Академический Документы
Профессиональный Документы
Культура Документы
The larger the box, the more dispersion in a set of data. Image: Seton Hall
University
Measures of Dispersion.
Coefficient of dispersion: A “catch-all” term for a variety of formulas,
including distance between quartiles.
Standard deviation: probably the most common measure. It tells you how
spread out numbers are from the mean,
Index of Dispersion: a measure of dispersion commonly used with nominal
variables.
Interquartile range (IQR): describes where the bulk of the data lies (the
“middle fifty” percent).
Interdecile range: the difference between the first decile (10%) and the last
decile (90%).
Range: the difference between the smallest and largest number in a set of
data.
Mean difference or difference in means: measures the absolute difference
between the mean value in two different groups in clinical trials.
Median absolute deviation (MAD): the median of the absolute deviations
from a data set’s median.
Quartiles: Numbers that split the data into four quarters (first, second, third,
and fourth quartiles).
In some processes, like manufacturing or measurement, low dispersion is
associated with high precision. High dispersion is associated with low precision.
However, the range (which gives you an idea of how spread out the entire set of
data is) is much larger for data set B (60) when compared to data set A (6). In
fact, nearly all measures of dispersion would be ten times greater for data set B,
which makes sense as the range is ten times larger. For example, take a look at
the standard deviations for the two data sets:
Standard deviation for A: 2.160246899469287.
Standard deviation for B: 21.602468994692867.
The figure for data set B is exactly ten times that of A.