Difficulty: Beginner | Prerequisites: Statistics Fundamentals and Sampling Methods notes (companion document)
Once you know what a population, sample, and variable are, the next question is: what kind of variable are you working with, and how was it measured? The answers determine which statistical tools you are allowed to use. You would not calculate a meaningful average of zip codes, for instance, even though they look like numbers. This set of notes covers the two major variable types (qualitative and quantitative) and the four levels of measurement (nominal, ordinal, interval, ratio). Getting this classification right at the start saves you from applying the wrong test later in the course.
Variables are either qualitative (categories, labels) or quantitative (numbers you can do maths with). Data is measured on one of four scales, from nominal (just labels) up to ratio (a true zero and meaningful ratios). The scale dictates which calculations make sense.
Qualitative variable (categorical variable)
A variable whose values are attributes, labels, or non-numerical entries. Think of it as anything you would describe with a word or category rather than a number you can add up.
Quantitative variable (numerical variable)
A variable whose values are numerical measurements or counts. In simple terms, these are numbers where arithmetic operations like addition and averaging make sense.
Nominal scale
A level of measurement where data is qualitative and has no natural order. Think of it as pure labels: names, colours, ID numbers. You can count how many fall into each category, but you cannot rank them.
Ordinal scale
A level of measurement where data can be meaningfully ordered or ranked, but the differences between values are not uniform or measurable. In simple terms, you know which is "more" or "less," but you cannot say by how much.
Interval scale
A level of measurement where data has a definite order and the differences between values are meaningful and consistent, but there is no true zero point. Think of it as: you can say "the difference between 20°C and 30°C is the same as between 30°C and 40°C," but 0°C does not mean "no temperature."
Ratio scale
A level of measurement that has all the properties of the interval scale plus a true zero point, meaning ratios between values are meaningful. In simple terms, you can say "twice as much" or "half as much" and it means something real. Examples include weight, height, and income.
Qualitative (categorical) variables
These consist of attributes, labels, or non-numerical entries. Examples from the lecture:
Car model (Toyota Camry, Honda Civic, etc.)
Birth year (used here as a label or category, not as a number to average)
Type of mobile phone (iPhone, Samsung Galaxy, etc.)
Zip code (looks like a number, but adding or averaging zip codes is meaningless)
The key test: does arithmetic on these values produce anything useful? If not, the variable is qualitative.
Quantitative (numerical) variables
These consist of numerical measurements or counts where arithmetic is meaningful. Examples from the lecture:
GPA (you can calculate an average GPA)
Height of students (measured in centimetres or inches, supports addition, subtraction, averages)
Combination lock codes (numerical and countable)
Number of text messages sent (a count, so you can total or average them)
The level of measurement tells you which statistical calculations are valid for a given variable. The four levels form a hierarchy: each level up includes all the properties of the one below it, plus something extra.
Nominal – labels only
Data is qualitative and unordered. You can count frequencies and find the mode, but ordering or averaging is meaningless.
Examples: eye colour, blood type, country of birth, jersey numbers
Ordinal – labels with a ranking
Data can be placed in a meaningful order, but the gaps between ranks are not necessarily equal.
Examples: satisfaction ratings (poor, fair, good, excellent), class rank (1st, 2nd, 3rd), education level (GCSE, A-level, degree, postgraduate)
You can say "excellent is better than good," but you cannot say the jump from good to excellent is the same size as the jump from poor to fair.
Interval – ordered with equal spacing, but no true zero
Differences between values are consistent and meaningful, but the zero point is arbitrary, so ratios do not work.
Examples: temperature in Celsius or Fahrenheit, calendar year, IQ scores
0°F does not mean "no heat." You cannot say 80°F is "twice as hot" as 40°F in any physically meaningful way.
Ratio – ordered, equal spacing, and a true zero
All arithmetic operations are valid, including ratios. Zero means the complete absence of the quantity.
Examples: weight, height, distance, age, income, number of children
100 kg is twice as heavy as 50 kg, and 0 kg means no weight at all.
A simple mnemonic: N-O-I-R (nominal, ordinal, interval, ratio), which also happens to be the French word for "black." Moving from N to R, each level adds a new property:
Nominal: classification
Ordinal: classification + order
Interval: classification + order + equal intervals
Ratio: classification + order + equal intervals + true zero
No formulas in this section, but the hierarchy diagram is worth sketching:
Nominal ⊂ Ordinal ⊂ Interval ⊂ Ratio
(labels) (+ order) (+ equal gaps) (+ true zero)
Each level inherits the properties to its left and adds one more.
