Lies, Damned Lies, and Statistics
As an investor, you can make a lot of money if you are adept with numbers. This skill can let you see what is really going on and create mathematical models that maximize your earnings. There is a little known story that Warren Buffett was once in a meeting for Blockbuster, the movie rental chain that is now bankrupt. During a presentation pitching the merits of the firm’s value as an investment, growth rate projections were given. Buffett was said to have flipped over an envelope, performed a few calculations, and then said something along the lines of (forgive me for not recalling the specific numbers), “You realize if you break down the compounding rate used to achieve your ending valuation, you are assuming that at the end of the period, every man, woman, and child will be renting at least 3 movies a week from you at present profitability, fertility, and immigration levels? I don’t think that is very likely.”
Such a disposition is a valuable tool. It allows you to avoid errors and protect yourself, your family, and even your country from misguided decisions. It provides a degree of insulation against deception. Unfortunately, in what has been increasing called an anti-intellectual culture here in the United States, scrutiny is often considered a vice, while faith is lauded as a virtue. Personally, I prefer the old chestnut from a former American President: “Trust, but verify.”
I firmly believe that each of us has a duty to avoid purposely manipulating data. One of the things I hate in life – and there really aren’t very many since I tend to adapt and roll with most situations – is when people tweak figures to make a point instead of letting the evidence be as impartial as possible, giving the viewer the opportunity to make up his or her own mind. If a philosophy, policy, or plan of action is good, you should be able to convince people of it, not have to trick them into doing it. Even worse, once people catch on that you can’t trust authorities, even good science or data is discarded or looked upon with suspicion.
People, even well-intentioned people, will try to manipulate you with numbers. It is going to happen. By learning to look for it, and to question everything you are told, you quickly realize that the world is not a simple place. Things that appear to have one basic underlying cause might really be the result of countless variables interacting with each other. Things that you think are true might not be true at all. Keeping an open, rational mind is the only intelligent way to approach the situation and to try and arrive at the best path.
Since I began a project reading through some sociological works from the 1970-1990 period, I’ve been thinking about how often statistics are used to deceive the general public, often with the complacency of the press for the sake of selling more newspapers or generating higher advertising revenues. Here are just a handful things that have made my notes in the past few days alone, demonstrating how pervasive it is. The only defense is to learn to look at the data yourself. It’s not as hard as it sounds and once you’ve done it, you’ll never again be able to accept a headline at face value.
Bad Numbers in the Domestic Violence and Rape Advocacy Movements
There is a certain fringe wing of the academic and non-profit community that is so obsessed with domestic violence and rape as a storytelling mechanism they are completely detached from reality to the point of being deranged. Consider the newspaper article Nearly 1 in 5 Women in U.S. Survey Say They Have Been Sexually Assaulted; a shocking statistic were it true but utterly meaningless and misleading (it’s also particularly telling that the findings about male victimization from the exact same source don’t warrant a headline – things like announcing more men are raped before the age of 10 or that, on page 39, it claims 1 in 4 men in the United States (28.5%) have experienced rape, physical violence, and/or stalking by an intimate partner at some point in their lifetime). Why? If you bother to go through the study itself [PDF], you find that the paper’s reporter wasn’t kidding when writing it “defines intimate partner and sexual violence broadly”.
You can say that, again. Broadly is hardly a sufficient adverb. The various categories of violence are, in some cases, so open-ended that one begins to suspect sneezing at the wrong time could be classified as an assault somewhere, in some chart.
Included in the definition of sexual violence is “coercion”, which, using their explanation of being “worn down”, could theoretically include sleeping with your husband or wife, despite being too tired or not wanting to do it, to make them happy. If you’ve ever done this at any point in your life, congratulations, you can count yourself among abuse victims of sexual violence. Or how about getting married to someone who then announces that, now that a ring is on the finger, they have no intention of ever sleeping with you again? If you, starved for intimacy, say that you can’t remain celibate for life and you’re leaving the relationship, you now count as the sexually violent aggressor because you’ve threatened to withdraw due to your needs not being met. Perhaps someone should send the memo to support groups like Dead Bedrooms. They must be told they are a veritable den of iniquity.
