Skip to main content

Questions tagged [quasi-likelihood]

In GLMs, quasi-likelihood estimation is a way to allow over- or under-dispersion by choosing an appropriate variance function.

Filter by
Sorted by
Tagged with
3 votes
1 answer
110 views

Validity of AIC When Comparing Models with Varying Dispersion Parameters

I'm currently making a binomial model with a logit link, which is parameterised as a quasibinomial since I'm allowing it to calculate the dispersion parameter. I was wondering, since changes to the ...
Daniel's user avatar
  • 33
7 votes
2 answers
246 views

How much dispersion is too much for quasipoisson regression?

Quasipoisson regression goes beyond standard poisson regression in taking into account overdispersion (whereby the dependent variable's variance is much greater than its mean). This is explained at ...
rolando2's user avatar
  • 12.8k
1 vote
0 answers
21 views

Calculate weight for GLM-quasi poisson model

I am running several models with the quasi-Poisson family. I am looking at data from vulture restaurants. Vulture count was modelled at each site as a function of either a linear or quadratic effect ...
Emeline AUDA's user avatar
0 votes
0 answers
16 views

How to derive standard errors of regressors in quasi-Poisson regression?

Suppose I want to relax a Poisson Regression to allow for overdispersion and apply a quasi-likelihood approach: $$E[y_i|x_i] = exp(x_i^T \beta)$$ $$Var[y_i|x_i] = \phi \cdot \mu_i$$ In other words, ...
Marlon Brando's user avatar
0 votes
0 answers
21 views

Diagnostics and assumptions for Fractional Logistic Regression

I am investigating the effects of Variable X on Y'. Y' is a bounded, non-negative integer. So I have divided Y' by its upper bound to obtain a fraction Y which is in [0,1]. I am following a study ...
NutellaMonster's user avatar
6 votes
3 answers
998 views

The statistical model equivalent to this R formula

This is how my model is written in R: glm(formula = prop ~ A * B * C * D , family = quasibinomial, data = data, weights = w) This is a quasibinomial ...
ie86's user avatar
  • 355
0 votes
0 answers
9 views

Overdispersion Mixed generalized linear model

I am running a mixed generalized linear model to analyze insect capture in a baited trap. The experiment consisted of 3 separate cages, in each one one treatment (C+, C- or T) and 10 insects were ...
Andrea88's user avatar
0 votes
0 answers
48 views

Given that quasibinomial/quasipoisson reg models overdispersion, why ever use normal bin/poiss regression if quasi is more flexible?

In reading about quasibinomial regression: The quasi-binomial distribution, while similar to the binomial distribution, has an extra parameter 𝜙 (limited to |𝜙|≤min{𝑝/𝑛,(1−𝑝)/𝑛} ) that attempts ...
JElder's user avatar
  • 1,037
0 votes
0 answers
24 views

Coefficients in quasibinomial regression and model prediction plots

I'm using two quasibinomial models. In the first model, the dependent variable is the proportion x of successes in experiment A. In the second model, the dependent variable is the proportion y of ...
statuser's user avatar
0 votes
0 answers
37 views

Profile (quasi-)likelihood score tests

Suppose I have a log-likelihood or quasi-log-likelihood for my data in terms of the parameter vectors $\theta$ and $\psi$: $$L(\theta;\psi)=\frac{1}{T}\sum_{t=1}^T{\log{f(y_t|\theta;\psi)}}.$$ (I am ...
cfp's user avatar
  • 535
1 vote
0 answers
74 views

Can I use quasi-binomial regression on proportion data in this way?

prop.pass = proportion of students who passed the exam num = number of students who sat the exam ...
Jess's user avatar
  • 21
0 votes
0 answers
67 views

Poisson fixed effects model in pglm estimating time-invariant predictor

I am analyzing panel data on various municipalities (id_mun) over several years using the pglm package in R. My dataset contains a variable "treatment" which is continuous but is time-...
Mark F's user avatar
  • 1
0 votes
0 answers
71 views

Quasibinomial GLMM with LASSO regularization in R

I am currently assessing drivers of deforestation using a GLM (generalised linear model) with LASSO regularization (using package glmnet in R). As the response variable is % of area deforested I have ...
user392406's user avatar
2 votes
0 answers
27 views

Does binomial regression weight some observations more heavily than other?

