# How to calculate standard deviation

00:00:03,476 --> 00:00:08,814
One of the most important things to know about data is how spread out the data points are.

00:00:08,814 --> 00:00:15,022
The summary statistic that we calculate to show this is called the standard deviation.

00:00:15,022 --> 00:00:33,139
In Walkthrough 2.1 you reformatted a table that looked like this. We've also added a new set of data on the right here. Note that both sets of data have the same mean value. We've added those underneath the table.

00:00:33,139 --> 00:00:48,273
We've also added two charts showing a frequency of observations for each value. On the left we can see that there's a wide range of values but on the right there are only three values, so the right-hand shot is narrower with higher columns.

00:00:48,273 --> 00:01:11,318
So eyeballing this data implies that the right-hand side data is less spread out - it should have a smaller standard deviation. So let's work out these standard deviations. First of all we're going to label a cell underneath the mean value to show that we're calculating the standard deviation.

00:01:11,318 --> 00:01:28,346
Next to it we type =STDEV.P and open the bracket. Next we select the data cells, close the bracket and hit Enter and that calculates the standard deviation for this data.

00:01:30,824 --> 00:01:44,830
We also do the same for the right hand table. As we can see the standard deviation is much larger for the left hand side data.

00:01:47,195 --> 00:01:55,331
There's a handy rule of thumb for standard deviation: 95% of your data should be within two standard deviations of the mean.  

00:01:55,331 --> 00:02:12,098
Now it's quite a small data set for either of these tables but let's do it for the left hand table. Twice the standard deviation is about nine or ten and the mean is eleven so that means that all the data should fall in the range between about 1 and 20 and as you can see that's true.