Re: Entering '<1' and 'BDL'

On Tue, 20 Apr 2010, Tom Keller wrote:

> Why distinguish "below detection limit" at the data level? I would use the
> ... value measured and enforce the ramification of the detection limit in
> subsequent analysis. Then use NULL for missing values only. or isn't that
> feasible?

Tom,

There is no measured value when it falls below the detection limit of the
instrument. Consider: you have a 6' measuring tape and you use that to
measure the length of your car. The car extends beyond the length of the
tape. What value do you enter for car length?

There are actually three situations that may occur with permit compliance
monitoring, and each must be recorded as distinct from the others. One, the
sample was not taken, or analyzed, or entered in the system. Two, the sample
was taken and analyzed, but the value falls outside the measuring range of
the instrument (or analytical method) used. Three, the sample could not be
taken because the stream channel was dry (ergo, no water to sample), the
pond frozen (ditto), or the machinery shut down and the plant not operating.
The latter two are expected and reasonable occurrences; the first one is a
no-no.

Rich