Chapter 2 data analysis - practical

DATA ANALYSIS - PRACTICAL

Analysis of Variance (ANOVA) using SPSS 14 Computer programme

Prof Dr. Md. Ruhul Amin

Problem No. 1

Litter weight(kg) in first kidding in three different breeds of goats were as follows. Performing ANOVA, test whether any difference exists among the breeds. Compare each pair of means. Jamnapari 3.1 4.2 5.2 4.8 4.0 3.9 4.2 4.6 3.5 3.8

Kambing katjang

2.2 1.9 2.0 2.1 2.0 1.5 2.2 2.3 1.5

Sirohi 2.3 2.0 2.5 3.8 2.7 2.8 2.6 2.5 2.9

ANOVA….

1. Go to SPSS 14 spread sheet2. Click on the Variable view at the bottom3. Define each of your variables i)Breed: label =3 ii) Breed: 1=jam, 2=kam, 3=sir (values) iii) Litter wt. iv) For typing data go to Data view sheet

SPSS spread sheet

Typing databreed Lwt

jam 3.1

jam 4.2

… …

kam 2.2

kam 1.9

… ….

sir 2.3

sir 2.0

… …

Data file

Analysis…

1. Go to “Analyze” button2. Select “compare means”3. Select “One-way ANOVA”4. Bring lwt as dependent variable5. Select “Post hoc”6. Click “OK”

ANOVA

Sum of squares

df Means squares

F Sig.

Between groups

23.172 2 11.586 49.084 .000

Within groups 5.901 25 .236

Total 29.072 27

ANOVA

Value of F is significant (p<0.001)There is significant difference (p<0.001) between

litter wt. of goats in 3 different breeds.

Post hoc Test (Duncan’s test)

LwtDuncan a,b

3 N Subset for alpha = .o51 2 3

kam 9 1.9667

sir 9 2.7000

jam 10 4.1300

sig 1.000 1.000 1.000 1.000

Kam 1.97c

Sir 2.70b

Jam 4.13a

Interpretation of post hoc test

Significantly (p<0.05) highest lwt. was observed in Jamnapari, lowest was in Kambing Katjang and the Sirohi ranked intermediate.

Interaction effect

Suppose records on litter weight of goats shown in Problem 1 were obtained randomly from 3 flock located on 3 states of Malaysia. Obviously “FLOCK” may have an effect like “BREED”. Again different combination of BREEDxFLOCK may behave differently. This is called interaction effect. Total no. of interaction will be 3x3 = 9.

Data file for 2-way (breedxherd) analysis

Two-way ANOVASV SS df MS F

Between breedsBetween flocks

BreedxFlock

Total

F value for interaction effect