So, we have some data by year for the past 5 years on the number of people who received a service in a given state. We’d like to group and summarize the data. Unadjusted for the eligible population, it’s just the average percents. However,
we are adjusting for eligible population using census estimates. Now we get a rate—cases per million, for instance, by year by state. We could average state by year rates or we could sum separately the numerators and the denominators and compute that rate.
I don’t have experience here. Can someone describe standard practices? Thanks, Gene Maguin 
Gene, Based on your description it sounds like option 2 will give you a weighted average of the annual rates, correcting for differences in the number of persons receiving services and the eligible population per year.
I would be concerned about simply averaging rates. Brian Brian Dates, M.A. From: SPSSX(r) Discussion [mailto:[hidden email]]
On Behalf Of Maguin, Eugene So, we have some data by year for the past 5 years on the number of people who received a service in a given state. We’d like to group and summarize the data. Unadjusted for the eligible population, it’s just the average percents. However,
we are adjusting for eligible population using census estimates. Now we get a rate—cases per million, for instance, by year by state. We could average state by year rates or we could sum separately the numerators and the denominators and compute that rate.
I don’t have experience here. Can someone describe standard practices? Thanks, Gene Maguin ===================== To manage your subscription to SPSSXL, send a message to
[hidden email] (not to SPSSXL), with no body text except the command. To leave the list, send the command SIGNOFF SPSSXL For a list of commands to manage subscriptions, send the command INFO REFCARD

In reply to this post by Maguin, Eugene
I don't speak as a public health administrator, but in good statistical presentations of any sort, I think you can use
averages of the state rates if they are /similar/  Averages are great for simple summaries of similar numbers;
they are trickier when they represent some sort of parameter instead of a shared, "central tendency."
Your computing or reporting is similarly confounded by similar versus different Ns: Especially when the means do
differ, an average of similarsized groups does not offer the scope of misrepresentation that is possible, in either direction, when combining groups with disparate Ns.  If a small group is extreme, it should not dominate the report, nor should it be buried.
 Rich Ulrich From: SPSSX(r) Discussion <[hidden email]> on behalf of Maguin, Eugene <[hidden email]>
Sent: Thursday, December 15, 2016 11:53:14 AM To: [hidden email] Subject: nonspss stats question So, we have some data by year for the past 5 years on the number of people who received a service in a given state. We’d like to group and summarize the data. Unadjusted for the eligible population, it’s just the average percents. However,
we are adjusting for eligible population using census estimates. Now we get a rate—cases per million, for instance, by year by state. We could average state by year rates or we could sum separately the numerators and the denominators and compute that rate.
I don’t have experience here. Can someone describe standard practices? Thanks, Gene Maguin 
Free forum by Nabble  Edit this page 