Corral-lations

Correlation isn’t causation. You’d think as an ex-fisheries scientist I would have grasped that immediately, but back in the day when I was trying to work out just why some books, some authors, took off… and others withered on the vine, I got it very wrong.  I knew almost nothing about publishing back then, and assumed that it was a merit based outcome, dependent purely on the quality and appeal of the book.

That led to a lot of spurious correlations in my analysis… because especially in Trad publishing factors beyond the merits of the book are substantial, and way out of control of the author. Besides it’s a very, very complicated fishery. We don’t all attract or capture readers in the same way, even if quite a lot are snagged by the wallet.

It’s hard enough to work out roughly how best to pursue getting readers of a group you know well, let alone as a foreigner, writing for an audience I knew nothing of that I hadn’t got from books or the media, or occasional expats (not a great sample), into an industry I was almost entirely wrong about too. It’s so complex, you can do everything ‘right’ and still hit bad luck, or the opposite.

Still, there are something you can work out. And the more you know of your target group, the less likely any success is just going to be pure luck.  But correlation is not causation!

To illustrate this I’m going to do something really stupid and write about American politics (not to take sides in a country that is not my own – I wish you all the very best with it, but it’s really none of my business, and I remain fairly ignorant of it. Not perhaps as ignorant as the guy who came up with this…)

Some obviously left wing ‘I have been trying to understand why Hillary didn’t win’ individual came across something from the UK Grauniad, lamenting Brexit, and blaming it all on the ‘Shit-life syndrome’. He concluded the same true for his party’s loss in 2016.

To summarize the ‘shit-life’ theory: the analyst looked at the areas that voted for Brexit (and second guy the areas in his chunk of middle America that voted for President Trump). The areas were all outside of large cities, were relatively poor, and, with lower median education levels and in the US anyway, with shitty minimum wage jobs, high levels of opioid abuse and suicide, particularly among white males.

Their (NOT MY) conclusion because of this correlation, being in the group that had shit-lives caused them to vote for Brexit or Donald Trump, instead of the parties those areas historically supported. Because he lied to them and promised to immediately make things better. And they’re stupid and believed.

Let’s leave the politics out this: But these are unsupported conclusions driven by absolutely no data.

“But… but… but muh correlation!”

The correlation, in fact, only shows that areas with these issues more (not all) people voted for a person or an issue that this guy didn’t agree with, than voted for the ones he did.  His conclusions based on the correlation are wishful thinking and a total lack of understanding of either the numbers or the source of the data. This applies as much to writing as politics, and has been fairly well displayed by Tor Books attempts at buying Mil Sf.

To find actual causes and reasons for such correlations one has dig a lot deeper and understand a lot more of the complexity of the situation – and not be misled by a few things that appeal to your preconceptions.  For example of dumb preconception: there are higher drug-abuse deaths: therefore ALL these people must all be stupid, drug addicts and desperate for immediate gratification, and therefore voting for someone he doesn’t like. Mathematically and logically that’s… like: WHUT?  Let’s leave politics out of it, and look at the math. Voting is voluntary. Participation often selects the more concerned, motivated and people who have actually bothered to think (some of them, some of the time, anyway). Opiate use is likewise not universal. What evidence does the guy who believes these shit-life people have voted for his enemy… that the groups meet at all, let alone being one and the same?

Look, my conclusion, with equally little statistical support (but at least I am not trying to pretend it is there) is these people would probably vote for someone who doesn’t sneer and talk down at them, and is offering – rather than the immediate gratification of extra welfare or better minimum wages for shitty jobs, so they can spend it on drugs — the possibility of better jobs, where they are valued and can have more self-esteem. I don’t know these people in particular, but I sure know a lot like them. I find a lot to admire in a fellow willing to battle rather than take a handout. And… given the vast increase in college participation, with the same IQ pool to draw from – 1) college students are dimmer on average than they were when that participation was smaller. 2) Non-participation in the college system… given that the debts are high, the quality iffy, and the outcomes for many courses not very rewarding and just as minimum wage shit-life…  certainly means that that someone who didn’t attend college could well be brighter than average. Once again I certainly know a few people like that.

One of the two of us could be right… or more likely, neither. But one of us is not jumping to conclusions, believing them supported by stats.

