This is just the minimum, lower quartile, median, upper quartile and the maximum. They are pretty good way to describe the distribution of continuous variable.
Details #
These are the numbers that define the ends of the whiskers (or the outlier points), the bottom and top of the box, and its waist in a box and whisker plot so can sometimes be read off, or estimated from such a plot even if the actual values are not given in a paper.
One interesting spin off from this is that, if you think the distribution might be near enough to Gaussian, you can estimate the mean and median of the same distribution from all of these values, or from some of them. That can also work if you think the distribution is something that can be transformed to near Gaussian, e.g. a “lognormal” distribution.
Why does this matter? I can help you estimate things that not reported in publications when you want to compare findings with your own, or across reports (“meta-analysis”).
Try also #
- Distribution
- Five number summary
- Gaussian distribution
- Maximum
- Mean
- Median
- Meta-analysis
- Minimum
- Quartiles
- Seven number summaries
- Standard Deviation (SD)
- Transforming data/variables
Chapters #
Not covered in the OMbook.
Online resources #
I will create a shiny app and probably an R function, using someone else’s much cleverer work, to make this estimation of means and SDs from five figure summaries easy for we mere mortals! May take a few days.
Dates #
First created 31.iii.25, links expanded 10.iv.25.
