Skip to content

Latest commit

 

History

History
795 lines (659 loc) · 28.2 KB

calculatePR_TE_H.md

File metadata and controls

795 lines (659 loc) · 28.2 KB

Calculate Page rank, Target Entropy, and Hide

Jesper Bruun & Adrienne Traxleer 3/28/2020

Required functions

#required functions
source("functions/tarEnt.r")
## 
## Attaching package: 'igraph'

## The following objects are masked from 'package:stats':
## 
##     decompose, spectrum

## The following object is masked from 'package:base':
## 
##     union
source("functions/searchInf.r")

Calculating Page rank

#pagerank
accPS_PR<-lapply(accPS,page.rank)
accCD_PR<-lapply(accCD,page.rank)
accICS_PR<-lapply(accICS,page.rank)
singlePS_PR<-lapply(weeksPS,page.rank)
singleCD_PR<-lapply(weeksCD,page.rank)
singleICS_PR<-lapply(weeksICS,page.rank)

Calculating Target Entropy

#target entropy
accPS_TE<-lapply(accPS,TargetEntropy)
accCD_TE<-lapply(accCD,TargetEntropy)
accICS_TE<-lapply(accICS,TargetEntropy)
singlePS_TE<-lapply(weeksPS,TargetEntropy)
singleCD_TE<-lapply(weeksCD,TargetEntropy)
singleICS_TE<-lapply(weeksICS,TargetEntropy)

Calculating Hide

#hide
accPS_S<-lapply(accPS,sInfMatrix)
accCD_S<-lapply(accCD,sInfMatrix)
accICS_S<-lapply(accICS,sInfMatrix)
singlePS_S<-lapply(weeksPS,sInfMatrix)
singleCD_S<-lapply(weeksCD,sInfMatrix)
singleICS_S<-lapply(weeksICS,sInfMatrix)

accPS_H<-lapply(accPS_S,colSums,na.rm=T)
accCD_H<-lapply(accCD_S,colSums,na.rm=T)
accICS_H<-lapply(accICS_S,colSums,na.rm=T)
singlePS_H<-lapply(singlePS_S,colSums,na.rm=T)
singleCD_H<-lapply(singleCD_S,colSums,na.rm=T)
singleICS_H<-lapply(singleICS_S,colSums,na.rm=T)

.RData file with calculations done

THE PRTEH.RData file contains all calculations Here, we load the file and make som comparisons.

load("data/PRTEH.RData")

We want to figure out whether there are differences in passing and failing that may be connected to other attributes. We use wilcoxon tests per default, since we do not assume distributions to be normal. There seems to be no gender differences.

wilcox.test(V(accPS[[1]])$grade[V(accPS[[1]])$gender==1],V(accPS[[1]])$grade[V(accPS[[1]])$gender==0])
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  V(accPS[[1]])$grade[V(accPS[[1]])$gender == 1] and V(accPS[[1]])$grade[V(accPS[[1]])$gender == 0]
## W = 2595, p-value = 0.1538
## alternative hypothesis: true location shift is not equal to 0
wilcox.test(V(accPS[[1]])$pass[V(accPS[[1]])$gender==1],V(accPS[[1]])$pass[V(accPS[[1]])$gender==0])
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  V(accPS[[1]])$pass[V(accPS[[1]])$gender == 1] and V(accPS[[1]])$pass[V(accPS[[1]])$gender == 0]
## W = 2594, p-value = 0.05464
## alternative hypothesis: true location shift is not equal to 0
wilcox.test(V(accPS[[1]])$justpass[V(accPS[[1]])$gender==1],V(accPS[[1]])$justpass[V(accPS[[1]])$gender==0])
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  V(accPS[[1]])$justpass[V(accPS[[1]])$gender == 1] and V(accPS[[1]])$justpass[V(accPS[[1]])$gender == 0]
## W = 485.5, p-value = 0.185
## alternative hypothesis: true location shift is not equal to 0

