Questions tagged [generalized-estimating-equations]

Stands for Generalized Estimating Equations which is an approach to estimating regression coefficients. GEE can be used on clustered / longitudinal data and has the attractive property that it provides consistent estimators of regression coefficients and unbiased inference even when the association structure within a cluster is mis-specified.

Filter by
Sorted by
Tagged with
1 vote
1 answer
18 views

Solving estimating equation using R

I am working with different types of data and comparing a variety of estimating equation approaches which share a multi-dimensional parameter $\beta$. Given a set of data $\boldsymbol{X}$, is there an ...
0 votes
0 answers
25 views

Are generalized estimating equations estimates affected by endogenous covariates (covariates correlated with model residuals)?

Are generalized estimating equations consistent or still okay with endogenous variables? The authors of GEE say here that their estimates remain consistent without any qualifications about endogeneity ...
2 votes
0 answers
32 views

separate models vs joint model

My goal is to estimate the association between children BMI and distance to the nearest fast food restaurants. The hypothesis is that children BMI increases with increasing proximity of fast food ...
0 votes
0 answers
13 views

Testing association in repeated measures with zero in contingency table

I have a small dataset of repeated measures data (each participant (id) has up to two samples taken) and would like to determine if there is an association between ...
  • 179
0 votes
0 answers
33 views

How to simulate data from the model $\eta_{i t}=\beta_0+\beta_1 t / 10$ for $Cov(\eta_{i})$, R= correlation of $Cov(\eta_i)$ $i$ is the subject? [closed]

We assume that each $X_i=\left(x_{i 1}, \ldots, x_{i, 10}\right)^{\prime}$ is generated from a distribution with mean $(0 \cdot 1,0 \cdot 2, \ldots, 1 \cdot 0)^{\prime}$ and finite covariance. $\eta_{...
0 votes
0 answers
21 views

If $(y\theta-a(\theta)+b(y))/\phi$ is the log partition-function, then what is the base-measure, sufficient statistic, and natural parameter?

The log partition function for an observation is $$\log f= (y\theta-a(\theta)+b(y))/\phi$$ Differentiate with respect to $\theta$ to get $$(y - a'(\theta))/\phi.$$ Taking expectations $$E[Y]-a'(\theta)...
1 vote
1 answer
32 views

Sample size calculation for a difference between two groups with clustered data

I want to compare proportions or means in two groups in a clustered sample. Suppose the sampled clusters are schools or hospitals or Census areas, and we sample from each of these an equal number in ...
  • 246
-1 votes
1 answer
52 views

Why does $E\left(y_{i t}\right)=a^{\prime}\left(\theta_{i t}\right)$? in the context of assuming some GEE marginal density?

In generalized estimating equations we have a glm-response variable. To establish notation, we let $Y_{i}=\left(y_{i 1}, \ldots, y_{i n_{i}}\right)^{\text {T }}$ be the $n_{i} \times 1$ vector of ...
0 votes
1 answer
30 views

Why is "design matrix of correlation parameters" a proxy for the "actual covariance matrix/working correlation matrix?

The example shows that knowing the design matrix of correlation parameters is sufficient to specify the working correlation. ...
0 votes
1 answer
30 views

Are the model residuals from GEE interpretable as residuals from simple linear regression?

Are the model residuals from GEE interpretable as residuals from simple linear regression, so that I may plot a residual versus fitted plot to determine whether there's heteroskedasticity? It is known ...
0 votes
0 answers
24 views

Fixed-effects regression with variable number of binary responses but only one success

I believe my problem requires the use of generalized estimating equations but there are a couple of characteristics of the data that make me question the suitability. The following data is ...
0 votes
0 answers
25 views

How to obtain "normalized" model residuals for Generalized Estimating Equations?

How does one compute the "normalized" model residuals based via geepack's geeglm/gee in R? The nlme package in R allows one to compute the normalized model residuals: (standardized ...
2 votes
0 answers
35 views

Does the sandwich-estimate eliminate heteroskedasticity in a model residuals versus fitted plot, or simply make the estimation robust to heteroskedas? [closed]

Does the sandwich-estimator/Huber-White/GEE eliminate heteroskedasticity in a residuals versus fitted plot, or simply make the model robust to it?
1 vote
0 answers
14 views

Why are fisher-scoring estimates of the fixed-effects not used to calculate empirical bayes estimates of random-effects? Are they in-admissible?

Why is it not practiced using estimates of fixed-effects from fisher scoring used to calculate GEE coefficients to estimate random-effects via empirical bayes? We have another estimate of $\theta$ in ...
1 vote
0 answers
84 views

Why GEE and GLM estimate the different standard errors?

I use the independence working correlation, the parameter estimates β in glm and GEE is identical,but the standard errors is different。 The standard error of both GEE and GLM are calculated using the ...
  • 11
2 votes
1 answer
45 views

Is the odds ratio not a measure of effect size?

