How to Read Scatter Plots
- 1). Trace both vertically and horizontally from the point until the axes, to find the values of any particular point on the scatter plot. Each point represents one person or item from the group being studied while the point's vertical position indicates one piece of data about that person and its horizontal position indicates another. For example, in a graph showing the relationship between shoe size and height, each point on the graph represents one person. Tracing down vertically, you find that the point is above the number 44 on the horizontal axis showing shoe size. Tracing horizontally, you find the point is level with 5'8" on the horizontal axis. This means that the person in question is 5'8" tall with size 44 shoes. Do the same with all the points on the graph.
- 2). Look at the shape the points make on the graph when investigating the correlation between factors. Draw a bubble around the points and note the slope that it forms. If the bubble is upwards sloping from left to right, then there is a positive correlation between the two factors; an increase in one factor is usually associated with an increase in the other. If it is downwards sloping from left to right, then there is negative correlation; a decrease in one factor is associated with an increase in the other.
- 3). Look at the tightness of the bubble to find the strength of the correlation. If the bubble around the points is very tight or forms an exact line, then the correlation is very strong; a fixed change in one factor is associated with a fixed change in the other. You would expect this kind of correlation between factors with a strong relationship; pressure and volume of a fixed quantity of gas, for example. If the balloon is very fat then the correlation is weak; a fixed change in one factor is associated with a variable change in the other. You would expect this with data which has a weak or distant causal relationship, such as the relationship between height and shoe size. If you cannot draw an oblong around the points, then they show no correlation.
- 4). Draw a line of best fit through the points on the graph. This is a straight or smoothly curving line which has an equal number of points on each side. The sum of the distances between the points and the line should, as far as possible, be equal for the points on each side of the line. You can then use the line of best fit to make predictions about the likely value of one type of data when only the other type of data is known. Read in a straight line from the appropriate point on the axis which contains the known data to the line of best fit. You can then read from the line of best fit, in a straight line, to the axis which contains the unknown data and obtain a probable value from the number you find on the axis. The best predictions are made when working with a strongly correlated set of data.