FCI NAs excluded

par(mfrow=c(2,2))
plot(table(V(accPS[[1]])$fci_pre[V(accPS[[1]])$pass==1]),ylab="N",xlab="score",main="FCI pre of passing students")
plot(table(V(accPS[[1]])$fci_pre[V(accPS[[1]])$pass==0]),ylab="N",xlab="score",main="FCI pre of failing students")
plot(table(V(accPS[[1]])$fci_pre[V(accPS[[1]])$justpass==1]),ylab="N",xlab="score",main="FCI pre of just passing students")
plot(table(V(accPS[[1]])$fci_pre[V(accPS[[1]])$justpass==0]),ylab="N",xlab="score",main="FCI pre of just failing students")

### FCI impute 0 if NA

par(mfrow=c(2,2))
plot(table(V(accPS[[1]])$fci_pre_0[V(accPS[[1]])$pass==1]),ylab="N",xlab="score",main="FCI pre of passing students")
plot(table(V(accPS[[1]])$fci_pre_0[V(accPS[[1]])$pass==0]),ylab="N",xlab="score",main="FCI pre of failing students")
plot(table(V(accPS[[1]])$fci_pre_0[V(accPS[[1]])$justpass==1]),ylab="N",xlab="score",main="FCI pre of just passing students")
plot(table(V(accPS[[1]])$fci_pre_0[V(accPS[[1]])$justpass==0]),ylab="N",xlab="score",main="FCI pre of just failing students")

### FCI impute random score selected from others with same grade if NA

par(mfrow=c(2,2))
plot(table(V(accPS[[1]])$fci_pre_s[V(accPS[[1]])$pass==1]),ylab="N",xlab="score",main="FCI pre of passing students")
plot(table(V(accPS[[1]])$fci_pre_s[V(accPS[[1]])$pass==0]),ylab="N",xlab="score",main="FCI pre of failing students")
plot(table(V(accPS[[1]])$fci_pre_s[V(accPS[[1]])$justpass==1]),ylab="N",xlab="score",main="FCI pre of just passing students")
plot(table(V(accPS[[1]])$fci_pre_s[V(accPS[[1]])$justpass==0]),ylab="N",xlab="score",main="FCI pre of just failing students")

FCI categories

par(mfrow=c(2,2))
plot(table(V(accPS[[1]])$fci_pre_c[V(accPS[[1]])$pass==1]),ylab="N",xlab="score",main="FCI pre of passing students")
plot(table(V(accPS[[1]])$fci_pre_c[V(accPS[[1]])$pass==0]),ylab="N",xlab="score",main="FCI pre of failing students")
plot(table(V(accPS[[1]])$fci_pre_c[V(accPS[[1]])$justpass==1]),ylab="N",xlab="score",main="FCI pre of just passing students")
plot(table(V(accPS[[1]])$fci_pre_c[V(accPS[[1]])$justpass==0]),ylab="N",xlab="score",main="FCI pre of just failing students")

