Bar graph and histogram can visualise the frequency information of data. There are other visualization visualizations to summarise a data set in different ways but understand the components of those visualisation we need to first talk about how to summarise the data with a few numbers. In practice any data set can be summarised with the two numbers one representing the centre of the data set and the other representing the spread of the data set.
In any publication it is pretty much a standard procedure to report a pair of one measure of central tendency and a measure of dispersion as a simple summary of data. So the simplest central tendency measure is the mode which is the most frequent observation. This can be calculated for any levels of measurement as counting is all that is required to calculate the mode.
So let's take a look at our previous visual acuity example. To compute the mode it is not really necessary to sort the data but it'll be much easier if you do so to count which datum occurs the most so if we we can just look at each datum just data by data to see how many times each datum occurs so for example this 0. 02 occurs once and once, once, once, twice, twice, once, three times so 0.
22 occurs three times . 24 occurs two times, two times, one, one, one, one, one, right so the mode for this data set so the most frequent observation is . 22 so mode is .
22 logMAR. Another central tendency measure is median which is literally means the middle score. By definition, median is the value that divides the data into two groups containing an equal number of observations.
Because you need to be able to rank order between the values to calculate the median you can only calculate this quantity for the ordinal or above levels of measurement and you cannot calculate medians for the nominal so what that means is that you cannot calculate the median for the nominal level of data and by the way please note that I'm going to use a different data set here than the previous visual acuity for the illustration purposes. So the data here are the number of facebook friends your facebook friends have so if you do have a facebook account then you probably know what this is so if you have an account then you can go to your facebook friends profile to find out how many facebook friends each of your facebook friend has so this is the data they're taken, this is kind of a hypothetical data say you know this is coming from my own facebook and then these are the so let's just say that I only have 11 facebook friends and these numbers represent how many facebook friends my facebook friends have. So to calculate the median then you need to sort the data first then find the score splitting the number of data into lower and upper equal halves.
So here we have an odd number of data right 11 so and in this case and well and 11 right so you only have to find the middle number and so the middle number here is 98 right that's because from 98 above and below you have five and five right so we have exactly the same number of data five and five above and below this value of 98 right so in this case the median so I will say this is a median one is so this is our median value 98 so now you think you know that the one of your friend is a kind of oddball as that person seems to have an unrealistic number of friends so I'm talking about this one. So over 2 000 facebook friends so it seems like that this friend is a celebrity or something so you want to calculate the median again after you remove this celebrity friend okay so you just remove this okay now then the number of facebook friend I mean, the size of the data right the the number of sample in this data set is now becomes 10 right so now we have even number of data then how do we calculate the median so in this case. .
. You need to find out two middle scores splitting the number of data into the same lower and upper halves and calculate the average of the two scores so if you just look at this data now these two scores are two middle scores. So from these two scores we have four and four here right so the median is just average between these two numbers divided by two okay so that's 191 two so that's 95.
5 so that is our median number two so that's just a median without this extreme score. But if you compare this value 95. 5 and 98 so regardless of the extreme value or the outlier in a data set the median doesn't really change much.
In fact, this is one of the properties of a median that you know this central tendency measure is relatively unaffected by the extreme values or outliers at either end of the data. So last central tendency measurement is the mean also known as the average and there are different kinds of means we can calculate but here we will only talk about the most commonly known arithmetic mean because this is probably the most frequently reported central tendency measure in statistics for interval and ratio level of measurement. I'm pretty sure that you all know how to calculate the mean.
Here we have a formula to calculate a sample mean so here the x bar so that's how you read it x bar as a sample mean okay and that equals to and this symbol here is a greek letter it's a it's kind of a same as it's a similar to the S and this is read sigma sig-ma right and in mathematics and the statistics this symbol sigma means adding or summing what's coming on the right so whatever comes on the right of this symbol means add all the numbers on the right so i here is index of data so add the data from the first data to the n which is the last member of the data so add from the first data to the end that's what it means, all these subscripts and superscripts so that has to do with the data index here. So from i equals 1 to and so here if we look at this data right up here this number of facebook friend so we have the first data i it's 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11. So we have 11 data so you add the first data which is 57 to the last n equals 11 which is 22.
So you add them all up to calculate the sample mean plus, plus, plus, plus, all right so you add all these up and then what do you do and to calculate the. . .
just erase this after you add those numbers now you divide this by the number of score which is 11. then you're going to get the sample mean so the sum is the sum of all data equals actually 3151 (meant to say). So you can use your calculator to see if this is correct but the total uh um sum of the score is 3151 and we have n equals 11 so x bar equals 3151 over 11 and this will be 286.
45. So that is how we calculate the mean. Now just for the sake of the illustration because we just did this to calculate the median now let's say you do not want to include that celebrity friend.
So you remove remove this friend and do the calculation again okay so you just add all the numbers except for that large number so you do this sum again this is sum1 and sum2 it'll become 811. now our n becomes 10 because we remove 1 out of 11 data so the n is now so that's x bar1 x bar2 becomes an 11 over 10. so that's 81.
1. So if you compare these two you can see that there's quite a lot of change, depending upon the existence of outlier or the extreme value. So unlike median, mean the mean is heavily influenced by extreme values for outliers.