Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
faq:how_can_i_test_an_interaction_effect_using_cluster-based_permutation_tests [2017/08/17 11:21]
127.0.0.1 external edit
faq:how_can_i_test_an_interaction_effect_using_cluster-based_permutation_tests [2018/09/19 20:10] (current)
66.44.30.195 [How to test an interaction effect using cluster-based permutation tests?]
Line 14: Line 14:
 From these 4 data structures, you now make 2 difference data structures in the following way: From these 4 data structures, you now make 2 difference data structures in the following way:
  
-  * Copy GA11 to GAdiff11_12 and perform the assignment GAdiff11_12.avg=GA11.avg-GA12.avg+  * Copy GA11 to GAdiff11_12 and perform the assignment GAdiff11_12.individual=GA11.individual-GA12.individual
-  * Copy GA21 to GAdiff21_22 and perform the assignment GAdiff21_22.avg=GA21.avg-GA22.avg.+  * Copy GA21 to GAdiff21_22 and perform the assignment GAdiff21_22.individual=GA21.individual-GA22.individual.
  
 The objective is now to statistically compare GAdiff11_12 and GAdiff21_22. Because we will be comparing two differences,​ we will be testing an interaction effect. Using a cluster-based permutation test, we have to choose the appropriate statfun, depending on whether this comparison involves a within-subjects or a between-subjects factor. In a full within-subjects design, it involves a within-subject factor, and in a mixed between-within-subjects design, it involves a between-subjects factor (remember that the first factor in the design is the between-subjects factor). In the form of a recipe: The objective is now to statistically compare GAdiff11_12 and GAdiff21_22. Because we will be comparing two differences,​ we will be testing an interaction effect. Using a cluster-based permutation test, we have to choose the appropriate statfun, depending on whether this comparison involves a within-subjects or a between-subjects factor. In a full within-subjects design, it involves a within-subject factor, and in a mixed between-within-subjects design, it involves a between-subjects factor (remember that the first factor in the design is the between-subjects factor). In the form of a recipe:
Line 33: Line 33:
 For k=1,...,K and l=2,​...,​L ​ For k=1,...,K and l=2,​...,​L ​
  
-  * Copy GAk1 to GAdiffk1_kl and perform the assignment GAdiffk1_kl.avg=GAk1.avg-GAkl.avg.+  * Copy GAk1 to GAdiffk1_kl and perform the assignment GAdiffk1_kl.individual=GAk1.individual-GAkl.individual.
  
 The objective is now to statistically compare the K difference arrays GAdiffk1_kl. Because we will be comparing differences,​ we will be testing an interaction effect. Using a cluster-based permutation test, we have to choose the appropriate statfun, depending on whether this comparison involves a within-subjects or a between-subjects factor. In a full within-subjects design, it involves a within-subject factor, and in a mixed between-within-subjects design, it involves a between-subjects factor (remember that the first factor in the design is the between-subjects factor). ​ The objective is now to statistically compare the K difference arrays GAdiffk1_kl. Because we will be comparing differences,​ we will be testing an interaction effect. Using a cluster-based permutation test, we have to choose the appropriate statfun, depending on whether this comparison involves a within-subjects or a between-subjects factor. In a full within-subjects design, it involves a within-subject factor, and in a mixed between-within-subjects design, it involves a between-subjects factor (remember that the first factor in the design is the between-subjects factor). ​