Section number

par(mfrow=c(1,2))
c1<-table(V(accPS[[1]])$pass[V(accPS[[1]])$cohort==1])/length(which(V(accPS[[1]])$cohort==1))
c2<-table(V(accPS[[1]])$pass[V(accPS[[1]])$cohort==2])/length(which(V(accPS[[1]])$cohort==2))
c3<-table(V(accPS[[1]])$pass[V(accPS[[1]])$cohort==3])/length(which(V(accPS[[1]])$cohort==3))
c4<-table(V(accPS[[1]])$pass[V(accPS[[1]])$cohort==4])/length(which(V(accPS[[1]])$cohort==4))
c5<-table(V(accPS[[1]])$pass[V(accPS[[1]])$cohort==5])/length(which(V(accPS[[1]])$cohort==5))
c6<-table(V(accPS[[1]])$pass[V(accPS[[1]])$cohort==6])/length(which(V(accPS[[1]])$cohort==6))
c10<-table(V(accPS[[1]])$pass[V(accPS[[1]])$cohort==10])/length(which(V(accPS[[1]])$cohort==10))
cohfail<-c(c1[1],c2[1],c3[1],c4[1],c5[1],c6[1],c10[1])
cohpass<-c(c1[2],c2[2],c3[2],c4[2],c5[2],c6[2],c10[2])
plot(cohpass,xlab="Section number",ylab="Fraction",ylim=c(0,1),main="Passing and failing per section",col="darkgreen")
points(cohfail,col="red")
cj1<-table(V(accPS[[1]])$justpass[V(accPS[[1]])$cohort==1])/length(which(V(accPS[[1]])$cohort==1))
cj2<-table(V(accPS[[1]])$justpass[V(accPS[[1]])$cohort==2])/length(which(V(accPS[[1]])$cohort==2))
cj3<-table(V(accPS[[1]])$justpass[V(accPS[[1]])$cohort==3])/length(which(V(accPS[[1]])$cohort==3))
cj4<-table(V(accPS[[1]])$justpass[V(accPS[[1]])$cohort==4])/length(which(V(accPS[[1]])$cohort==4))
cj5<-table(V(accPS[[1]])$justpass[V(accPS[[1]])$cohort==5])/length(which(V(accPS[[1]])$cohort==5))
cj6<-table(V(accPS[[1]])$justpass[V(accPS[[1]])$cohort==6])/length(which(V(accPS[[1]])$cohort==6))
cj10<-table(V(accPS[[1]])$justpass[V(accPS[[1]])$cohort==10])/length(which(V(accPS[[1]])$cohort==10))
cohjfail<-c(cj1[1],cj2[1],cj3[1],cj4[1],cj5[1],cj6[1],cj10[1])
cohjpass<-c(cj1[2],cj2[2],cj3[2],cj4[2],cj5[2],cj6[2],cj10[2])
plot(cohjpass,xlab="Section number",ylab="Fraction of whole section",ylim=c(0,1),main="Just passing vs just failing",col="darkgreen")
points(cohjfail,col="red")

## Problem solving layer ### Page rank difference

par(mfrow=c(2,2))
hist(accPS_PR[[7]]$vector[V(accPS[[7]])$pass==1],main="Pagerank of passing")
hist(accPS_PR[[7]]$vector[V(accPS[[7]])$pass==0],main="Pagerank of failing")
hist(accPS_PR[[7]]$vector[V(accPS[[7]])$justpass==1],main="Pagerank of just passing")
hist(accPS_PR[[7]]$vector[V(accPS[[7]])$justpass==0],main="Pagerank of just failing")

t.test(accPS_PR[[7]]$vector[V(accPS[[7]])$pass==1],accPS_PR[[7]]$vector[V(accPS[[7]])$pass==0])
## 
##  Welch Two Sample t-test
## 
## data:  accPS_PR[[7]]$vector[V(accPS[[7]])$pass == 1] and accPS_PR[[7]]$vector[V(accPS[[7]])$pass == 0]
## t = 4.2921, df = 82.187, p-value = 4.807e-05
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  0.001165521 0.003179134
## sample estimates:
##   mean of x   mean of y 
## 0.006521376 0.004349049
wilcox.test(accPS_PR[[7]]$vector[V(accPS[[7]])$pass==1],accPS_PR[[7]]$vector[V(accPS[[7]])$pass==0])
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  accPS_PR[[7]]$vector[V(accPS[[7]])$pass == 1] and accPS_PR[[7]]$vector[V(accPS[[7]])$pass == 0]
## W = 3359, p-value = 0.0003693
## alternative hypothesis: true location shift is not equal to 0
t.test(accPS_PR[[7]]$vector[V(accPS[[7]])$justpass==1],accPS_PR[[7]]$vector[V(accPS[[7]])$justpass==0])
## 
##  Welch Two Sample t-test
## 
## data:  accPS_PR[[7]]$vector[V(accPS[[7]])$justpass == 1] and accPS_PR[[7]]$vector[V(accPS[[7]])$justpass == 0]
## t = 2.0144, df = 62.863, p-value = 0.04825
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  1.086488e-05 2.730883e-03
## sample estimates:
##   mean of x   mean of y 
## 0.006243573 0.004872699
wilcox.test(accPS_PR[[7]]$vector[V(accPS[[7]])$justpass==1],accPS_PR[[7]]$vector[V(accPS[[7]])$justpass==0])
## 
##  Wilcoxon rank sum test
## 
## data:  accPS_PR[[7]]$vector[V(accPS[[7]])$justpass == 1] and accPS_PR[[7]]$vector[V(accPS[[7]])$justpass == 0]
## W = 690, p-value = 0.0679
## alternative hypothesis: true location shift is not equal to 0
par(mfrow=c(2,2))
hist(accPS_TE[[7]][V(accPS[[7]])$pass==1],main="Target Entropy of passing")
hist(accPS_TE[[7]][V(accPS[[7]])$pass==0],main="Target Entropy of failing")
hist(accPS_TE[[7]][V(accPS[[7]])$justpass==1],main="Target Entropy of just passing")
hist(accPS_TE[[7]][V(accPS[[7]])$justpass==0],main="Target Entropy of just failing")

