R-ticulate

8Working with Continuous Data

In Chapter 1 we have seen different types of variables and learnt how to confidently distinguish between them. Continuous variables, or, more broadly, continuous data contain numeric values that can take on any real number. Examples were given in Chapter 1, but to quickly remind ourselves – ‘body height’ for instance is a continuous variable, as it can take on any number with infinite decimals (within reasonable boundaries of course). Once we know that we are dealing with continuous variables, we need to know (i) if they will act as predictor or response variables in our dataset (Chapter 1), and (ii) what distribution they resemble (Chapter 3). This will inform our subsequent analysis. Box 12.2 can be used to help determine what type of analysis will give us the answers we are looking for.

In this chapter, we will cover some basic concepts that relate to continuous data and then have a detailed look at correlation analysis. Just like in the other chapters, we generally assume that our data follow a normal distribution (at least approximately). Because the real world shows us that this is often not the case, we have three options: (i) transform the data hoping they will then meet the assumption, (ii) use a so‐called ‘non‐parametric’, rank‐based method, or (iii) turn to entirely different, more modern methods that can account for non‐normality. Option (iii) has become by far the most common, as it leaves us with maximum flexibility. Nevertheless, ...