3dMVM multiple covariates question

AFNI Message Board
1

Hello!

I am getting this error with 3dMVM. The issue seems to be related to the two quantitative covariates (age and anxiety score). According to other posts with this same error, if I remove Anxiety and Age from the gltCode lines it should resolve. However, I am unsure how to specify multiple analyses with and without these covariates if I simply remove them from the gltCode lines. In other words, if the way to resolve the error is just having glt lines 1 through 8 (no qVars in gltCode), how will I get the other anaylses I’m looking for in the remaining lines?

I have also copied the error below my code.


3dMVM 																											                                \
 -prefix 3dMVM.FB.Doors.Anxiety.012720																	                                        \
 -jobs 11																										                                \
 -wsVars 'Task*Outcome'																							                                \
 -bsVars 'Age*Anxiety'																									                        \
 -qVars 'Age,Anxiety'																									                        \
 -num_glt 32																									                                \
-gltLabel 1 DislikeGames 						            -gltCode 1 'Task : +1*DislikeGames'								                    \
-gltLabel 2 DislikeGames_Lose 					            -gltCode 2 'Task : +1*DislikeGames Outcome : +1*Lose'			                    \
-gltLabel 3 DislikeGames_Win 					            -gltCode 3 'Task : +1*DislikeGames Outcome : +1*Win'			                    \
-gltLabel 4 DislikeGames_Win.V.Lose				            -gltCode 4 'Task : +1*DislikeGames Outcome : +1*Win -1*Lose'	                    \
-gltLabel 5 LikeGames 							            -gltCode 5 'Task : +1*LikeGames'								                    \
-gltLabel 6 LikeGames_Lose 						            -gltCode 6 'Task : +1*LikeGames Outcome : +1*Lose'				                    \
-gltLabel 7 LikeGames_Win 						            -gltCode 7 'Task : +1*LikeGames Outcome : +1*Win'				                    \
-gltLabel 8 LikeGames_Win.V.Lose 				            -gltCode 8 'Task : +1*LikeGames Outcome : +1*Win -1*Lose'		                    \
-gltLabel 9 DislikeGames_Age 					            -gltCode 9 'Task : +1*DislikeGames Age : '						                    \
-gltLabel 10 DislikeGames_Lose_Age 				            -gltCode 10 'Task : +1*DislikeGames Outcome : +1*Lose Age : '			            \
-gltLabel 11 DislikeGames_Win_Age 				            -gltCode 11 'Task : +1*DislikeGames Outcome : +1*Win Age : '			            \
-gltLabel 12 DislikeGames_Win.V.Lose_Age		                    -gltCode 12 'Task : +1*DislikeGames Outcome : +1*Win -1*Lose Age : '	            \
-gltLabel 13 LikeGames_Age 						            -gltCode 13 'Task : +1*LikeGames Age : '								            \
-gltLabel 14 LikeGames_Lose_Age 				            -gltCode 14 'Task : +1*LikeGames Outcome : +1*Lose Age : '				            \
-gltLabel 15 LikeGames_Win_Age 					    -gltCode 15 'Task : +1*LikeGames Outcome : +1*Win Age : '				            \
-gltLabel 16 LikeGames_Win.V.Lose_Age 			            -gltCode 16 'Task : +1*LikeGames Outcome : +1*Win -1*Lose Age : '		            \
-gltLabel 17 DislikeGames_Anxiety  			                    -gltCode 17 'Task : +1*DislikeGames Anxiety : '						                \
-gltLabel 18 DislikeGames_Lose_Anxiety 				    -gltCode 18 'Task : +1*DislikeGames Outcome : +1*Lose Anxiety : '			        \
-gltLabel 19 DislikeGames_Win_Anxiety 				    -gltCode 19 'Task : +1*DislikeGames Outcome : +1*Win Anxiety : '			        \
-gltLabel 20 DislikeGames_Win.V.Lose_Anxiety		            -gltCode 20 'Task : +1*DislikeGames Outcome : +1*Win -1*Lose Anxiety : '	        \
-gltLabel 21 LikeGames_Anxiety 						    -gltCode 21 'Task : +1*LikeGames Anxiety : '								        \
-gltLabel 22 LikeGames_Lose_Anxiety 				            -gltCode 22 'Task : +1*LikeGames Outcome : +1*Lose Anxiety : '				        \
-gltLabel 23 LikeGames_Win_Anxiety 					    -gltCode 23 'Task : +1*LikeGames Outcome : +1*Win Anxiety : '				        \
-gltLabel 24 LikeGames_Win.V.Lose_Anxiety 			    -gltCode 24 'Task : +1*LikeGames Outcome : +1*Win -1*Lose Anxiety : '		        \
-gltLabel 25 DislikeGames_Age_Anxiety 			            -gltCode 25 'Task : +1*DislikeGames Age : Anxiety : '						        \
-gltLabel 26 DislikeGames_Lose_Age_Anxiety 			    -gltCode 26 'Task : +1*DislikeGames Outcome : +1*Lose Age : Anxiety : '			    \
-gltLabel 27 DislikeGames_Win_Age_Anxiety 			    -gltCode 27 'Task : +1*DislikeGames Outcome : +1*Win Age : Anxiety : '			    \
-gltLabel 28 DislikeGames_Win.V.Lose_Age_Anxiety		    -gltCode 28 'Task : +1*DislikeGames Outcome : +1*Win -1*Lose Age : Anxiety : '	    \
-gltLabel 29 LikeGames_Age_Anxiety 					    -gltCode 29 'Task : +1*LikeGames Age : Anxiety : '								    \
-gltLabel 30 LikeGames_Lose_Age_Anxiety 				    -gltCode 30 'Task : +1*LikeGames Outcome : +1*Lose Age : Anxiety : '				\
-gltLabel 31 LikeGames_Win_Age_Anxiety 				    -gltCode 31 'Task : +1*LikeGames Outcome : +1*Win Age : Anxiety : '				    \
-gltLabel 32 LikeGames_Win.V.Lose_Age_Anxiety 		    -gltCode 32 'Task : +1*LikeGames Outcome : +1*Win -1*Lose Age : Anxiety : '		    \
-dataTable																		                                                                \
Subj        Task		    Age         Outcome	    Anxiety       InputFile				     						                                            \
etc.