Or how about “intimate partner violence”? Included in those numbers is a category called “psychological aggression” which includes a woman “demanding to know [her husband’s] whereabouts”. So your coworker with the gambling addiction? That one time he went on a 72 hour bender in Vegas and left his family at home with no idea of where he was? Yeah, when she finally got him on the phone, or he stumbled in through the front door drunk, and she demanded some answers, she was the aggressor and he now counts as having experienced intimate partner violence.
How about you walk in to find your college girlfriend in bed with your best friend? You turn around to leave the room, she jumps up and tries to stop you, refusing to let you through the door. To get free, you finally push her aside, leave, and never speak to her or your best friend again, devastated. Guess what? With no qualifications or exclusions, she now counts as one of 32,793,000 female “pushed or shoved” victims of intimate partner physical violence that helps create the headline number. (Lest you feel left out, men, there are 21,953,000 male victims included, too.)
Practically no one in the real world, outside of the ivory walls of think tanks and tenured departments, thinks this is a reasonable definition of intimate partner violence, which conjures very specific images and concepts. In fact, I’d go so far as to say it is highly offensive to real victims of violence who have to survive hell on Earth, fearing for their life, having cigarettes put out on them, or being stalked.
None of those objections seem to matter because journalists either never checked the source or feel like they are achieving some ideological ends by looking the other way. The result is a bad study getting picked up and quoted, referenced and highlighted, perhaps even used to pass new laws, with almost no one paying attention to how stupidly far the definitions utilized differ from common vernacular. (I should be careful. Using the word “stupid” counts as a form of intimate violence, too. I might have just created several more victims.)
Bad Numbers in the Guns Debate
The gun debate is often fertile ground for bad or misleading data. Consider anti-gun lobbyists who sometimes focus on total gun deaths rather than gun homicides, subtly suggesting the murder rate from firearms is nearly triple the real figure without ever actually lying. The first number is hardly meaningful as it includes an overwhelming number of suicides. (When someone wants to end their life quickly, guns are one of the primary preferred methods if they are available.) Or those on the pro-gun side who point this manipulation out and then fail to explain that even adjusting for it, gun murders are still significantly higher in the United States than other countries.
But even that isn’t the entire story. Sociologists have known for a long time that most crime is committed by boys and men between the ages of 14 and 25 against other boys and men between the ages of 14 and 25 who are disproportionately their victims. The numbers are so stark, the U.S. Department of Justice’s Bureau of Justice Statistics breaks out data for individual ages through this period, while bracketing other ages together. Look at the specific number of crimes committed by each age then compare them to the population of folks in that age bracket in Patterns & Trends Arrest in the United States, 1990-2010 by Howard N. Snyder, Ph.D., BJS Statistician [PDF] and you’ll quickly realize giving an intentional homicide rate per capita without adjusting for the median age of the population is misleading as two identical societies, one slightly older, will have very different crime rates despite identical policies, rights, and laws. The United States is much younger than countries like Japan or Denmark, hence we have a higher murder rate. All else equal, this is to be expected. It would be almost impossible to get our rate to that of the older countries.
Even the gun ownership rate data is frequently used in a misleading way. Look at those who tout “guns per capita” as a de facto meaningful statistic. The United States is far and away the world leader at 97.0 guns per 100.0 citizens. Yet, a few notches below us is Switzerland at 45.7 guns per 100.0 citizens. However, most of the guns in the United States are owned by collectors (it’s estimated only 24% of individuals, and 37% of households, in the country actually own a gun), whereas in Switzerland, practically all of the males are given a gun around the time they turn 20 years old, along with comprehensive training. Thus, the probability of a person or household owning a gun is actually higher in Switzerland than it is in the United States, whereas in the United States, someone who owns a gun likely owns a lot of guns, increasing the per capita figures. As this Time Magazine article posits, one of the reasons for such low crime rates in Switzerland is because the gun culture is simultaneously tied to responsibility with children as young as 12 signed up for shooting classes. They aren’t toys, they are deadly weapons and tools, treated with the respect they deserve.
Consider those who compare an enormous country like the United States with smaller countries. There are several psychological reasons for this but if you lived in Copenhagen, Denmark, would you lie awake at night fearful over murders happening in New Delhi, India? Of course not. It’s 3,631 miles (5,844 kilometers) away and you don’t even speak the same language or share the same culture. Yet, if you were living in New York City, New York, you might watch a murder unfold in real time, in your native language, as it happened in Honolulu, Hawaii, which is 4,957 miles (7,978 kilometers) away. This creates a very distorted image of violence rates in people’s subconscious.
