.-
help for ^mhodds^                                               (STB-27: ssa8)
.- 

Calculation of odds ratios
--------------------------

    ^mhodds^ fail xvar [varlist] [^if^ exp] [^in^ range] [fweight] 
        [^,^ ^b^inomial^(^varname^)^ ^by(^varlist^)^ ^c^ompare^(^val1^,^val2^)^ ^l^evel^(^#^)^]


Description
-----------

^mhodds^ estimates the ratio of the odds of failure for categories ^val1^ and
^val2^ of xvar.  The optional ^varlist^ contains categorical variables which
are to be controlled for using stratification. Strata are defined by
cross-classification of these variables and the odds ratio estimate is
combined over strata using the Mantel-Haenszel method. Using the ^by^ option,
the variation of the combined odds ratio with further categorical variables
can be explored. Matched case-control studies can be analysed by controlling
for a categorical variable which takes a different value for each matched
set.

For individual records, the variable fail is coded 0/1 and the ^binomial^
option is not used.  For frequency records the variable fail is coded 0/1,
the ^binomial^ option is not used, and the frequencies are supplied using
frequency weights.  For records which contain the number of failures out of a
number of subjects, the variable fail contains the number of failures and the
variable containing the number of subjects is supplied by the ^binomial^
option.


Options
-------

^binomial(^varname^)^ indicates the variable that contains the number of subjects
    for binomial frequency records.  This option is not used for either 
    individual or simple frequency records. 

^by(^varlist^)^ specifies categorical variables by which the odds ratio is to be
    tabulated.  A separate odds ratio is produced for each unique combination
    of categories, and a test for unequal separate odds ratios is performed.

^compare(^val1^,^val2^)^ specifies the levels of xvar to be compared; ^val1^ defines
    the numerator and ^val2^ the denominator.  When ^compare^ is absent and there
    are only two categories, the second is compared to the first; when there
    are more than two categories the odds ratio for a unit increase in ^xvar^,
    controlled for specified confounding variables, is given.

^level(^#^)^ gives the level for the confidence intervals (default 95).


Examples
--------

 . ^mhodds d smoke age, c(2,1) by(sex)^         -- Individual records
      					           Controlled for age, by sex

 . ^mhodds D smoke age, b(N) c(2,1) by(sex)^    -- Binomial frequency records
					           Controlled for age, by sex

 . ^mhodds d smoke age sex [fweight=f], c(2,1)^ -- Simple frequency records
						   Controlled for age and sex


Authors
-------

        David Clayton, MRC Biostatistical Research Unit, Cambridge
        david.clayton@@mrc-bsu.cam.ac.uk
        Michael Hills, London School of Hygiene and Tropical Medicine, London
        mhills@@lshtm.ac.uk


Also see
--------

    STB:  STB-27 ssa9
On-line:  help for @tabodds@