hi everyone got a video Lesson here for you about precision versus accuracy and random versus systematic error and these are some really important Concepts in any scientific subject that you're studying so it's really important to have a good understanding of them so let's have a look at what they're all about these are the six key terms that we're going to look at in this video they are Precision scatter accuracy true value systematic error and random error so let's look at the definition of four of these terms the first one is Precision Precision refers to the spread of a set of results measurements are precise when there is less scatter in the results to understand what that means we need to understand what scatter is well scatter is a measurement of how close the values are to each other in a set of data so what we're saying with Precision is if the values are all close together then they are precise but if they're far apart then they're not very precise and if they're close together that means that there is not much scatter whereas if they're far apart that means there's a lot of scatter accuracy is a bit different accuracy refers to how close the measurements are to the True Value now to understand that we need to understand what a true value is well the true value is the accepted or agreed or nominal value basically the true value is what the actual value is or what the value should be so for example if I asked you what is the boiling point of water well what is it hopefully you said it's 100° C because that's right the boiling point for water is 100 de C that is the true value but if you were to do an experiment and measure the temperature at which water boils you may not get that value there's a range of reasons why and that's going to lead into our discussion about precision and accuracy and random and systematic error a Target here to visualize our the results of an experiment let's imagine that the bullseye of this target is like the True Value for a particular measurement that we're trying to take now we might do a range of measurements and get some results that look like this we might have one point here one down here one here and one all the way over here now I've got a title up here explaining that we would call this low precision and low accuracy low Precision because as you can see the results are very scattered they're not all near each other and low accuracy because the average of these four points the average air like if we were to take an average I guess it would be sort of around this point here which is certainly not in the middle at the bullseye which is where the True Value would be so we would call this low precision and low accuracy so in this example we're going to look at low precision and high accuracy so to have low precision and high accuracy we might have results that will say something like this so four different results taken obviously low Precision because they're scattered all over the target but if you think about where the average result for those four points would be it would be right smack bang in the middle here right on the True Value which is of course high accuracy so you can have a situation where you have low precision and high accuracy have higher precision and low accuracy for a set of results well yes you definitely can if we had our four results that were all right around there that is very high Precision because they're very closely um grouped together there's not much scatter at all but it's certainly not high accuracy CU they're all way out away from from the True Value so it's low accuracy and high precision and then of course this is what all scientists would be striving for you can have a situation where you have high accuracy and higher Precision so what causes those examples where Precision is affected or um accuracy is affected well it's something to do with systematic error and random error now systematic error is error that affects all measurements in the same way systematic errors are caused by a floor in the system that's a good way to remember them the example of the sort of thing that would cause a systematic error is something like not calibrating your measuring instrument so it's out by the same amount every time so in this example here where we had high Precision but low accuracy this is a classic Telltale sign of a systematic error all the results are out by EX L the same amount and there's usually some sort of flaw in the system and an error that's occurred with the device that we're using random errors are errors that occur randomly funny that and they affect measurements in an unpredictable manner random errors may occur due to carelessness or lack of concentration you sometimes hear them referred to as human error but what you need to realize is random errors are random and no matter how careful you are with your experiment whenever a measurement is taken a random error will occur it's the magnitude of the random error that we can control but random errors will always be present and we can actually we can minimize the effect of random errors but we can't stop random errors from happening and we'll look at that more a little bit later so this is the sort of example of what you could get from random error because there doesn't seem to be any pattern with these results they're all missing the Target and they're missing the Target in totally different areas so um it's it's low precision and precision is what can be affected by random error so now we're going to have a look at graphing with a systematic error so this is the line of best fit that we would see for these points here which are the True Value so imagine that line of best fit represents the data that you would see for the True Value but the red points are what this student actually got from their measurements and if we draw a line of best fit for those points we see that sort of uh shape now look at the difference between those lines of best fear as we move along the graph well it's the same all the way along it's consistent the entire time which suggests that we'd be looking at systematic error here in this example we've got a graph where we've got our true value so let me redraw that line of best fit for our true value we've also got some results where you can see there's a fairly significant amount of random error and The Telltale sign is because those different data points are scattered all around either side of the True Value if we were to draw a line of best fit for those points we'd probably get something a little bit like that and you can see certainly we don't have a consistent difference between those two lines of best fit the other thing that you might notice is if those lines if those data points in red were a little bit different and Scattered a little bit differently we could almost end up with a line of best fit exactly the same as our true value which means regardless of whether you have random error or not you can still end up with results very close to the True Value but with systematic error it's impossible to ever get results that are exactly the True Value because they'll always be out by a consistent amount final example we've got here is some results from four students who did an experiment on finding the bo the temperature of the boiling point for water so as we talked about before we know the true value for the boiling point for water is 100° C but you won't always get that value when you're doing an experiment because of random and systematic error so we got four students here they are here they did four tests then they took an average which is the mean and they looked at the range in their results and let's have a look first of all at student one student one has taken their four tests here their average is 98.5 and their r range is 6.9 so comparing those to the others they're not very accurate and they're certainly not very precise because their points their data points are all over the shop one's above 100° and three are well below 100° so we would call that low precision and low accuracy so now looking at student 2's results student 2's got their four tests they're all up above 100° and they're all 102 their average is 102.5 so that's not very good in terms of accuracy but their range which refers to the scatter of the results is really low which means there's not much scatter and have a look at these points they're all very close to each other so this is a case of high Precision but low accuracy and so we know that this must be affected by a systematic error and then looking at student 3's results they've got an average of 100.1 which is good accuracy and they've got a range of 0.5 which is fairly good in terms of scatter so this one we would say is good accuracy and good precision and in student 4's example we've got a range of five so they've got a lot of scatter so low precision but have a look at their average it's right on 100° C so this is an example of low precision and high accuracy so we've learned all about precision and accuracy and random and systematic error but what do we actually need to do about them in experiments well we need to be able to identify if systematic errors have occurred the way to do that is by repeating the experiment with different apparatus and different materials usually in a different place at a different time that will allow us to identify possible sources of systematic error because we'll compare our results from our first experiment to our second if we get exactly the same thing we'd be confident that there probably wasn't systematic error if we get two very different results then that starts to suggest that there may be systematic error and of course we'd need to repeat again and again and again to work out which one is the experiment where the systematic error did affect the results so what do we do about random error well all we can do about random error is reduce the effect of random errors on our overall results to do that we need to repeat the experiment several times and average the results we call this increasing the sample size so in the temperature of boiling water experiment the students boiled water four times if you wanted to really reduce the effects of random error on that experiment you could boil water 10 times 50 times 100 times the more you do it and take an average of the results the less those random errors will be affecting the results remember increasing the sample size doesn't rid the experiment of random error it simply reduces the effects of the random errors on the results so that's the end of this lesson thanks very much for watching and we'll see see you next time