So to loop back to writing: and my point here. You may think this is ridiculous, but I can’t tell you how many times I’ve hit it. Where a point of correlation is assumed to mean causation. “Fantasy (like Harry Potter or Game of thrones) sells big. I don’t actually read it or know anyone who does, and I think people who do are weird and dumb (but I did watch Game of Thrones) but because these books are a success, I am going to write Epic Fantasy.”  Or – as I mentioned earlier — a publishing house with editors that don’t have the background or know the readers of the same buying ‘Military SF’… because it sells… Or a Romance editor crashing an entire line because she knew romance sold but had no idea why. She thought it was the branding, not the content, and that if she made it various ‘non-binary’ erotica with none of this romantic hetero-masculine hero and happily ever after rubbish it would sell even better…

If there is one thing I have learned: Stick to your knitting.  ‘Such-and-such is popular’ doesn’t mean you will be if you write it, unless you write as well the author who is popular – who probably knows and loves that niche. Write for the kind of people you know, and like and respect, in a genre you like and are familiar with. Trying to manipulate the buying habits of readers you don’t know in genre you don’t like for your own gain, on the basis of assumptions that suit your daydreams… probably works as well as the guy with the ‘shit-life’ theory of why people didn’t vote the way he wanted them to’s idea of a remedy. He was going to tell them all they were lied to.

I’m SURE that’ll work. The utter shock of being told a politician was mendacious is bound to make everyone drop dead.  Mind you, that might be the plan. I gather the dead are far more reliable voters in America. /s

The image is a supply of bull in a corral.
Image by Fabiano Capeletti from Pixabay

45 comments

    1. The problem with “write what you know” is that it results in far too many books about writers navel-gazing about the problems of being a writer (the reason I’ve stopped reading Stephen King). I think I like Ashley’s formula better.

      1. But, write what you know assumes that the writer has developed SOME area of expertise:
        – Hemingway – war, reporting, ambulance driver
        – Michener – Navy vet
        – Tom Wolfe – literate Southerner in NYC, a tremendous listener, someone unafraid to challenge conventional wisdom
        – Fitzgerald – mediocre ad man, writer, college dropout. Came across hard times, and learned to write for magazines, movies. Life experience with mental illness.

        Hell, a writer could be a construction worker, domestic laborer, waitress, bartender, salesman, firefighter (Denis Leary), cop – the job doesn’t matter as much as the ability to soak in the atmosphere, and later, write about it.

        It’s why professors and students seldom write interesting novels. Their life experience is too narrow.

        1. “It goes without saying that you will not write a good novel unless you possess the sense of reality; but it will be difficult to give you a recipe for calling that sense into being. Humanity is immense and reality has a myriad forms; the most one can affirm is that some of the flowers of fiction have the odour of it, and others have not; as for telling you in advance how your nosegay should be composed, that is another affair. It is equally excellent and inconclusive to say that one must write from experience; to our supposititious aspirant such a declaration might savour of mockery. What kind of experience is intended, and where does it begin and end? Experience is never limited and it is never complete; it is an immense sensibility, a kind of huge spider-web, of the finest silken threads, suspended in the chamber of consciousness and catching every air-borne particle in its tissue. It is the very atmosphere of the mind; and when the mind is imaginative–much more when it happens to be that of a man of genius–it takes to itself the faintest hints of life, it converts the very pulses of the air into revelations. The young lady living in a village has only to be a damsel upon whom nothing is lost to make it quite unfair (as it seems to me) to declare to her that she shall have nothing to say about the military. Greater miracles have been seen than that, imagination assisting, she should speak the truth about some of these gentlemen. I remember an English novelist, a woman of genius, telling me that she was much commended for the impression she had managed to give in one of her tales of the nature and way of life of the French Protestant youth. She had been asked where she learned so much about this recondite being, she had been congratulated on her peculiar opportunities. These opportunities consisted in her having once, in Paris, as she ascended a staircase, passed an open door where, in the household of a pasteur, some of the young Protestants were seated at table round a finished meal. The glimpse made a picture; it lasted only a moment, but that moment was experience. She had got her impression, and she evolved her type. She knew what youth was, and what Protestantism; she also had the advantage of having seen what it was to be French; so that she converted these ideas into a concrete image and produced a reality. Above all, however, she was blessed with the faculty which when you give it an inch takes an ell, and which for the artist is a much greater source of strength than any accident of residence or of place in the social scale.” Henry James

  1. One can find correlations where one looks for correlations. Conversely, one does not find correlations where one does not look for correlations.

