Each of these words is a synonym for measurement error. Notice that we are NOT talking about mistakes, or doing something wrong. Measurement error in the present context is an inherent variation or imprecision that exists in the measurement process. You can't eliminate it - a perfect measurement is not possible.
This kind of error is called "random error"for the reason that the actual measurement can be a bit high or a bit low, randomly. If you take a number of measurements and then average them, the result will be close to the "true" value.
The other type of error is called "systematic" error. When this type of error is present, all measurements are off - in the same direction. They will be all low or all high. Consider a bathroom scale which is not set correctly; maybe it reads 5 pounds low. Makes you feel better, but that is a systematic error. A clock that is fast exhibits systematic error; you can read the time as PRECISELY as you want from it, but it is not ACCURATE.
Prof. Scalise described the relation of accuracy and precision by using a rifle and target analogy. A hit pattern that is scattered out but roughly centered on the bullseye is accurate (average is in the center) but not very precise (scattered out). A very tight hit pattern that os off-center is precise but not accurate. Obviously a hit pattern that is smalland tight and is located in the bullseye is both precise and accurate. Finally a scattered pattern that is way off-center is neither precise nor accurate.
There is one more pattern to mention - the Texas Sharpshooter fallacy. In this flawed approach, the sharpshooter goes out and carefully shoots at a spot on the side of a barn, then locates the center of the pattern and paints the bullseye right over it! This is setting the target AFTER you have made the measurements and is NOT scientific. It looks really good, however. Every single pattern is dead center on the target.
This relates to the technique of making "ballpark" estimates, which are rough calculations using rounded numbers to get an idea of the magnitude of a result. Prof. Scalise used a real wienie of an example: how many Ballpark franks (hot dogs without buns) would be required to fill up Texas Statium? LOTS!! But how many? A million? A billion??? Don't just give up and say you don't know; take what you DO know and make a rough estimate.
A simple histogram is a plot of data values and the number of times each value appears. It can, by visualizing the data, give an insight into the distribution of measurements.
The average of a number of measurements tells you where the "center of mass" is. The standard deviation gives an indocation of how far from the average the measurements are. For example, you might like to live in a place where the annual average temperature is sventy five degrees F. Sounds fine, but we still need to know the standard deviation. If that standard deviation is thirty degrees, the annual high is well over one hundred and the lows are near freezing.
There are a LOT of factors that can introduce error. They can include such things as:
A common bathroom scale was used an an example. If you get of the scale several times, you may notice a slight variation in the reading due to mechanical friction. If you lean to one side, you can get a parallax error. If the scale is not properly zeroed, all weights will be off; this will be s systematic error.
Remember this one! The term "human error" does not describe ANYTHING. It means you haven't analyzed the error mechanisms closely enough to understand what is happening.
Prof. Scalise illustrated what are know as "error bars" in science. The error bars show the uncertainty in a measurement. They outline a little box; the actual value is almost certainly in that box SOMEWHERE.
Our message here is that no measurement is prefect; in fact some measurements are necessarily of rather low precision. Whenever you see some measurement quoted to a large number of places WITHOUT any error bars drawn or error ranges stated, be careful.