WSM 201 – MODULE III DETAILED NOTES
These detailed notes are prepared for MSc Forestry / Wood Science students for exam
preparation. The notes include theory, derivations, formulas, experimental designs,
assumptions, examples, and MCQs.
1. Analysis of Variance (ANOVA)
ANOVA is a statistical method developed by R.A. Fisher. It is used to test whether the means
of three or more groups are significantly different.
WHY ANOVA IS NEEDED:
If we use multiple t-tests repeatedly, the probability of error increases. ANOVA solves this
problem by testing all means simultaneously.
BASIC IDEA:
Total variation in data is divided into:
1. Variation due to treatments
2. Variation due to random error
ASSUMPTIONS OF ANOVA:
• Observations are independent
• Data are normally distributed
• Variances are homogeneous
DERIVATION OF F-RATIO:
F = Mean Square Due to Treatments / Mean Square Due to Error
Where:
MS Treatment = SST / (t − 1)
MS Error = SSE / (N − t)
If calculated F > table F value → reject null hypothesis.
APPLICATIONS:
• Fertilizer trials
• Tree growth studies
• Wood quality comparison
• Forestry experiments
2. One-Way ANOVA – Detailed
One-way ANOVA studies the effect of one factor on a response variable.
EXAMPLE:
Compare heights of seedlings under 4 fertilizers.
HYPOTHESES:
H₀: μ1 = μ2 = μ3 = μ4
H₁: At least one mean differs
ANOVA TABLE:
Source df SS MS F
Treatment t−1 SST MST MST/MSE
Error N−t SSE MSE
Total N−1 TSS
FORMULAS:
Correction Factor (CF) = (Grand Total)^2 / N
Total Sum of Squares:
TSS = ΣX² − CF
Treatment Sum of Squares:
SST = Σ(Treatment Total² / r) − CF
Error Sum of Squares:
SSE = TSS − SST
3. Two-Way ANOVA
Two-way ANOVA studies two factors simultaneously.
EXAMPLE:
Effect of fertilizer and irrigation on plant growth.
ADVANTAGES:
• Saves time
• Studies interaction effect
• More information obtained
SOURCES OF VARIATION:
• Factor A
• Factor B
• Interaction A×B
• Error
IMPORTANT:
Interaction means the effect of one factor changes according to another factor.
4. Analysis of Covariance (ANCOVA)
ANCOVA combines ANOVA and regression.
PURPOSE:
To remove the effect of unwanted variables (covariates).
EXAMPLE:
Comparing tree height while controlling soil fertility.
ADVANTAGES:
• Reduces experimental error
• Improves precision
• Gives adjusted treatment means
ASSUMPTIONS:
• Linear relationship exists
• Homogeneity of regression slopes
5. Tests for Homogeneity and Linearity
HOMOGENEITY:
Means equality of variances.
WHY IMPORTANT?
ANOVA assumes equal variance among groups.
TESTS:
• Bartlett’s Test
• Levene’s Test
LINEARITY:
Relationship between variables follows straight line.
EXAMPLE:
Height increases linearly with age.
6. Research Problem and Research Design
RESEARCH PROBLEM:
A scientific question requiring investigation.
CHARACTERISTICS:
• Clear
• Specific
• Practical
• Researchable
RESEARCH DESIGN:
A blueprint for data collection and analysis.
OBJECTIVES:
• Reduce bias
• Increase reliability
• Improve accuracy
7. Principles of Design of Experiments
1. RANDOMIZATION:
Treatments assigned randomly.
ADVANTAGE:
Removes personal bias.
2. REPLICATION:
Repeating treatments several times.
ADVANTAGE:
Increases precision.
3. LOCAL CONTROL:
Grouping similar units together.
ADVANTAGE:
Reduces error variation.
8. Completely Randomized Design (CRD)
In CRD, treatments are assigned completely randomly.
SUITABLE FOR:
Homogeneous experimental material.
ADVANTAGES:
• Simple
• Flexible
• Easy analysis
DISADVANTAGES:
• Less efficient if variation exists
LAYOUT EXAMPLE:
T1 T2 T3 T1 T2 T3
9. Randomized Block Design (RBD)
Experimental units are divided into blocks.
PURPOSE:
Reduce variation due to environmental effects.
ADVANTAGES:
• More accurate than CRD
• Controls local variation
DISADVANTAGES:
• Analysis becomes complex with many treatments
10. Latin Square Design (LSD)
Used when two directional variations exist.
FEATURES:
• Number of rows = columns = treatments
• Each treatment appears once in every row and column
ADVANTAGE:
Controls two sources of variation simultaneously.
LIMITATION:
Number of treatments must equal rows and columns.
11. Factorial Experiments
Two or more factors are studied together.
EXAMPLE:
3 fertilizers × 2 irrigation levels.
ADVANTAGES:
• Study interactions
• Economical
• Efficient
INTERACTION EFFECT:
Combined effect differs from individual effects.
12. Statistical Software
Common software:
• SPSS
•R
• SAS
• MS Excel
USES:
• ANOVA
• Regression
• Graph plotting
• Data analysis
13. Important Viva Questions
• Define ANOVA.
• What is F-ratio?
• Difference between CRD and RBD.
• Explain replication.
• What is randomization?
• Define ANCOVA.
• Advantages of factorial experiments.
14. MCQs
1. ANOVA compares:
a) Variances
b) Means
c) Ratios
Answer: b
2. F-ratio is:
a) MSE/MST
b) MST/MSE
Answer: b
3. RBD controls:
a) Local variation
b) Total variation
Answer: a
4. ANCOVA combines:
a) ANOVA + Regression
b) ANOVA + t-test
Answer: a
5. Latin Square Design controls:
a) One source
b) Two sources
Answer: b