help
help > Any effect vs. Simple main effect
Feb 22, 2017 06:02 PM | Jenna Traynor - McMaster University
Any effect vs. Simple main effect
Hi Alfonso,
I am running a within-group bivariate correlation analysis looking at the association between functional connectivity and 6 different behavioural scores, using Conn V.17. I have a question regarding my findings:
When I set up the findings to look for any effect among my behavioural variables 1-6, I select my patient group and the six variables and set up the contrast as so: [0 1 0 0 0 0 0; 0 0 1 0 0 0 0; 0 0 0 1 0 0 0; 0 0 0 0 1 0 0; 0 0 0 0 0 1 0; 0 0 0 0 0 0 1]. I then select all of my seeds and enter into results explorer, which shows that there is no significant effect at all.
However, if while in the second level results window, I select the simple main effect of each of my behavioural scores separately i.e. [0 1], and all of my seeds, and then enter into results explorer, I get an abundance of significant values for each behavioural score.
My experimental question is: is there an association between FC and each of my behavioural scores - does this mean that I can look at the simple main effects even if the overall F test was not significant? If so, how can I explain this statistically? I know there is some practice of looking at simple main effects despite overall ANOVA being insignificant and I am wondering if that applies here?
Thank you,
Jenna
I am running a within-group bivariate correlation analysis looking at the association between functional connectivity and 6 different behavioural scores, using Conn V.17. I have a question regarding my findings:
When I set up the findings to look for any effect among my behavioural variables 1-6, I select my patient group and the six variables and set up the contrast as so: [0 1 0 0 0 0 0; 0 0 1 0 0 0 0; 0 0 0 1 0 0 0; 0 0 0 0 1 0 0; 0 0 0 0 0 1 0; 0 0 0 0 0 0 1]. I then select all of my seeds and enter into results explorer, which shows that there is no significant effect at all.
However, if while in the second level results window, I select the simple main effect of each of my behavioural scores separately i.e. [0 1], and all of my seeds, and then enter into results explorer, I get an abundance of significant values for each behavioural score.
My experimental question is: is there an association between FC and each of my behavioural scores - does this mean that I can look at the simple main effects even if the overall F test was not significant? If so, how can I explain this statistically? I know there is some practice of looking at simple main effects despite overall ANOVA being insignificant and I am wondering if that applies here?
Thank you,
Jenna
Threaded View
Title | Author | Date |
---|---|---|
Jenna Traynor | Feb 22, 2017 | |
Alfonso Nieto-Castanon | Mar 1, 2017 | |
Jenna Traynor | Mar 5, 2017 | |
Alfonso Nieto-Castanon | Mar 7, 2017 | |
Jenna Traynor | Mar 13, 2017 | |