Hi FS experts and
users,
I am trying to
implement independent sample t-test and correlation analysis.
Here is my fsgdf
design:
-----------------------------
GroupDescriptorFile 1
Title lh_ttest
Class CON_IA
Class PAT_IA
Variables score1 score2
Input CON_IA01 CON_IA 0 0
Input CON_IA03 CON_IA 0 0
Input CON_IA05 CON_IA 0 0
Input CON_IA06 CON_IA 0 0
Input CON_IA07 CON_IA 3 7
Input CON_IA09 CON_IA 0 0
Input CON_IA14 CON_IA 0 0
Input CON_IA15 CON_IA 0 0
Input CON_IA16 CON_IA 7 12
Input CON_IA19 CON_IA 6 15
Input CON_IA24 CON_IA 6 11
Input CON_IA27 CON_IA 3 3
Input CON_IA32 CON_IA 7 21
Input CON_IA33 CON_IA 7 21
Input CON_IA35 CON_IA 0 0
Input CON_IA38 CON_IA 7 15
Input CON_IA39 CON_IA 2 3
Input CON_IA40 CON_IA 6 8
Input CON_IA41 CON_IA 2 4
Input PAT_IA01 PAT_IA 8 13
Input PAT_IA02 PAT_IA 9 38
Input PAT_IA03 PAT_IA 9 33
Input PAT_IA04 PAT_IA 9 27
Input PAT_IA06 PAT_IA 9 26
Input PAT_IA08 PAT_IA 8 16
Input PAT_IA09 PAT_IA 9 33
Input PAT_IA16 PAT_IA 9 35
Input PAT_IA18 PAT_IA 9 39
Input PAT_IA21 PAT_IA 9 32
Input PAT_IA22 PAT_IA 9 31
Input PAT_IA23 PAT_IA 9 35
Input PAT_IA24 PAT_IA 9 40
Input PAT_IA25 PAT_IA 9 29
Input PAT_IA27 PAT_IA 9 30
Input PAT_IA29 PAT_IA 9 30
Input PAT_IA30 PAT_IA 8 32
Input PAT_IA33 PAT_IA 9 13
Input PAT_IA34 PAT_IA 9 21
------------------------------------------------
Contrasts for independent sample t-test:
1,CON_IA>PAT_IA:1 -1 0 0 0 0
2,PAT_IA>CON_IA:-1 1 0 0 0 0
Contrasts for correlation analysis:
1,possitive correlation analysis between pial
thickness and score1:0 0 0.5 0.5 0 0
2,negtive correlation analysis between pial
thickness and score1:0 0 -0.5 -0.5 0 0
3,positive correlation analysis between pial
thickness and score2:0 0 0 0 0.5 0.5
4,negtive correlation analysis between pial
thickness and score2:0 0 0 0 -0.5 -0.5
Are my designs appropriate?
If I want to implement regression analysis(i.e.
dependent variable is thickness and independent variables are score1 and
score2), how to design my contrast?
Thanks.
All the best.
Rujing Zha
2014-02-20
charujing123