Measuring between and within group variance.

[Sonnen Blume]

I just came across an interesting analysis that reports between and within-group variance (of healthcare costs between hospitals). I've data that could do similar things, so I'd like to know if this doable in Stata 16.

With the hierarchical data structure
(births clustered within hospitals), each model partitioned
the total variance of cost into two components: ‘between-
hospital’ variance and ‘within-hospital’ variance.

https://obgyn.onlinelibrary.wiley.co...471-0528.15007

Thank you for sharing your comments.

[Clyde Schechter]

Yes, you can do that with the -mixed- command.

mixed outcome_variable || hospital:

applied to a data set where each outcome is a birth, and outcome_variable is replaced by whatever variable you want to partition the variance of. At the end of the regression output table you will see a variance component for the hospital level, and a residual level variance component. The residual level component is the within-hospital variance, and the hospital-level one is the between hospital variance.

[Sonnen Blume]

Yes, you can do that with the -mixed- command.

mixed outcome_variable || hospital:

applied to a data set where each outcome is a birth, and outcome_variable is replaced by whatever variable you want to partition the variance of. At the end of the regression output table you will see a variance component for the hospital level, and a residual level variance component. The residual level component is the within-hospital variance, and the hospital-level one is the between hospital variance.

Thanks so much Clyde! I'll do an analysis and will give a feedback here soon...