topic Re: Should I use Proc VARMAX for my multiple multivariate time series dataset? in Statistical Procedures

Should I use Proc VARMAX for my multiple multivariate time series dataset?

evance — Tue, 12 Sep 2023 16:56:32 GMT

data WORK.ANALYTIC;
  infile datalines dsd truncover;
  input fips:$7. year:32. ptb:32. entropyindex:32. mua:32. rural:32.;
datalines;
01001 2007 15.4 0.6408070774 1 0 
01001 2008 14.9 0.6347032529 1 0 
01001 2009 14.6 0.5946361175 1 0 
01001 2010 13.6 0.4998764511 1 0 
01001 2011 13.7 0.4841341568 1 0 
01001 2012 13.5 0.4451194338 1 0 
01001 2013 13.2 0.4636948987 1 0 
01001 2014 11 0.5613597794 1 0 
01001 2015 10.5 0.4616523638 1 0 
01001 2016 10.6 0.5859625243 1 0 
01001 2017 10.2 0.5429811399 1 0 
01001 2018 11.3 0.7012052158 1 0 
01001 2019 11.9 0.5924427314 1 0 
01003 2007 16.3 0.5890522164 1 0
01003 2008 15.6 0.6327763934 1 0

Hello,

I am trying to look at the relationship between my exposure variable (ptb) and outcome variable (entropyindex) over the time period 2007-2019 for every county in the US (each county has a unique "fips" code). There are over 3,000 counties each year. Additionally, I have two covariates - mua and rural.

A snippet of my data is provided above.

For data structured this way, would the best way to run a time series analysis be to use proc varmax? Please let me know if I can provide any additional information!

Re: Should I use Proc VARMAX for my multiple multivariate time series dataset?

ballardw — Tue, 12 Sep 2023 15:06:44 GMT

I no nothing about VARMAX but if your MUA variable has something to do with municipal status and Rural designates something about rural/urban status they a going to be very highly correlated with FIPS and need some caution.

Some counties will change status slowly with population changes.

Note: post code in a text box opened on the forum with the </> icon that appears above the main message window. The main message box reformats text and may not run as expected. Your example doesn't create a very usable data set when I copy it from the message window and run because of the DSD with no specified delimiter. So I get values for FIPS like "01001 2" and the other variables are all missing.

Re: Should I use Proc VARMAX for my multiple multivariate time series dataset?

evance — Tue, 12 Sep 2023 16:54:04 GMT

Thank you for your response! MUA is medically underserved area.

I'm not sure I understand the second half of your reply because this is my first experience posting on this website.

Re: Should I use Proc VARMAX for my multiple multivariate time series dataset?

evance — Tue, 12 Sep 2023 16:57:01 GMT

I think I fixed it. Please let me know if that helps.