Need proof? Ask someone how many of the approximately 313,000,000 Americans are murdered each year with a gun. Make them guess. Have them throw a number out there, writing it down if possible. The odds are good they won’t come anywhere near reality because the total is only 11,068 people.
That isn’t a typo, the number is only 11,068 people out of 313,000,000 in a country bigger than the entire geographic landmass of Europe. Even that creates a horribly misleading picture in terms for the risk for the average member of society as a whole lot of those are due to either inter-gang violence in confined areas such as the South Side of Chicago or drug trafficking on the Mexican border. (At this point, you may wonder why the mental image of gun violence is a handful of mass-shootings in suburban areas. Simply put, despite being a very rare phenomenon, it sells better. A disproportionate percentage of the 11,068 gun murders happen to, and are committed by, poor black boys in the inner city to the point that black Americans make up only 13% of the population but account for ~55% of firearm homicide victims. At the risk of sounding like Kayne West – God help me, I never thought that is a sentence I’d have to write but I’ve now typed twice in my life – it seems like nobody cares when they die because society considers them expendable. Presidents don’t give speeches. Congressmen don’t line up for interviews. Yet, shoot a handful – a statistically meaningless rounding error relative to the population – of affluent white folks in the suburbs and it will be in the news cycle for weeks with bills passed through the legislature and money raised for political action committees. The result is people who are least likely to be victims of gun violence becoming the public face of gun violence.)
In reality, the gun homicide rate in this country has been falling off a cliff for decades, collapsing a jaw-dropping 49% since the 1993 peak but, as Pew Research shows in that impartial analysis of the figures, the public is entirely unaware of this phenomenon. People believe, truly wholeheartedly are convinced, gun murder is increasing. The irrationality on the subject is so great, sometimes on both sides, that it receives a bizarrely disproportional amount of attention. Consider that alcohol kills around 88,000 Americans each year between liver disease, drunk drivers, etc, making it nearly 8x more dangerous than guns. You don’t see people calling alcohol drinkers “alcohol nuts” on one side or those on the other carrying bottles of Jack Daniels around with them as political statement.
Personally, I think a lot of manipulation on the anti-gun side, in particular, is committed by those with good intentions as they seem to be under the delusion that things like the Sandy Hook murders wouldn’t happen in a gun-free society, which requires an almost adorable level of naïveté. One of the worst events in American history was the 1927 Bath School disaster in which a Michigan school board treasurer killed his wife, burned down his farm, and detonated a bomb at a local elementary school, ultimately murdering 36 school children and two teachers before committing suicide. Crazy people who want to kill a lot of innocent men, women, and children do not require bullets to achieve their twisted objectives. Guns are the tool sometimes used but were they unavailable, it would do nothing to stop them. The Boston Marathon attack that maimed or severely injured 170 people, with 2 being killed, was achieved with a Fagor 6-quart pressure cooker that you can buy at J.C. Penney for between $50 and $99 (and the only reason the death count was so low is it was specifically loaded to maximize injuries rather than death to make the scene as gruesome as possible).
Bad Numbers in the Marriage Equality Movement
The same temptation to engage in bad studies or tout misleading data happens in other political movements, too. Consider one of the issues of the political zeitgeist the past decade, marriage equality. There was a recent study that was highly trumpeted in newspapers and on television that demonstrated same-sex couples are better parents. The problem? Yes, the numbers prove it to be true as a class but it’s deceptive; mathematically meaningless because for a same-sex couple to have a baby requires an incredible degree of selection bias creating an apples-to-oranges comparison. If you are married to a spouse of the same sex, you have to want to be a parent to have children. Accidents can’t happen. You have to devote time, money, and resources to having a child, which is now the new normal (think of the first married gay celebrities who come to mind and nearly all of them these days have their own biological, genetic children – Neil Patrick Harris, Matt Bomer, Ricky Martin, Elton John, etc. and even more have adopted children. So much so that The New York Times did a piece highlighting the fact that parents now pressure and expect their gay kids to settle down, get married, and produce grandchildren). Having a child through surrogacy or adoption takes money, which often requires a higher than average degree of education, social capital, and stability when compared to the general population.
