Matematisk statistik CTH/GU
Optimal number of measurements needed to prove reduction of exposure
Consider a situation where workers in a working place are
exposed to certain airborne chemicals. If a measure is taken
to reduce the exposure (such a measure can be ventilation,
change of substances etc), it is desirable to have a method
to verify that a change has indeed ocurred as a result
of the measure taken. One interesting problem in this context
is: How many observations do we have to make before and
after the preventive measure is taken?
Examples of such a situation are reported in the references
below. In a large number number of cases, the exposures are
measured as the average concentrations in the breathing zone
of workers during a shift. Often the variation is so high and
the effect of the measure taken is so low, that statistical
methods have to be used. The particular distribution used
is the lognormal distribution.
This problem can be stated as a problem of testing a statistical
hypothesis (that the taken measure has no effect vs that it does
have an effect) and the power of the test can be used to
decide how many observations are needed before and after.
References
E. Olsen (1994) Analysis of exposure using a logbook method. Appl. Occup.
Hyg. 9, 712-722.
S.M. Rappaport (1991) Assessment of long-term exposures to toxic
substances in air. Ann. Occup. Hyg. 35, 61-121.
Kontaktperson
Ziad Taib, tel: 772 3530, epost: ziad@math.chalmers.se
Last modified: Thu Feb 18 12:04:34 MET 1999