Designing the experiment - Variables and variability factors
Henriėtte L. de Kock
University of Pretoria, South Africa
riette.dekock@up.ac.za
Company X is a large corporate group with many subsidiary divisions, including bakery, meat and sauce production plants. The company recently purchased Burger Bar Inc., a popular chain of fast food outlets. Burger Bar has been a national icon for over 50 years with a slogan: Nothing can beat the BB taste experience! Company X calculated that they could save a lot of money if they source more of the various food components used by BB internally within the group, rather that purchasing it from the traditional BB suppliers. The brand manager will only accept changes if proof is provided that consumers will not notice changes. How would you approach this case study?
An experiment is a systematic search for a cause and effect relationship. During this presentation we will discuss practical guidelines for selecting sensory test variables and consider how to optimize the value of dependent or response variables and reduce experimental variability by using case studies. We prefer to use a six steps approach to Sensory Evaluation.
Step 1: Identify the sensory problem or question.
During this step information is gathered which leads to a hypothesis (your best guess at what the cause and effect relationship is). This includes the identification of test variables - an educated guess about what types of things may affect the system that you are working with. You then test the hypothesis during a controlled experiment.
Step 2: Define the objective of the test
The main objective of a sensory experiment is to determine the effect of certain treatment/s (independent variables) on the response of a group of people.
Step 3: Select the test conditions
The primary obstacle in accomplishing this objective is that responses of sensory groups are extremely variable making the detection of true differences in response due to treatments difficult. In addition the material that are investigated may also be variable. It is therefore vital to carefully plan the experiment so as to control sources of variation. Four major decisions are made when planning a sensory test, i.e. selection and presentation of test samples, a test method, the testing environment and the test panel. These four aspects may all introduce various sources of variation.
Step 4: Run the sensory test
Step 5: Analyse the test result
The statistical analysis of test results is directly dictated by the way the experiment was designed.
Step 6: Take action
Conclusions from the results will allow you to predict the results of future cause and effect relationships..
Download presentation