I have a question regarding the odds ratio. A paper that I submitted was sent back for revision after peer review. However, one of the authors asked a question that has left me quite confused. In this ...
  • 83
0 votes
0 answers
25 views

Comparing treatment effect using mixed effect model/gee/offset? Which one is better?

I think I have seen various posts indicating that comparison of treatment effect in observation studies should employ either longitudinal mixed effect model or GEE. Suppose each individual is given ...
  • 1,149
3 votes
1 answer
60 views

Calculation of degrees of freedom for B-splines

I am confused about how the degrees of freedom in a B-spline are calculated in the package splines. The documentation for the B-spline function can be found here: https://www.rdocumentation.org/...
1 vote
1 answer
57 views

Calculate a permutation-based p-value for a risk ratio

Summary How do you calculate a two-sided p-value for a risk ratio/relative risk (obtained from a GEE logistic regression via predicting risks with and without a treatment) based on a permutation test, ...
0 votes
0 answers
28 views

GEE for continuous outcome, continuous covariates

I'm curious about using GEE to analyze the relations between a single continuous outcome variable measured pre- and post-, and three continuous covariates also measured pre- and post-. I have no ...
1 vote
1 answer
50 views

GEE correlation structure's number of parameters

I am kind of confused on GEE correlation structure's number of parameters. Say I have 10 students(or clusters) and I measure their physical strength 10 times for each of them with their corresponding ...
  • 1,149
2 votes
0 answers
36 views

Generalized estimating equation (GEE) for multilevel data

Can Generalized estimating equation (GEE) handle multilevel data? and how?
0 votes
0 answers
16 views

How to do generalized estimating equations with zero inflated poisson regression in R?

I did not find a package to do zero inflated Poisson/negative binomial regression with generalized estimating equations. Is there such a package available? How to do generalized estimating equations ...
  • 1,149
1 vote
0 answers
110 views

Serious Coding Error in QIC function in geepack?

I believe the QIC function in geepack has a significant error. The function appears to incorrectly specify the independence model, which is needed to calculate QIC. The function will therefore often ...
3 votes
1 answer
54 views

Is it possible that GEE and mixed effect GLM give contradicting answers? If so, which one should be trusted?

Is it possible that GEE and mixed effect GLM give contradictive answers in significance of covariates? I assume both GEE and GLM selects same covariates. If so, which one should be trusted? From ...
  • 1,149
2 votes
1 answer
138 views

Relationship between log-linear (e.g. Poisson) GLM and GLMM with random intercepts

In their textbook "Analysis of Longitudinal data" Diggle et al. state (p. 137 in the second edition) that: In log-linear models for counted data, random effects and marginal parameters can ...
  • 23
0 votes
1 answer
68 views

AIC and BIC in GEE

Can I apply Akaike’s information criterion (AIC) and Quasi-likelihood under the independence model information criterion (QIC) in GEE?
0 votes
0 answers
17 views

If I analyse the pre-post data, should the "pre" (time=0) be included or excluded from the reponse?

Let's assume I analyse repeated data, recorded at t0, t1...t3. I want to analyse the response itself and then check various contrasts, for example change from baseline or consecutive. If the model is: ...
0 votes
0 answers
15 views

Empirical risk minimization for relu/max loss function

Classical risk minimization (RM) minimizes the expected loss over the training distribution $p_{\mathrm{train}}(x)$, $$\theta^*_{RM} = \arg \min_\theta E_{p_{\text{train}}}[\ell(x, \theta)].$$ As the ...
  • 101
1 vote
0 answers
28 views

If the normality assumption in the for the GLS estimation fails, would you switch to GEE?

I want a marginal model, ideally fit via GLS. But the normality of residuals doesn't hold. It isn't much skewed, I don't want any transformations. It's just non-normal in shape. Yet still reporting ...
0 votes
0 answers
64 views

Why is the baseline x time interaction included in clinical trials models (MMRM)?

I noticed, that many, many times, when the longitudinal studies are analysed for efficacy, the following components are taken as fixed effects: time treatment time : treatment (interaction) baseline ...
0 votes
0 answers
14 views

Is the Generalized Estimating Equtions method a good non-parametric replacement for the Generalized Least Squares?

I want to use GLS on my longitudinal data, but it turns out that residuals are non-normal and is a non-easy way. Not just "transformable" skewness, no known relatioship between mean and ...
1 vote
0 answers
38 views

In a typical clinical longitudinal trial, why do marginal models (MMRM, GEE) are much more common than conditional ones (mixed)?

I read statistical analysis plans and reports of over 80 longitudinal trials. I noticed, that in 70% they were analysed using a marginal model, namely the MMRM approach (mixed-model repeated measures)....
0 votes
0 answers
49 views

Will I be likely criticized, if use the independence covariance structure for the GEE method?

