Assignment 8 VICTOR LEON
A) What are the correct hypothesis?
H0 = The mean in all groups are equal
HA = At least there is one difference amongst the means
library('asbio')## Warning: package 'asbio' was built under R version 3.4.2## Loading required package: tcltkdata(life.exp)
attach(life.exp)
#B) Calculate SSE and SSa by hand.
nn85mean <- mean(lifespan[ treatment == "N/N85"])
nr50mean <- mean(lifespan[ treatment == 'N/R50'])
rr50mean <- mean(lifespan[ treatment == 'R/R50'])
nr40mean <- mean(lifespan[ treatment == 'N/R40'])
nn85<- life.exp[treatment=='N/N85',1]
nr50<- life.exp[treatment=='N/R50',1]
rr50<- life.exp[treatment=='R/R50',1]
nr40<- life.exp[treatment=='N/R40',1]
allnn85 <- sum((nn85-nn85mean)^(2))
allnr50 <- sum((nr50-nr50mean)^(2))
allrr50 <- sum((rr50-rr50mean)^(2))
allnr40 <- sum((nr40-nr40mean)^(2))
k <- 4
df_between <- k-1
N<- length(nn85)+length(nr50)+length(rr50)+length(nr40)
df_within <- N-k
df_total <- df_between + df_within
all_sum<- sum(nn85,nr50,rr50,nr40)
grandmean <- all_sum/N
allnn85gm <- sum((nn85-grandmean)^(2))
allnr50gm <- sum((nr50-grandmean)^(2))
allrr50gm <- sum((rr50-grandmean)^(2))
allnr40gm <- sum((nr40-grandmean)^(2))
SST <- (allnn85gm+allnr50+allrr50gm+allnr40gm)
SSW <- (allnn85+allnr50+allrr50+allnr40)
SSB <- sum( ((nn85mean-grandmean)^2*length(nn85)),
((nr50mean-grandmean)^2*length(nr50)),
((rr50mean-grandmean)^2*length(rr50)),
((nr40mean-grandmean)^2*length(nr40)))
#Also SSB is SST-SSW
MSB <- SSB/df_between
MSW <- SSW/df_within
F <- MSB/MSW# (C)
Df<- cbind(df_between,df_within)
Mean_SQ <- cbind(MSB,MSW)
F_value<- F
P_value <- df(F,df_between,df_within)
table_a <- cbind(NULL,'treatment','Residuals')
table_big <- rbind(table_a,Df,Mean_SQ,F_value,P_value)# (D)
summary(life.exp)## lifespan treatment
## Min. :17.90 N/N85:57
## 1st Qu.:35.27 N/R40:60
## Median :42.30 N/R50:71
## Mean :40.88 R/R50:56
## 3rd Qu.:47.70
## Max. :54.60summary.lm(lm(lifespan~treatment,data=life.exp))##
## Call:
## lm(formula = lifespan ~ treatment, data = life.exp)
##
## Residuals:
## Min 1Q Median 3Q Max
## -25.5167 -2.9339 0.6615 5.2364 9.6088
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 32.6912 0.8886 36.788 < 2e-16 ***
## treatmentN/R40 12.4254 1.2409 10.013 < 2e-16 ***
## treatmentN/R50 9.6060 1.1932 8.051 3.82e-14 ***
## treatmentR/R50 10.1945 1.2623 8.076 3.25e-14 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 6.709 on 240 degrees of freedom
## Multiple R-squared: 0.3278, Adjusted R-squared: 0.3194
## F-statistic: 39 on 3 and 240 DF, p-value: < 2.2e-16anova(lm(lifespan~treatment,data=life.exp))## Analysis of Variance Table
##
## Response: lifespan
## Df Sum Sq Mean Sq F value Pr(>F)
## treatment 3 5267 1755.68 39.004 < 2.2e-16 ***
## Residuals 240 10803 45.01
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1# (E)
lmq <- lm(lifespan~treatment,data=life.exp)
plot(lmq, which=1:4)
bartlett.test(lifespan~treatment, data=life.exp)##
## Bartlett test of homogeneity of variances
##
## data: lifespan by treatment
## Bartlett's K-squared = 10.096, df = 3, p-value = 0.01777# (F)
tk <- aov(lifespan~treatment, data = life.exp)
TukeyHSD(tk)## Tukey multiple comparisons of means
## 95% family-wise confidence level
##
## Fit: aov(formula = lifespan ~ treatment, data = life.exp)
##
## $treatment
## diff lwr upr p adj
## N/R40-N/N85 12.4254385 9.215018 15.6358589 0.0000000
## N/R50-N/N85 9.6059554 6.519071 12.6928401 0.0000000
## R/R50-N/N85 10.1944862 6.928685 13.4602879 0.0000000
## N/R50-N/R40 -2.8194831 -5.863260 0.2242942 0.0804712
## R/R50-N/R40 -2.2309523 -5.456039 0.9941344 0.2808243
## R/R50-N/R50 0.5885309 -2.513604 3.6906659 0.9610849plot(TukeyHSD(tk,'treatment'))
#Summarize conclusions ( F )