The FACTEX Procedure


EXAMINE Statement

  • EXAMINE <options>;

You use the EXAMINE statement to specify the characteristics of the design that are to be listed in the output.

The options are remembered by the procedure; once specified, they remain in effect until you submit a new EXAMINE statement with different options or until you turn off all EXAMINE options by submitting just

examine;

The following options are available.

ABERRATION
AB

displays the aberration vector for the design, summarizing the confounded interactions. See the section Minimum Aberration for more information.

ALIASING <( <d> <UNITS <= ONE | ALL>> )>
A <( <d> <UNITS <= ONE | ALL>> )>

lists the alias structure of the design, which identifies effects that are confounded with one another and are thus, indistinguishable.

You can specify d in parentheses immediately after the ALIASING option for a listing of the alias structure with effects up to and including order d. For example, the following statement requests aliases for up to fourth-order effects (for example, A*B*C*D):

examine aliasing(4);

Each line of the alias structure is listed in the form

effect = effect == effect

for as many effects as are aliased with one another.

The default value for d is determined automatically from the model as follows:

  • If you specify the model with a resolution number r in the MODEL statement, then d is the integer part of $(r+1)/2$.

  • If you specify the model with a list of effects in the MODEL statement, then d is the larger of the following, where main effects have order 1, two-factor interactions have order 2, and so on:

    • one plus the largest order of an effect to be estimated

    • the largest order of an effect considered to be nonnegligible

If your design involves unit effects, then you can also specify UNITS in parentheses immediately after the ALIASING option to display for each treatment effect alias string the top unit effect with which it is aliased. Using the UNITS option can give you information about which error stratum can be used to estimate the background error variance for each estimable treatment effect. If unit effects are nested, as they typically are in complex split-plot designs, then treatment effects can be aliased with more than one unit effect. In this case, specify (UNITS=ALL) to see all unit effects with which each treatment effect is aliased.

For details about aliasing, see the section Alias Structure.

CONFOUNDING
C

lists the confounding rules used to construct the design. For the definition of confounding rules, see the sections Confounding Rules and Suitable Confounding Rules.

DESIGN
D

lists the points in the design in standard order with the factor levels coded. For a description of the randomization and coding rules, see the section OUTPUT Statement.

SUMMARY <(<d>)>
S <(<d>)>

displays the modeling summary for the design, summarizing how many interactions of each order are estimable and how many are clearly estimable (that is, unaliased with any other interactions of interest).

You can specify d in parentheses immediately after the SUMMARY option to display the modeling summary accounting for effects up to and including order d. The default value for d is determined automatically from the model as it is for the ALIASING option.