 Sometimes, we need to choose what average to use to analyze data in order to get the most believable result. There are a lot of averages, and you can always construct your own for your problem. But here are all three most common ones, mean, median, and mode. Mean is just a sum of all values in a data set divided by their number. For example, take a look at this set. These are the ages of children who took part in a maths competition. Let's find the mean. To do that, we add all the numbers together and divide them by the total number of values. So our answer in this case is 13. Median is the middle value in an ordered data set. Let's try to find the median age of children from our maths competition. Here, we can see that 14 is the median value of the set. Mode is the most frequent score of a set. In the case of our children, the mode is also 14. As you can see, average values can be very similar or even the same. But there are situations where some of them fit the purpose better than others. For example, if we are trying to find a measure of central tendency for categorical data, we can only use mode. That's because we don't have values to calculate the mean or to order them in a set to find the median. Consider this. Students from the same class were asked what their favorite pet was. And here are their answers. We can easily notice that most of the students preferred dogs. So we can conclude that this is the mode of their pet preference. If our variable is ordinal, we can order values to find the median and it will serve as a much better estimate than the mode. For example, if we look at the results of a questionnaire with the ages of respondents, we can see the tendency by ordering them and finding the median of an ordered set. So here, our median is teenager. Although there are more children than teenagers and the mode would be child. But as we can see, the median is closer to the real tendency than the mode in this example. And this will be the case for most of the ordinal variables. Now, if we have an interval variable, we have a possibility of calculating it. So the most credible average here is the mean. Let's imagine in the previous questionnaire, we asked for an exact age and got the following results. So our mean here is 13. But the mean has one bug. It's too sensitive for the outliers. Imagine that in the previous example, the ages of adults were 71, 74 and 82. Not 21, 24 and 22, so much older. If we count the new average now, we get 29, not 13 as the first time. This situation is almost the same as the previous, but our mean has changed so much. That's why if our set is skewed, it's better to use the median instead of the mean, which in our case is 12.5. There are many other situations and their corresponding averages. But in most cases, for a fast evaluation, you can apply the following rules that we've learned from this lesson. For the categorical variables, use mode. For the ordinal variables, use median. For interval variables that are not skewed, use mean. And finally, for skewed interval variables, use median.