Ask Mallard
Core concept

Choosing a count model

5 min read

Choose a count model by how events arise, how much counts vary, and whether a separate process can produce zeros.

Fit

When it matters, and when it does not

It matters when

  • You count events over a defined observation period.
  • You need to choose between Poisson, negative binomial and zero-inflated models.

It matters less when

Before you start

What to check in your study

Decide these first
  1. What is countedoutcomeTypeDefine events and the time or area in which they can occur.
  2. Time at riskUnequal follow-up needs an offset: a known term that scales the expected count.
  3. Why a zero can occurDistinguish a chance absence of events from a group unable to have the event.
It rests on these
  1. The chosen distribution fits counts after allowing for the predictors.How to check: Check fitted counts, zeros and residual spread. A large raw variance alone does not decide the model.
  2. Repeated counts from one person are related.How to check: Use a mixed model or another approach that accounts for that dependence.

Plain language

How it works

Poisson links predictors to the expected count. Its variance equals its mean after conditioning on those predictors. Negative binomial adds a parameter for extra spread. It can produce many zeros without a separate zero process.

The choice at a glance
Count process: Fitted spread matches the conditional mean leads to Poisson: fitted spread matches mean; Extra spread: Extra spread remains after predictors leads to Negative binomial: allow extra spread; Separate zero process: A zero-only group is plausible leads to Zero-inflated: zero-only plus count group.Count processFitted spread matches the conditional meanPoisson: fitted spread matches meanUse the study question and assumptions together.Extra spreadExtra spread remains after predictorsNegative binomial: allow extra spreadUse the study question and assumptions together.Separate zero processA zero-only group is plausibleZero-inflated: zero-only plus count groupUse the study question and assumptions together.
Conceptual diagram. It shows the structure of the idea, not data from any study.
Reading the choice

A zero-inflated model mixes a group whose count must be zero with an at-risk count distribution. The at-risk group can also have zero events. Its count distribution may be Poisson or negative binomial.

What decides it

Choose from the clinical process and model checks together. Check the exposure definition before changing the distribution. A hurdle model instead separates any event from a positive count. It answers another two-part question.

Worked example

A worked example

Invented for illustration
Setting
A clinic counts urgent visits during equal follow-up.
Question
Do many patients with no visits require a zero-inflated model?
Approach
Of 100 patients, 60 have no visits. A fitted Poisson model predicts a mean of 2 visits per patient.
Result
There are 60 of 100 zeros, or 60 in 100. That Poisson model predicts about 14 in 100 zeros.
What it means
The mismatch prompts a fit check. It does not establish a group that cannot visit; negative binomial variation may also explain extra zeros.

In practice

Recent studies PubMed search · not vetted

Searching PubMed…

Watch out

Common mistakes

The one to remember

Many zeros alone do not establish a separate zero-generating group. Check fit after accounting for exposure time and covariates.

  • Choosing negative binomial from the raw variance alone.Instead: Assess spread after fitting relevant predictors and time at risk.
  • Reading a zero-inflated count coefficient as a population rate ratio.Instead: The count component describes the latent at-risk group. A population mean also depends on the zero component.

Inside Mallard

How Mallard uses this

Mallard lists Choosing a count model under Core research concepts.

No Mallard code example points to this page.
Atlas node concept.choosing_count_model, the id a plan cue records.

Related Learn pages

Related designs and methods

Evidence

Sources & further reading

  1. Semiparametric analysis of zero-inflated count data (opens in a new tab) · Biometrics · doi 10.1111/j.1541-0420.2006.00575.x
  2. A score test for overdispersion in zero-inflated poisson mixed regression model (opens in a new tab) · Stat Med · doi 10.1002/sim.2616
  3. The analysis of count data: a gentle introduction to poisson regression and its alternatives (opens in a new tab) · J Pers Assess · doi 10.1080/00223890802634175
Prepared with AI assistance. Source titles, authors, publication years and available DOIs were verified against PubMed on 2026-11-06. No independent professional approval is claimed. Lam and colleagues develop the structural-zero mixture. Xiang and colleagues test extra spread beyond a zero-inflated Poisson fit. Coxe and colleagues explain the wider count-model choices.
Last source review: 2026-11-06
AI-generated educational content
Mallard uses reputable methodological and reporting sources, but this material may contain errors or omit important context. Verify important methodological decisions with the cited sources and, when appropriate, a qualified research methodologist or statistician.