Page 282 - Ai_V3.0_c11_flipbook
P. 282
Levels of Measurement Levels of
The method used to measure a collection of data Measurement
is known as the level of measurement. Not all data
can be handled in same manner. It makes sense to
classify data sets according to several criteria. Some Quantitative Qualitative
are quantitative, others are qualitative. Some datasets
are continuous, whereas others are discrete. Qualitative
data might be either nominal or ordinal. Quantitative
data can also be categorised into: interval and ratio. Discrete Continuous Nominal Ordinal
The four levels of measurement:
Nominal Ordinal Interval Ratio
Categorises and labels variables
Ranks categories in order
Has known, equal intervals
Has a true or meaningful zero
Note that a true zero refers to a scale where 0 indicates the absence of something.
Nominal
In nominal measurement, the numerical values represent a unique “name” of the attribute. The cases may be ordered in
any manner. For example, jersey numbers in cricket are measured at the nominal level. A player with the number 20 is
not better than a player with the number 3 and is certainly not twice better whatever number 10 represents, instead the
numbers act as a label to identify different players.
Nominal variables are like labels or categories—think car brands or seasons. They can’t be ranked or used in calculations.
Examples include eye colour, gender, or smartphone brands. Even if numbers are involved, like a player’s jersey number,
they’re just identifiers, not for calculations or comparisons. True zero point does not exist in nominal data.
Examples:
What is your hair colour?
1 – Brown Where do you live?
What is your gender?
M – Male 2 – Black A – North of the equator
F – Female 3 – Blonde B – South of the equator
4 – Grey C – Neither in the international space station
5 – Other
Ordinal
In ordinal measurement, attributes can be ordered. The distance or interval between attributes is irrelevant here.
For example, in a survey, you can code educational qualification as, 0 = secondary; 1 = senior secondary; 2 = graduation; 3
= post-graduation; 4 = PhD. In this level of measurement, higher numbers mean more education. However, is the distance
from 0 to 1 equal to 3 to 4? Of course, no. The interval between the values cannot be interpreted as an ordinal measure.
Ordinal data consists of categories arranged in a specific order, like rating a meal from “unpalatable” to “delicious.”
Although words, not numbers, are used, there’s a clear progression from negative to positive. However, the actual
difference between each category can’t be measured. Like nominal data, ordinal data can’t be used in calculations. True
zero point does not exist in ordinal data.
280 Touchpad Artificial Intelligence (Ver. 3.0)-XI

