# Shared Concepts and Topics

### Syntax: SLICE Statement

• SLICE model-effect </ options>;

You can specify all options of the LSMEANS statement in the SLICE statement. The philosophy of the SLICE statement is to apply the analysis according to the options to the subsets of the matrix that correspond to chosen partitions.

The following behavior differences between the SLICE and the LSMEANS statement are noteworthy:

• The specification of the model-effect is optional in the LSMEANS statement and required in the SLICE statement.

• Only a single SLICE model-effect can be specified before the option slash (/). However, you can specify multiple partitioning rules with the SLICEBY option.

• The MEANS option is the default for most procedures in the LSMEANS statement. For the SLICE statement, the default is the NOMEANS option.

Also, the three generalized linear modeling options: EXP , ILINK , and ODDSRATIO in the SLICE statement are additionally supported by PROC GLIMMIX and by PROC PLM when it is used to perform statistical analyses on item stores that were created by PROC GLIMMIX.

In addition to the options in the LSMEANS statement, you can specify the following options in the SLICE statement after the slash (/):

SLICEBY <=> slice-specification
SIMPLE <=> slice-specification
SLICEBY(slice-specification <, slice-specification <, >>)
SIMPLE(slice-specification <, slice-specification <, >>)

determines how to construct the partition of the least squares means for the model-effect. A slice-specification consists of an effect name followed by an optional list of formatted values. For example, the following statements creates partitions of the `A`*`B` interaction effect for all levels of variable `A`:

```class a b;
model y = a b a*b;
slice a*b / sliceby=a;
```

The following statements produces two partitions of the interaction:

```class a b;
model y = a b a*b;
slice a*b / sliceby(b='2' a='1') diff;
```

In the first partition the variable `B` takes on formatted value '2'. In the second partition the variable `A` takes on the formatted value '1'.

NOF

suppresses the F test for testing the mutual equality of the estimable functions in the partition.