t.test(accPS_TE[[7]][V(accPS[[7]])$pass==1],accPS_TE[[7]][V(accPS[[7]])$pass==0])
## 
##  Welch Two Sample t-test
## 
## data:  accPS_TE[[7]][V(accPS[[7]])$pass == 1] and accPS_TE[[7]][V(accPS[[7]])$pass == 0]
## t = 3.6327, df = 69.377, p-value = 0.0005333
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  0.2800503 0.9621424
## sample estimates:
## mean of x mean of y 
##  2.344856  1.723759
wilcox.test(accPS_TE[[7]][V(accPS[[7]])$pass==1],accPS_TE[[7]][V(accPS[[7]])$pass==0])
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  accPS_TE[[7]][V(accPS[[7]])$pass == 1] and accPS_TE[[7]][V(accPS[[7]])$pass == 0]
## W = 3378, p-value = 0.0002788
## alternative hypothesis: true location shift is not equal to 0
t.test(accPS_TE[[7]][V(accPS[[7]])$justpass==1],accPS_TE[[7]][V(accPS[[7]])$justpass==0])
## 
##  Welch Two Sample t-test
## 
## data:  accPS_TE[[7]][V(accPS[[7]])$justpass == 1] and accPS_TE[[7]][V(accPS[[7]])$justpass == 0]
## t = 1.8215, df = 62.265, p-value = 0.07333
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -0.04025175  0.86753459
## sample estimates:
## mean of x mean of y 
##  2.327374  1.913732
wilcox.test(accPS_TE[[7]][V(accPS[[7]])$justpass==1],accPS_TE[[7]][V(accPS[[7]])$justpass==0])
## 
##  Wilcoxon rank sum test
## 
## data:  accPS_TE[[7]][V(accPS[[7]])$justpass == 1] and accPS_TE[[7]][V(accPS[[7]])$justpass == 0]
## W = 713, p-value = 0.03365
## alternative hypothesis: true location shift is not equal to 0
par(mfrow=c(2,2))
hist(accPS_H[[7]][V(accPS[[7]])$pass==1],main="Hide of passing")
hist(accPS_H[[7]][V(accPS[[7]])$pass==0],main="Hide of failing")
hist(accPS_H[[7]][V(accPS[[7]])$justpass==1],main="Hide of just passing")
hist(accPS_H[[7]][V(accPS[[7]])$justpass==0],main="Hide of just failing")

t.test(accPS_H[[7]][V(accPS[[7]])$pass==1],accPS_H[[7]][V(accPS[[7]])$pass==0])
## 
##  Welch Two Sample t-test
## 
## data:  accPS_H[[7]][V(accPS[[7]])$pass == 1] and accPS_H[[7]][V(accPS[[7]])$pass == 0]
## t = -3.3365, df = 87.884, p-value = 0.001245
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -292.77130  -74.19516
## sample estimates:
## mean of x mean of y 
##  1748.350  1931.833
wilcox.test(accPS_H[[7]][V(accPS[[7]])$pass==1],accPS_H[[7]][V(accPS[[7]])$pass==0])
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  accPS_H[[7]][V(accPS[[7]])$pass == 1] and accPS_H[[7]][V(accPS[[7]])$pass == 0]
## W = 1512, p-value = 0.0004091
## alternative hypothesis: true location shift is not equal to 0
t.test(accPS_H[[7]][V(accPS[[7]])$justpass==1],accPS_H[[7]][V(accPS[[7]])$justpass==0])
## 
##  Welch Two Sample t-test
## 
## data:  accPS_H[[7]][V(accPS[[7]])$justpass == 1] and accPS_H[[7]][V(accPS[[7]])$justpass == 0]
## t = -1.4266, df = 63.834, p-value = 0.1586
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -239.48379   39.95149
## sample estimates:
## mean of x mean of y 
##  1798.392  1898.158
wilcox.test(accPS_H[[7]][V(accPS[[7]])$justpass==1],accPS_H[[7]][V(accPS[[7]])$justpass==0])
## 
##  Wilcoxon rank sum test
## 
## data:  accPS_H[[7]][V(accPS[[7]])$justpass == 1] and accPS_H[[7]][V(accPS[[7]])$justpass == 0]
## W = 403, p-value = 0.06987
## alternative hypothesis: true location shift is not equal to 0

