0todd0000
commented
8 years ago Hi, I'll answer each point below...
1: I wonder if my interpretation is correct based on the attached plot: Yb was significantly greater than yA at 8-11% and yA was significantly different than yB at 95-98%. Therefore, the null hypothesis was rejected.
That's correct. To expand on the reason for rejection the following might work: the null hypothesis was rejected because the t statistic field traversed the critical threshold at alpha=0.05.
2: I want to know if the area above the mean (Zero) is related to yA and if the SPM[t] passes the t-critical then the yA is significantly different than the yB.
The t statistic is computed basically as (meanA - meanB) / variance, so areas above zero imply (meanA > mean B), and areas below zero imply the opposite: (meanA < meanB). If SPM{t} crosses either threshold you can conclude significant difference.
3: Is that determined based on which value comes first when they are placed in SPM1 = spm1d.stats.ttest2(yA, yB); ???
Yes, the sign of the t statistic is the same as the sign of (meanA - meanB). Thus spm1d.stats.ttest2(yA, yB) and spm1d.stats.ttest2(yB, yA) will yield opposite SPM{t} curves.
4: How could I make the red dotted line (t-critical) thicker??
The easiest way is probably to use the handles returned by the "plot" function like this:
handles = SPMi1.plot(); % 1x5 graphics array (i.e. array of handles)
set(handles(3), 'linewidth', 5) %upper critical threshold
set(handles(4), 'linewidth', 10) %lower critical thresholdThe "plot" returns handles to the main plotted elements including (in approximate order):
- SPM{t}
- test statistic field, the zero datum line
- upper critical threshold
- lower critical threshold, and
- supra-threshold cluster patches.
The order of elements 1-3 won't change, but elements 4 and 5 will change depending on: (a) whether or not two-tailed inference is conducted, and (b) whether or not there are any supra-threshold clusters.
PouyanMehryar
m-a-robinson
Hi Todd,
I used the below code to investigate the difference between the two variables SPM1 = spm1d.stats.ttest2(yA, yB);
SPMi1 = SPM1.inference(0.05, 'two_tailed', true); disp(SPMi1)
The graph below shows the SPM and SPMi1 results of yA and yB.
1: I wonder if my interpretation is correct based on the attached plot: Yb was significantly greater than yA at 8-11% and yA was significantly different than yB at 95-98%. Therefore, the null hypothesis was rejected.
2: I want to know if the area above the mean (Zero) is related to yA and if the SPM[t] passes the t-critical then the yA is significantly different than the yB.
3: Is that determined based on which value comes first when they are placed in SPM1 = spm1d.stats.ttest2(yA, yB); ???
4: How could I make the red dotted line (t-critical) thicker??
Thank you in advance for your response.