The text included in the book is the outcome from the need realized by the authors over past decade, during the interactions with the scientists’, researchers’, doctoral and master’s scholars of Forest Research Institute(FRI), Dehradun and its University, Banaras Hindu University, (BHU) as well as Sher- e-Kashmir University of Agricultural Sciences and Technology of Jammu (SKUAST-Jammu) and the experiences gained during student life by first author. This is an introductory textbook dealing with statistical designs and analysis of real-life forestry and agricultural experiments. This book has been designed especially for the researchers and scholars of applied biological research, who have to use statistical principles for the collection, analysis and interpretation of experimental data. All the basic techniques, which are essential in experiments for comparative purposes have been included in the book with examples. Various sources including web pages etc. are used to extract the materials for the book. Due to inadequate space, some sources related to long time project work are overlooked and not included in references. We acknowledge the same. The excellent work by Statisticians and others have been used as source of materials incorporated in the book. All these are duly acknowledged, where ever possible. The materials for illustrative examples were noted from various books, laboratory records, and personal discussions with the forestry research scholars of UHF, Nauni; PAU, Ludhiana; SKUAST-Jammu; Henvanti Nandan Bahuguna Central University, Srinagar, Banaras Hindu University, Varanasi and FRI, Dheradun. We shall appreciate the readers, if they bring suggestions for improvement in this work.

Experimentation is fundamental to scientific investigation because it provides a systematic basis for studying natural and controlled systems and drawing conclusions from observations. In forestry and horticulture, experiments are often affected by developmental, environmental, genetic, sampling, and experimental variations. This chapter introduces the principles required for planning, conducting, analysing, and interpreting such experiments. It explains comparative experiments, the need for experimentation, experimental units, sampling units, plots, blocks, treatments, factors, levels, and sources of variation. Particular emphasis is given to the basic principles of experimental design—randomisation, replication, and local control—which help improve validity, precision, and coverage of experimental results.

Hypothesis testing provides a statistical framework for making inferences about populations from sample observations obtained through random experiments. This chapter introduces the scientific logic underlying statistical hypotheses and explains the formulation of null and alternative hypotheses. It discusses one-tailed and two-tailed tests, levels of significance, Type I and Type II errors, sampling distributions, test statistics, critical regions, p-values, and decision-making. The chapter also explains statistical significance and the interpretation of test results. Applications of Student’s t-test and related procedures are illustrated through forestry and biological examples, enabling researchers to determine whether observed differences between means can reasonably be attributed to treatments rather than random sampling variation.

Analysis of Variance (ANOVA) is an important statistical technique for separating variation attributable to different sources in an experiment. The chapter explains how total variation can be partitioned into controlled and uncontrolled components and how experimental error is used to assess treatment differences. The F-test provides the basis for testing the homogeneity of several treatment means simultaneously. The chapter discusses the assumptions underlying ANOVA, including independence, normality, additivity, and homogeneity of variance. It also introduces fixed- and random-effects models and explains their relevance to experimental analysis. Forestry and agricultural examples demonstrate the application of ANOVA for interpreting treatment and experimental effects.

Transformation of data is useful when the assumptions required for analysis of variance are not adequately satisfied. Forestry experiments may produce observations showing skewness, unequal variances, non-normality, or departures from additivity. This chapter explains the need for transforming observations to an appropriate scale before statistical analysis. It discusses the characteristics of an ideal transformation, including stabilisation of variance, improvement of normality, efficient estimation of means, and achievement of linear and additive effects. Various transformation techniques used in forestry experimental analysis are introduced with suitable examples. The chapter emphasizes that appropriate transformation can improve the validity, sensitivity, and interpretability of statistical tests.

When ANOVA indicates significant differences among treatment means, further analysis is required to determine which particular means differ significantly. This chapter deals with post-ANOVA procedures used for detailed comparison of treatment means. It introduces methods such as critical difference, least significant difference, and multiple comparison procedures. Particular attention is given to Scheffé’s method, Newman–Keuls method, and Duncan’s New Multiple Range Test. The chapter explains the principles, calculations, and interpretation of these techniques using forestry examples, including comparisons among tree species and provenances. It also discusses considerations for selecting an appropriate multiple-range or comparison method according to the objectives and number of comparisons involved.