Concept Discussion layer

Page rank difference

par(mfrow=c(2,2))
hist(accCD_PR[[7]]$vector[V(accCD[[7]])$pass==1],main="Pagerank of passing")
hist(accCD_PR[[7]]$vector[V(accCD[[7]])$pass==0],main="Pagerank of failing")
hist(accCD_PR[[7]]$vector[V(accCD[[7]])$justpass==1],main="Pagerank of just passing")
hist(accCD_PR[[7]]$vector[V(accCD[[7]])$justpass==0],main="Pagerank of just failing")

t.test(accCD_PR[[7]]$vector[V(accCD[[7]])$pass==1],accCD_PR[[7]]$vector[V(accCD[[7]])$pass==0])
## 
##  Welch Two Sample t-test
## 
## data:  accCD_PR[[7]]$vector[V(accCD[[7]])$pass == 1] and accCD_PR[[7]]$vector[V(accCD[[7]])$pass == 0]
## t = 4.3191, df = 124.76, p-value = 3.164e-05
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  0.001339729 0.003606025
## sample estimates:
##   mean of x   mean of y 
## 0.006590177 0.004117300
wilcox.test(accCD_PR[[7]]$vector[V(accCD[[7]])$pass==1],accCD_PR[[7]]$vector[V(accCD[[7]])$pass==0])
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  accCD_PR[[7]]$vector[V(accCD[[7]])$pass == 1] and accCD_PR[[7]]$vector[V(accCD[[7]])$pass == 0]
## W = 3247, p-value = 0.00174
## alternative hypothesis: true location shift is not equal to 0
t.test(accCD_PR[[7]]$vector[V(accCD[[7]])$justpass==1],accCD_PR[[7]]$vector[V(accCD[[7]])$justpass==0])
## 
##  Welch Two Sample t-test
## 
## data:  accCD_PR[[7]]$vector[V(accCD[[7]])$justpass == 1] and accCD_PR[[7]]$vector[V(accCD[[7]])$justpass == 0]
## t = 2.5086, df = 62.886, p-value = 0.01471
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  0.0003636771 0.0032125096
## sample estimates:
##   mean of x   mean of y 
## 0.005891567 0.004103474
wilcox.test(accCD_PR[[7]]$vector[V(accCD[[7]])$justpass==1],accCD_PR[[7]]$vector[V(accCD[[7]])$justpass==0])
## Warning in wilcox.test.default(accCD_PR[[7]]$vector[V(accCD[[7]])$justpass == :
## cannot compute exact p-value with ties

## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  accCD_PR[[7]]$vector[V(accCD[[7]])$justpass == 1] and accCD_PR[[7]]$vector[V(accCD[[7]])$justpass == 0]
## W = 711, p-value = 0.03645
## alternative hypothesis: true location shift is not equal to 0
par(mfrow=c(2,2))
hist(accCD_TE[[7]][V(accCD[[7]])$pass==1],main="Target Entropy of passing")
hist(accCD_TE[[7]][V(accCD[[7]])$pass==0],main="Target Entropy of failing")
hist(accCD_TE[[7]][V(accCD[[7]])$justpass==1],main="Target Entropy of just passing")
hist(accCD_TE[[7]][V(accCD[[7]])$justpass==0],main="Target Entropy of just failing")