To put it bluntly, it is not the fact they are gay that makes them better parents as a class, it’s the fact that this particular sub-class of people (those who require surrogacy or adoption to procreate) are better educated, wealthier, and want to be parents with those unable to meet this threshold being unable to have children. (Exceptions do exist, especially among older people who were pressured into straight marriages and later got divorced after having children.) When you take this population and compare it to a population that includes the 16-year old straight couple who got drunk and knocked up in the backseat of a car behind the bowling alley on a Friday night or the meth-addicted prostitute who forgot to use a condom with one of her johns, of course there are differentials. In other words, when you control for other socioeconomic variables, sexual orientation is meaningless. Comparably situated families with similar demographics lead to comparable outcomes.
Despite being true, that doesn’t make for good headlines nor does it serve as ammunition for those who want to see equal rights in all fifty states. Even though I (obviously) support marriage equality, I can’t stand it because it’s dishonest. Those announcing these findings are hoping that the reader draws an incorrect inference that somehow because of their sexual orientation gay parents are better, providing a sort of inoculation against the extremist rhetoric on the other side. As a rationalist, though, it’s simply not true. The data does not support such a conclusion. There is no noticeable effect, either positive or negative, and acting like there is, even for the right reasons, leaves me uneasy.
(Before some of you send me messages, yes, I know it was done as a counterbalance to the now thoroughly discredited (and laughed out of multiple courts), nearly career-ending hackwork done by the so-called Regnerus study at the University of Texas Austin but I still don’t like it. The Regnerus study was all but backed by right-wing extremists with a final product so twisted, and ineptly managed for the sake of arriving at a foregone conclusion that gay parents were inferior to traditional families with a mom and a dad that, that, as one of the Sixth Circuit Court of Appeals judges pointed out earlier today in oral arguments, even the man’s own department had to issue a statement denouncing it as bad science. It was a professional embarrassment to the school as well it should have been. If you want to see how a scholar can destroy an otherwise good reputation in a single misguided act of intellectual dishonesty, study his career.)
Bad Numbers in Demographic Counts for Political Purposes
Related directly to the previous point on marriage equality, the most recent iteration of deception in data: Population counts for the gay rights movement, which have been hotly debated in the major newspapers this past month. Both sides manipulate population statistics and the dishonesty is inexcusable; perhaps doubly so because the question itself is irrelevant as the population size should have no influence on whether or not a right is extended to a group (e.g., Jewish Americans are only 1.7% to 2.6% of the U.S. population yet we don’t curtail their rights because of their relatively small numbers), making it a distraction to the underlying merits of the discussion.
First, a little background in the chance this isn’t an area you follow regularly. For several decades, academics, polling agencies, and other statistical researchers have attempted to measure the percentage of the population that is primarily same-sex attracted. It took them a lot of effort, and many refinements, to develop a very deep body of research based on a variety of factors, controlling for social bias and all sorts of other things. Perfecting these models was difficult because, especially in the past, people might say they don’t know a gay person, oblivious to the fact their successful son, off living in New York as a doctor, was not, in fact, waiting for the “right girl”; that he was settled down, owned a home, and had integrated his life with someone. Or the gay kid might be in the presence of other family members and not answer honestly. Imagine you’re 22 years old, raised by fundamentalists who have no idea that you are gay, and approached by a poll taker on the street or telephone asking if any of you are LGBT. Or people like Ben Hobson, who were married with children and no one, even their wife, has a clue they are gay. Or that the people who most loudly denounce gays and identify as completely straight are likely to be gay themselves and have grown up in a home or environment with authoritarian role models who indicated such things should be suppressed. What do you think the response rate is going to be? The level of denial is so great that researchers routinely come across a not-insignificant percentage of men who haven’t slept with a woman in years, exclusively had sexual and romantic relationships with other men, and report no opposite sex attraction, yet still consider themselves “straight”. It sounds loony but these folks exist. They might need to watch the Princess Bride.