It can be read, that the GEE method is robust to covariance structure misspecification. One can choose even the... independence structure for repeated observations, and still it may work. I understand,...
  • 51
2 votes
0 answers
53 views

Why do some poeple claim, that choice of the working correlation in GEE doesn't affect the marginal coefficients?

I found this discussion: GEE: choosing proper working correlation structure Cite: Correlation structure in GEE, unlike mixed models, does not affect the marginal parameter estimates (which you are ...
  • 51
1 vote
0 answers
32 views

Is GLS really a special case of the GEE?

I was told, that the GLS is a special case of the GEE, if the conditional distribution is gaussian and the link is identity. How is that possible? GLS is a two (or more - IWLS) stage procedure. It ...
  • 51
0 votes
1 answer
47 views

What is the purpose to have the "independent" covariance structure in GEE or GLS?

The methods of estimation like GLS or GEE are especially helpful, when there are clusters of data, like repeated observations, many per cluster=subject. Such observations are naturally correlated in ...
  • 51
0 votes
0 answers
13 views

Why is anova not significant when analysis of the str error multiple regression model suggests that it is? [duplicate]

I'm a bit confused as the output of my model in R. I have built a generalised estimating equasion glm model aiming to see the effect of time (here coded as timestrat) on a variable called new1804. I ...
1 vote
0 answers
47 views

Does the GEE estimation and mixed models handle data missing inside a series?

I'm wondering, if either GEE or mixed models can handle all the 4 cases of missing data: T0 T1 T2 T3 m + + + (missing at t0, late entry) m + + (missing inside a series, here at t1) m (missing ...
  • 51
0 votes
0 answers
24 views

If "conditional to random effect" in mixed models a kind of averaging or calculated differently?

I read practically every discussion in this forum, and still I don't get it. I'm sorry, but all the explanations still don't tell me how to interpret it. Please don't cite other articles, as I saw all ...
  • 51
0 votes
0 answers
72 views

Why is it said, that GEE is not likelihood based, IF the estimating equations are derived exactly thru d(log(L)/dB?

I found this article: https://sakai.unc.edu/access/content/group/2842013b-58f5-4453-aa8d-3e01bacbfc3d/public/Ecol562_Spring2012/docs/lectures/lecture22.htm#choosing and most of it is showing the ...
  • 51
1 vote
0 answers
58 views

How is that possible that SAS and R can test for main and interaction effects for the GEE if it has no likelihood?

I was taught, that GEE, being not likelihood based, has no way to compare models. That we cannot assess the main and interaction effects the way we do with ordinary GLM, OLS, GLS, mixed models and so ...
  • 51
0 votes
0 answers
77 views

correlation matrix from geeglm function

I'm using geepack::geeglm() to fit a gee model. I can't figure out how to get the correlation matrix from the output though it does not show up when ...
3 votes
1 answer
67 views

Show Cox regression estimating equation is unbiased

Let $N(x) = I(X \leq x, \Delta=1)$ be the counting process for observed failure events, where $X = \min\{T,C\}$ and $\Delta = I(T \leq C)$, for censoring time $C$ and failure time $T$. Assume that $T$ ...
0 votes
0 answers
23 views

Which test or method can I use to assess trend in dependent proportions over time?

Let's assume there are 4 time points, t0 (baseline) ... t3 At each time point I assess a % of successes of some outcome: p0, p1...p3. The denominators vary over time, so the counts itself do not ...
0 votes
0 answers
12 views

the coefficient of a variable is the average of the coefficients of its two parts in generalised estimating equation models

I used the z-score of each variable in generalized estimating equation models. one independent variable A was divided into two parts, and the z-score of each part was taken as an independent variable. ...
  • 143
0 votes
1 answer
68 views

GEE - ROC Curves the Same Using Different Cluster Variables

I am creating multiple GEEs with the same covariates, but I am testing different clustering variables. The outcome is a binary yes/no variable and both VAR1 and VAR2 are binary, as well. Patients can ...
0 votes
0 answers
38 views

GEE Explained Variance is Negative

I am creating a series of GEE models and a couple of statistics I normally report are the ICC and the explained variance for both Level 1 and Level 2 of the model. However, I am calculating that for ...
1 vote
1 answer
30 views

Multivariate Regression with Two Different Types of Response

Problem Setting: I have an interesting question related with longitudinal study and multivariate regression. I found that in lots of biomedical studies, multiple discrete and continuous endpoints are ...
  • 21
0 votes
0 answers
53 views

Inferring effect & effect modification from simulation data

I have a "black box" system (computer simulation), which takes inputs: $x_1 \in [0,1]$, $x_2 \in [0,1]$, and $N_i$ others $\vec\theta = \{\theta_1, \dots, \theta_{N_i}\}$, and produces an ...

1
2 3 4 5
7