t.test(accCD_TE[[7]][V(accCD[[7]])$pass==1],accCD_TE[[7]][V(accCD[[7]])$pass==0])
## 
##  Welch Two Sample t-test
## 
## data:  accCD_TE[[7]][V(accCD[[7]])$pass == 1] and accCD_TE[[7]][V(accCD[[7]])$pass == 0]
## t = 3.6897, df = 79.541, p-value = 0.0004093
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  0.2419638 0.8086826
## sample estimates:
## mean of x mean of y 
## 1.4699234 0.9446003
wilcox.test(accCD_TE[[7]][V(accCD[[7]])$pass==1],accCD_TE[[7]][V(accCD[[7]])$pass==0])
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  accCD_TE[[7]][V(accCD[[7]])$pass == 1] and accCD_TE[[7]][V(accCD[[7]])$pass == 0]
## W = 3268, p-value = 0.00129
## alternative hypothesis: true location shift is not equal to 0
t.test(accCD_TE[[7]][V(accCD[[7]])$justpass==1],accCD_TE[[7]][V(accCD[[7]])$justpass==0])
## 
##  Welch Two Sample t-test
## 
## data:  accCD_TE[[7]][V(accCD[[7]])$justpass == 1] and accCD_TE[[7]][V(accCD[[7]])$justpass == 0]
## t = 1.7455, df = 64.773, p-value = 0.08564
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -0.04985913  0.74114727
## sample estimates:
## mean of x mean of y 
##  1.362747  1.017103
wilcox.test(accCD_TE[[7]][V(accCD[[7]])$justpass==1],accCD_TE[[7]][V(accCD[[7]])$justpass==0])
## Warning in wilcox.test.default(accCD_TE[[7]][V(accCD[[7]])$justpass == 1], :
## cannot compute exact p-value with ties

## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  accCD_TE[[7]][V(accCD[[7]])$justpass == 1] and accCD_TE[[7]][V(accCD[[7]])$justpass == 0]
## W = 675.5, p-value = 0.1006
## alternative hypothesis: true location shift is not equal to 0
par(mfrow=c(2,2))
hist(accCD_H[[7]][V(accCD[[7]])$pass==1],main="Hide of passing")
hist(accCD_H[[7]][V(accCD[[7]])$pass==0],main="Hide of failing")
hist(accCD_H[[7]][V(accCD[[7]])$justpass==1],main="Hide of just passing")
hist(accCD_H[[7]][V(accCD[[7]])$justpass==0],main="Hide of just failing")

t.test(accCD_H[[7]][V(accCD[[7]])$pass==1],accCD_H[[7]][V(accCD[[7]])$pass==0])
## 
##  Welch Two Sample t-test
## 
## data:  accCD_H[[7]][V(accCD[[7]])$pass == 1] and accCD_H[[7]][V(accCD[[7]])$pass == 0]
## t = -0.31999, df = 50.891, p-value = 0.7503
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -330.6557  239.7433
## sample estimates:
## mean of x mean of y 
##  1544.629  1590.085
wilcox.test(accCD_H[[7]][V(accCD[[7]])$pass==1],accCD_H[[7]][V(accCD[[7]])$pass==0])
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  accCD_H[[7]][V(accCD[[7]])$pass == 1] and accCD_H[[7]][V(accCD[[7]])$pass == 0]
## W = 1960, p-value = 0.06988
## alternative hypothesis: true location shift is not equal to 0
t.test(accCD_H[[7]][V(accCD[[7]])$justpass==1],accCD_H[[7]][V(accCD[[7]])$justpass==0])
## 
##  Welch Two Sample t-test
## 
## data:  accCD_H[[7]][V(accCD[[7]])$justpass == 1] and accCD_H[[7]][V(accCD[[7]])$justpass == 0]
## t = -1.0346, df = 53.214, p-value = 0.3055
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -543.4374  173.5570
## sample estimates:
## mean of x mean of y 
##  1448.344  1633.284
wilcox.test(accCD_H[[7]][V(accCD[[7]])$justpass==1],accCD_H[[7]][V(accCD[[7]])$justpass==0])
## Warning in wilcox.test.default(accCD_H[[7]][V(accCD[[7]])$justpass == 1], :
## cannot compute exact p-value with ties

## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  accCD_H[[7]][V(accCD[[7]])$justpass == 1] and accCD_H[[7]][V(accCD[[7]])$justpass == 0]
## W = 350, p-value = 0.01291
## alternative hypothesis: true location shift is not equal to 0

In Class Social layer

Page rank difference

par(mfrow=c(2,2))
hist(accICS_PR[[7]]$vector[V(accICS[[7]])$pass==1],main="Pagerank of passing")
hist(accICS_PR[[7]]$vector[V(accICS[[7]])$pass==0],main="Pagerank of failing")
hist(accICS_PR[[7]]$vector[V(accICS[[7]])$justpass==1],main="Pagerank of just passing")
hist(accICS_PR[[7]]$vector[V(accICS[[7]])$justpass==0],main="Pagerank of just failing")

t.test(accICS_PR[[7]]$vector[V(accICS[[7]])$pass==1],accICS_PR[[7]]$vector[V(accICS[[7]])$pass==0])
## 
##  Welch Two Sample t-test
## 
## data:  accICS_PR[[7]]$vector[V(accICS[[7]])$pass == 1] and accICS_PR[[7]]$vector[V(accICS[[7]])$pass == 0]
## t = 5.8955, df = 126.33, p-value = 3.195e-08
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  0.001804640 0.003628276
## sample estimates:
##   mean of x   mean of y 
## 0.006645936 0.003929478
wilcox.test(accICS_PR[[7]]$vector[V(accICS[[7]])$pass==1],accICS_PR[[7]]$vector[V(accICS[[7]])$pass==0])
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  accICS_PR[[7]]$vector[V(accICS[[7]])$pass == 1] and accICS_PR[[7]]$vector[V(accICS[[7]])$pass == 0]
## W = 3508, p-value = 3.566e-05
## alternative hypothesis: true location shift is not equal to 0
t.test(accICS_PR[[7]]$vector[V(accICS[[7]])$justpass==1],accICS_PR[[7]]$vector[V(accICS[[7]])$justpass==0])
## 
##  Welch Two Sample t-test
## 
## data:  accICS_PR[[7]]$vector[V(accICS[[7]])$justpass == 1] and accICS_PR[[7]]$vector[V(accICS[[7]])$justpass == 0]
## t = 3.1029, df = 63.577, p-value = 0.002859
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  0.0007085704 0.0032712550
## sample estimates:
##   mean of x   mean of y 
## 0.006376039 0.004386126
wilcox.test(accICS_PR[[7]]$vector[V(accICS[[7]])$justpass==1],accICS_PR[[7]]$vector[V(accICS[[7]])$justpass==0])
## Warning in wilcox.test.default(accICS_PR[[7]]$vector[V(accICS[[7]])$justpass
## == : cannot compute exact p-value with ties

## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  accICS_PR[[7]]$vector[V(accICS[[7]])$justpass == 1] and accICS_PR[[7]]$vector[V(accICS[[7]])$justpass == 0]
## W = 759, p-value = 0.006903
## alternative hypothesis: true location shift is not equal to 0
par(mfrow=c(2,2))
hist(accICS_TE[[7]][V(accICS[[7]])$pass==1],main="Target Entropy of passing")
hist(accICS_TE[[7]][V(accICS[[7]])$pass==0],main="Target Entropy of failing")
hist(accICS_TE[[7]][V(accICS[[7]])$justpass==1],main="Target Entropy of just passing")
hist(accICS_TE[[7]][V(accICS[[7]])$justpass==0],main="Target Entropy of just failing")

