Young Survival from Marked Adults
The Lukacs (2004) Young Survival from Marked Adults data type provides estimates of the survival rate of young monitored by checking a marked adult. Typically, the adult is a female bird, and the number of chicks are counted on each visit. The model allows imperfect detection of chicks. One critical assumption of the model is that chicks are not adopted by the adult female, i.e., the actual brood size cannot increase, but can only decrease due to mortality of chicks.
The encounter histories for this data type consist of counts of chicks on each occasion. From 0 to 99 chicks can be counted, although the model does not produce precise estimates of survival for more than 5 or so chicks unless the chick detection rate is very close to 1. As an example, suppose the count of chicks for 6 occasions were 5, 4, 5, 3, 3, and 1. The encounter history would then be:
where only one group is observed. That is, the number of broods with this encounter history is 1. Dots can be used to designate a missing count, e.g.,
would be used if the 4th count in the above example was missing. Dots can also be used at the start of the encounter history to indicate that the brood entered the study at some occasion after the first one.
Parameters for the model are phi (apparent survivial) and p (probability of detection for a single chick). These parameters can be time specific, but the last phi and p estimates will be confounded, and only a product of these 2 parameters can be estimated.
For the case of a single young, this model reduces to the regular Cormack-Jolly-Seber model.