Section
3:
Means
and Medians
Trial
Award Patterns
Probability
Standard
Deviation
Normal
Distributions
Logarithms
Awards Test


Testing Whether Jury and Judge
Mean Punitive Awards Differ
Finally, one can now test whether the difference between jury and
judge punitive awards was likely to occur by chance. That is, just as
in the case of 2 x 2 tables, one can ask whether the observed difference
was likely to occur by chance.
To calculate the probability that the jury and judge punitive pattern
occurred by chance, we need six numbers: mean, standard deviation, and
number of observations for both jury and judge punitive awards. All
of these numbers appear in Table 2.
Table 2. Jury and Judge Trial Award Patterns
(in log dollars)


Mean

Number
of Plaintiff Awards

Standard
Deviation


Jury

Judge

Jury

Judge

Jury

Judge

Compensatory
damages 
4.70

4.36

3,001

1,375

0.891

0.808

Punitive
damages 
4.66

4.36

121

55

1.241

1.044

Use this new handy calculator, and select "Enter mean,
SD and N":
Graphpad.com
calculators
The calculation yields a pvalue of just over .10. That is,
there is a probability of more than one in ten that we would observe
by chance differences this large or larger between mean judge and jury
punitive awards if they came from the same population. Hence, the observed
difference is not statistically significant.
