CONTRAST Subcommand (POWER ONEWAY ANOVA command)

The optional CONTRAST subcommand specifies the contrasts and pairwise differences.

COEFFICIENTS = values

The optional keyword specifies the contrast coefficients and requests the contrast test. The keyword is omitted by default. When the keyword is specified, the number of specified values must be equal to that for /PARAMETERS GROUP_SIZES or /PARAMETERS GROUP_MEANS. All of the specified values must be summed up to 0, otherwise the last value is automatically adjusted.

TEST = {NONDIRECTIONAL}{DIRECTIONAL}

The optional keyword determines whether the contrast test is one or two-sided. NONDIRECTIONAL, or a two-sided test, is the default setting. When DIRECTIONAL is specified, power is computed for a one-sided test. The keyword takes effect when /CONTRAST COEFFICIENTS is specified.

PAIRWISE = boolean

The optional keyword controls whether or not to estimate the power of testing for the pairwise differences. FALSE is the default setting, which suppresses output for the pairwise differences. When TRUE is specified Table 3 is required and displays. In cases where COEFFICIENTS is omitted and PAIRWISE = FALSE, a warning message is created that indicates:

COEFFICIENTS is required

ADJUST = {BONFERRONI} {SIDAK} {LSD}

The optional keyword determines the adjustment of multiple comparisons. BONFERRONI is the default setting, which uses the Bonferroni correction in estimating the power of pairwise differences. SIDAK and LSD use the Sidak correction and the least significant difference, respectively. The keyword takes effect when /CONTRAST PAIRWISE = TRUE is specified.

HALFWIDTH = values

The optional keyword estimates the sample size based on the specified confidence interval half-widths. The parameter takes effect when both POWER and POOL_SD are specified (/PARAMETERS ES cannot be specified, otherwise POOL_SD is ignored). The keyword is ignored otherwise. Note that when /CONTRAST TEST = DIRECTIONAL is specified, the test with the user-supplied contrast is ignored, but the overall test and test for pairwise differences (if applicable) are always valid. At least one value must be specified for the keyword. When multiple values are specified, they are separated by a blank space (or blank spaces). All valid values must be greater than 0. Duplicate values are ignored.

ES = value

The optional keyword specifies the effect size of the contrast test, which is measured by δ. When ES is specified for /PARAMETERS (to formulate a valid contrast test), ES must be also be specified for /CONTRAST. Only a single numeric value is allowed. When /CONTRAST TEST = NONDIRECTIONAL, the specified value must be ≥ 0, otherwise, the value can be any real number. When POWER is specified, the specified value cannot be 0.