Matching goals with variables.
When designing a statistical study, it is important to understand and match your goals with each variable’s measurement type: nominal, ordinal, interval, and ratio.
Each type contains distinct characteristics and provides differing benefits.
Nominal variables are also qualitative which represent data usually assuming categorical values, such as a person’s ethnicity – a distinct value from a predefined list of possibilities.
This type of variable represents data where a clearly-defined order exists, but not a mathematical difference between them. For example, if a study was measuring the order in which a person consumed medications, the list would contain a list of medications in the order they self-administered their medication, but the time in between the order would be irrelevant.
When measuring data which contains a discernable order with measurable distances between values, but no natural zero exists or it contains no meaning relevant to the study, that data would constitute an interval variable.
Interval variable are often found in studies of time. The middle of the day (12pm) subtracted from the evening (6pm) would produce a statically-significant value, but no “zero time” value exists in that a time of zero (in military time) does not constitute an absence of time.
Ratios are used for ordered-values with measurable distances between them and when a zero value indicates the lack of an amount when compared to the rest of order. A common ratio example would be a person’s financial account balance – zero would indicate they have no balance.
One thought on “Statistics: Variables – Measuring Values”
Pingback: Statistics | TechMbaBi