7 votes
Accepted

Small Sample Sizes and Zero Inflated Count Data in R

The problem you are observing with lack of significance of GA has nothing to do with zero inflation or with random effects. It is simply a limitation of Wald tests for count models. If you replace the ...
Gordon Smyth's user avatar
  • 12.9k
6 votes
Accepted

What causes the parameter phi (precision) to be very small in beta regression (by betareg in R)?

Your $Y$ only takes the value $0.001, 0.5$ and $0.999$. That is not a good fit a beta-Regression which models a continuum of proportions. Also 0.5 to 0.999 is a huge spread, so the precision has to be ...
Lukas Lohse's user avatar
  • 2,557
5 votes
Accepted

Are ordinal models technically a GLM?

The ordinal model that you state is called the proportional odds ordinal logistic regression model (POLR), and was popularized by Peter McCullagh (McCullagh 1980). Yes, it is a generalized linear ...
Gordon Smyth's user avatar
  • 12.9k
4 votes

What causes the parameter phi (precision) to be very small in beta regression (by betareg in R)?

In addition to Lukas's excellent point about your Y value, I think you are onto something about your x values. One is categorized age. It would be better to use age in years (and maybe add a spline). ...
Peter Flom's user avatar
  • 120k
3 votes

Constructing a generalized linear model when the dependent variable has a Exponentially modified Gaussian distribution

Generalized linear models assume that the response variable follows an exponential dispersion model (EDM) distribution. For fixed dispersion value, an EDM is a linear exponential family (LEF). As ...
Gordon Smyth's user avatar
  • 12.9k
2 votes

Using generalized linear models over OLS and why do estimates differ?

About your minor sub questions 1) In case ... the errors are homoscedastic ... why would one prefer a glm? GLM does two things: model different link functions model different conditional error ...
Sextus Empiricus's user avatar
2 votes

How to adjust for exposure time in a binomial model

You observe if an event occurred or not, but the length of the observation interval varies. If you model the events as a Poisson process, but you only observe if there are zero or more events, but not ...
kjetil b halvorsen's user avatar
2 votes

What does one do when the coefficient for the log of the rate estimator for a poisson rate model is very different from 1?

For your dataset, I don't think this can be answered without more context. Why do your variable days vary? If we start with some dataset with individual accidents, ...
kjetil b halvorsen's user avatar
2 votes
Accepted

How to create a regression model object from intercept and coefficients values only (without the database) in R

You lack information about the values of the explanatory variables. Exploring this issue reveals some useful things to know about logistic regression. For instance, you can go far beyond simple rules ...
whuber's user avatar
  • 323k
2 votes

Assessing and interpreting the results of a multinomial logistic regression in R

In general, odds ratios tend to provide the most straightforward interpretation when presenting to a non-statistical or clinical audience and are widely. Currently your estimates are on the log-odds ...
Jack's user avatar
  • 344
2 votes

Choosing the right modeling procedure

With ordered outcomes like your integer values, ordinal regression is a useful choice. It makes no assumptions about the distribution of outcome values (or of residuals around model estimates). It can ...
EdM's user avatar
  • 92.5k
2 votes

How to find the appropriate family for a GLM

I divided the amount of vigilant behaviours in one survey by the length of the survey (most 30 mins) If your outcome is a count variable, and it is a count per some unit time, you can use a Poisson (...
Demetri Pananos's user avatar
2 votes
Accepted

Is it correct to evaluate response histogram to decide between a normal regression and other GLMS?

You are correct that evaluating a histogram to choose the likelihood is ill advised. The reason is because the likelihood concerns the conditional distribution of the outcome and a histogram ...
Demetri Pananos's user avatar
1 vote

Estimate a GLM where the intercept is known

A quick workaround is to use nonlinear least squares estimation and bound the S-shaped curve to the (0.1,1) range. ...
SmartDataProducts's user avatar
1 vote

What information is the "intercept" presenting in the IRR column when using parameters package for glmer

You're right that this is confusing. The intercept term is an incidence rate, while the rest of the terms are incidence rate ratios. It may (?) help to think about the definition of the model for a ...
Ben Bolker's user avatar
  • 43.7k
1 vote

Working out/Derivation of Standard Error of Coefficient in Logistic Regression

In logistic regression, as with most nonlinear models, you need to rely on asymptotic normality to characterize the sampling distribution of the coefficient. The variance of this distribution is given ...
Durden's user avatar
  • 1,351
1 vote
Accepted

Modeling no-show rates of attendees in R

You need the actual numbers, not only the rates. Present those as a contingency table of counts, with five rows, one for each weekday. Then there are two columns, one with the number attending, the ...
kjetil b halvorsen's user avatar
1 vote

Interpolation of errors from model predictions over time-series

Your predictions and associated prediction intervals (not really errors as such) will only vary according to the value of the sr variable. So you cannot really get ...
Nicholas Clark's user avatar
1 vote
Accepted

Fixed effects: is there more?

Short answer: No, that's all there is. Longer answer: All of the documentation of fixest stresses that it is fast, user-friendly, and has features for exporting the results easily. I didn't search all ...
Peter Flom's user avatar
  • 120k
1 vote

Choosing the right modeling procedure

These sound like very complex causal questions, where you seem to be ignoring many important aspects (e.g. what diseases do the people have, is the diagnosis probability the same for all of them even ...
Björn's user avatar
  • 32.3k
1 vote

Counterintuitive phenomenon in a Poisson regression

Your data do not follow a Poisson distribution conditional on $X_2$, but only conditional on $X_1$ and $X_2$, so you can't expect valid tests. More specifically, adding $\beta_1X_1$ to the linear ...
Thomas Lumley's user avatar
1 vote

How to improve my glm model, with a lot of potential predictors and millions of observations?

First, don't use p values or any kind of stepwise, forward, or backward to do this. Using AIC instead is a little better, but still problematic. Splitting into train and test, as Roland suggested, ...
Peter Flom's user avatar
  • 120k
1 vote

Difference Between GLM predict Output in R vs. GLM predict Output in Python (statsmodels) Logistic Regression

Two issues: (minor): Python's StandardScaler differs slightly from R's scale. While scale ...
Igor F.'s user avatar
  • 9,139

Only top scored, non community-wiki answers of a minimum length are eligible