ETS-Lindgren 5411 GTEM! Test Cell - Manual (Page 80)

80 |
Radiated Emissions Test Performance of GTEM!
S
TATISTICAL
C
OMPARISONS
 Pearson’s Correlation Coefficient—Pearson’s correlation coefficient and
linear regression equation coefficients were calculated for data (the first
three sets of dipole data and the first two sets of personal computer data) in
which values from different distances were combined for overall evaluation.
The correlation helps to show when the data is not independent and can
properly be combined for further analysis.
 Meaning of the Value of Pearson’s Correlation Coefficient—Values of
Pearson’s correlation coefficient between +6.0 and +1, and regression line
slopes between +.05 and +1.5 indicate a strong relationship between the
samples of data.
 Student’s-t Statistic—Student’s-t statistic for paired, sample variables was
used to analyze the comparison data for both the dipoles and the personal
computers. This approach allowed testing the null hypothesis for the
difference between the GTEM! data and the OATS data.
 Meaning of Student’s-t Statistic—The interval of Student’s-t distribution
between –t and +t represents a region in which with a specific probability, all
sets of samples of data are from the same population, and therefore are the
same even though their means and sample variances appear to be different.
The hypothesis that the sample mean is no different than the population
mean is called the null hypothesis, Ho, and it is accepted or rejected by virtue
of whether the sample mean lies within the interval of –t and +t. The
confidence that the sample mean and the population mean are the same is
100 (1-a) percent. If T lies outside of the interval –t to +t, the null hypothesis
must be rejected.
Student’s-t statistic is tabulated for various degrees of freedom (d.f.). The
tabulation must be entered with the d.f. and the confidence level to find the
limits of the interval to be used. For paired, related data, the d.f. are one less
than the number of pairs of data; and for paired independent data,
the d.f. are one less than the sum of the number in each set of sample data.
Student’s-t statistic for non-independent (related) paired sample variables
was used for the first two sets of personal computer data since the same test
object and configuration were used for measurements both in the GTEM! and
the OATS.
Student’s-t Statistic for independent paired sample variables was used for
the first three sets of dipole data, since a different signal generator was used
on the OATS from that used in the GTEM! measurement.