Scatter Plots

Scatter Plot in Psychology Definition

Scatter plots are used to determine if there is a relationship between the two variables being studied. If there is a relationship between the two variables, it will be shown on the scatter plot. In order for data to be shown on a scatter plot, it has to be measured in numerical values.

On the x-axis of the scatter plot graph is one variable and on the y-axis is the other variable. Remember, these variables have to be numerical! You can’t have a scatter plot that has one numerical variable and one non-numerical variable.

When looking at a scatter plot, look left to right to determine if there is a relationship between the variables. There will be points scattered around the graph but if there is a relationship, you will see the majority of the points either trending upward or downward.

Line of Best Fit

So there’s a bunch of points that look like they’re going to the top right or bottom right corner of the graph but you’re not fully sure of the direction of all the points. What do you do?

Line of best fit to the rescue!

Think of the line of best fit as an estimate. The line of best fit will most often not actually go through every single point on the scatter plot; otherwise, it wouldn’t be a line! It’s just showing the general idea of how the points would look if they were averaged. In some cases, the points on the scatter plot and the line of best fit are all very close together which is great because that means there’s not a whole lot of variation in your data. However, it does happen that your line of best fit has some points around it, but a couple of points that are nowhere near it. That’s okay! Those are just your outliers.

Types of Scatter Plot in Psychology

There are three types of scatter plots that give researchers more information about their data.

Positive Scatter Plots in Psychology

The first type of scatter plot is a positive or rising scatter plot. Here, when looking from left to right at the data, the points will, for the most part, trend upwards. If you were to insert a line of best fit in a positive scatter plot, the slope would be positive. Easy enough to remember!

Data that results in a positive scatter plot has both variables either increasing or decreasing together. They have to move in the same direction!

Negative Scatter Plots

As you might guess from the name, the points for a negative scatter plot will trend downward when looking at the graph from left to right. The slope of the line of best fit would also be negative.

Data that results in a negative scatter plot has to have one variable increasing and one variable decreasing.

Fg. 2 Negative scatter plot, commons.wikimedia.

No Relationship

No correlation means that the points are all over the place. When you look at a scatter plot of data that shows no correlation, there will be no method to the madness of the points. The points will be randomly over the graph and you will not be able to tell if they all tend to go upwards or downwards. There is no relationship between the two variables.

Fg. 3 No relationship scatter plot, commons.wikimedia

Examples of Scatter Plots in Psychology

Psychologists use scatter plots most often for correlational research. Correlational research collects data from two different variables and determines the relationship between the two variables. Unlike experimental research, correlational research does not manipulate variables, but rather simply looks at or collects data for the variables.

By putting the correlational data in a scatter plot, researchers can look at the data and determine if there is a relationship. When variables are correlated, they change at the same time (either they both increase together or one increases while the other decreases).

Remember with correlational research, correlation does not equal causation! Just because two variables are correlated, it does not mean that one variable causes the change in the other. Correlational studies can only show that variables are related. If they are, researchers would then have to conduct an experimental study to determine causation.

All of the examples above would be considered correlational research.

Scatter Plot Correlation Coefficient

In scatter plots and correlational research, there is an important value called the correlation coefficient.

A scatter plot whose line of best fit has a negative slope will have a negative correlation coefficient. When there is a scatter plot with a positive slope, the correlation coefficient will be positive.

A correlation coefficient of -1.00 or 1.00 would be the strongest possible correlations. -.90 or .90 would still be strong correlation coefficients -- anything below -.80 and above .80 would be considered significant correlation coefficients. For the coefficients that fall outside of that range, they might show correlation, but it would not be considered statistically significant.