The core experimental process at the heart of a project will typically involve the recording of one or more response (independent) variables under different experimental conditions which are described by factor (dependent) variables.
The overall experiment design involves identifying the variables that describe each of the factors that you wish to consider, and then choosing the range of experimental conditions by selecting different values (also called levels) of the factor variables.
ACTION: As a first step in selecting the appropriate experimental design and methods for data analysis, you need to
identify all the response and factor variables that you might record in your experiment, and
decide on the type of each of those variables.
Types of variable
Scale (quantitative) variables:- have a direct relationship with the quantity being measured
Continuous variable – any value within the range can be recorded, depending on the precision of the measurement e.g. refractive index, pH.
Discrete variable – the recorded values are restricted to a specific set, e.g.
Frequency – counting events, e.g. radioactivity, number of fibres
Qualitative variables:- do not have a scaled relationship with the quantity being measurement
Ordinal variable – the variable describes the rank order of the quantity being described, e.g. ranking 0 to 5 for the quality of a fingermark from poor to excellent.
Note that the rank order does NOT give any scaled information, e.g. the time between the first and second person in a race could be minute or very long.
Nominal variable – the ‘variable’ only names each possible category of the quantity, e.g. identifying different fabrics as silk, wool, polyester.
Derived variables:- the values are normally calculated from other direct measurements
Proportion, e.g. ratio of values of continuous variables.
Probability, e.g. ratio of frequencies