Fisher FERITSCOPE FMP30 Operator's Manual page 219

termined using the standard deviation of the sample mean values (since
the sample mean values can be classified as having normal distribu-
tion). u
Whether a quantity can be classified as having normal distribution can
be checked in the normal probability chart, since a straight line in the
normal probability chart indicates normal distribution. When evaluating
the current application with the FERITSCOPE
Smirnov test is performed for small random samples (up to 40 measure-

ments) and a 2 test is performed for large random samples (more than
40 measurements) to check whether the measurements can be classified
as having normal distribution.
Offline
State of a peripheral device (printer or PC) connected to the instrument
that does not allow it to receive data.
One-point calibration
calibration)
Online
State of a peripheral device (printer or PC) connected to the instrument
that allows it to receive data; the connected instrument is ready for op-
eration.
Outlier measurements
Measurements that are considerable larger or considerable smaller than
the other measurements of the measurement series and therefore can be
considered as unexpected or unacceptable. With outlier rejection en-
abled ( Grubbs Test;
FERITSCOPE® FMP30 indicates outlier measurements by two short au-
dible signals immediately after the measurement acquisition and the si-
multaneous appearance of
Outlier rejection
Is used to prevent the distortion of the measurement results by outlier
measurements. With the FERITSCOPE
enabled or disabled. Measurements recognized as outliers will not be
included in the evaluation. Two methods are available for outlier rejec-
tion:
Grubbs Test

Sigma outlier rejection (Specification of a known spread)

Page 214
Corrective calibration (one-point or two-point
Sigma outlier rejection Outlier rejection)the
and
® FMP30, outlier rejection can be
® FMP30, a Kolmogorov-
on the display.
Operator's Manual FERITSCOPE
®
FMP30