harold negative bold comparison for Friday review

harold_voxPropNeg_thresh.Rsq95pct.md

Mean proportion of negative bold voxels is greater in young group than oldOld group.

NB: polling only suprathreshold voxels where rsq > (.95 * participantRsqMax)

1. Significant (p=0.03) one-way ANOVA comparing group means of voxPropNeg:

---

            Df Sum Sq Mean Sq F value Pr(>F)  
GROUP        2 0.6191 0.30953   3.939 0.0303 *
Residuals   30 2.3574 0.07858  


              mean        sd data:n
young    0.5488425 0.2602795     11
youngOld 0.4405093 0.2296574     11
oldOld   0.2196954 0.3394930     11

2. In ANOVA decomposition, young and oldOlD groups are significantly different (p=0.02):

---

Multiple Comparisons of Means: Tukey Contrasts

Fit: aov(formula = VOXPROPNEG ~ GROUP, data = top.95pct)

Linear Hypotheses:
                       Estimate Std. Error t value Pr(>|t|)  
youngOld - young == 0   -0.1083     0.1195  -0.906   0.6405  
oldOld - young == 0     -0.3291     0.1195  -2.754   0.0261 *
oldOld - youngOld == 0  -0.2208     0.1195  -1.847   0.1716  

corresponding 95% family-wise confidence level:

                       Estimate lwr      upr     
youngOld - young == 0  -0.10833 -0.40302  0.18636
oldOld - young == 0    -0.32915 -0.62384 -0.03446
oldOld - youngOld == 0 -0.22081 -0.51550  0.07388

3. Group-wise descriptive statistics for voxPropNeg:

---

              mean        sd        cv   skewness    kurtosis data:n
young    0.5488425 0.2602795 0.4742335 -0.1323639 -1.17932039     11
youngOld 0.4405093 0.2296574 0.5213451  0.7427953  0.50421132     11
oldOld   0.2196954 0.3394930 1.5452892  1.3390400 -0.01180097     11