Harvardx course Practical Improvement Science in Health Care with the Institute for Healthcare Improvement.
Have you ever wondered about the relationship between two measures. For example, what happens to patient satisfaction when people have longer wait times? A scatter plot is a graphical tool used to examine the association between two measures. It’s a powerful tool for helping us understand relationships and associations between two variables. Scatter plots can be helpful for doing PDSA testing to understand variables that might affect the project measures. Scatter plots help us to see unusual patterns, data affected by special causes, and interesting clustering of data points. It also can show how changes in process measures are affecting outcomes.
Here’s an example of a scatter diagram showing a negative relationship. So let’s imagine again that we’ve got waiting time in– we’ll put time down here — in minutes and we’re doing 10, 20, 30, 40, 50. Make it straight arrow. So this is the x-axis. And then on the y-axis, we’re going to put satisfaction. So we’re doing 1, 2, 3, 4, 5, 6, 7, 8, 9, 10. So this is our scatter plot. All right. So imagine as we’re looking at this and we start collecting data from our patients that we start to see a series of dots. And as — basically — as time increases, we see patient satisfaction decrease and creates kind of a picture like this. And we see data look this way, right? So as the time is increasing, the score — the patient satisfaction score — is dropping. We call this a negative relationship. So if satisfaction scores increased as time increased, we would have a positive correlation. If the dots are randomly displayed like a shotgun blast, we’d say there’s no correlation.
One important thing to note is relationship or correlation is not the same as causation. There are statistical tools and methods to determine whether relationship is due to chance alone or statistically important. Most commercial spreadsheet programs have an option to create a scatter plot. Can you think of measures related to a process you’re trying to improve that may benefit from doing a scatter plot?
If this was helpful, share and include me @DaveWilliamsATX. Sign up here to receive a monthly email from me that includes all my blog posts and other Improvement Science resources I think you’d appreciate.