HGLM Likelihood Test for Fixed Terms

This dialog allows you to assess the significance of the fixed terms in a Hierarchical Generalized Linear Models (HGLM) using a likelihood test done by the HGFTEST procedure.

Fixed term

The drop-down list specifies which fixed term in the model to test. Only terms that are not marginal to other terms in the model will be in the drop down-list. If All feasible is selected, then all fixed terms in the model that are not marginal to other terms in the model will be tested. For example in the model A + B + C + A.B + A.C + B.C, only the terms A.B, A.C, B.C can be tested as they are not marginal to other terms. In this model, the terms A, B and C are marginal to the two-way interactions. If the full 3-way model A*B*C was fitted, the only non-marginal term would be A.B.C, as all other terms are contained with A.B.C.

Display

This controls what output is printed from the analysis:

Tests

test statistics for the selected fixed terms

Maximum number of iterations

This gives maximum number of iterations of the hierarchical generalized linear algorithm. If the model in the likelhood test does not converge, try increasing this number.

Save

This lets you save results from the from the HGFTEST analysis in Genstat data structures. After selecting the appropriate boxes, you need to type the names for the identifiers of the data structures into the corresponding In: fields.

Test statistic	Scalar or Pointer	Test statistic for each selected fixed term
Degrees of freedom	Scalar or Pointer	Degrees of freedom for each selected fixed term

If there is a single random term being tested, the saved structures will be scalars, otherwise they will be pointers.

Display in spreadsheet

The saved results will be displayed in a new spreadsheet.

Action Icons

	Pin	Controls whether to keep the dialog open when you click Run. When the pin is up the dialog will remain open, otherwise when the pin is down the dialog will close.
	Clear	Clear all fields and list boxes.
	Help	Open the Help topic for this dialog.