Dtjytyk

Drawing ramified coverings with tikz

Some numbers are more equivalent than others

List of people who lose a child in תנ"ך

Reply 'no position' while the job posting is still there

Flux received by a negative charge

A Permanent Norse Presence in America

Should I stop contributing to retirement accounts?

Remove Expired Scratch Orgs From VSCode

How can Trident be so inexpensive? Will it orbit Triton or just do a (slow) flyby?

Engineer refusing to file/disclose patents

Do the concepts of IP address and network interface not belong to the same layer?

Bob has never been a M before

Does the Mind Blank spell prevent the target from being frightened?

Did US corporations pay demonstrators in the German demonstrations against article 13?

Can someone explain how this makes sense electrically?

If a character with the Alert feat rolls a crit fail on their Perception check, are they surprised?

Greatest common substring

Why did the HMS Bounty go back to a time when whales are already rare?

How should I respond when I lied about my education and the company finds out through background check?

What's the difference between 違法 and 不法?

Could solar power be utilized and substitute coal in the 19th Century

Gibbs free energy in standard state vs. equilibrium

Translation of Scottish 16th century church stained glass

What major Native American tribes were around Santa Fe during the late 1850s?

Rank ANCOVA is a bad idea?

Robust ANCOVA literatureWhen would least squares be a bad idea?Is skewness always bad?How to choose between t-test or non-parametric test e.g. Wilcoxon in small samplesRank deficient bootstrap resamplesWilcoxon signed-rank testUnderlying Idea behing Non Parametric TestSurvival analysis with multiple factorsWhat non parametric test? multiple dependent variables, between subject variable, and a within subject variableAlgebraic Manipulations in Mann-Whitney-Wilcoxon Test Statistics

3

1

$begingroup$

I am currently working on a project looking at the best way to handle missing data (such as quality of life) due to death. One approach is to use categories/ranks for the outcome, where death is the worst category. One approach that my colleagues wish to consider is rank ANCOVA. However I am pleased to say that they also wish to consider proportional odds models and continuation ratio models.

I found an R help page that says “This [rank ANCOVA] has been shown to yield unreliable analyses. Use the more formal proportional odds ordinal logistic model. This is a generalization of the Wilcoxon-Mann-Whitney-Kruskal-Wallis statistic. This is implemented in the rms package and elsewhere”.

My question is this: Could anyone tell me where this has been shown? In particular the part about rank ancova being unreliable?

regression nonparametric

share|cite|improve this question

asked Mar 14 at 9:48

Dan JacksonDan Jackson

161

$endgroup$

add a comment | 

3

1

$begingroup$

I am currently working on a project looking at the best way to handle missing data (such as quality of life) due to death. One approach is to use categories/ranks for the outcome, where death is the worst category. One approach that my colleagues wish to consider is rank ANCOVA. However I am pleased to say that they also wish to consider proportional odds models and continuation ratio models.

I found an R help page that says “This [rank ANCOVA] has been shown to yield unreliable analyses. Use the more formal proportional odds ordinal logistic model. This is a generalization of the Wilcoxon-Mann-Whitney-Kruskal-Wallis statistic. This is implemented in the rms package and elsewhere”.

My question is this: Could anyone tell me where this has been shown? In particular the part about rank ancova being unreliable?

regression nonparametric

share|cite|improve this question

asked Mar 14 at 9:48

Dan JacksonDan Jackson

161

$endgroup$

add a comment | 

$begingroup$

I am currently working on a project looking at the best way to handle missing data (such as quality of life) due to death. One approach is to use categories/ranks for the outcome, where death is the worst category. One approach that my colleagues wish to consider is rank ANCOVA. However I am pleased to say that they also wish to consider proportional odds models and continuation ratio models.

I found an R help page that says “This [rank ANCOVA] has been shown to yield unreliable analyses. Use the more formal proportional odds ordinal logistic model. This is a generalization of the Wilcoxon-Mann-Whitney-Kruskal-Wallis statistic. This is implemented in the rms package and elsewhere”.

