for DV, higher numbers = more green choices

summary(aov(DV ~ default, data=data))
##              Df Sum Sq Mean Sq F value  Pr(>F)   
## default       2   58.6   29.31   6.881 0.00121 **
## Residuals   279 1188.5    4.26                   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
br<-subset(data, default=="brown")
gr<-subset(data, default=="green")
no<-subset(data, default=="none")

#default manipulation worked
t.test(br$DV, no$DV)
## 
##  Welch Two Sample t-test
## 
## data:  br$DV and no$DV
## t = -1.8631, df = 177, p-value = 0.0641
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -1.22822966  0.03532606
## sample estimates:
## mean of x mean of y 
##  5.347368  5.943820
t.test(gr$DV, no$DV)
## 
##  Welch Two Sample t-test
## 
## data:  gr$DV and no$DV
## t = 1.8517, df = 180.97, p-value = 0.0657
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -0.03314852  1.04346725
## sample estimates:
## mean of x mean of y 
##   6.44898   5.94382
t.test(gr$DV, br$DV)
## 
##  Welch Two Sample t-test
## 
## data:  gr$DV and br$DV
## t = 3.5738, df = 174.15, p-value = 0.0004553
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  0.4932291 1.7099932
## sample estimates:
## mean of x mean of y 
##  6.448980  5.347368

and the lay theory manipulations worked - the manipulation checks were affected as expected

change<-subset(data, changecheck==2)
stable<-subset(data, stablecheck==1)
t.test(change$MC, stable$MC)
## 
##  Welch Two Sample t-test
## 
## data:  change$MC and stable$MC
## t = -7.8373, df = 277.45, p-value = 9.893e-14
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -0.6336167 -0.3792173
## sample estimates:
## mean of x mean of y 
##  3.119737  3.626154
t.test(change$certain, stable$certain)  
## 
##  Welch Two Sample t-test
## 
## data:  change$certain and stable$certain
## t = -2.0067, df = 280, p-value = 0.04574
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -0.46479033 -0.00447418
## sample estimates:
## mean of x mean of y 
##  5.674342  5.908974

default condition and lay theory manipulation interact in the opposite way we expected

summary(aov(DV ~ lay*default, data=data))
##              Df Sum Sq Mean Sq F value  Pr(>F)    
## lay           1    6.1   6.058   1.444 0.23053    
## default       2   61.9  30.930   7.373 0.00076 ***
## lay:default   2   21.3  10.662   2.542 0.08058 .  
## Residuals   276 1157.9   4.195                    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#default condition was effective for people who read that preferences are stable
summary(aov(DV ~ default, data=stable))
##              Df Sum Sq Mean Sq F value   Pr(>F)    
## default       2   75.6   37.81   8.429 0.000365 ***
## Residuals   127  569.6    4.49                     
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#but not for people who read that preferences are malleable
summary(aov(DV ~ default, data=change))
##              Df Sum Sq Mean Sq F value Pr(>F)
## default       2    7.6   3.787   0.959  0.386
## Residuals   149  588.3   3.948

but if we use measured preference certainty, then we get the pattern that we expected

summary(aov(DV ~ certain*default, data=data))
##                  Df Sum Sq Mean Sq F value   Pr(>F)    
## certain           1   58.5   58.48  14.672 0.000158 ***
## default           2   55.3   27.66   6.939 0.001147 ** 
## certain:default   2   33.3   16.66   4.181 0.016266 *  
## Residuals       276 1100.0    3.99                     
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
summary(data$certain)
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   1.833   5.167   5.833   5.783   6.667   7.000
certain<-subset(data, certain>5.8)
uncertain<-subset(data, certain<=5.8)

#no default effect when preferences are certain
summary(aov(DV~default, data=certain))
##              Df Sum Sq Mean Sq F value Pr(>F)
## default       2   15.8   7.891   1.755  0.176
## Residuals   158  710.4   4.496
#but there is an effect when they are uncertain
summary(aov(DV~default, data=uncertain))
##              Df Sum Sq Mean Sq F value   Pr(>F)    
## default       2   64.6   32.28   9.789 0.000117 ***
## Residuals   118  389.2    3.30                     
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1