I am performing a quasibinomial regression, where each subject has an unfixed number of trials. So one subject may have had 5 trials while another had 90. In R the regression equation follows: glm(...
Matt's user avatar
  • 91
0 votes
0 answers
139 views

Computing log-posterior for large variance priors

Let's say that some quantity is modelled by a time-dependent Poisson distribution, $$ y(t) \sim \text{Pois}(\mu(t)) $$ where $$ \mu(t) = \alpha_0 \exp(-\alpha_1 e^{-\alpha_2 t}) $$ and $\alpha_k > ...
Mathieu Rousseau's user avatar
0 votes
0 answers
62 views

Approximating profile likelihood confidence intervals when I only have a score function and not a likelihood

I'm working on a modeling problem where I can define a score function that looks a lot like a binomial likelihood, but the model isn't really binomial. I'd like to use profile likelihood to estimate ...
famulare's user avatar
  • 362
2 votes
1 answer
66 views

Likelihood ratio test for model specification with boundary Null

I am interested in understanding the asymptotic distribution of Likelihood ratio (LR) test statistic for model specification. I am focusing on the case in which the null hypothesis is of the form (i.e....
Alfonso's user avatar
  • 21
2 votes
3 answers
775 views

Poisson model for non-integer

I have a GLM with (quasi)poisson family. My dataset has 3 variables: rate_data rate_benchmark X So fitting the model: ...
Overkill123's user avatar
0 votes
1 answer
54 views

Why does using proportion DV (vs. count) change the results?

I'm running a regression model using a count variable as the DV and I'm trying to understand what changes when using it as a count outcome vs. a proportion. REF: https://ademos.people.uic.edu/...
llbia's user avatar
  • 361
2 votes
1 answer
45 views

Can I apply quasi-Poisson regression on discontinuous data?

I'm trying to relate the concentrations of certain air pollutants to the incidence of a disease, but I don't have continuous data on that pollution. I have daily data on cases and deaths from the ...
user379040's user avatar
1 vote
0 answers
242 views

Interpretation of p-values in glm() and sjPlot::tab_model() for quasi-poisson regression

I am running a quasi-poisson regression predicting a count outcome from 4 variables of interest. I understand that these estimates need to be exponentiated to correctly interpret the results. I used ...
stilesb's user avatar
  • 31
2 votes
1 answer
158 views

Comparing two quasipoisson glm models

Have these results but not sure how to interpret them to pick the best fitting model, I think the high p value suggests I should go with model 1? Also in terms of the order that I put the models in, ...
user avatar
1 vote
1 answer
119 views

Bnomial (Logit) regression for proportion/percentage data

I have run a binomial (logit) regression on some proportion data as the dependent variable in an Interrupted Time Seies Analysis [see below]: ...
j.rahilly's user avatar
2 votes
0 answers
333 views

What are the assumptions of fractional response regression model?

I employed fractional response regression proposed by Wooldridge and Papke for my study as the dependent variable is a proportion (remaining/whole) and is between 0 to 1 (including 0 and 1). I want to ...
Laiy's user avatar
  • 245
2 votes
0 answers
55 views

Is Quasi-Likelihood As "Strong" As The Standard Likelihood?

I have been trying to learn more about the differences between Quasi Likelihood compared to the standard Likelihood. When learning about this, the following points come to mind: The Quasi-Likelihood ...
stats_noob's user avatar
1 vote
1 answer
212 views

WHy is the over dispersion in this poisson and quasi-poisson the same?

I have a zero inflated count data, on which I have run a poisson and quasi poisson reg using glm(). The output from a poisson model is as follows: ...
Rabin KC's user avatar
1 vote
1 answer
64 views

Terminology question regarding a certain "partial maximum likelihood" which approximates the marginal likelihood

Suppose that we have a model with many parameters, which we'll partition into two subvectors called $\theta$ and $\lambda$. In this situation, $\lambda$ corresponds to those parameters that are really ...
Mike Battaglia's user avatar
2 votes
0 answers
597 views

Fitting a regression model on bounded data

Typically, in a regression equation, the dependent variable is unbounded i.e. follows normal distribution. But sometimes it may happen that dependent variable is bounded i.e. dependent variable is ...
Bogaso's user avatar
  • 871
1 vote
0 answers
32 views

How to interprete the outcome for quasi-binomial models

I am currently trying to understand the quasi-binomial model and have a rather basic question, which I seem not to be able to see an answer. The outcomes in binomial models are binary. So, 0 vs 1, car ...
RGG's user avatar
  • 73
2 votes
0 answers
407 views

Serious Coding Error in QIC function in geepack? [closed]

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 ...
Brian M.'s user avatar
0 votes
1 answer
98 views

quasi-likelihood estimates of beta

working through Peter McCullagh's glm book and having a hard time with understanding quasi-likelihood. I'm working on this question below and I think I need to find the quasi-likelihood estimates and ...
neuroandstats's user avatar
1 vote
1 answer
212 views

quasi likelihood for ungrouped binary data

I read in one of the textbooks that for ungrouped binary data the dispersion parameter should always be $\phi = 1$. Do you know why it is the case?
shani's user avatar
  • 681
0 votes
0 answers
56 views

Find the optimal linear combination of the following estimating function / quasi-score

Consider a random variable $Y_1 \sim Bin(n,\theta)$ and $Y_2 \sim Bin(n,\theta^2)$ where $Y_1$ is independent from $Y_2$. Consider the residuals $R_1 = Y_1 - n\theta$ and $R_2 = Y_2 - n\theta^2$. Find ...
malek-il-mathematik's user avatar
0 votes
2 answers
881 views

