When we had two categorical variables, we could produce a crosstable to see how these two variables were related. If we have two continuous variables, we may use something called a scatterplot instead. Each dot in the scatterplot represents one individual in our data. We may also include a reference line here, to see if we have a pattern in our data (this will be discussed later).
The scatterplot can thus be used to illustrate how two continuous variables co-vary – or “correlate” – in their pattern of values. If increasing values of one variable correspond to increasing values of another variable, it is called a positive correlation. If increasing values of one variable correspond to decreasing values of another variable, we have a negative correlation. In the graph below, different types of correlation are presented. The letter “x” stands for x-axis (horizontal axis) and the letter “y” stands for y-axis (vertical axis).
Note While not addressed here, patterns can of course also be non-linear (in contrast to the positive and negative correlations shown in the graphs above).
In the scatterplot above, we display gpa on the y-axis (vertical axis) and cognitive on the x-axis (horizontal axis). We can see a quite clear positive correlation here: the higher the cognitive test scores, the higher the grade point average. This is also illustrated by the fitted regression line.
Note You can use the Graph Editor (see Graphs) to further edit the scatterplot.
