missing boxes in boxplot graph

[Brayden King]

I'm trying to create a boxplot comparing a variable across two categories (see below for descriptive statistics). For one category the box shows up as expected, but for the other I see nothing. I am able to see the outsider values, but no media or intervals. When I add the option "nooutside", one of the categories is completely blank. What could be happening here?

Just for info sake, here are the distributions of the variable by each category:
. sum totspendSB_postJan if boycottintent==0

Variable Obs Mean Std. Dev. Min Max

to~B_postJan 945 7.714138 27.84039 0 414.91

. sum totspendSB_postJan if boycottintent==1

Variable Obs Mean Std. Dev. Min Max

to~B_postJan 121 6.295454 19.66838 0 130.24

[Brayden King]

missing boxplot.tifhere's a picture of the graph with the missing box

[Nick Cox]

Please do read FAQ Advice #12 on posting .png (not .tif) and on showing the graph command in question.

Here there's an easy guess. So many values are 0 that it's the minimum -- and the lower quartile -- and possibly the median too.

For such data bare box plots don't work well. I suggest quantile plots or spikeplots instead.

Check the results of

count if totspendSB_postJan == 0 & boycottintent==0

count if totspendSB_postJan == 0 & boycottintent==1

Alternatively, omit the zeros, report how many there are and show positive spending on a log scale.

https://stats.stackexchange.com/ques...sing-a-whisker tells a similar story.