- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
I'd like to create histograms for about 50 continuous variables with bins based on the <25th percentile, 25th-50th percentile, 50th-75th percentile, and > 75th percentile. Is there a way to do this with proc univariate?
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
The pre-programmed histograms in SAS don't allow you to set bin limits at 25 percentile, 50 percentile, 75 percentile and so on. With some programming I think you could make that happen.
Of course, a box plot sounds exactly like what you are talking about, it shows 25 percentile, 50 percentile, 75 percentile and other statistics, and allows you to compare the "distributions" (as displayed by the boxplot) of 50 variables visually.
Paige Miller
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
That's what I thought...I started programming these IQRs for some frequency distributions along with the histograms and hoped there was an easier way!
Thanks!
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Meant to say along with the box plots
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
You might find this example helpful: http://support.sas.com/kb/35/171.html
Go to the Results Tab to see the graph, the Full code has the Proc Template code and uses a standard SAS data set to display.
It has an extra element of a small fringe plot to provide a little more detail of distribution that the bare histogram provides.
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
In a histogram the bins based on percentiles would produce a flat line, wouldn’t it?
Perhaps you mean something different?
@simkinm2 wrote:
I'd like to create histograms for about 50 continuous variables with bins based on the <25th percentile, 25th-50th percentile, 50th-75th percentile, and > 75th percentile. Is there a way to do this with proc univariate?