For a set of
observations classified by two variables, with
and
levels respectively, a two-way table of frequencies with
rows and
columns can be computed.
To measure the association between the two classification variables two statistics that can be used are:
Where
are the fitted values from the model that assumes the effects due to the classification variables are additive, i.e., there is no association. These values are the expected cell frequencies and are given by,
Under the hypothesis of no association between the two classification variables, both these statistics have, approximately, a
distribution with
degrees of freedom. This distribution is arrived at under the assumption that the expected cell frequencies,
, are not too small. For a discussion of this point see
Everitt (1977). He concludes by saying, ‘`... in the majority of cases the chi-square criterion may be used for tables with expectations in excess of
in the smallest cell’'.
In the case of the
table, i.e.,
and
, the
approximation can be improved by using Yates' continuity correction factor. This decreases the absolute value of
by
. For
tables with a small value of
the exact probabilities from Fisher's test are computed. These are based on the hypergeometric distribution and are computed using
nag_hypergeom_dist (g01blc). A two-tail probability is computed as
, where
and
are the upper and lower one-tail probabilities from the hypergeometric distribution.
- 1:
– IntegerInput
-
On entry: the number of rows in the contingency table, .
Constraint:
.
- 2:
– IntegerInput
-
On entry: the number of columns in the contingency table, .
Constraint:
.
- 3:
– const IntegerInput
-
On entry: the contingency table, must contain , for and .
Constraint:
for and .
- 4:
– IntegerInput
-
On entry: the stride separating matrix column elements in the arrays
nobst,
expt,
chist.
Constraint:
.
- 5:
– doubleOutput
-
On exit: the table of expected values, contains , for and .
- 6:
– doubleOutput
-
On exit: the table of contributions, contains , for and .
- 7:
– double *Output
-
On exit: if
,
and
then
prob contains the two-tail significance level for Fisher's exact test, otherwise
prob contains the significance level from the Pearson
statistic.
- 8:
– double *Output
-
On exit: the Pearson statistic.
- 9:
– double *Output
-
On exit: the likelihood ratio test statistic.
- 10:
– double *Output
-
On exit: the degrees of freedom for the statistics.
- 11:
– NagError *Input/Output
-
The NAG error argument (see
Section 3.7 in How to Use the NAG Library and its Documentation).
For the accuracy of the probabilities for Fisher's exact test see
nag_hypergeom_dist (g01blc).
Multidimensional contingency tables can be analysed using log-linear models fitted by
nag_glm_binomial (g02gbc).
The data below, taken from
Everitt (1977), is from 141 patients with brain tumours. The row classification variable is the site of the tumour: frontal lobes, temporal lobes and other cerebral areas. The column classification variable is the type of tumour: benign, malignant and other cerebral tumours.
The data is read in and the statistics computed and printed.