METHOD subcommand (CROSSTABS command)
METHOD
displays
additional results for each statistic requested. If no METHOD
subcommand is specified, the standard
asymptotic results are displayed. If fractional weights have been
specified, results for all methods will be calculated on the weight
rounded to the nearest integer. This subcommand is available only
if you have the Exact Tests add-on option installed, which is only
available on Windows operating systems.
MC. Displays an unbiased point estimate and confidence interval based on the Monte Carlo sampling method, for all statistics. Asymptotic results are also displayed. When exact results can be calculated, they will be provided instead of the Monte Carlo results.
CIN(n). Controls the confidence level for the
Monte Carlo estimate. CIN
is
available only when /METHOD=MC is specified. CIN
has a default
value of 99.0. You can specify a confidence interval between 0.01
and 99.9, inclusive.
SAMPLES. Specifies the number of tables sampled
from the reference set when calculating the Monte Carlo estimate of
the exact p value. Larger sample
sizes lead to narrower confidence limits but also take longer to calculate.
You can specify any integer between 1 and 1,000,000,000 as the sample
size. SAMPLES
has a default value
of 10,000.
EXACT. Computes the exact significance level
for all statistics in addition to the asymptotic results. EXACT
and MC
are mutually exclusive alternatives (you cannot specify both on
the same command). Calculating the exact p value can be memory-intensive. If you have specified /METHOD=EXACT and find that you have insufficient
memory to calculate results, you should first close any other applications
that are currently running in order to make more memory available.
You can also enlarge the size of your swap file (see your Windows
documentation for more information). If you still cannot obtain exact
results, specify /METHOD=MC to
obtain the Monte Carlo estimate of the exact p value. An optional TIMER
keyword is available if you choose /METHOD=EXACT.
TIMER(n). Specifies the maximum number of minutes
allowed to run the exact analysis for each statistic. If the time
limit is reached, the test is terminated, no exact results are provided,
and the program begins to calculate the next test in the analysis. TIMER
is available only when /METHOD=EXACT is specified. You can specify
any integer value for TIMER
.
Specifying a value of 0 for TIMER
turns the timer off completely. TIMER
has a default value of 5 minutes. If a test exceeds a time limit
of 30 minutes, it is recommended that you use the Monte Carlo, rather
than the exact, method.
Example
CROSSTABS TABLES=FEAR BY SEX
/CELLS=ROW COLUMN EXPECTED RESIDUALS
/STATISTICS=CHISQ
/METHOD=MC SAMPLES(10000) CIN(95).
- This example requests chi-square statistics.
- An unbiased point estimate and confidence interval based on the Monte Carlo sampling method are displayed with the asymptotic results.