Statistical Analysis Stack Exchange
Statistical Analysis Stack Exchange

Questions tagged [quasi-likelihood]

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

In GLMs, quasi-likelihood estimation is a way to allow over- or under-dispersion by choosing an appropriate variance function. It's often used for binary or count data, e.g. in quasi-binomial or quasi-Poisson models; there it does not correspond to any actual count distribution.

Wikipedia has an article https://en.wikipedia.org/wiki/Quasi-likelihood with further information and references.

136 questions
51
votes
2 answers

What is quasi-binomial distribution (in the context of GLM)?

I'm hoping someone can provide an intuitive overview of what quasibinomial distribution is and what it does. I'm particularly interested in these points: How quasibinomial differs to the binomial distribution. When the response variable is a…
luciano
  • 12,197
  • 30
  • 87
  • 119
24
votes
1 answer

Why is the quasi-Poisson in GLM not treated as a special case of negative binomial?

I'm trying to fit generalized linear models to some sets of count data that might or might not be overdispersed. The two canonical distributions that apply here are the Poisson and Negative Binomial (Negbin), with E.V. $\mu$ and variance $Var_P =…
user28400
  • 413
  • 1
  • 4
  • 9
21
votes
4 answers

How does a Poisson distribution work when modeling continuous data and does it result in information loss?

A co-worker is analyzing some biological data for her dissertation with some nasty Heteroscedasticity (figure below). She's analyzing it with a mixed model but is still having trouble with the residuals. Log-transforming the response variables…
N Brouwer
  • 1,786
  • 2
  • 18
  • 32
19
votes
1 answer

What is the difference between logistic regression and Fractional response regression?

As far as I know, the difference between logistic model and fractional response model (frm) is that the dependent variable (Y) in which frm is [0,1], but logistic is {0, 1}. Further, frm uses the quasi-likelihood estimator to determine its…
newbie
  • 375
  • 1
  • 2
  • 9
19
votes
2 answers

Poisson or quasi poisson in a regression with count data and overdispersion?

I have count data (demand/offer analysis with counting number of customers, depending on - possibly - many factors). I tried a linear regression with normal errors, but my QQ-plot is not really good. I tried a log transformation of the answer: once…
Antonin
  • 409
  • 2
  • 5
  • 10
18
votes
2 answers

Are over-dispersion tests in GLMs actually *useful*?

The phenomenon of 'over-dispersion' in a GLM arises whenever we use a model that restricts the variance of the response variable, and the data exhibits greater variance than the model restriction allows. This occurs commonly when modelling count…
Ben
  • 91,027
  • 3
  • 150
  • 376
17
votes
1 answer

How to deal with overdispersion in Poisson regression: quasi-likelihood, negative binomial GLM, or subject-level random effect?

I've come across three proposals to deal with overdispersion in a Poisson response variable and an all fixed-effects starting model: Use a quasi model; Use negative binomial GLM; Use a mixed model with a subject-level random effect. But which to…
Bryan
  • 741
  • 1
  • 4
  • 15
14
votes
1 answer

Identical coefficients estimated in Poisson vs Quasi-Poisson model

In modeling claim count data in an insurance environment, I began with Poisson but then noticed overdispersion. A Quasi-Poisson better modeled the greater mean-variance relationship than the basic Poisson, but I noticed that the coefficients were…
Frank H.
  • 566
  • 1
  • 4
  • 16
11
votes
1 answer

Comparison negative binomial model and quasi-Poisson

I have run negative binomial and quasi-Poisson models based on an hypothesis testing approach. My final models using both methods have different covariates and interactions. It seems that there are no patterns when I plot my residuals in both cases.…
Elena Spark
  • 375
  • 3
  • 12
10
votes
2 answers

Vector calculus in statistics

I'm teaching a class on integration of functions of several variables and vector calculus this semester. The class is made up most of economics majors and engineering majors, with a smattering of math and physics folks as well. I taught this class…
Paul Siegel
  • 221
  • 1
  • 7
10
votes
1 answer

What is the difference between beta regression and quasi glm with variance = $\mu(1-\mu)$?

First let me give some background; I will summarize my questions at the end. The Beta distribution, parameterized by its mean $\mu$ and $\phi$, has $\operatorname{Var}(Y) = \operatorname{V}(\mu)/(\phi+1)$, where $\operatorname{V}(\mu) = \mu(1-\mu)$…
Andrew Milne
  • 105
  • 1
  • 8
8
votes
1 answer

Models for Generalized Estimating Equation?

From Wikipedia, Generalized Estimating Equation (GEE) is a method to estimate the parameters of a generalized linear model (with an exponential family distribution for the response). By reading other references online, I am confused whether GEE is…
Tim
  • 1
  • 29
  • 102
  • 189
8
votes
1 answer

Definition of dispersion parameter for quasipoisson family

I try to model quasi-poisson family in bugs language, to handle overdispersion. According to Introduction to WinBUGS for ecologists, this is done by: $log(\lambda_i) = f(x_i) + \epsilon_i$ $N_i \sim Poiss(\lambda_i)$ where $\epsilon_i$ does the…
Tomas
  • 5,735
  • 11
  • 52
  • 93
7
votes
1 answer

Quasibinomial vs negative binomial and hurdles

I have some over-dispersed data and am trying to decide which model would best suit the data. The data are usually counts of symptoms or number of correct items on some cognitive tasks. As an example: set.seed(69) g1<-rnorm(700,30,9);…
Matt Albrecht
  • 3,213
  • 1
  • 24
  • 32
7
votes
1 answer

GLM analogue of weighted least squares

The short version: I can fit a model using Weighted Least Squares, given a diagonal matrix of weights $W$, by solving $(X^TWX)\hat{\beta}=X^TWy$ for $\hat{\beta}$. Is there a GLM analogue? if so, what is it? There seems to be a GLM analogue, e.g.…
shadowtalker
  • 11,395
  • 3
  • 49
  • 109
1
2 3
…
9 10