Descriptive Statistics > Forest Plot
What is a Forest Plot?
A forest plot (sometimes called a blobbogram) is a graph that compares several clinical studies. The graph has several components:
- A vertical line in the center. This is the line of equality, or no effect. If the blobbogram is a relative risk ratio, the vertical line will have no effect at zero; if it shows a mean difference, the null is 0. For ratios (e.g. the odds ratio) the line is at 1.
- A horizontal line representing each study. Sometimes this will be a horizontal bar. The width of the bar represents the confidence interval, usually the 95% interval. This is the range where the true value is likely to fall. The diamond/point/square in the center of the line is a point estimate of the true value. The bigger the shape, the larger the sample size. This is the most likely value out of the range of possible values; values towards the end of the line are less likely.
- An outline of a diamond at the base of the graph. This usually represents a weighted average for all studies but it can also be an odds ratio. This should be clearly labeled as either statistic. In the image above, it’s an odds ratio.
- An adjacent table provides more information about the study. The table can include:
1. A Simple Forest Plot Example
The horizontal line is perhaps the most important part of the graph. When a line representing a study crosses the vertical line, it represents no difference. If all of the horizontal lines cross the vertical line, it’s a sign that all of the studies were in agreement. What’s really of interest is if a horizontal line doesn’t cross the vertical — that’s an indication there were statistically significant differences between studies.
The line is a representation of mean difference (a statistic that measures the absolute value of two results) or standardized mean difference. This is sometimes called a “weighted mean difference,” but this is a bit of a misnomer as no weighting is actually involved in the calculations. The length of the line indicates the confidence interval: longer lines mean more uncertainty. Optional tick marks represent the 95% and 99% confidence intervals. Comments are now closed for this post. Need help or want to post a correction? Please post a comment on our Facebook page and I'll do my best to help!