Within PROC GLM, the interaction between the two categorical predictors, Income and Gender, was shown to be significant. An item store was saved from the GLM analysis. Which statement from PROC PLM…

Question

Options

• Asliceby Gender / adjust=tukey;
• Bslice Income*Gender / sliceby=Gender adjust=tukey;
• Cslice Income*Gender / sliceby=Income adjust=tukey;
• Dsliceby Income / adjust=tukey;

• A
  3% (1)
• B
  12% (4)
• C
  15% (5)
• D
  71% (24)

Explanation

Option C is actually the correct answer here, not D - the stated answer key appears to contain an error. In PROC PLM, the SLICE statement syntax requires both the interaction term and a SLICEBY= option: slice Income*Gender / sliceby=Income adjust=tukey; - the sliceby=Income option tells SAS to hold Income constant and test the effect of the remaining factor (Gender) within each Income level, while adjust=tukey corrects for multiple comparisons across those tests.

Why the distractors are wrong:

• A (sliceby Gender / adjust=tukey) is not valid PROC PLM syntax - it omits the slice keyword and the interaction term entirely.
• B (slice Income*Gender / sliceby=Gender adjust=tukey) reverses the logic - sliceby=Gender would test Income within each level of Gender, not the other way around.
• D (sliceby Income / adjust=tukey) is syntactically incomplete - it drops the slice keyword and the interaction effect Income*Gender, making it invalid PROC PLM code.

Memory tip: Think of SLICEBY= as "hold THIS variable fixed, test everything else." To test Gender within Income levels, Income is being held fixed → sliceby=Income. The full statement always starts with slice <interaction> /.

Note: The stated correct answer of D appears to be a typo or keying error in the exam. Based on SAS documentation and syntax, C is the correct answer.

No community discussion yet for this question.