### completeness

(measure of how many of a survey's objects of a certain magnitude have been detected)

The term **completeness** is used regarding the limits of an observation's
or survey's detections: for a given apparent magnitude,
a survey's completeness is the fraction of objects of that magnitude
that are actually detected, the others presumably lost by noise, such
as that inherent in the instrument. For example, one might say for
a particular survey that for magnitude 20, its completeness is 95%.
Generally, the larger the magnitude (i.e., the fainter the object
appears in the sky), the smaller the completeness fraction.
The general issue stems from the desire to get and use
absolutely every bit valid information possible up to the
limits of the instrument.

A **completeness limit** for a survey or a dataset derived from it,
and for a given limiting fraction, is the magnitude at which
completeness has fallen to that fraction. One could describe
a dataset as having a 95% completeness limit of magnitude 20,
meaning that at least 95% of any object with magnitude smaller
than 20 is included.

Knowing a survey's completeness depends upon knowing what the survey
hasn't seen, so it is estimated. If data from a survey with more
sensitivity is available, that can be used. Otherwise, the
general approach is to create mock data based upon the distribution
seen at closer distances, assuming some uniformity across time and space,
and to calculate what is likely to be out there.

Another issue for surveys, termed **contamination** is sufficient
noise to appear like an object where there actually is none. The
ratio of such occurrences to real objects is termed its **purity**
(at a given magnitude) and analogous **purity Limits** can be estimated
for a specific survey.

(*measure,surveys*)
http://en.wikipedia.org/wiki/Malmquist_bias

https://www.aanda.org/articles/aa/full_html/2016/08/aa28142-16/T3.html

index