    You can flip the example by looking at urban, poor, drug and gang drenched regions – strong correlation with Hillary voters.

    Any actual correlation that might indicate causation between uneducated* and voting is between the conditions and the message received.

    * “Educated” and “college graduate,” if diagrammed a la Venn, is two overlapping circles – with the overlap getting thinner and thinner every year.

    1. Oh so true. Recently had a fellow decide that his Ph.D. made him educated evidently on ALL subjects and someone who graduated the School of Hard Knocks (I refused to try to get into a credentialist war of any sort – I know what BS smells like up close, after all.) wasn’t worth his time at all, and any non-“academic” source of information was likewise unworthy. I really hope this person doesn’t work in any branch of engineering or any of the hard sciences. This is one time I really hope that degree is in something like Art History where minimal real-world damage is likely to result.

  2. I am still somewhat amazed by nitwits who fail to understand that government actions can have negative outcomes.
    .
    It’s a category error to confuse “unlike my opponent, I will not actively attempt to make your life worse” with “vote for me, and I shall treat you as my treasured cosset”.
    Nor is it inexplicable that people whose lives are being directly affected by government policy, would like to some accountability from those making the policy.
    (Nor… Forget it, Jake. It’s Chinatown.)

  3. My brain keeps going “dude, do you realize how numbers even work?”

    The high point of the drug over-dose time was roughly twice as high as automotive deaths. That’s ignoring how many were deliberate murders or suicides, or that most of the people who OD’d on fet probably didn’t even know they were taking it, or that the death rate is likely higher in rural areas because that’s a good place to smuggle stuff and you don’t get medical aid as quickly… good heavens, the guy wants to try to imagine that something like “died in a car accident” is THAT BIG of a consideration?

    1. “My brain keeps going “dude, do you realize how numbers even work?””

      I used to think that. “Are you kidding me?!” escaped my lips frequently, followed by lots of bad words while reading articles.

      But then I went through the -entire- medical literature on gun control. The studies I read were appalling from a statistical perspective, normal basic statistics were either wrong or were completely ignored in the conclusions. My personal all-time favorite was the one that proclaimed that owning a dog was almost as highly correlated with a death in the home by shooting as owning a gun. Within 5% if memory serves.

      Given the credentials of all those authors, there’s zero chance they got it wrong. No way.

      They’re not stupid. They’re lying. Most don’t even take the trouble to tell a clever lie, they push a shoddy one because its good enough. In the 90s any anti-gun paper in the medical literature was rewarded with grant money and CDC approval, didn’t matter how bad it was. Once you accept that they’re lying for money and career points, all of it makes sense.

      This guy of Dave’s with the “shit-life” theory doesn’t have to do any work, he only needs to proclaim any vaguely plausible conjecture and the choir will sing “AMEN!”

      1. Reminds me of the first time I figured out they were screwing with gun stats. I, too, read the actual study, on having guns meaning you’re more likely to die from one. And, of course, they very carefully didn’t correct for if the gun was legally owned— because if you removed career criminals (AKA, felon in possession of a firearm) then owning a firearm meant you were waaaaaay less likely to be shot at all, at home or not.

        Why? Because most non suicide gun deaths are criminals shooting each other.

        Speaking of, I mentioned the possibility of “overdose” murders to my husband, and he said it’s a known Cartel tactic. They brag about killing folks who are related to someone they’re mad at with drugs….

      2. I will disagree with you in one place:

        Given the credentials of all those authors, there’s zero chance they got it wrong. No way. They’re not stupid. They’re lying.

        Having known a lot of these people with credentials, yes, there’s every chance they got it wrong. The level of stupidity most people have with statistics, including people with STEM degrees (never mind those in the Econ and social sciences), is amazing. I’m not immune from this either, I just like to think I’m a little more aware of my own weakness.

        You can throw in some malice with the stupidity, and I won’t disagree with you; these people are obviously taking the statistics that tell them what they want to hear. But don’t underestimate the power of dumb either.

