SOLVE USING R STUDIO
In a study of memory enhancement, 50 lab rats were randomly assigned to one of the following five treatments: ginkgo, phos-phatidylserine (PS), phosphatidylcholine (PC), vinpocetine, and control. They were fed an appropriate dose of these supplements for 6 weeks, then run in a series of maze tests to evaluate memory. The performance score on the maze (higher score = better performance) is given below. Assume technical assumptions of ANOVA have been met.
i. Is there evidence for an effect of these drugs on
memory?
ii. Design a set of orthogonal contrasts and test the following
sub-hypotheses. Provide test statistic, p-value, and df
corresponding with these tests altogether in a unified ANOVA
table.
a. Do the drugs, generally as a group, affect memory compared to the control group?
b. Do the plant-based drugs, ginkgo and vinpocetine, differ from the two synthetic drugs (PS and PC)?
c. Do PS and PC differ from one another?
d. Do ginkgo and vinpocetine differ from one another?
d=scan("clipboard")
d=matrix(d,ncol=5,byrow=TRUE)
Ginkgo=d[,1]
PS=d[,2]
PC=d[,3]
Vinopocetine=d[,4]
Control=d[,5]
Ginkgo=c(30.3,29.7,34.5,28.7,35.1,29.8,39.1,31.4,33.5,26.9)
PS=c(63.0,60.6,33.5,51.3,49.7,44.7,63.2,47.0,37.0,51.3)
PC=c(39.2,55.4,66.7,61.8,51.1,47.6,49.4,56.6,60.8,37.2)
Vinopocetine=c(35.8,31.3,42.6,40.7,20.1,23.8,26.8,17.1,39.6,28.9)
Control=c(21.0,21.2,21.4,28.5,17.6,23.7,15.4,19.0,15.4,22.6)
scores=c(Ginkgo,PS,PC,Vinopocetine,Control)
dru=rep(1:5,each=10)
dru
[1] 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 4 4
4 4 4 4 4 4 4 4 5 5 5 5 5 5 5 5 5 5
drugs=factor(dru)
drugs
[1] 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 4 4
4 4 4 4 4 4 4 4 5 5 5 5 5 5 5 5 5 5
Levels: 1 2 3 4 5
ANOVA=aov(scores~drugs)
ANOVA
Call:
aov(formula = scores ~ drugs)
Terms:
drugs Residuals
Sum of Squares 7506.807 2738.054
Deg. of Freedom 4 45
Residual standard error: 7.800362
Estimated effects may be unbalanced
s=summary(ANOVA)
s
Df Sum Sq Mean Sq F value Pr(>F)
drugs 4 7507 1876.7 30.84 2.23e-12 ***
Residuals 45 2738 60.8
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
##H0: drugs do not have effect on memory
##From summary, p-value=2.23e-12=0, we Reject H0.
##YES, there is effect of these drugs on
memory.
##A set of orthogonal contrasts and test the following
sub-hypotheses.
##Assigning
coefficients:
#If two groups of equal size are to be compared, assign
coefficients of +1 to the members of one group and -1 to those of
the other group. It does not matter which group is assigned the
positive coefficient.
#In comparing groups containing different numbers of treatments,
assign to the first group, coefficients equal to the number of
treatments in the second group, and to the second group,
coefficients of the opposite sign equal to the number of treatments
in the first group.
#sum of coefficients of contrast equals to zero i.e.
sum (aj)=0.
Samplemeans=c(mean(Ginkgo),mean(PS),mean(PC),mean(Vinopocetine),mean(Control))
Samplemeans
#[1] 31.90 50.13 52.58 30.67 20.58
##Consider vector of length 5: Ginkgo=value-1, PS=value-2, PC=value-3, Vinopocetine=value-4, Control=value-5
##Sub-Hypotheis:
#a. Do the drugs, generally as a group, affect memory compared to
the control group?
aj=c(1,1,1,1,-4)
sum(aj) #[1]0
#b. Do the plant-based drugs, ginkgo and vinopocetine, differ
from the two synthetic drugs (PS and PC)?
bj=c(1,-1,-1,1,0)
sum(bj) #[1]0
#c. Do PS and PC differ from one another?
cj=c(0,1,-1,0,0)
sum(cj) #[1]0
#d. Do ginkgo and vinopocetine differ from one another?
