Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 5

Project#2:TheProductivityofDairyCows

DueDate:March9th,2015byMidnight.
SubmissionMethod:D2LDropbox.

Instructions:

Usethisfiledirectlytoimplementtheproject.
WriteyournameandCWIDnumberinthefollowingtable.

Name:NiNyomanSulastri
CWID:11701373

ProjectDescription:

Theproductivityofdairycowsisreducedbyheatstress.Thisexperiment
concernstheeffectonbodytemperatureofseveraltypesofcooling
treatments.Twowhiteandtwoblackcows,identifiedbythevariableCowid,
wererandomlyassignedtotwoexposuresoffourtreatments:fan,shade,
mist,andsun(control).Theexperimentswereconductedattimesforwhich
environmentalconditionswereessentiallyidentical.Tworesponseswere
measured:surfaceandrectaltemperature.ThedataareattachedinExcel
file(productivity_of_dairy_cows).
Performseparateanalysesforthetworesponsevariablesfortheeffectof
cowcolorandshadetreatment.Notethatthisisasplitplotwithcowcoloras
themainploteffect.
Requirements:
Writetheappropriatemodelforthisdesign.
Analyzethedataaccordingtothechosendesign.
Commentonyourfindings.
Theminimumnumberofpagesis3pages.
ThefontisTimeswithsizeof13.
Linespacingis1.5.
Includeplotsifanyinanappendix.

Answer
Statistical Model
Consider the model for the completely randomized split-plot design with two
levels of color (white and black), four levels of treatment-trt (fan, shade, mist,
sun). Cow id is considered repetition in color factor. If yijk denotes the kth response
for the ith level of factor color, jth level of factor treatment (trt), then
y ijk=+ i + ik + j+ ij + ijk

Where:

: overall mean

: fixed effect for ith level of factor color

: fixed effect for jth level of factor trt

ij : fixed effect for ith level of factor color, jth level of factor trt
ik

: random effect for the kth whole plot receiving the ith level of factor color

ijk

: a random error associated with the response from the ith experimental unit
receiving treatment i

Hypothesis
Surface
WholeplotAnalysis:
Ho1 : There are no significant differences the surface temperature of cow body
between the two color of cow body
Ha1 : There are significant differences the surface temperature of cow body between
the two color of cow body
SubplotAnalysis:
Ho2 : There are no significant differences the surface temperature of cow body
between four types of cooling treatments
Ha2 : There are significant differences the surface temperature of cow body between
four types of cooling treatments

Ho3 : There are no significant interaction the surface temperature of cow body between
the levels of body color and the levels of cooling treatments
Ha3 : There are significant interaction the surface temperature of cow body between
the levels of body color and the levels of cooling treatments

Rectal
WholeplotAnalysis:
Ho1 : There are no significant differences the rectal temperature of cow body between
the two color of cow body
Ha1 : There are significant differences the rectal temperature of cow body between the
two color of cow body
SubplotAnalysis:
Ho2 : There are no significant differences the rectal temperature of cow body between
four types of cooling treatments
Ha2 : There are significant differences the rectal temperature of cow body between
four types of cooling treatments
Ho3 : There are no significant interaction the rectal temperature of cow body between
the levels of body color and the levels of cooling treatments
Ha3 : There are significant interaction the rectal temperature of cow body between the
levels of body color and the levels of cooling treatments

Anova table
> data<-read.table('project2.txt',header=TRUE)
> attach(data)
> data
> av<-aov(surface~trt*color+Error(color/cowid),data=data)
> summary(av)
Response: surface
Sources
color
whole plot error
treatment(trt)
trt:color
Subplot error

Df
1
2
3
3
22

SS
0.0028
1.7856
73.54
3.18
32.48

MS
0.0028
0.8928
24.514
1.059
1.476

F
0.003

P-value
0.96

16.606
0.717

7.31e-06
0.552

Signif. codes: 0 *** 0.001 ** 0.01 * 0.05 . 0.1 1

> av<-aov(rectal~trt*color+Error(color/cowid),data=data)

***

> summary(av)
Response: rectal
Sources
color
whole plot error
treatment(trt)
trt:color
Subplot error

Df
1
2
3
3
22

SS
0.0613
0.5713
1.253
0.521
5.129

MS
0.06125
0.28563
0.4175
0.1738
0.2331

F
0.214

P-value
0.689

1.791
0.745

0.178
0.537

Signif. codes: 0 *** 0.001 ** 0.01 * 0.05 . 0.1 1

Based on two anova table above, only cooling treatment have significant effect on
cow body temperature especially the body surface, since the P-value is lower than
0.05. Next the pairwise comparisons only focus on the cooling treatment.
Tukey HSD post hoc test
> av1=aov(surface~trt,data=data)
> TukeyHSD(av1)
Comparison
mist-fan
shade-fan
sun-fan
shade-mist
sun-mist
sun-shade

Diff
-0.3500
1.4625
3.5000
1.8125
3.8500
2.0375

95% Confident Interval


Lower Limit
Upper Limit
-1.9286228
1.228623
-0.1161228
3.041123
1.9213772
5.078623
0.2338772
3.391123
2.2713772
5.428623
0.4588772
3.616123

P-Value
0.9295224
0.0769294
0.0000091
0.0197636
0.0000018
0.0076213

Based on Tukey HSD results and profile plot (Appendix 1), mist and fan treatment
are insignificantly different and resulted in the lowest temperature for body
surface, while the control treatment (sun) resulted in the highest temperature for
cow body surface. Therefore, to increase the production of dairy milk of the cow,
mist and fan treatment should be applied to reduce the heat stress.

Appendix 1
>with(data,interaction.plot(color,trt,surface,type="b",pch=19,fixed=T,xlab="cow
color",ylab="surface temperature",ylim=range(35,40),main="Profile Plot"))

You might also like