Getting the measurement level right matters outside the classroom too. A data analyst who averages zip codes or treats Likert-scale survey responses (strongly disagree to strongly agree) as interval data without justification may draw conclusions the data does not support. In healthcare, confusing ordinal pain scales (1–10 self-reported) with true interval data can lead to flawed dosing models. Knowing your variable type is the first sanity check in any analysis.
Students often assume that anything expressed as a number is quantitative. Zip codes, phone numbers, and jersey numbers are numerical in form but qualitative in nature, because arithmetic on them is meaningless.
Birth year can be tricky. When used as a label ("born in 1998"), it is categorical. When used to calculate age differences, it behaves more like interval data. Context matters.
Many students believe ordinal data can be averaged. Technically, computing a mean of ordinal values (like averaging "good = 3" and "excellent = 4") assumes equal spacing between categories, which ordinal data does not guarantee. Some researchers do it anyway with justification, but exam answers typically expect you to flag the limitation.
The difference between interval and ratio trips people up. The deciding factor is whether zero means "none of the thing." 0°C does not mean no temperature (interval). 0 kg means no mass (ratio).
⚠️ Expect questions that give you a variable and ask you to classify it as qualitative or quantitative, and to name its level of measurement. These are easy marks if you practise.
⚠️ Zip codes and birth years are favourite trick examples. Know why they are qualitative despite looking numerical.
⚠️ Be ready to explain why you cannot compute a meaningful average of nominal or ordinal data (or at least why doing so requires extra assumptions for ordinal data).
⚠️ The interval vs ratio distinction often appears as "Can you say one value is twice as much as another?" If yes, ratio. If no, interval.
True or false: A zip code is a quantitative variable.
Fill in the blank: The _______ scale has a true zero point and supports meaningful ratios.
True or false: Ordinal data can always be averaged without any additional assumptions.
Fill in the blank: Data measured on the _______ scale has a definite order but no true zero point.
True or false: "Type of mobile phone" is measured on the nominal scale.
Answers: 1. False (it is qualitative, because arithmetic on zip codes is meaningless). 2. Ratio. 3. False (averaging ordinal data assumes equal spacing, which is not guaranteed). 4. Interval. 5. True.
Q: A survey asks respondents to rate their dining experience as "poor," "acceptable," "good," or "outstanding." What type of variable is this, and what is its level of measurement?
A: It is a qualitative (categorical) variable measured on an ordinal scale, because the responses have a meaningful order but the gaps between categories are not necessarily equal.
Q: A researcher records the weights (in kilograms) of 50 packages. Identify the variable type and level of measurement.
A: Quantitative variable, ratio scale. Weight has a true zero (0 kg means no weight), and ratios are meaningful (a 10 kg package is twice as heavy as a 5 kg package).
Q: Why is temperature in Fahrenheit considered interval rather than ratio?
A: Because 0°F does not represent the absence of heat; it is an arbitrary point on the scale. You cannot say 80°F is "twice as hot" as 40°F. The differences between values are equal and meaningful, but ratios are not.
Q: Give an example of a variable that looks numerical but is actually qualitative, and explain why.
A: A phone number. Although it consists entirely of digits, performing arithmetic on phone numbers (adding, averaging) produces no meaningful result. The digits serve as identifiers, not measurements.
Q: A student claims that because class ranks (1st, 2nd, 3rd) are numbers, they can calculate the average class rank to compare two schools. What is wrong with this reasoning?
A: Class ranks are ordinal. The difference between rank 1 and rank 2 is not necessarily the same as the difference between rank 50 and rank 51. Averaging ordinal values assumes equal spacing, which is not guaranteed, so the resulting "average rank" may be misleading.
Variable types and measurement levels connect directly to choosing the right descriptive statistics (covered next in most syllabi): you use mode for nominal data, median for ordinal, and mean for interval or ratio. Later, when you reach hypothesis testing, the measurement level determines whether you use a chi-square test (categorical data), a t-test (interval/ratio), or a non-parametric alternative (ordinal). Getting comfortable with these classifications now pays dividends throughout the course.
qualitative variable, categorical variable, quantitative variable, numerical variable, nominal scale, ordinal scale, interval scale, ratio scale, levels of measurement, types of data, types of variables, NOIR mnemonic, measurement scales, STA 2023, Principles of Statistics, UCF, discrete variable, continuous variable, data classification