All – A statistical question:
I have the frequency distributions for hourly ozone concentrations in
the Smokies over several years. One set is from the peak ozone years
(1997-2002) and another is from 2012-2016 (the recent years).
The frequencies are the number of hours of ozone in 10 ppb concentration
bins, from 0 up to 140 ppb. For each time period, n = 5, and, I have 6
different sites in the Smokies with these data.
I want to statistically compare the peak years to the recent years. How
would I compare the two sets of frequency distributions?
Note: simply knowing that they are different is insufficient – I’d like
to know exactly which bins differ between the two time periods. Without
doing formal statistics on this, I can tell you that the frequencies of
high concentrations have been greatly reduced, and the low- to
mid-concentrations enhanced.
I thought of doing a bunch of t-tests, one for each bin, but I'm sure
that violates some statistical property.
FYI, the air in the Smokies is now cleaner than it has ever been since
measurements began. Thank the U.S. EPA’s Clean Air Act while it is still
here!
If you do reply, please provide some mechanics for how to actually do
the analysis (e.g., in SAS, my preferred analysis program; sorry, I
don’t know R....yet).
Thanks in advance for any advice on this.
Best, Howie Neufeld
--
Dr. Howard S. Neufeld, Professor
Mailing Address:
Department of Biology
572 Rivers St.
Appalachian State University
Boone, NC 28608
Tel: 828-262-2683; Fax 828-262-2127
Websites:
Academic: http://biology.appstate.edu/faculty-staff/104
Personal: http://www.appstate.edu/~neufeldhs/index.html
Fall Colors Academic: http://biology.appstate.edu/fall-colors
Fall Colors Facebook: https://www.facebook.com/FallColorGuy