Precision is one of those concepts that is not really taught in standard econometric or data analysis classes. But as long as we are using data, it is present at the very heart of all the analysis we do. In this guide, we will learn about precision, what it is, why is it important, and how it affects our day-to-day data work.
So what is it? Precision deals with how we store our variables and perform operations — add, subtract, multiple, divide — across them. While Stata performs all calculations at the highest level of precision, the information that is saved back into the dataset can have lower levels of accuracy. You might have seen storage types, e.g. integers, floats, long, double, and strings, show up on screen. They display the level of detail that exists for each variable. These storage options are also set when we generate new variables or import data. And if we do not specify the storage type, Stata makes a judgement call on our behalf. Most of the time these choices are fine, but they might not be the best under certain circumstances. But if some storage types are always considered more accurate than other, then why don’t we always go for the best option? Because, even though higher precision implies better accuracy, it also…
