalized group and they are publishing them in apparent proportion to demand is actually pretty positive but you can only get there with a realistic estimation of what the real numbers are rather than the impressions that people have.

You see a similar confusion regarding race. African Americans are about 13% of the population and Hispanics are about 15%. But what do people think the numbers are? Dated data from 2001 but at that time average respondents to Gallup thought that 33% of the population (overestimating by nearly a factor of three) was African American. Similarly, the average respondent thought that Hispanics made up 29% of the population (overestimating by a factor of two). In sum, the average American thinks that the US is 62% African American/Hispanic whereas it is 65% White. Kind of a material misunderstanding of the numbers.

Is it important that so many people so dramatically misestimate? Probably. Most of the issues I see being heatedly discussed regarding disparate impact often are based on an incorrect understanding of the numbers.

But really, I think the underlying issue is whether people are interested in truth. Facts will never be able to make much of dent on conviction. It is much easier to generate real numbers than it is to generate real interest in the truth.