Have you ever noticed that when you are on vacation, your mind relaxes but starts thinking of work-related stuff?
It isn’t a bad thing. I know I am supposed to forget about work when I am away but by relaxing, interesting things come to mind at the weirdest times.
That is what happened while remote camping back in the summer..
One of the things I liked to do while out in the wilderness was practice shooting my bow in Archery. Archery was one of my favourite ways to pass time because I wanted to be as good as I could at bow shooting for hunting season.
So, I set up my target and started shooting my bow from 30 yards to start. I took my three shots and had excellent grouping, but my aim was off. This resulted in the picture below.
.png?width=348&height=463&name=Untitled%20design%20(6).png)
.png?width=377&height=363&name=Untitled%20design%20(5).png)
This happens with laboratory measurements as well. The method is good and is working but their ‘sights’ are off. They can make the same measurement over and over again, but it just isn’t the actual measurement.
If you were to send them duplicate samples, they would report similar values for that duplicate. In other words, they would ‘pass’ the test since the values would be the same for both duplicates. Unbeknownst to the users of the data, the values are wrong but reproducible.
How does this happen? That is topic for another blog but..
What does this mean for your results?
It means you can reproducibly get the wrong results..👇
1. Your data is wrong.
2. You might have results that are too high or..
3. You might have results that are too low.
You probably won't even know..
In the hunting world, I would miss the animal I am shooting at. Or worse, injure and not kill it. Not good.. not good at all.
This is why all projects need an assessment of both accuracy and precision if you want your results to be reliable. By building this into your program, it provides bullet proof validation that your data is exactly as you have represented it as. This is very important for litigation cases. Because who doesn’t want the best data for their litigation cases?
Later blogs will describe some techniques at validating your data to ensure that you have the best for your case.
