Chapter 11 Flashcards
The Origin of a Lognormal Distribution
When factors work in a multiplicative way. • Double vs cut in half are equally likely
The logarithms of values from a lognormal
distribution will have a
Gaussian distribution.
Geometric Mean
In this sample from a lognormal distribution, the mean is not a good measure of central tendency.
The geometric mean
is the antilog of the mean of
the logarithms. Transform all values to their
logarithms. Compute the mean of those logarithms.
Transform that mean back to the original units of the
data (calculate the antilog).
Common Mistakes
Trying to take the log of zero or negative values.
• Misinterpreting a lognormal distribution as a
Gaussian distribution with an outlier.
• Difficult: need a theoretical reason for
interpreting as a lognormal.
• Inconsistently using different bases for logarithms.
• Be consistent: whichever log base is used, the
geometric mean with have exactly the same
value.