Computing p-values for svyolr objects

[gjones1219]

Hello! I have written on this issue before, but am coming back to it because I do not think it was addressed before. You responded to my comment previously, and I am appreciative and apologize for not responding until now, but I do not think the functionality I was looking for was added.

I am writing to see if it would be possible for you all to assess p-values for svyolr objects. This functionality seems to be missing from the package, and would be invaluable for significance testing for ordinal logistic regression models. Attaching reproducible code for your review:

#######

library(broom)
library(survey)

data(api)
dclus1<-svydesign(id=~dnum, weights=~pw, data=apiclus1, fpc=~fpc)
dclus1<-update(dclus1, mealcat=cut(meals,c(0,25,50,75,100)))

m<-svyolr(mealcat~avg.ed+mobility+stype, design=dclus1)
m

broom::tidy(m, conf.int = FALSE,
conf.level = 0.95,
exponentiate = FALSE,
p.values = TRUE)

#Usually provides an error message saying this doesn't run.

Thank you, and please let me know if I can further clarify this request!

Thanks,
gjones1219

[simonpcouch]

Could you please rework your reproducible example to use the reprex package? That makes it easier to see both the input and the output, formatted in such a way that I can easily re-run in a local session.

Related to #1061.

[gjones1219]

Sorry about that! Here you go!

library(broom)
library(survey)
#> Loading required package: grid
#> Loading required package: Matrix
#> Loading required package: survival
#>
#> Attaching package: 'survey'
#> The following object is masked from 'package:graphics':
#>
#> dotchart
library(reprex)

data(api)
dclus1<-svydesign(id=~dnum, weights=~pw, data=apiclus1, fpc=~fpc)
dclus1<-update(dclus1, mealcat=cut(meals,c(0,25,50,75,100)))

m<-svyolr(mealcat~avg.ed+mobility+stype, design=dclus1)
m
#> Call:
#> svyolr(mealcat ~ avg.ed + mobility + stype, design = dclus1)
#>
#> Coefficients:
#> avg.ed mobility stypeH stypeM
#> -2.6999217 0.0325042 -1.7574715 -0.6191463
#>
#> Intercepts:
#> (0,25]|(25,50] (25,50]|(50,75] (50,75]|(75,100]
#> -8.857919 -6.586464 -4.924938

broom::tidy(m, conf.int = FALSE,
conf.level = 0.95,
exponentiate = FALSE,
p.values = TRUE)
#> Error in UseMethod("extractAIC"): no applicable method for 'extractAIC' applied to an object of class "svyolr"
Created on 2022-06-08 by the reprex package (v2.0.1)

[simonpcouch]

Thanks! A slightly more minimal reprex:

library(broom)
library(survey)
#> Loading required package: grid
#> Loading required package: Matrix
#> Loading required package: survival
#> 
#> Attaching package: 'survey'
#> The following object is masked from 'package:graphics':
#> 
#>     dotchart

data(api)
dclus1 <- svydesign(id = ~dnum, weights = ~pw, data = apiclus1, fpc = ~fpc)
dclus1 <- update(dclus1, mealcat = cut(meals, c(0, 25, 50, 75, 100)))

m <- svyolr(mealcat ~ avg.ed + mobility + stype, design = dclus1)

tidy(m, p.values = TRUE)
#> Error in UseMethod("extractAIC"): no applicable method for 'extractAIC' applied to an object of class "svyolr"

^{Created on 2022-06-08 by the reprex package (v2.0.1)}

The tidy.svyolr() tidier is a light wrapper around tidy.polr(). However, the implementation for p-value calculation in tidy.polr() is both computationally intensive and specific to that model, so the p.values argument to tidy.svyolr() is now ignored and will raise a warning when passed.

I appreciate you noting this!

[gjones1219]

Hi Simon,

Thanks for responding however I am not sure I understand your answer. Also, I have the latest version of broom (0.8.0, I believe; please let me know if there might be a newer one) but do not see the tidy.svyolr() or tidy.poly() functions. Can you please advise here? Thanks!

Grant

[simonpcouch]

Sure thing—tidy.svyolr and tidy.poly are S3 methods. You can read about those methods with ?tidy.svyolr or ?tidy.poly, and learn more about them here. :)

[github-actions]

This issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with a reprex: https://reprex.tidyverse.org) and link to this issue.