Orthogonal designs provide an efficient framework for arranging treatments and sources of variation so that their effects can be estimated independently. This chapter introduces contrasts, orthogonal contrasts, and the concept of orthogonality in experimental design. Orthogonality helps prevent the effects of different factors from becoming entangled and facilitates systematic statistical analysis. The chapter discusses three basic experimental designs: Completely Randomized Design, Randomized Complete Block Design, and Latin Square Design. Their principles, applications, and suitability for forestry and agricultural experiments are explained. Examples involving seedlings, tree species, fertilizers, chemicals, wood pieces, and seeds demonstrate how appropriate design selection can reduce heterogeneity and improve experimental precision.

Factorial experiments provide an efficient method for studying two or more factors simultaneously at different levels. This chapter introduces the concept and notation of factorial arrangements and explains treatment combinations in experiments such as 2 × 2 and higher-order factorials. It distinguishes among simple effects, main effects, and interaction effects, emphasizing the importance of interactions in interpreting biological responses. Factorial experiments allow researchers to examine the effects of individual factors as well as their combined influence. The chapter also explains that factorial treatments may be arranged using different basic experimental designs, such as Completely Randomized, Randomized Block, or Latin Square designs, according to experimental objectives and conditions.

Loss of experimental observations is a common practical problem in forestry and agricultural experiments because seeds, seedlings, plants, and trees may be damaged by natural events, animals, anthropogenic activities, or other causes. This chapter presents the missing plot technique for estimating unavailable observations so that an experiment can continue to be analysed. It explains why missing observations can reduce the information and precision of an experiment and discusses alternative approaches for handling incomplete data. The chapter focuses particularly on the method of least squares and the iteration method. Applications include estimation of missing values in Randomized Block Designs and other experimental layouts used in forestry research.

Analysis of Covariance (ANCOVA) combines features of analysis of variance and regression analysis to account for the influence of a concomitant or covariate variable. In forestry and horticultural experiments, initial measurements or other associated characteristics may influence the response variable and contribute to experimental variation. This chapter explains how treatment effects can be evaluated after adjusting for the linear relationship with a covariate. It presents the ANCOVA framework, calculation of sums of squares, regression coefficients, adjusted error and treatment variation, F-tests, and adjusted treatment means. Examples involving seedling growth and physiological responses demonstrate how covariance adjustment can provide more precise comparisons among treatments.

Split-plot design is particularly useful when certain biological treatments cannot conveniently be applied to small experimental units. This chapter explains how large experimental plots can be divided into smaller sub-plots to accommodate subsidiary treatments. Main-plot treatments are randomly allocated to larger units, while sub-plot treatments are independently allocated within them. The design allows researchers to study main effects, sub-plot effects, and interactions, although greater precision is generally obtained for sub-plot comparisons than for main-plot comparisons. Examples include wood preservation and nursery experiments involving sowing methods and seasons. The chapter also explains the two error terms required for appropriate analysis of split-plot experiments.

Series of experiments are conducted at different locations, times, or combinations of locations and years to evaluate the spatial and temporal performance of treatments. Such experiments are especially important in forestry and agricultural research when conclusions are intended to apply across wider geographical or environmental conditions. This chapter explains combined analysis of a series of experiments after individual experiments have first been analysed according to their respective designs. Particular attention is given to testing the homogeneity of error variances, including Bartlett’s test. When variances are homogeneous, pooled analysis can assess treatment effects, location or year effects, and treatment-by-location or treatment-by-year interactions, thereby evaluating consistency of treatment responses.

Augmented designs are useful when a large number of standard varieties or treatments are available for replication but new or promising materials are available only in limited quantities. This chapter introduces augmented designs as modifications of standard experimental designs in which additional, usually unreplicated, entries are incorporated alongside replicated standard treatments. Applications include testing new varieties, seedlings, medicinal plants, growth regulators, seedling carriers, and progeny under forestry and agricultural conditions. The chapter explains the principles of augmented designs and their applications in experimental research. It also discusses augmented Completely Randomized Design, including its layout, analysis of variance, and comparison of replicated varieties and unreplicated seedlings.

References