home *** CD-ROM | disk | FTP | other *** search
-
- TWO BY TWO CONTINGENCY TABLES
-
-
- Frequency data are often categorized
-
- by two criteria. For example, a
-
- persons' sex and his political prefer-
-
- ence, or the type of medical treatment
-
- lung cancer patients receive (new or
-
- old) and whether or not one survives
-
- (alive or died).
-
- In the case where each criterion has
-
- exactly two possibilities, a
-
- convenient way to represent the data
-
- is in a 2X2 table.
-
- Suppose that 42 people are called at
-
- random to determine whether there is
-
- any relationship between a person's
-
- sex and whether he will vote
-
- Democratic or Republican in the next
-
- election.
-
- Calls are made at random and the
-
- results tallied as follows:
-
- Of the 25 men contacted, 15 will
-
- vote Democratic and 10 will vote
-
- Republican. Of the 17 women
-
- contacted, 7 will vote Democratic and
-
- 10 will vote Republican. These data
-
- can be viewed in the 2X2 table below.
-
-
- PARTY
-
- Dem Rep
- ------------------
- ! ! !
- male ! 15 ! 10 ! 25
- SEX ------------------
- ! ! !
- female ! 7 ! 10 ! 17
- ------------------
- ! ! !
- ! 22 ! 20 ! 42
- ! ! !
-
-
-
- The question of whether there is a
-
- difference between the voting
-
- preferences of males and females based
-
- on the findings represented by the re-
-
- sults from a small sample is handled
-
- in a field of mathematics called
-
- inferential statistics.
-
- The program included with this
-
- article will calculate several useful
-
- statistics from a 2X2 table.
-
- I assume that the reader understands
-
- the concept of statistical
-
- significance and knows how to test an
-
- hypothesis.
-
- If you don't know how, but you would
-
- like to find out, here are some books:
-
- Statistical Theory and Methodology In
-
- Science and Engineering by Brownlee.
-
- Published by John Wiley and Sons, New
-
- York.
-
- Bancroft's Introduction to Biosta-
-
- tistics by Ipsen and Feigl. Published
-
- by Harper & Row, New York.
-
- Go to your library and find any book
-
- on statistics. Look up CHI-SQUARE.
-
- ----< continued in next article >-----
-
-