Number of Clients at 2004 vs 2012 Significant test

HI Guys,

I have a data-set of health centers. I have a variable that counts the number of clients the center served. I want to see if the centers increase significantly the number of clients served from 2004 to 2012. I have the same agencies for all the years. What would be the correct test to perform. Also should my data be in long or wide format?

Thank you,


Less is more. Stay pure. Stay poor.
Most likely this sounds like a case of wide data where you subtract the 2004 number from 2012 numbers to get the differences in the number of patients between the time periods. Then you would run a one sample ttest comparing the differences to a change of 0 patients, given the tests assumptions are met.
Thank you very much.. just to make it more complicate, I would love to test this. I have also a funding variable: how much money each agency received for each year. So if I want to test the same hypothesis but controlling for "funding (in dollars)", what should I do? Thank you in advance.


Less is more. Stay pure. Stay poor.
You may be able to run a simple linear regression model with funds predicting difference. I believe if the intercept is significant that means the difference is not equal to zero (representing a change). Now you can look at the generated ttest in the model to see if funds are a significant predictor.

However I don't think this approach may control for funds when interpreting the differences. We can see if anyone else chimes in to help out.
Just to let you know. I have two groups my controls and the agencies pf interested. I was trying to read and do some research in Repeated Measure ANOVA but it seems that there are few information. I posted a couple of times questions related to ANOVA and nobody answer. I guess it is complicated.


TS Contributor
Well, meanwhile you've started about 50 threads or so, asking for help?
Perhaps doing academical and/or professional work by using an online
forum so frequently isn't the optimal solution. There are consultants out

Just my tuppence.
I have data for agencies funded consistently from 2003 to 2012. I have 4 agencies in Massachusetts (agencies of interests) and 40 controls (other agencies). My data is in long format. I have 4 variables:

Year= 2204 to 2012
Total_Clients= ##
MA= controls or Massachusetts agencies (1 or 0)
Funding: How money the agency received

I know I have to conduct repeated measure ANOVA but I don’t know how to do it in Stata, especially if I have two control for other factors. The first thing I want to do is:

1. If there was a significant increase in number of clients served from 2004 vs 2012
1. To see if more funding increase the number of clients
2. To see if Massachusetts agencies increase from 2004 to 2012 the number of clients more than the control controlling for funding.

I am attaching a subset of my dataset. I know this can be complicated but I would appreciate any comments.
Thank you in advance.