Conversation
idblr
added
enhancement
|
Thanks, @huntermills707. To confirm, you compared population-weighted NDIQuint variables derived from In the meantime, I've changed the base branch to a new dev branch |
idblr
added
question
Yes! I used the August 2022 updated values from NHLBI ADOPT. |
|
@huntermills707 PR accepted and merged into |
* 🔨 Prepare for PR (#32) * `Hmisc` now Imports * 🐛 Fix to `powell_wiley()` categorical outcome (PR #32) * Updated population-weighted quantile method from `stats::quantile` to `Hmisc::wtd.quantile` * Updated import statement in `powell_wiley()` * 📝 Additional documentation for PR #32 * Prepare for CRAN v0.2.0 release * 🐛 Fixed NOTE in CRAN checks to provide package anchors for Rd `\link{}` targets not in the package itself and the base packages within 'ndi-package.Rd' * 📝 In 'cran-comments.md': added valid URLs, updated test environments, and OK'd author formatting * 📛 Updated DOI badge * 📝 Updated vignettes with latest package version --------- Co-authored-by: Hunter Mills <huntermills707@gmail.com>
* 🔨 Prepare for PR (#32) * `Hmisc` now Imports * 🐛 Fix to `powell_wiley()` categorical outcome (PR #32) * Updated population-weighted quantile method from `stats::quantile` to `Hmisc::wtd.quantile` * Updated import statement in `powell_wiley()` * 📝 Additional documentation for PR #32 * Prepare for CRAN v0.2.0 release * 🐛 Fixed NOTE in CRAN checks to provide package anchors for Rd `\link{}` targets not in the package itself and the base packages within 'ndi-package.Rd' * 📝 In 'cran-comments.md': added valid URLs, updated test environments, and OK'd author formatting * 📛 Updated DOI badge * 📝 Updated vignettes with latest package version * Prepare for CRAN v0.2.1 release * 🔗 Fixed broken URLs in 'theil.Rd', 'ndi-package.Rd', 'ndi2.html', README, and NEWS --------- Co-authored-by: Hunter Mills <huntermills707@gmail.com>
* 🔨 Prepare for PR (#32) * `Hmisc` now Imports * 🐛 Fix to `powell_wiley()` categorical outcome (PR #32) * Updated population-weighted quantile method from `stats::quantile` to `Hmisc::wtd.quantile` * Updated import statement in `powell_wiley()` * 📝 Additional documentation for PR #32 * Prepare for CRAN v0.2.0 release * 🐛 Fixed NOTE in CRAN checks to provide package anchors for Rd `\link{}` targets not in the package itself and the base packages within 'ndi-package.Rd' * 📝 In 'cran-comments.md': added valid URLs, updated test environments, and OK'd author formatting * 📛 Updated DOI badge * 📝 Updated vignettes with latest package version * Prepare for CRAN v0.2.1 release * 🔗 Fixed broken URLs in 'theil.Rd', 'ndi-package.Rd', 'ndi2.html', README, and NEWS * Additional preparation for CRAN v0.2.1 release * 🔗 Fixed addition broken URLs in 'ndi2.html' --------- Co-authored-by: Hunter Mills <huntermills707@gmail.com>
2 participants
Andrews et al. (2020) and Slotman et al. (2022) were not very descriptive about the method used to "Population weighted NDI quintiles."
As currently implemented, Quintile cutoffs are determined by the product of the calculated NDI, and the logarithmic transform of the Total Population.
This does yield similar quintile results as the NDIQuint provided by Slotman et al. (2022) when rederiving NDIQuint using the provided NDI values joined with ACS5 Total_Population (B01001_001), but there are some discrepancies. See below:
Another definition would be to use a weighted quantile function, were the values are NDI, and weights are ACS5 Total_Population. I.E. distribute the NDI to each person in the tract and determine cutoffs for the entire population. Below is the comparison of this method and NDIQuint provided by Slotman et al. (2022):
This is a relatively minor change, feel free to disagree/reject. I am a python developer, and am unable to fully test the build (Ex roxygen2). I also did not see a appropriate dev branch to fork from.