.-
help for ^recap^ written by Matthias an der Heiden, RKI Berlin, AnderHeidenM@rki.de
.-
Capture-Recapture analysis for 3 Sources
with Goodness-of-Fit based confidence intervals
------------------------------
^recap var_freq var_A var_B var_C [if] [in]^
var_A: a variables with values 0 or 1, that desribes the belonging to list A
var_B: a variables with values 0 or 1, that desribes the belonging to list B
var_C: a variables with values 0 or 1, that desribes the belonging to list C
var_freq: a non-negative variable the describes the frequency of observations in
the combination of lists given by var_A, var_B and var_C. The unknown
frequency of cases occuring in none of the lists is missing.
Description
-----------
^recap^ provides an estimate of the total size of a certain well-defined
population.
To apply this method you need three different sources containing members
of this population, such that every member of the population has in principle
the same change to be "captured" by a source. The model is flexible enough to
allow dependences between the sources, such that an individual caught by source
1 has a smaller probability to be caught by source 2.
An important point is that an individual that is present in two or three
sources can be correctly identified. All in all there are eight possible
situations: An individual can be present in all three sources, in two out of
three, in one out of three or in none of them. By knowing the frequencies for
seven of these eight possible combinations we can estimate the number of
individuals that are captured by none of our sources.
To account for possible interactions between the three sources eight different
models are compared in the analysis. The simplest one is the independent model,
then there are three models that allow for an interaction between two of the
sources. More complicated are models with two pair-interactions present and the
so called saturated Model, that adjusts for interaction between all three pairs
of sources. Note that all these models assume the absence of a triple interaction
between the sources.
To compare the different models the Akaike’s Information Criterion (AIC) and
Bayesian Information Criterion (BIC) are provided.
^recap^ also provides Goodness-of-Fit based 95% confidence intervals, see
R.R. Regal and E.B. Hook "Goodness-of-Fit based confidence intervals for
estimates of the size of a closed population", Statistics in Medicine, Vol. 3,
287-291 (1984).
Options
-------
None.
Example
-------
freq A B C
-------------------
5 1 1 1
15 1 0 1
3 0 1 1
20 0 0 1
7 1 1 0
37 1 0 0
14 0 1 0
. 0 0 0
. ^recap freq A B C^
calculates the capture-recapture estimates for this table.
Also see
--------
Manual: [R] poisson
Online: help for @poisson@