How to Reject Outliers in Data
Whenever you are examining experimental data, you need to make sure that it is checked thoroughly for any outliers so that there is no effect on the overall result. You need to follow three simple steps to reject any outliers in the data you are examining. Calculate the mean of the data, compute the standard deviation for each data point, and then eliminate all points that you feel are varying from the result that you have calculated. This technique is helpful as experimental data keeps coming in and you will have to adjust it according to the latest values.
Instructions
-
1
Calculate the mean
First of all, you must calculate the mean of all the data points. The data should also include the outliers that are suspected or the ones you feel should be rejected. -
2
Calculate the standard deviation
The second step should be to calculate the standard deviation of the data that you have. This will determine the data that is deviating from the mean. -
3
Compute the number of standard deviations for each data point
After you have calculated the standard deviation, you need to note down the number of standard deviations for each data point in a separate column. First, calculate the area under the normal distribution curve between z and ∞ and then compute the area under the normal distribution -∞ and z. This totally depends on the value of z. If z is greater than 0, use the first method and if z is less than 0, use the second method. -
4
Reject the outliers
Once you have computed all the data considering the values of z, you need to see the data points or the entries that are visibly different from all the other values. You can reject all the outliers that are different from the data standard. However, you need to make sure that you are perfectly right when you reject a value as the overall value may be altered if the decision is not right. The end result may be either underestimated or overestimated if you have not compiled the correct result.
