Yet Another Potential Statistical Blunder


Those of you who regularly read this blog probably know that I’m skeptical of any of the statistics posted about who is #1 in LIMS, who has the greatest readership, which product is the most liked, and similar statistics.

Considering that many of the people who get their hands on numbers seem to think they can easily create meaningful statistics from them, it is no wonder that we hear so many contradictory claims.

Just now, I was looking at the raw numbers for my blog readership and noticed that the number of readers from the UK was five times higher than that of the US. If I had the urge to try to “make something” of those numbers, I could begin to make claims that my blog is five times more popular in the UK than the US. I could say that, being so well-read in the US, it was a shock to find that I’m yet five times more pored-over in the UK. I could start a new campaign in the UK to point-out that they should jump at the chance to immediately use my consulting services, as I’m obviously a household word there.

On the other hand, when I looked at my blog numbers, this morning, it was halfway through the work day for most people in the UK, while many in the US were just starting work. While I do not know this for a fact, I suspect the numbers were higher in the UK merely because more people had a chance to wake-up and read things. I could have made a huge blunder by making a lot of claims about these numbers, if I were the sort of person to do that.

Yet, these are the types of evaluations that are made on numbers in our industry on a daily basis. In some cases, by what people have told me, I know for a fact that they’ve jumped to wild conclusions about the raw numbers they have obtained. In other cases, merely by reading some of the claims, most of us can tell that they are rather silly claims. Yet, the people making them continue because, if they can convince just a single person with their claims, they can obtain a new customer. They’re getting results and they’re happy, so they’re not going to change what they’re doing to please the rest of us.

My own customers express frustration about this and ask me what they can do to get around these confusing numbers. I always tell them to just ignore every number they see. If they’re interested in a product or service, I tell them they should go look at it, talk to someone, do a little digging to find out the details behind it, but don’t read any marketing promotion without a grain of salt.

Or, if you do want to read some promotional material to believe, here it is:  GeoMetrick Enterprises is #1 in LIMS. We’re the best at everything, doing 100% only what 10 out of every 10 customers needs. We believe so strongly in this that we give a total guarantee that we’re just so great that 15 out of 15 randomly-selected of our customers is almost certainly smiling right at this very moment.  😉

Gloria Metrick
GeoMetrick Enterprises
http://www.GeoMetrick.com/


3 responses on “Yet Another Potential Statistical Blunder

  1. Very true about stats. They are used with impunity in marketing and politics. Nothing like using math to tell lies. It has a sort of irony you just can’t buy.

    Here is a for instance… In the LIMS/LIS industry, companies are always going on about how they are number 1. I think everyone in this world can be number 1 at something. I can claim that Gloria Metrick is #1 in the LIMS industry. Without more qualifying facts, the claim is meaningless.

    All of that being said, I do like raw facts. For instance, there are over 200,000 clia certified labs in the US and I can give you a free and open list of every single one of them in the US and its territories. How many clients does Labware have? Between 500 and 1000? The same can be asked of any LIMS/LIS vendors and their number of clients are dinky compared to the market and almost to the point to make even the largest in the group to be statistically meaningless. There is hardly a leader between any of them given the shear huge size of the lab market but still, everyone is #1 at something.

  2. I see what you’re saying. Let me add just a few more of my own comments on statistics:

    For one, I think people think too much about what they WANT the statistics to mean then find ways to make them mean what they want. Some of this is probably due to pressure from their employer to find the right numbers. In my first job after college, I worked writing mathematical software for scientific applications and much of my work involved someone bringing me a whole lot of numbers and asking me to “do something” with them. Occasionally, there was nothing I could “do” and they would then be quite angry, even though the problems usually stemmed from poorly run experiments, contamination, and other issues that prevented them from proving whatever they had planned to prove.

    Just this past Sunday, I got pulled into a great discussion about the fact that, merely because two things happen at the same time, it doesn’t prove any correlation between them – it doesn’t necessarily mean that one causes the other. Many times, though, those people working with the numbers are so excited by their results that they don’t want to find out if there’s something else involved.

    My last comment comes way back from when I was in high school. As a lesson in how you can be truthful and misleading at the same time, we were given a sheet full of facts about tomatoes, leading the reader to believe that tomatoes are quite dangerous. Did you know, for example, that 100% of all people who eat tomatoes die?!?!?! Yes, it’s true. We all die so, of course, people who eat tomatoes aren’t an exception. This sounds like a silly exercise, but it was full of these types of statements and it was a good illustration about how purely truthful statements can scare the reader even when there is no danger to be found.

  3. Science is riddled with this type of stuff. It is all part of being human. It is very hard to find pure and absolute truth in the world of humans and if you think deeply on this topic, it will lead you into some very interesting places and it brings up questions about things you took as unshakeable. So one can argue that most everything is taken on some degree of faith where fact and faith form a relative percentage to one another. Can that ratio be measured and quantified? Probably not. It would need to be a probability I suspect.

    You might be surprised about how much “faith” science depends on. Theories are based on that combination of faith and fact. Ergo, religion does not have the monopoly on faith; science depends on quite a bit of it. Of course it all depends on what you call religion but I will stop at that point.

Leave a Reply

Your email address will not be published. Required fields are marked *