Where does count collide?

In statistics a count is the value of a count variable: the number of times a discrete event occurs in a fixed window, so it is a non-negative whole number with no fixed upper bound, such as clinic visits per patient or defects per part. Because such tallies are bounded below by zero and are often skewed, they are modeled by distributions like the Poisson rather than by the normal curve. Coxe, West, and Aiken's 2009 The Analysis of Count Data gave a plain introduction to Poisson regression and its alternatives for exactly this kind of variable. The referent is a numeric quantity produced by tallying and analysed with special models, and it is not a charge in an indictment, a clinical measurement of cells, or a title of nobility.[1]

Unit

A non-negative integer: how many events fell in a fixed window

Recovery means

Not applicable

Measured by

Obtained by tallying occurrences, then modeled with Poisson or negative binomial regression

In clinical medicine a count is the number of a given kind of cell or particle in a measured volume of a sample, such as the red cell count, white cell count, or platelet count reported in a complete blood count. It is one of the most basic measurements in medicine, read as cells per litre or per microlitre, and a shift in a count helps flag infection, anaemia, or a risk of bleeding. Buttarello and Plebani's 2008 Automated Blood Cell Counts reviewed how modern analysers perform these counts and where they can err. The referent is a concentration measured in a fluid sample, not a tally of events in a data set, a charge in a court, or a noble title.[2]

Unit

Cells or particles per unit volume of a sample

Recovery means

Not applicable

Measured by

Measured by automated haematology analysers, or by manual counting under a microscope