        1. “Having known a lot of these people with credentials, yes, there’s every chance they got it wrong. The level of stupidity most people have with statistics, including people with STEM degrees (never mind those in the Econ and social sciences), is amazing.”

          If it was only physicians I’d agree with you. Doctors are not masters of statistics, and the papers you see by docs reflect that. Lots of really bad study design and wrong stats.

          But we’re talking about guys with their PhD in Public Health, which is -all- statistics. Men of considerable standing in the medical biz, in more than one country. You see it in Canada, Britain and Australia too.

          When the Big Kahuna of Public Health at Harvard Medical School puts his name to a paper where the study design is wrong, the sample is so skewed a house painter (me) can tell, the conclusions don’t match the numbers IN THE STUDY, and they can’t/won’t produce the raw data on demand… and you look at 100+ studies stretching over 30 years and they’re always the same (I found 6 (six) out of about 120 I looked at which met the most basic requirements of a public health study)… and the NIH and the United States Congress -both- conclude the entire body of literature is without scientific merit…

          …plus Reality(tm) resolutely refuses to match their assertions…

          … then incompetence can no longer explain it. That’s a propaganda campaign, and a very well funded one. Leftists lying for money.

          1. Left field related, apparently the “won’t give their data” stuff is relatively long running.

            I found out recently that the big famous study of the Shroud of Turin…they refused to release the data. Until forced by law, decades later.

            The “proof” was as bad as the gun studies. (Which folks figured out since using a technique for stuff that’s been buried for the whole time won’t work with stuff that has been in common use, but ouch.)

            1. I have a particular hatred of people who lie in journals. There’s a lot of them these days, and they do it all the f-ing time. You keep seeing the same lying study cited over and over for years and years after the lie is revealed and put down. Like e-coli spreading over a culture dish, it keeps popping up.

              Somebody refuses to release their data, there should be consequences from the scientific community. There used to be, after all.

          2. Amen, Brother, Amen.

            I majored in History – but specialized in quantitative analysis of historical data. My minor was in Science. Had picked up an Associate’s degree in Business.

            Later, worked as a computer programmer and network admin. In every work I did, I found that those who actually understood stats were as rare as Honest Politicians.

            And, it’s getting worse – the grad degrees, most of which require some stats coursework, are skim milk compared to even 10 years ago.

            BTW, Olivia Jade – that kid whose own parents thought a college needed to be bribed to accept her, because she was too dumb – has re-launched her career on social media, and seems to be getting back to her previous success. She doesn’t deserve her rep for being dumb; what she deserved was better parents.

            1. I do not pretend to understand statistics, because I am math challenged. But I do understand study design and how statistics are meant to be used as part of finding out if you got a result that was signal and not just noise. (Funny how a musical connection gets you through a lot of things.)

              When you look at a study of an area that was losing population over the study period, and they didn’t control for that by using incidents per 100,000, it makes you laugh.

              That’s the level of statistical dishonesty we’re talking about here. Yes, you had 10% less incidents in 2012 than you did in 2009, but the population declined 15% over the same period. It doesn’t take Albert Einstein to know that’s not how you do it, even I the mathematically challenged physical therapist can see that one.

  4. Dave said: “But… but… but muh correlation!”

    Yeah. We were talking about Psych 101 the other day, and how the “scientific” underpinnings of those survey courses have been washing away like a sand castle at high tide since the 1970s, but the courses persist in teaching the debunked theories as if they were true.

    Having seen this up close and personal in Real Life quite a bit, the reason American (and Canadian, and Brit etc.) intellectuals cling to those debunked theories is that they WANT them to be true.

    Most people in my estimation don’t really like science. Its inconvenient. Nobody likes it when their favorite theory gets tested and blown out of the water. This is particularly true in the social sciences, when a plausible sounding theory like “shit-life” can form the basis of a whole career.

    Going and -measuring- things never even occurs to most of these intellectuals. The entire edifice of Intersectionality has never been exposed to experiment, its a collection of conjectures and unexamined assumptions. Measuring it would be a catastrophe. No more basis to scold the rest of us when the measurement comes to zero.

