Data Transformation.ppt

Data Transformation

1. “Strengthening the Research Capacity through
Statistical Interventions”
Dr. JitendraKumar Chaudhary
Assistant Professor
Department of Animal Genetics & Breeding,
CVSc. & A.H., Central Agricultural University (CAU),
Selesih, Aizawl, Mizoram-796014
[email protected]

2. DataTransformation:
Ifameasurementvariabledoesnotfitanormal
distributionorhasgreatlydifferentstandard
deviationsindifferentgroups,youshouldgofordata
transformation.
NormalityTest:
1)Skewness&Kurtosis
2)ShapiroWilktest
3)Histogram

3. ProceduretotestnormalityinSPSS
Gotoanalyze,descriptiveexplore,takemeasurement
independentandgroupinfactorboxtickonplots,tickon
histogramandnormalityplotswithtest,continueandok.
Results
Dividetheskewnessandkurtosisbystd.errorandseeifthe
valuesliesbetween-1.96to+1.96thenthedataarenormally
distributed(aminordifferenceisalsonormallydistributed).
SeethattheShapirowilktestissignificantornot,ifitisnon-
significantmeansdataisnormalotherwisedataisnotnormal.

4. Thefollowingarethethreetransformations,
whicharebeingusedmostcommonly,inbiological
research
a)LogarithmicTransformation
b)SquarerootTransformation
c)ArcsineorAngularTransformation

5. Logarithmictransformationforwholenumbercounts
withwiderange
Thistransformationissuitableforthedatawherethe
varianceisproportionaltothesquareofthemeanorthe
coefficientofthevariation(S.D./Mean)isconstantor
whereeffectsaremultiplicative.
Theseconditionsaregenerallyfoundinthedatathatare
wholenumbersandcoverawiderangeofvalues.
Forexample-numberofinsectperplot,numberofegg
massperplantorperunitareaetc.

6. Naturallogarithmofapositivenumber=LN(Givecell
numberforwhichtransformationtobedone),
Naturallogarithmisbasedontheconstante
(2.718281828845904)
Logarithmofapositivenumberatbase10=Log
10(Give
cellnumberforwhichtransformationtobedone),OrLog
(number,10)

7. Square root transformation
Thetransformationisappropriateforthedatasetswhere
thevarianceisproportionaltothemean.
Herethedataconsistsofsmallwholenumbers,for
exampledataobtainedincountingrareevents,suchas
numberofinfestedplantsinaplot,thenumberofinsects
caughtintraps,numberofweedsperplot.

8. Square root transformation
ThesedatasetsgenerallyfollowthePoissondistribution
andsquareroottransformationapproximatesPoissonto
normaldistribution.
Squareroottransformation=SQRT(Givecellnumberfor
whichtransformationtobedone)

9. Arcsine Transformation for Proportions or Percentage
Thetransformationisappropriateforthedataon
proportioni.e.dataobtainedfromacountandthedata
expressedasdecimalfractionandpercentage.
Thedistributionofpercentageisbinomialandthis
transformationmakesthedistributionnormal.
Arcsine Transformation=ASIN (Cell Identification e.g.
(0.05)*(180)*(7/22) result will be in degrees.