Biostatistics II - Exercise Day 1

Data analyse

# load pakken der kan indlæse statafiler (pakken hedder foreign)

library(foreign)
data<- read.dta(file.path("C:/Users/Vision/Desktop/STATISTIK/BIOSTATISTIK2/repeatedANOVA7-1.dta"))

#øvrige pakker
#install.packages("ez")
library(ez)
library(reshape2)
library(ggplot2)

# her er der af en eller anden grund en ekstra datalinje uden data (na) -  denne fjernes med komandoen na.omit
data <- na.omit(data)

# nu skal tiden laves om til hastighed i stedet for disse tilføjes som nye kolonner i det eksisterende datasæt

data$FNspeed<-(1000/(data$FormNormal)) 
data$FCspeed<-(1000/(data$FormCongruent))
data$FIspeed<-(1000/(data$FormIncong)) 
data$CNspeed<-(1000/(data$ColorNormal)) 
data$CCspeed<-(1000/(data$ColorCong)) 
data$CIspeed<-(1000/(data$FormIncong)) 

# Nu deles datasættet op i 3 datasæt:

data.orig<-data[,c(1,2,3,4,5,6,7,8)]
data.log<-data[,c(1,2,9,10,11,12,13,14)]
data.speed<-data[,c(1,2,15,16,17,18,19,20)]

# Data skal være i long format  - så de skal konverteres

data.orig.long <- melt(data=data.orig, id.vars=c("subject", "field"))
data.log.long <- melt(data=data.log, id.vars=c("subject", "field"))
data.speed.long <- melt(data=data.speed, id.vars=c("subject", "field"))

# Lige lidt ændring i kolonnenavne

colnames(data.orig.long) <- c("subject", "field", "strooptype", "value")
colnames(data.log.long) <- c("subject", "field", "strooptype", "value")
colnames(data.speed.long) <- c("subject", "field", "strooptype", "speed")
#print first 10 rows of data.orig.long
head(data.orig.long, n=10)

   subject             field strooptype value
1        1 Field-Independent FormNormal   191
2        2 Field-Independent FormNormal   175
3        3 Field-Independent FormNormal   166
4        4 Field-Independent FormNormal   206
5        5 Field-Independent FormNormal   179
6        6 Field-Independent FormNormal   183
7        7 Field-Independent FormNormal   174
8        8 Field-Independent FormNormal   185
9        9 Field-Independent FormNormal   182
10      10 Field-Independent FormNormal   191

#print first 10 rows of data.log.long
head(data.log.long, n=10)

   subject             field strooptype    value
1        1 Field-Independent       lnFN 4.043051
2        2 Field-Independent       lnFN 3.713572
3        3 Field-Independent       lnFN 3.465736
4        4 Field-Independent       lnFN 4.276666
5        5 Field-Independent       lnFN 3.806662
6        6 Field-Independent       lnFN 3.891820
7        7 Field-Independent       lnFN 3.688879
8        8 Field-Independent       lnFN 3.931826
9        9 Field-Independent       lnFN 3.871201
10      10 Field-Independent       lnFN 4.043051

#print first 10 rows of data.speed.long
head(data.speed.long, n=10)

   subject             field strooptype    speed
1        1 Field-Independent    FNspeed 5.235602
2        2 Field-Independent    FNspeed 5.714286
3        3 Field-Independent    FNspeed 6.024096
4        4 Field-Independent    FNspeed 4.854369
5        5 Field-Independent    FNspeed 5.586592
6        6 Field-Independent    FNspeed 5.464481
7        7 Field-Independent    FNspeed 5.747126
8        8 Field-Independent    FNspeed 5.405405
9        9 Field-Independent    FNspeed 5.494505
10      10 Field-Independent    FNspeed 5.235602

# visualisering

ggplot(data=data.orig.long, aes(x=strooptype, y= value, fill=strooptype)) + 
  geom_boxplot() + 
  facet_grid(.~field) +
  theme(axis.text.x = element_text(angle = 90, hjust = 1)) +
  ggtitle("Oprindelige Data")

ggplot(data=data.log.long, aes(x=strooptype, y= value, fill=strooptype)) + 
  geom_boxplot() + 
  facet_grid(.~field) +
  theme(axis.text.x = element_text(angle = 90, hjust = 1)) + 
  ggtitle("Log Transformerede Data")

ggplot(data=data.speed.long, aes(x=strooptype, y= speed, fill=strooptype)) + 
  geom_boxplot() + 
  facet_grid(.~field) +
  theme(axis.text.x = element_text(angle = 90, hjust = 1)) +
  ggtitle("Speed-konverterede Data")

Visuelt ser det ud som om konvertering til speed fungerer lige så godt som logtransformationen

# Så skal der laves en RM-ANOVA test på hvert af de sidste to datasæt:

anova1 <- ezANOVA(data = data.log.long,
            dv = value,               #Dependent variable
            wid = subject,            #Identifikation
            between = field,          #Between subject variables
            within = strooptype,      #Within subject variables
            detailed = TRUE)          #Øger detaljegraden af output

Warning: Converting "subject" to factor for ANOVA.

anova2 <- ezANOVA(data = data.speed.long,
            dv = speed,
            wid = subject,
            between = field,
            within = strooptype,
            detailed = TRUE)

Warning: Converting "subject" to factor for ANOVA.

???