The level of measurement refers to the classification that shows the data in the figures assigned to variables. According to Engelhard (2017), Stevens, a renowned psychologist, developed the best four levels of classification- nominal, ordinal, interval, and ratio. He extracted his framework of distinguishing levels of measurement from psychology. However, Stevens's work has met critics from scholars of different disciplines. This paper summarizes Steven's four levels of measurement proposed in 1951 and how a researcher utilizes the levels in examining the outcomes of a study.
In the nominal scale, labels are used to identify categories of items. However, numbers can be used to refer to variables, but they lack numerical value. Nationality, language, ethnicity, among others, form good examples of classification in nominal. In the aspect of the English language, parts of speech can be classified in conjunctions, pronouns, nouns adjectives, adverbs, among others. From a statistical point of view, nominal marks the lowest level of measurement due to its lack of specific numerical value. In mathematical operations, the level uses only equality and inequality to apply to its sets of membership. Besides, the mode is only the measure of central tendency in nominal (Engelhard, 2017).
Delegate your assignment to our experts and they will do the rest.
Ordinary scale permits the ranking of data in order from the first to the last, but there is no relationship showing a degree of difference between them. For instance, when measuring health, values like healthy versus sick are used. In a school set up, John’s position can be put at 10 and James’s at 20, but it does not imply that John’s performance is twice that of James. Therefore, statements of greater than and less than are used (Bruschi, 2016). Median and mode are the only two measures of central tendency applicable in an ordinary scale.
Quantitative factors show the degree of difference in the interval scale without including the ratios between them. Temperature is an example of an interval that uses the scale of Celsius and two points-boiling and freezing, with 100 intervals. Interval scale permits arithmetic mean, mode, and median to be performed as a central tendency. Furthermore, standard deviation and range are allowed as its measure of statistical dispersion.
Besides, there is a ratio scale that has a meaningful zero value. It derives its name from the fact that it is a relative ratio between the sizes of a progressing quantity unit size of the same type. Fields like science and engineering manipulate measurements using ratio scales. Examples are duration, mass, and energy. As opposed to other scales of measurement, ratios non-arbitrary zero point marks to meaningful as one can relate that a particular object is thrice the other (Engelhard, 2017). Measures for central tendencies applied here are harmonic and geometric mean. Others are arithmetic mean, mode, and median. Besides, the coefficient of variation and all other statistical are permitted in ratio scale.
The diagram of the relationship between precision and level of measurements
Knowing the level of measurement aids researchers in interpreting the data from the variables. For instance, when the researchers identify that the measure is nominal, they understand that short codes are represented by numerical values, which signifies long names. Furthermore, when the level of measurement is identified, the researcher is in a position to decide on the type of appropriate statistical analysis regarding the assigned values ( Engelhard, 2017). For example, if researchers know that the measure is nominal, then they understand that the average of the outcomes cannot be averaged.
References
Bruschi, A. (2016). Measurement in social research: Some misunderstandings. Quality & Quantity , 51 (5), 2219-2243. https://doi.org/10.1007/s11135-016-0383-5
Engelhard, G. (2017). Cogitations on invariant measurement . Springer.