My question is this: Could anyone tell me where this has been shown? In particular the part about rank ancova being unreliable?

regression nonparametric

share|cite|improve this question

asked Mar 14 at 9:48

Dan JacksonDan Jackson

161

$endgroup$

share|cite|improve this question

asked Mar 14 at 9:48

Dan JacksonDan Jackson

161

share|cite|improve this question

asked Mar 14 at 9:48

Dan JacksonDan Jackson

161

asked Mar 14 at 9:48

Dan JacksonDan Jackson

161

asked Mar 14 at 9:48

Dan JacksonDan Jackson

161

1 Answer
1

active

oldest

votes

4

$begingroup$

This paper may be a good one demonstrating problems with rank transform anova: https://www.tandfonline.com/doi/abs/10.1080/01621459.1995.10476644
https://www.jstor.org/stable/2291530?seq=1#metadata_info_tab_contents

Just transforming data with ranks and using ordinary regression/anova/ancova, in cases other than the k-sample problem, i.e., attempting to use this when adjusting for covariates or assessing interactions, is not fully based on good statistical principles, and results are hard to interpret.

share|cite|improve this answer

answered Mar 14 at 11:23

Frank HarrellFrank Harrell

55.6k3108244

$endgroup$

add a comment | 

Your Answer

StackExchange.ifUsing("editor", function () {
return StackExchange.using("mathjaxEditing", function () {
StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix) {
StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
});
});
}, "mathjax-editing");

StackExchange.ready(function() {
var channelOptions = {
tags: "".split(" "),
id: "65"
};
initTagRenderer("".split(" "), "".split(" "), channelOptions);

StackExchange.using("externalEditor", function() {
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled) {
StackExchange.using("snippets", function() {
createEditor();
});
}
else {
createEditor();
}
});

function createEditor() {
StackExchange.prepareEditor({
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: false,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: null,
bindNavPrevention: true,
postfix: "",
imageUploader: {
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
},
onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
});


}
});

draft saved

draft discarded

Sign up or log in

StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});

Sign up using Google

Sign up using Facebook

Sign up using Email and Password

Post as a guest

Name

Email

Required, but never shown

StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstats.stackexchange.com%2fquestions%2f397462%2frank-ancova-is-a-bad-idea%23new-answer', 'question_page');
}
);

Post as a guest

Name

Email

Required, but never shown

4

$begingroup$

This paper may be a good one demonstrating problems with rank transform anova: https://www.tandfonline.com/doi/abs/10.1080/01621459.1995.10476644
https://www.jstor.org/stable/2291530?seq=1#metadata_info_tab_contents

Just transforming data with ranks and using ordinary regression/anova/ancova, in cases other than the k-sample problem, i.e., attempting to use this when adjusting for covariates or assessing interactions, is not fully based on good statistical principles, and results are hard to interpret.

share|cite|improve this answer

answered Mar 14 at 11:23

Frank HarrellFrank Harrell

55.6k3108244

$endgroup$

add a comment | 

4

$begingroup$

This paper may be a good one demonstrating problems with rank transform anova: https://www.tandfonline.com/doi/abs/10.1080/01621459.1995.10476644
https://www.jstor.org/stable/2291530?seq=1#metadata_info_tab_contents

Just transforming data with ranks and using ordinary regression/anova/ancova, in cases other than the k-sample problem, i.e., attempting to use this when adjusting for covariates or assessing interactions, is not fully based on good statistical principles, and results are hard to interpret.

share|cite|improve this answer

answered Mar 14 at 11:23

Frank HarrellFrank Harrell

55.6k3108244

$endgroup$

add a comment | 

$begingroup$

This paper may be a good one demonstrating problems with rank transform anova: https://www.tandfonline.com/doi/abs/10.1080/01621459.1995.10476644
https://www.jstor.org/stable/2291530?seq=1#metadata_info_tab_contents

Just transforming data with ranks and using ordinary regression/anova/ancova, in cases other than the k-sample problem, i.e., attempting to use this when adjusting for covariates or assessing interactions, is not fully based on good statistical principles, and results are hard to interpret.

share|cite|improve this answer

answered Mar 14 at 11:23

Frank HarrellFrank Harrell

55.6k3108244

$endgroup$

share|cite|improve this answer

answered Mar 14 at 11:23

Frank HarrellFrank Harrell

55.6k3108244

answered Mar 14 at 11:23

Frank HarrellFrank Harrell

55.6k3108244

answered Mar 14 at 11:23

Frank HarrellFrank Harrell

55.6k3108244