    With writing, I’m in a bit of a quandary regarding the data. We don’t really -have- any data worth the name. Sales figures that trickle down to us are second or third derivative from the cash register, as it were. If we could get some register-level information on seasonal and general sales by genre etc. I’m sure the resulting curves would be… interesting shall we say.

    One gets the feeling that the Big Five have a lot of people whose jobs rest on concealing the truth at the behest of their political and ideological fellow travelers. But I don’t KNOW that, its just a suspicion based on general observations of the world. A correlation between related things.

    The difference between me and the general issue Leftist is that I’m keen to find out the truth, so that I can use it to further my work. Leftists are keen to conceal it, so that the truth doesn’t come out about theirs.

    Here endeth today’s discontent. 😡

      1. Maybe, but if they are it isn’t in the “Amazon Rankings” thing so beloved of certain Lefty fruitbats. I’m unaware of any actual -sales- numbers being available to anyone except the author of the book themselves.

      2. If you are in Amazon’s own in house publisher, fully share all your numbers (I’m not, but a friend is). Not public. Several of the larger houses have ‘author portals – which allow the author to access real time numbers. Numbers, such as they have that is. Most of their systems rely heavily bookscan, apparently. That captures (bookscan’s own claim) about 1/3 of outlets. I suspect that 1/3 varies by where that kind of book sells (I gather, for example, that on military base shops are not in their count).

    1. Thing is, I don’t think it’s anything so grand as concealing the numbers for political or ideological reasons.

      I think they’re concealing the numbers because they want to keep their jobs.

      1. Given what happens every time the published conclusion doesn’t match what it “should” say, even if it’s 180 from what is actually in the document?

        I think you’re right.

        I can even see it being justified, if I imagine myself on the inside….the folks who believe the BS will believe it no matter what, and the actual data shows the opposite.
        Just like that study with the death rate for owning guns. They didn’t SAY it, but there were enough pieces left that I could figure out “criminals are most of the folks dying.”

        1. That study was a classic.

          The sample was skewed, being taken in Cook County which is the shit part of Chicago. Other factors like criminality were ignored without being eliminated from the sample. The study did not make a distinction if the gun kept in the home was used in the shooting, or if it was a different gun brought by the attacker. Et cetera.

          But the crowning glory came when the lead author could not (or would not) produce his raw data for a Congressional inquiry. Chances are high the whole thing was clouds and dew-drops.

          Still being cited as definitive in the medical literature in 2019.

      2. The comic book industry does have sales figures available online, and I’m sure they wish it was otherwise. Every time they come out with the newest Woke title their sales numbers are horrific. They’ve taken to blaming the fans for the failure of titles like “Black Panther & The Crew,” the nerds are too racist/sexist/transphobic to appreciate Teh Aht of the amazing super-duper comics.

        You can see TOR coming out with statements deploring the Deplorables if the real sales figures for Hugo Award Winning Author sucked. As it is, they can hide and pretend everything is peachy-keen.

        1. Anyone who makes it big has to make it on sales outside of what we’d call fandom, deplorable or not.

          It’s just figuring out how to do that.

          If someone does, let us all know?

        2. I was reminded today of the incident where Superman declares his solidarity with illegal immigrants.

          Sometimes I want to cry.

          1. Thing is, fun fact. Superman would be covered under the “foundling” rule, where if a baby is found within the United States and there is no evidence that it was born elsewhere and it reaches a certain age without anyone claiming it, it recieves United States citizenship.

            ‘Merica, y’all.

            1. Unless you’re in the John Byrne continuity, in which he was actually born in the US. (He was carried, the duration of the gestation, in vitro, in a “birthing matrix” which was launched from Krypton and landed in Kansas and there delivered the baby.)

  5. I’ve noticed that more and more often, I look at the margin of error. I’ve read polls where the margin of error was over twice the gap between the results. That, to me, screams of weakness. Ditto “[source] Best Seller” tags that happily elide the very low number of reporting points that they draw from. If all your reporting sales points are in NYC, DC, Chicago, Seattle, LA, and San Francisco, your results might be a wee bit off from looking at what moves the fastest at Walmart and Costco.

  6. Good post and good points… Sigh… And if I actually wrote what I know, I’d be going to Leavenworth… Dammit… Most of the polling is like everything else, ‘pushes’ so that they get the answer they want. Figures lie and liars figure…

Comments are closed.