ERROR


** Error:
   Inappropriate coding in test No.43!

Warning messages:
1: In if (is.na(code[[n]][QVpos + 2])) { :
  the condition has length > 1 and only the first element will be used
2: In if (QVpos == 1) outList[[1]][[n]] <- NA else outList[[1]][[n]] <- glfConstr(code[[n]][-(QVpos:(QVpos +  :
  the condition has length > 1 and only the first element will be used
3: In QVpos:(QVpos + 2) :
  numerical expression has 2 elements: only the first used
4: In QVpos:(QVpos + 2) :
  numerical expression has 2 elements: only the first used
5: In gl_Constr(lop$num_glt, lop$gltCode, lop) :
  NAs introduced by coercion

Thank you in advance for any guidance!

Best,
Camille

2

All those specifications that involve both Age and Anxiety at the same time don’t make sense. To be clear, notations in gltCode such as “Anxiety :” are meant to make inferences about the slope effect. If you want to control a quantitative variable at its center value, do not include the variable in the specification.

3

Thank you so much!

We had included both Age and Anxiety at the same time to for a Task x Outcome x Age x Anxiety interaction. If I delete the erroneous lines where I have both Age and Anxiety in the gltCode, how do I analyze the full interaction - one that includes interactive effects of age and anxiety (using continuous measures): Task x Outcome x Age x Anxiety?

Thank you again!
Camille

4

Camille,

Since both Age and Anxiety are quantitative variables, to model their interaction properly, do the following:

  1. properly center Age and Anxiety;
  2. create a new variable called, for example, ‘AA’, that is the product of centered Age and Anxiety;
  3. add ‘AA’ to the data table;
  4. change the following lines

-bsVars ‘Age*Anxiety’ \
-qVars ‘Age,Anxiety’ \

to

-bsVars ‘Age+Anxiety+AA’ \
-qVars ‘Age,Anxiety,AA’
–qVarCenters ‘0,0,0’ \

  1. formulate appropriate contrast through -gltCode