# Understanding and Interpreting letter value boxplots

I have read through the R documentation of lvplots, but it didn't provide enough material on how I would interpret these plots. I want to just have an intuitive idea on interpreting these plots. I know for k=2 , the IQR and median are represented.

1)What boxes are created when increasing the value of k > 2 , what does the length of box represent?

2)What does the width of the box represent?

Below is a particular example for k=3. • You should try to elevate this question above its R context. People here have different favourite software and so the focus should be on the underlying statistical issue without assuming that they want to read software documentation to find out what you are asking. (There is a meta issue about how much hand-holding is expected from R documentation, which I won't address.) – Nick Cox Sep 3 '17 at 9:05

The key term is letter-value (box)plots and the key reference is now

Hofmann, Heike, Wickham, Hadley and Kafadar, Karen. 2017. Letter-value plots: Boxplots for large Data. Journal of Computational and Graphical Statistics 10.1080/10618600.2017.1305277 http://dx.doi.org/10.1080/10618600.2017.1305277

Earlier versions of this paper can easily be found on-line.

As I understand it the width of each box just indicates how a box is defined. The fattest box is between letter values that are (approximate) quartiles, the next fattest boxes stretch between (approximate) quartiles and the (approximate) octiles beyond in either tail, and so on. Positively, this is just an extension of the common box plot convention that each box indicates that it is the interval between quartiles and the width is otherwise just a conventional choice. (Only occasionally are boxes shown that indicate the number of values in each.)

A little more negatively, people have to learn that the width of the box is otherwise arbitrary. It's not, for example, a boxy version of a density plot.

But the interpretation is otherwise similar to that of box plots, e.g. the central half of a sample is within these limits; the central three-quarters within these limits; and so on. Are groups or variables similar or different in distribution?

For a survey of letter values with different emphasis, see

Cox, N. J. 2016. Speaking Stata: Letter values as selected quantiles Stata Journal 16(4): 1058-1071. http://www.stata-journal.com/article.html?article=st0465 (Subscription required for access until 2019Q4.)