Appropriate Regression Model for Proportions and Rate Data

I have a problem where my dependent variable is given as a click-through rate and thus bounded [0,1]. While I have the traffic for each sample (a combination of design factors) and could reconstruct a ...
doublea's user avatar
2 votes
0 answers
511 views

Comparing performance of Quasi-binomial model and Beta-binomial model

I read some books in biostatistics about fitting binary date with Beta-Binomial regression model and Quasi-Binomial regression model. It proposes a setting: Setting: Assuming we have a sequence of ...
KPMGGMC's user avatar
  • 31
4 votes
1 answer
307 views

Correct GLM or NLS to model exponential model with response variable with positive and negative values

I have been struggling to find the right way to model this dataset, this is a Data Frame with the dataset: ...
Derek Corcoran's user avatar
1 vote
1 answer
106 views

How to calculate model weight for GLM-quasi poisson model

I am running several models with the quasi-Poisson family, I have calculated QAIC for each model but I wanted to know the weight of each individual model. I tried AICcmodavg but it did not work. Is ...
Babu Rao's user avatar
1 vote
1 answer
130 views

Quasi-likelihood can't be generated by any valid probability distribution

I am learning about quasi-Poisson and i'm stuck at the concept of quasi-likelihood function. In wikipedia, it is said that: The term quasi-likelihood function was introduced by Robert Wedderburn in ...
InTheSearchForKnowledge's user avatar
2 votes
1 answer
181 views

Mean-variance relationship in the quasi-likelihood

I have some questions regarding the quasi-likelihood model of GLM: I understand that one reason to use quasi-likelihood in GLM is over-dispersion. This seems to justify using the quasi-Poisson, or ...
Maverick Meerkat's user avatar
2 votes
1 answer
209 views

How to compute the gradient for a GARCH with the package rugarch in R

I am estimating a GARCH(1,1) with external regressors and the package rugarch allows me to do it easily. However, to compute QMLE robust standard errors, I need the ...
Julian Pineda's user avatar
3 votes
2 answers
774 views

How to compute the sandwich variance ML estimator in R

I'm currently estimating a DCC-type model by maximum likelihood. Im using the command solnp and it return an object where I can compute the Hessian H evaluated at ...
Julian Pineda's user avatar
0 votes
1 answer
181 views

Test for significant differences for data between 0 and 1

I have to test for significant differences between scenarios. Data consist of the length of a segment divided by the total length of the network. They are distributed between 0 (never equal to 0) and ...
Mauro Carolli's user avatar
0 votes
1 answer
265 views

Output in two-part Fractional Response Model (FRM) package incorrect?

I was experimenting with the Fractional Response Model (FRM) package, and decided to replicate the results using the base GLM package to better understand the theory. I am able to replicate the ...
Econometrics33's user avatar
3 votes
1 answer
428 views

Quasi-likelihood function

I got stuck in the derivation of the quasi-likelihood function. Namely, given an i.i.d sample $\{Y_i,X_i \}_{i=1}^n$ with $n$ the sample size, let the conditional mean and variance functions be ...
Rico's user avatar
  • 152
0 votes
0 answers
60 views

Comparing the marginal effect of a GLM to the OLS estimates

My question is, whether there is any way to (somewhat) compare the marginal effect of a GLM estimate to an OLS estimate. As in, "since the OLS and GLM results are very similar, I will favour OLS ...
Tom's user avatar
  • 528
2 votes
1 answer
1k views

Comparing the marginal effects of glm output to polr output

I have a dependent variable that is technically ordinal, so I ran a ordered probit model (polr). However, an ordered probit model does not produce any residuals ...
Tom's user avatar
  • 528
1 vote
0 answers
202 views

How flexible is Stata's ivpois? Could I use it for a (quasi) binomial distribution?

According to this post on statalist, Stata's ivpois (an instrumental variable approach) is pretty flexible, with very little assumptions. The problem mentioned in ...
Tom's user avatar
  • 528
1 vote
0 answers
741 views

How to do a Control Function (CF) / Two Stage Residual Inclusion (2SRI) with an ordinal dependent variable in the first stage and a glm in the second

I am trying to use a Control Function (CF) / Two Stage Residual Inclusion (2SRI) approach, because the modeled relationship that I am trying to estimate is non-linear (my dependent variable has a ...
Tom's user avatar
  • 528
1 vote
2 answers
579 views

Getting the (Stata) margins from fractional regression (=glm with family quasibinomial) for an ordinal variable in R

I first found this really nice Stata video on fractional regression (the dependent variable is a proportion including 0 and 1). I am especially interested in how he applies the margin approach to ...
Tom's user avatar
  • 528
3 votes
2 answers
877 views

What is the interpretation of a glm coefficient on a dependent variable that has a % interpretation

I have a dependent variable that takes on values between 0 and 1, including 0 and 1. The variable signifies a proportion (0 = nothing, 1 = all). I am running a model of the type: ...
Tom's user avatar
  • 528