# Demographic Estimates Across Geo-Units

**URL:** https://forum.ipums.org/t/demographic-estimates-across-geo-units/4162
**Category:** USA
**Created:** [June 2, 2021, 4:20pm UTC](https://forum.ipums.org/t/demographic-estimates-across-geo-units/4162 "2021-06-02T16:20:46Z")
**Posts on this page:** 1
**Showing post:** 4

<div class="post-metadata">

### Author: ![Cela](https://avatars.discourse-cdn.com/v4/letter/c/f19dbf/32.png) [@Cela](https://forum.ipums.org/u/Cela)
#### Post date: [June 10, 2021, 8:05pm UTC](https://forum.ipums.org/t/demographic-estimates-across-geo-units/4162/4 "2021-06-10T20:05:49Z")

</div>

Hi Matthew,

First off - thank you! I appreciate the sample code and additional reading.

Does IPUMS have suggestions on when to use person weights vs. replicate weights (or is that not even a valid comparison?).

For instance, I was using code from this IPUMS thread provided by an IPUMS Staff Member - [Does anyone have sample code for using svydesign function in R? - #2 by gfellis](https://forum.ipums.org/t/does-anyone-have-sample-code-for-using-svydesign-function-in-r/2444/2)

Example:

> svy ← svydesign(~CLUSTER, weights = ~PERWT, strata = ~STRATA, data = data, nest = TRUE, check.strata = FALSE)  
> svymean(~HISPAN, svy)

This also gives you standard errors - but it’s not clear to me when you want to use either method.

Thanks!

---

_[View the full topic](https://forum.ipums.org/t/demographic-estimates-across-geo-units/4162)._