> caffeine1.rbd <- aov(endtime ~ subject_c + dose_c)
> 
> summary(caffeine1.rbd)
            Df Sum Sq Mean Sq F value    Pr(>F)    
subject_c    8 5558.0  694.75 13.2159 4.174e-07 ***
dose_c       3  933.1  311.04  5.9168  0.003591 ** 
Residuals   24 1261.7   52.57                      
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 
> 
> TukeyHSD(caffeine1.rbd, "dose_c")
  Tukey multiple comparisons of means
    95% family-wise confidence level

Fit: aov(formula = endtime ~ subject_c + dose_c)

$dose_c
           diff       lwr       upr     p adj
5-0  11.2366667  1.808030 20.665303 0.0153292
9-0  12.2411111  2.812474 21.669748 0.0076616
13-0 11.7088889  2.280252 21.137526 0.0110929
9-5   1.0044444 -8.424192 10.433081 0.9909369
13-5  0.4722222 -8.956414  9.900859 0.9990313
13-9 -0.5322222 -9.960859  8.896414 0.9986162

> 
> interaction.plot(subject_c,dose_c,endtime)
> 
> caffeine.res <- residuals(caffeine1.rbd)
> 
> hist(caffeine.res)
> 
> friedman.test(endtime ~ dose_c | subject_c)

        Friedman rank sum test

data:  endtime and dose_c and subject_c 
Friedman chi-squared = 14.2, df = 3, p-value = 0.002645