Rejecting Data

1
(No Transcript)
2
Rejecting Data
For small number of samples one uses the Q-test
to reject data. For the calculation of molarity
where we have many data points other methods
should be used. Here I present one such
method. Given a random distribution of data, the
probability of a single measurement of being /-d
of the mean is Pd, where Pd is the red area under
the Gaussian curve shown on the next transparancy.
3
Lets assume that red area is Pd.
2d
4
If we make N measurements, then the probability
of them all being /-d of the mean is (Pd)N.
What does it mean if (Pd)N .95?
It means that we are 95 sure that all the data
should fall within /-d of the mean. If a data
point is outside the range /-d we can throw is
away with 95 confidence that it does not belong
to the statistical distribution.
How do we calculate d? Assume N90.
If (Pd)N .95 then Pd(.95)1/N.99943.
5
If F0.99943/2.49971, what is y?
We find that d3.43s where s is standard
deviation.
6
Binning assignment is due with the KHP lab
sheets.