t.test(accICS_TE[[7]][V(accICS[[7]])$pass==1],accICS_TE[[7]][V(accICS[[7]])$pass==0])
## 
##  Welch Two Sample t-test
## 
## data:  accICS_TE[[7]][V(accICS[[7]])$pass == 1] and accICS_TE[[7]][V(accICS[[7]])$pass == 0]
## t = 3.9962, df = 74.091, p-value = 0.0001505
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  0.3677683 1.0991996
## sample estimates:
## mean of x mean of y 
##  2.105811  1.372328
wilcox.test(accICS_TE[[7]][V(accICS[[7]])$pass==1],accICS_TE[[7]][V(accICS[[7]])$pass==0])
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  accICS_TE[[7]][V(accICS[[7]])$pass == 1] and accICS_TE[[7]][V(accICS[[7]])$pass == 0]
## W = 3398, p-value = 0.0002019
## alternative hypothesis: true location shift is not equal to 0
t.test(accICS_TE[[7]][V(accICS[[7]])$justpass==1],accICS_TE[[7]][V(accICS[[7]])$justpass==0])
## 
##  Welch Two Sample t-test
## 
## data:  accICS_TE[[7]][V(accICS[[7]])$justpass == 1] and accICS_TE[[7]][V(accICS[[7]])$justpass == 0]
## t = 2.5943, df = 58.27, p-value = 0.01197
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  0.1327624 1.0292943
## sample estimates:
## mean of x mean of y 
##  2.146499  1.565471
wilcox.test(accICS_TE[[7]][V(accICS[[7]])$justpass==1],accICS_TE[[7]][V(accICS[[7]])$justpass==0])
## Warning in wilcox.test.default(accICS_TE[[7]][V(accICS[[7]])$justpass == :
## cannot compute exact p-value with ties

## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  accICS_TE[[7]][V(accICS[[7]])$justpass == 1] and accICS_TE[[7]][V(accICS[[7]])$justpass == 0]
## W = 735, p-value = 0.01652
## alternative hypothesis: true location shift is not equal to 0
par(mfrow=c(2,2))
hist(accICS_H[[7]][V(accICS[[7]])$pass==1],main="Hide of passing")
hist(accICS_H[[7]][V(accICS[[7]])$pass==0],main="Hide of failing")
hist(accICS_H[[7]][V(accICS[[7]])$justpass==1],main="Hide of just passing")
hist(accICS_H[[7]][V(accICS[[7]])$justpass==0],main="Hide of just failing")

t.test(accICS_H[[7]][V(accICS[[7]])$pass==1],accICS_H[[7]][V(accICS[[7]])$pass==0])
## 
##  Welch Two Sample t-test
## 
## data:  accICS_H[[7]][V(accICS[[7]])$pass == 1] and accICS_H[[7]][V(accICS[[7]])$pass == 0]
## t = -2.7948, df = 74.151, p-value = 0.006607
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -376.2145  -63.0535
## sample estimates:
## mean of x mean of y 
##  1378.832  1598.466
wilcox.test(accICS_H[[7]][V(accICS[[7]])$pass==1],accICS_H[[7]][V(accICS[[7]])$pass==0])
## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  accICS_H[[7]][V(accICS[[7]])$pass == 1] and accICS_H[[7]][V(accICS[[7]])$pass == 0]
## W = 1487, p-value = 0.000283
## alternative hypothesis: true location shift is not equal to 0
t.test(accICS_H[[7]][V(accICS[[7]])$justpass==1],accICS_H[[7]][V(accICS[[7]])$justpass==0])
## 
##  Welch Two Sample t-test
## 
## data:  accICS_H[[7]][V(accICS[[7]])$justpass == 1] and accICS_H[[7]][V(accICS[[7]])$justpass == 0]
## t = -1.7922, df = 56.457, p-value = 0.07846
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -400.04102   22.21097
## sample estimates:
## mean of x mean of y 
##  1361.292  1550.207
wilcox.test(accICS_H[[7]][V(accICS[[7]])$justpass==1],accICS_H[[7]][V(accICS[[7]])$justpass==0])
## Warning in wilcox.test.default(accICS_H[[7]][V(accICS[[7]])$justpass == : cannot
## compute exact p-value with ties

## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  accICS_H[[7]][V(accICS[[7]])$justpass == 1] and accICS_H[[7]][V(accICS[[7]])$justpass == 0]
## W = 367, p-value = 0.02325
## alternative hypothesis: true location shift is not equal to 0