The medical literature began using a different phrase, “men who have sex with men”, or “MSM” for short, to get around these issues as it made it easier to gather and analyze data regardless of personal identity. (It was a necessary adjustment after studies like the one in New York a few years ago showed almost 39% of men who slept with other men wouldn’t even tell their doctors, in private, when protected by medical privacy laws!) In nearly every significant examination of the facts across countless institutions and researchers over the past few decades, the percentage of MSM in the general population of the United States is no less than 4% and no more than 10% with the most common number tending to come in right around 5% to 7%. (There is only one major methodology that has ever yielded higher results of which I am aware and it was done by a new research method Pew used to veil answers so people could respond without giving away their identity or specific answer, which showed both higher same-sex attraction in the general population, higher same-sex experimentation, and higher anti-gay beliefs when people could speak honestly without fear of retribution. Given that it is an outlier, I think judgment should be reserved before it is accepted or the results can be replicated. I prefer conservatism in my numbers.)
The amount of men and women who will acknowledge this to a stranger during the research gathering phase depends on things such as:
- Geographic area – Safer places like New York and California vs. Mississippi and Alabama.
- Age – Younger people who grew up in a world where it is no big deal are coming out a decade earlier and in greater numbers than their parents and grandparents.
- Anonymity – Knowing the answers can’t be traced back to them, someone isn’t watching them take the test, and a few other items.
The numbers are so clear that even laymen who look at the figures come to roughly the same estimate ranges. Last Christmas, Seth Stephens-Davidowitz wrote a piece in The New York Times studying data from search engines, pornography sites, and social networks, concluding the number was somewhere around 5%. In this 9-page summary, one research firm found the average age at which gay men realized they weren’t straight is 10 years old, knew for sure at 15 years old, told someone for the first time at 18 years old, and finally told a parent at 20 years old. This demonstrates a pattern of hiding for the sake of self-preservation until they are out of the house. Getting an accurate headcount is damn near impossible so long as there is reason to stay in the shadows given that even immediate family is kept in the dark. (For those of you who are data junkies, you can dive into all 160 pages of methodology and results.)
Gallup and other data sources figured this out and began studying it routinely. Using their methodology, they ran a survey back in 2012 that found 3.4% of the population identified as LGBT, 4.4% said they “didn’t know” what they were or refused to answer the question, and 92.2% said they were not. Reasonable inference put that within the MSM percentage of no less than 4% and no more than 10% that most researchers had found over the years.
To expand on the second item in that list, what was most interesting was that when they looked at the response rate by generation, younger people were far more likely to be honest because they grew up with nowhere near the degree of social isolation and prejudice. The 65+ year old crowd had only 1.9% who would confirm they were LGBT while 6.5% refused to answer or didn’t know. Meanwhile, the 18 to 29 crowd had 6.4% who confirmed they were LGBT and had 3.5% who refused to answer or didn’t know. This shouldn’t be a surprise given the older generations were indoctrinated with “educational” videos like this classic 1950’s film shown to elementary school students.
Despite the almost irrefutable numerical boundaries of the population set, there are some fringe gay rights activists who will twist the definition of “gay” to the point of lying, increasing the numbers. They will sometimes quote figures as high as 17%, indicating that nearly 1 in 5.88 men is either gay or bisexual. It’s hogwash. If you think that 1 in 5.88 men is gay, you are delusional.
How do they get away with this assertion? It comes from certain studies that show that somewhere between 15% and 30% of men in the United States have had at least one sexual encounter with a person of the same sex at any point in their lifetime, varying wildly as this is not exactly an area people want to discuss, particularly to strangers. Nobody of reasonable intelligence will conclude that a guy who once got drunk 40 years ago, fooled around with a buddy while in the middle of the Pacific Ocean on a ship, and never before, or after, had any desire to ever be emotionally, romantically, or physically intimate with a man is gay. He’s not. Yes, he engaged in homosexual behavior, but his sexual orientation is heterosexual, not homosexual. He is exclusively emotionally, romantically, and physically attracted to women. He’s straight. (Likewise, no rational person believes that a man who is exclusively emotionally, romantically, and physically attracted to men, yet married a woman out of social pressure, had children, and grew old together, is actually straight. Yes, his behavior is heterosexual, but his sexual orientation was, is, and always will be gay.)
