Numbers are liars. Many of you have heard me say this or read my posts on why numbers mislead us. Lately, I have yet more reasons to remind everyone.

Numbers Are Liars – Some Background

With my numeric background, of course I want to look at the numbers. This is the case for my blog, my articles, or the performance of any presentation I give at a conference. These are a few examples. And, over the years, I’ve noticed all types of odd variations. I can quickly explain some of them. Other bits can be explained upon some research or upon reoccurrence. Others are never fully explained.

With that, I’ve found that I’m able to come up with plausible explanations for the largest of variances. This is true even when I can’t prove them. Here, we’re talking about articles and not research data so this is sufficient. After all, my efforts to explain these variances aren’t my actual job but just one part of what I do to support my job.

What I’ve found more than anything else, is that the numbers don’t necessarily mean what we’d like them to mean. Every time someone tells me they’re “number 1 in LIMS” or that their publication has “one zillion readers,” I happen to know that whatever numbers they used to “prove” that don’t prove it in the way that the listener would expect. Too often, numbers are used to lie to us. This is especially true when we’re talking about a marketing or circulation situation.

I’ve occasionally written about how numbers mislead us. In Top Posts of 2019 for This Blog, I talk about this in a way that I hope helps readers understand this issue. Most of the numbers we’re sent are flawed, unless they’re based on actual scientific research, and that includes most of the survey data we get.

A Big Surprise – a HUGE Jump in Reader Numbers

With all that said, I now claim to have really hit the “big time” with a post that really went “viral” and “captivated the masses.” It was such a “hit out of the park” that I can’t believe that industry people aren’t flocking to me to get my autograph, at this point. Yes, that’s how severely popular that post was.

Let me step back a moment to explain something – posts tend to have slightly different cycles. Some posts quickly attract readers, then reading quickly drops off. Other posts seem to build slowly with regard to readers.

Or, was it?

Well, here’s the surprise part – my last post’s numbers skyrocketed past anything else I’ve ever written by an order of magnitude and is still climbing. In checking against various resources with different tools and looking at many aspects, at this point, I’m totally stumped. I have no idea why that post’s numbers are so extreme. Over the years, I’ve come up with all types of explanations for big jumps in the numbers. However, this jump is beyond explanation, even when you consider my vivid imagination for these types of things.

The bad news is that things that can’t be explained can’t be replicated. The good news is that people really want to read that post for whatever reason they have for it.

Back to The Lies When Numbers are Liars

So, just to be clear, while some people would look at the numbers and declare victory and start marketing themselves as the most wonderful LIMS services vendor, ever, here at GeoMetrick Enterprises, we’ll just remain a bit skeptical that these numbers mean that. I will hold my Marketing people back from publishing these types of claims if you readers will remain skeptical, as well.

And Yet More Lies – COVID-19 Data – More Numbers are Liars

Let me start out by saying there is a lot of great research and data-crunching going on with regard to COVID-19. Some of it will be junk science but most of it (let’s hope) is well-produced.

With that said, there’s also the contingent of people (myself included) who scrape all that data and process it, on our own for our own purposes. Most of what we’re doing isn’t useful for any true predictions. Actually, I refuse to tell anyone much about what I’m doing only because I don’t want anyone to think it’s validated in any way nor that it has any real value.

In any case, some people have been merely watching the numbers and using them for a variety of predictive purposes. Every week, I seem to find an occasion to say to one person or another that “numbers are liars” as they watch the numbers go up and down and make declarations based on them.

As I claim that numeric increases don’t mean we’re all about to die any more than seeing them decrease means we’re all cured, they often become incredibly annoyed when they ask what those numbers then do mean and I respond with, “Nothing. They mean nothing.”

It’s Just Like Everything Else With Numbers

As I’ve been saying, for years, you can’t just look at numbers and believe them. You need their context. You need some assurance that the number-crunching was valid.

In any case, I found something particularly interesting about these COVID-19 numbers. That is that having this discussion with ordinary people isn’t much different from having this conversation with the people in our industry who come up with untested explanations for their numbers. Once, again, when a little number-crunching helps you do your job in some way, that’s fine – but it doesn’t mean it’s something worth sharing with the rest of us, either.

Finally – Numbers Are Liars

Between the one post with the astronomical and unexplainable numbers, and also my dealings with people trying to make sense of the COVID-19 numbers, I felt a huge burden of lies streaming down upon me. By that, I mean all those numbers and all the misleading explanations of them.

In any case, any time someone starts a sentence with “The numbers prove…” I’m quick to respond that those number don’t prove anything because numbers are liars!

4 Thoughts to “Numbers Are Liars – As Usual”

  1. So true! Everybody runs around touting a “FACT!”. If they have a number it is a “scientific FACT!” especially if the wear a white coat and hang a stethoscope around their neck.

    I recently re-read Professor Hardy Cross’ essays where he cautioned against claiming “science” when discussing questions that are too complex and have too many variables to be amenable to a valid experiment under the “scientific method”.

    I recently re-watched Richard Feynman’s lecture discussing experimental results – to paraphrase: you can never prove a theory – the best you can say is these results are not inconsistent with the theory. Now there are umpty-gazillion experiments demonstrating the predictions of the theory of gravity (ignoring the relativistic effects) so I think we are safe relying on gravity, but taking drastic action based on one experiment (especially if it can’t be replicated) is a lie in the name of science.

    It is like a tradeshow demo “we got it to work once” (especially if we always enter the same data) so let’s sell it”.

    I am not an expert statistician but I have worked with some great ones. I/they never say “this proves …” more like “there is less than a 5% chance this happened by random chance so there is a reasonable likelyhood it is real” or “perhaps something interesting happened here”. But many people feel you must not question their statements if they waive a number in your face.

  2. Daen De Leon

    Numbers on their own are like the letters of the alphabet : individually, letters are devoid of context and intent. Grouped together, they begin to form statements and propositions, take on meaning and context – – and can disseminate truth or lies. So it is with numbers.

    When real science uses numbers, they are always embedded in a precise and accurate context, with a rigorous statistical analysis for all but the most trivial data sets. This is why much scientific literature can appear opaque and incomprehensible to most lay people. It’s why, with very few exceptions, popular science journalism is abysmal. And it’s why the unscrupulous will cherry pick and recontextualize numbers that suit their own agendas.

    It’s not that the numbers lie, no more so than the alphabet lies.

    People lie.

  3. Wasn’t it Mark Twain who quipped:
    Figures often beguile me, particularly when I have the arranging of them myself; in which case the remark attributed to Disraeli would often apply with justice and force: “There are three kinds of lies: lies, damned lies and statistics.”

  4. […] frequently write about how we misinterpret numbers. Just one example is Numbers are Liars, As Usual. When we get marketing numbers, we insist they mean things that we have no proof of. We see an […]

Comments are closed.