What Normalization is: The process of bringing data values into alignment, for the purpose of comparison. In its simplest form, normalization involves adjusting values to fit on a common scale. More complex forms of normalization use adjustments to bring probability distributions or normal distributions into alignment.
What Normalization is not: A standard score, which is a method for expressing a value’s relationship to the mean.
Why Normalization matters: Typically, normalizing data involves the removal outliers and bad data points. This creates a smoother set of data and allows statistically valid trends to emerge. This is essential for comparison and analysis but the danger is that removing of outliers can sometimes hide critical information.
Find this post useful? Download our complete guide of key testing concepts with over 120 terms!