dj=c(1,0,0,-1,0)
sum(dj) #[1]0
##Sub-Hypotheis:
#H0a/ H01: the drugs, generally as a group, do not affect
memory compared to the control group
#H0b/ H02: the plant-based drugs, ginkgo and vinopocetine, do not
differ from the two synthetic drugs (PS and PC)
#H0c/ H03: PS and PC do not differ from one another
#H0d/ H04: ginkgo and vinopocetine do not differ from one
another
##Two contrasts aj and bj are orthogonal if
sum(aj*bj)=0.
sum(aj*bj)
sum(aj*cj)
sum(aj*dj)
sum(bj*cj)
sum(bj*dj)
sum(cj*dj)
#All summations equal to zero. Therefore, all contrasts are
orthogonal to each other.
RESULT:
##SSContrast= n*((sum(cj*Xjbar))^2)/sum(cj^2)
n=10
SSaj=n*(sum(aj*Samplemeans)^2)/sum(aj^2)
SSaj
#[1] 3441.181
SSbj=n*(sum(bj*Samplemeans)^2)/sum(bj^2)
SSbj
#[1] 4028.049
SScj=n*(sum(cj*Samplemeans)^2)/sum(cj^2)
SScj
#[1] 30.0125
SSdj=n*(sum(dj*Samplemeans)^2)/sum(dj^2)
SSdj
#[1] 7.5645
SSRes=2738.054 #From ANOVA Table
MSaj=SSaj/1 #[1] 3441.181
MSaj
MSbj=SSbj/1 #[1] 4028.049
MSbj
MScj=SScj/1 #[1] 30.0125
MScj
MSdj=SSdj/1 #[1] 7.5645
MSdj
MSRes=SSRes/45 #df of residuals=45
MSRes
#[1] 60.84564
Ftab=qf(0.95,1,45) ##Since, alpha=5%=0.05
Ftab
#[1] 4.056612
Faj=MSaj/MSRes
Faj ##[1] 56.55591
Fbj=MSbj/MSRes
Fbj ##[1] 66.20111
Fcj=MScj/MSRes
Fcj ##[1] 0.4932563
Fdj=MSdj/MSRes
Fdj ##[1] 0.1243228
##Calculated values of F-statistic (Faj,Fbj) are greater than
tabulated F-value=4.056612, we Reject sub-null-hypothesis 1 & 2
i.e. H01 & H02..
##Calculated values of F-statistic (Fcj,Fdj) are less than
tabulated F-value=4.056612,we do not Reject H03 & H04.
## Contrasts 1 & 2 are significant. Contrasts 3 & 4
are not significant.
##Difference in pairs
is also tested as:
TukeyHSD(ANOVA,'drugs')
Tukey multiple comparisons of means
95% family-wise confidence level
Fit: aov(formula = scores ~ drugs)
$drugs
diff lwr upr p adj
2-1 18.23 8.317807 28.1421932 0.0000412
3-1 20.68 10.767807 30.5921932 0.0000039
4-1 -1.23 -11.142193 8.6821932 0.9965651
5-1 -11.32 -21.232193 -1.4078068 0.0179478
3-2 2.45 -7.462193 12.3621932 0.9549078
4-2 -19.46 -29.372193 -9.5478068 0.0000126
5-2 -29.55 -39.462193 -19.6378068 0.0000000
4-3 -21.91 -31.822193 -11.9978068 0.0000012
5-3 -32.00 -41.912193 -22.0878068 0.0000000
5-4 -10.09 -20.002193 -0.1778068 0.0441940
##Clearly pairs (4-1, 3-2)have p-values (0.9965651 ,0.9549078)respectively; which are greater than alpha=0.05. We do not Reject H0.
##Conclusions of Sub-Hypotheis:
#a. YES, the drugs, generally as a group, do affect memory compared
to the control group.
#b. YES, the plant-based drugs, ginkgo and vinopocetine, do differ
from the two synthetic drugs (PS and PC).
#c. NO, PS and PC do not differ from one another.
#d. NO, ginkgo and vinopocetine do not differ from one
another.
SOLVE USING R STUDIO In a study of memory enhancement, 50 lab rats were randomly assigned to one ...