245 / 762
Chapter 6: Categorical Data
245
Definition:
A conditional distribution for a (first) variable given another (second)
variable is the percentage of items for the first variable that is contained in
the second variable.
Notes:
The conditional distribution of the row variable, given the column
variable, is obtained by dividing the frequency values in the original
table by the column totals.
This is usually expressed as a percentage.
Notes:
The conditional distribution of the column variable, given the raw
variable, is obtained by dividing the frequency values in the original
table by the row totals.
This is usually expressed as a percentage.
Note:
These conditional distributions will be referred to as
conditional
probabilities
in a later chapter.
Example 6-2:
Compute the conditional distributions for the age
classifications (row) given the degree classifications (column) for the data
given in
Example 6-1
. Use two decimal places.
Solution:
From the original two-way distribution, we need to compute each
frequency entry as a percentage of the respective column totals.
For the class of 29 and younger, the conditional distribution for the entry
4286, given the BS degree classification, will be (4286/21527)
100 % =
19.91%. For the class of 60 and older, the conditional distribution for the
entry 261, given the Ph.D degree classification, will be (261/1026)
100 % =
25.44%. One can continue in this manner to compute the remaining