Research in Motion

This blog is in the sequence of a previous blog on understanding biostatistics. Here, the discussion will be focused on Measurement and measurement scales that results in the categorization of measurements according to their nature.

Measurement

It is defined as “the process by which things are differentiated”. Or may be defined as the assignments of numbers to objects or events according to a set of rules. The scales of measurement results from these different sets of rules while measuring a particular variable. 

If we consider measurement in a broader sense we all use some form of measurement in our daily lives. For example when we buy stuff from the grocery shop, or when we calculate daily expenses or when we look at how much time is left for our work to finish. Measurement is integral to our life. 

When we measure changes in  a variable we are measuring some property or attribute of the variable. This property should be clearly identified and is usually defined as an operational definition of dependent variable in research. The characteristic of the variable that is to be measured should reflect the theoretical aspects of the Independent variable. For example if we are working on an intervention that improves the range of motion of a patient then measuring the amount of time taken to complete a particular activity might be an inappropriate measurement property. 

Physiotherapists have to measure different things to reach a final clinical decision and also to comment on our patient’s status whether something is good or bad. We also comment upon when to intervene or when to discontinue treatment. We measure different variables in clinics or in research studies. 

Measurement Scales

There are basically four levels of measurement described in literature. It’s important for the researcher/clinician to get a hang of these four levels of measurement. The levels are based on the extent to which a measure has the properties of a real number system. A real-number system is characterized by order, distance, and origin.

Order means that higher numbers represent greater amounts of the characteristic being measured.

Distance means that the magnitude of the differences between successive numbers is equal or logarithmic (i.e., the distance between successive numbers is an equal multiple).

Origin means that the number 0 represents an absence of the measured quality.

Nominal Scales 

Lowest measurement scale is nominal. It does not possess any of the properties of the real number system. This scale basically presents us a way of classifying the data. Examples include gender-male/female, married-not married and under 65 years of age-65 and over. The common practice of assigning numbers to distinguish between different medical diagnosis is a measurement on a nominal scale. 

Ordinal Scales 

When a measurement can be ranked according to a criteria then we say we used an ordinal scale. Ordinal scales have only one of the three properties of a real number system: order or rank. For example we can classify patients on their socioeconomic status as low, medium or high. Asking patients to rate whether their health condition has improved or deteriorated over time. One such scale which we commonly use is Global rating of change (GRC). 

Interval Scales 

Interval scales have the real-number system properties of order and distance, but they lack a meaningful origin. Interval scales allow us not only to order measurements but also identify distance between any two measurements. The commonest cited example is temperature measured on  a degree scale. 

Ratio Scales 

The highest level of measurement is the ratio scale. Ratio scales exhibit all three components of a real- number system: order, distance, and origin. Fundamental to the ratios scale is a true zero point. Examples can be height and weight.