The anti-gay side lies, too. Witness the Centers for Disease Control debacle a few weeks ago. It was so bad, it was outright mocked by researchers who do this sort of thing for a living. The CDC released the National Health Statistics Report Number 77, July 15, 2014, Sexual Orientation and Health Among U.S. Adults: National Health Interview Survey, 2013 [PDF]. The methodology was atrocious. Despite academics knowing for a long time that 1.) you are less likely to get a correct answer if you ask someone their orientation face-to-face as opposed to anonymously and 2.) you are less likely to get a correct answer if you ask someone to identify their orientation as opposed to their attractions given that certain words have baggage for certain people (I know of one couple, both men, who have been together for years, are exclusively monogamous, plan to be together for the rest of their lives, and yet do not identify as gay because they don’t like the political associations of the word), the CDC had interviewers sit face-to-face with people, entering their responses into a laptop, and point-blanked asked them to confirm whether they are, “Gay”, “Straight, that is not gay”, “Bisexual”, “Something else”, and “I don’t know the answer.” They experimented in some of the responses with more private questionnaire methods, including allowing the person to enter the data on their own, but only with a minority of the responses.
Membership in Mensa is not required to ascertain the results are fairly useless. Any survey designed as such would result in Olympian Johnny Weir – yes, this Johnny Weir – at one time in his life being counted among the straights. I think that’s a good indicator as to its accuracy. (That is not a swipe at Johnny Weir nor is it in any way negative. He accomplished more before the age of 25 than most people will by the time they are 80. It’s just an observation as to the absurdity of the methodology. If you think Johnny Weir was ever straight, you probably should avoid email as you will likely find yourself wiring money to a Nigerian Prince as part of an inheritance scheme.)
Regardless of the CDC findings failing to match up with any of the other statistical data done in the past few decades, and the criticisms of its methodology being so grave they should be cause for institutional embarrassment, the anti-equality side shouted from the mountaintops, “Gays represent only 1.6% of the population. This tiny minority is controlling all of society”. Just as the 1 in 5.88 figure is absurdly too high, you have to be delusional if you think only 1 in every 62.5 men is gay.
(Personally – and this is entirely conjecture – I think there is some indication that the CDC may be influenced by a major pharmaceutical giant, and that this entire thing was part of a 48-month chain of events that could very well end up increasing the sales of one of the firm’s HIV medications by an order of magnitude or two at the expense of both taxpayers and insurance customers. It would require its own post outlaying the connections but the evidence seems to at least indicate enough of a question that if I were a prosecuting attorney, I’d be looking into it quietly. There are too many coincidences and advisories that happen to coincide, buried in the footnotes. I’d want to know who knew whom, who worked where, and if any promises had been made. It’s also possible the CDC was just incompetent in this particular study area (it wouldn’t be the first time) and the pharmaceutical corporation in question used it to its advantage, which may be more likely.)
I do not understand how either side – those inflating or deflating the numbers – can find this behavior acceptable. How can they look at themselves in the mirror? Data must be trustworthy so policy discussions can be well-informed. I would feel such a sense of shame if my work were inaccurate for the sake of supporting a particular conclusion, even if I thought the outcome justified the means. I’d lose my self-respect because it came at the expense of my intellectual integrity.
The Lesson In All of This
Henry Ford said it best when he reminded us, “That which can be measured can be managed.” Lots of numbers are routinely published by various sources in an attempt to change the law, earn some sort of special treatment or tax break, or influence public policy. Scrutiny is not only to your advantage, it is a moral obligation. Never trust a headline number or pithy quote without looking into it first, even if you otherwise agree with or support a cause. Otherwise you get all sorts of nonsense, like the myth that domestic abuse is higher on Super Bowl Sunday, which is completely false; made-up; absolutely, totally, in every conceivable way untrue despite some lazy reporter occasionally repeating it in print.
We need to establish a culture of critical thinking where people are trained from childhood that questioning is a moral obligation. My own state’s motto, “Show Me”, is appropriate. Everything should be subject to examination. There should be no sacred cows or ego. Truth, as much as it can be objectively concluded, should be the only goal. I’m not optimistic at the moment, despite my usual temperament that tends toward the upbeat. A friend of mine is a school teacher and she was greatly upset as her district no longer allows her to grade students on correct grammar or spelling, but rather things like their “emotional expressiveness” in writing. If I didn’t know otherwise, I’d swear it was a plot to keep the masses in place. Then again, this presents an opportunity for those who are rational thinkers. It reduces competition. There’s the silver lining.