Experimental designs

EXPERIMENTAL DESIGNS

Commonly the experiments are designed by 

  • CRD (Completely Randomised Designs) 
  • RBD (Randomised Block Designs), 
  • Factorial Design and
  • Latin Square Design (LSD).
  • In Factorial experiments, two factors like protein (16% and 18%) and Energy (3000 kcal and 3300 kcal) at two levels are compared. 

Use of statistiacl methods nutrition experiments

  • In a feeding trial certain factors, such as the amounts and quantities of feed, the time and method of feeding and the general care and management, can be definitely fixed. Certain other factors, inherent in the animals used, cannot be controlled.
  • The object of a well-planned experiment is to reduce these uncontrollable factors to minimum by giving attention, in the selection of animals used, to genetic and nutritional history as well as to such factors as age, size, vigor etc. Even though this is effectively done, there still remain inherent variables which cause two individuals to respond somewhat differently though treated exactly alike in an experiment.
  •   The effect of the inherent variables cannot be measured, but the probability that the observed differences in experimental results could arise from the uncontrollable variables alone, can be estimated and taken into account. This is done by a statistical analysis of the data obtained. Such an analysis  helps the investigator to decide whether the results from a given comparison reflect a real difference in response to the treatments or may have occurred simply because of inherent variations in the animals used.
  • Statistical methods have become an essential tool of the investigators of nutrition and some knowledge of them is helpful to all students in this field as an aid in the evaluation of published research.      

Overall conclusion on feeding expeiments

  • No single method is suitable for the solution of all types of nutrition problems.
  • The effective investigator must select his method in accordance with his problem, frequently employing more than one method.
  • He must interpret his results with a full consideration of the advantages and limitations of the methods used.
Last modified: Friday, 30 March 2012, 9:45 AM