Design of Experiment

Mandate

To develop statistical designs and methodologies for analysis of data relating to field and laboratory
experimentation in agriculture and animal science.

Thrust Areas

• Cropping system research
• Information system for agricultural and animal experiments
• Experimental designs for agricultural, animal, agro-forestry and fisheries research

Projects

Current Projects

Institute funded projects:

Outside Funded Projects:

Completed Projects:

Research Highlights

(a) Applied Research

A number of applied type studies have been undertaken. These include:

• Analysis of long term experiments planned on fixed rotations and conducted at different
  co-operating centres under AICRP on Long term fertilizer experiments.
• Planning, designing, analysis and interpretation of data collected from experiments on
  Cropping System Research (CSR) and cultivators’ field trials under the Project Directorate for CSR.
• Analysis of data from experiments on Agroforestry through various statistical techniques and
  development of suitable model and design for Agroforestry research experiments.
• Development of yardsticks for major cereal crops for different agroclimatic regions.
• Statistical studies to determine the extent to which nitrogen could be substituted through organic sources.
• Development and identification of suitable model based on soil test values for estimating the optimal fertilizer response and economic gains.
• Quantification of drought threshold rainfall values for droughts of various duration using rainfall and crop-yield data.
• Assessment of changes in crop production in command areas of river valley projects.
• Analysis of data from uniformity trials on several crops with a view to finding the relationship between the plot and block sizes on the one hand and variation on the other for arriving at the optimum sizes and shapes of plots and blocks.
• Investigation on the use of designs with mixtures and their analysis in experiments involving split
  applications of inputs like fertilizer, irrigation, etc. to obtain optimum splits.
• Analysis of unbalanced data for obtaining Genotypic variance, phenotypic variance etc.
• Analysis of Agroforesry experimental data
• Software development for experimental designs
• Analysis of repeated measures data using profile analysis
• Web generation of experimental designs
• Fitting of on-linear models for the growth (repeated measures) data on pigs
• Combined analysis of data having homogenity/heterogeneity of error variances in agricultural field
  experiments

(b) Basic Research

The Institute has been actively engaged in developing more efficient designs of experiments and further methods of analysis of experimental data. In this regard, the important contributions include:

• Construction of balanced confounded factorial experiments for symmetrical as well as asymmetrical factorial experiments, providing flexibility in the choice of factors and their levels. Procedures of analyzing these designs have also been suggested
• For screening experiments or when the higher order interactions are negligible or the experimental resources are not enough for one complete replication, fractional factorial designs and particularly orthogonal resolution plans have been obtained
• For the experimental designs where it is desired to study the relationship between quantitative level of factors and the response,various types of designs for fitting response surfaces including rotatable, group divisible rotatable have been developed and analytic procedures have been suggested. These designs can also beused to obtain economic optimum combination of levels of factors.
• To study the rate of change of response with respect to factors, in biological populations or chemical experiments, slope rotatable designs have been developed.
• For the experimental situations, where it is desired that total dose of factors is fixed but response depends on various proportions of different factors, designs for experiments with mixture have been developed.
• For single factor experiments incomplete block designs viz. variance balanced, efficiency balanced, partially efficiency balanced and partially balanced block designs have been developed
• Sometimes, within blocking factors, there may be more sources of heterogeneity. To deal with such  situations block designs with nested rows and columns have been developed.
• Designs for bio-assays.
• Designs for making test treatments and control treatment(s) comparisons.
• Designs for studying competitive effects of treatments.
• Change-over designs.
• Designs for experiments on cultivators’ fields.
• Multi-dimensional designs for more than one set of non-interacting treatments.
• Designs for diallel cross experiments.
• Trend-free designs in block design set up (BIB, PBIB, Variance balance and Nested BIB designs and for optimum designs for diallel cross experiments.
• Linear trend-free designs for 2 – level with and without  confounding factorial experiments, and linear trend-free fraction factorial experiments.
• Cataloguing of the above designs has also been undertaken. Some further studies on obtaining efficient designs with varying replications and unequal block sizes were undertaken. Besides construction of designs, the optimality and robustness aspects of the designs have been investigated. This includes optimality studies under fixed and mixed effects models under homo-and hetero-scedastic set-ups in block designs, row-column designs, designs for making test treatment control comparisons etc., and robustness of designs against loss of data, outliers, interchange and exchange of treatments. In the case of analysis of experimental data the contributions include those for missing data, non-orthogonal data, analysis of covariance, intercrop data, data on diallel crosses, asymmetrical factorial designs, groups of experiments with common treatments, use of non-parametric methods.
• Designs for bioequivalence trials
• Neighbor balanced designs
• Designs for experiments with spatially correlated observations

(c) Information on Agricultural/Animal Experiments

• The Institute has collaborated in this regard with the All-India Co-ordinated Research Projects on Agronomy (Project Directorate for Cropping Systems Research), Long-term fertiliser experiments, AICRP on STCR , agroforestry etc. Agricultural Field Experiments Information System was developed for storage with provision for selective retrieval of information on the agricultural and animal experimentation in the country. The data from this Information System are being utilized in carrying out a number of methodological studies.