User:PeterLFlomPhD/Dynamite plot sandbox
A dynamite plot is a variation on a bar plot in which lines and whiskers are added to indicate either the standard deviation, the standard error, a confidence interval or some other measure of dispersion. Sometimes the lines go only upward from the bar, other times they go in both directions. Dynamite plots are not a standard statistical graphic - they are not mentioned in Cleveland's books [1], [2] nor in Everitt's Dictionary of Statistics. [3] but they are commonly used [4]
Uses[edit]
Critiques[edit]
Dynamite plots have been critiqued for
- Providing too little information, [4] scoring low on Edward Tufte's data-ink ratio.
- Not showing sample sizes [4]
- Not showing the data [4] not following Cleveland's rule to show the data. [1]: 29
Alternatives[edit]
Recommended alternatives depend on the sample size, both overall and in each bar.
When N is small[edit]
When N is large[edit]
Implementation[edit]
References[edit]
Short items[edit]
Long items[edit]
- ^ a b Cleveland, William (1994). The elements of graphing data. Murray Hill, N.J: AT & T Bell Laboratories. ISBN 0963488414.
- ^ Cleveland, William (1993). Visualizing data. Murray Hill, N.J. Summit, N.J: At & T Bell Laboratories Published by Hobart Press. ISBN 0963488406.
- ^ Everitt, Brian (1998). The Cambridge Dictionary of Statistics. Cambridge, UK New York: Cambridge University Press. ISBN 0521593468.
- ^ a b c d http://biostat.mc.vanderbilt.edu/wiki/pub/Main/TatsukiKoyama/Poster3.pdf
- ^ http://biostat.mc.vanderbilt.edu/wiki/Main/DynamitePlots
- ^ http://emdbolker.wikidot.com/blog:dynamite
- ^ http://biostat.mc.vanderbilt.edu/wiki/pub/Main/TatsukiKoyama/Poster3.pdf