Levels of Measurement (AQA A-Level Psychology): Revision Notes
Levels of Measurement
Understanding levels of measurement
In research methods, levels of measurement refer to the different forms by which data can be collected and organised. Understanding these levels is essential because the type of data you collect determines which statistical tests can be appropriately used to analyse your results.
The level of measurement you choose has a direct impact on the validity and reliability of your research findings. Selecting an inappropriate measurement level can lead to incorrect statistical analyses and potentially invalid conclusions.
There are three main levels of measurement that researchers work with, arranged from least to most informative. The helpful mnemonic NOIR (Nominal, Ordinal, Interval/Ratio) can help you remember this progression.
The three types of measurement data
Nominal data
Nominal data involves counting the frequency of occurrence within different categories. This represents the most basic level of measurement, providing the least detailed information about your variables.
Key characteristics of nominal data:
- Data is organised into distinct categories or groups
- Categories cannot be meaningfully ranked or ordered
- Focuses on counting how often different categories occur
- Often displayed using tally charts or frequency tables
Example: Weather Pattern Study
If you were studying weather patterns, you might count how many days in a week were rainy versus sunny:
- Rainy days: 4
- Sunny days: 3
Although you can determine that there were more rainy days than sunny days, you cannot say that "rainy" is somehow greater than or better than "sunny" - they are simply different categories.
Ordinal data
Ordinal data involves ranking information in order, typically using rating scales. This level provides more detailed information than nominal data because it shows relative positioning between data points.
Key characteristics of ordinal data:
- Data can be arranged in rank order from highest to lowest (or vice versa)
- Shows which values are better or worse than others
- The gaps between different ranks may vary considerably
- Cannot determine the exact differences between individual data points
Example: Race Finishing Positions
Consider finishing positions in a race - first, second, third, fourth place, and so on:
- 1st place: Runner A
- 2nd place: Runner B
- 3rd place: Runner C
While you know that first place performed better than second place, you cannot determine from the ordinal data alone whether the gap between first and second place was the same as the gap between second and third place. One athlete might have won by several seconds, whilst another might have lost by mere milliseconds.
Interval/ratio data
Interval/ratio data uses standardised units of measurement and represents the most informative and precise form of measurement. This type of data uses equal measurement intervals throughout the scale.
Key characteristics of interval/ratio data:
- Uses standardised measurement units (such as time, weight, temperature, or distance)
- Equal intervals between all measurement points
- Provides the most accurate and detailed information
- Allows for precise mathematical calculations and comparisons
Example: Precise Measurements
Examples of interval/ratio data include:
- Measuring reaction times: 245ms, 267ms, 198ms
- Recording temperatures: 23°C, 18°C, 31°C
- Measuring distances: 5.2m, 3.8m, 7.1m
Unlike the previous levels, interval/ratio data tells you not just that one measurement is greater than another, but exactly how much greater it is.
The distinction between interval and ratio data relates to whether there is an absolute zero point. Ratio data has an absolute zero (like weight or distance), whilst interval data has an arbitrary zero point (like temperature in Celsius, where zero degrees does not mean "no temperature").
Connection to statistical analysis
The level of measurement you use directly influences which statistical tests are appropriate for analysing your data. Researchers must identify their measurement level before selecting analytical techniques, as different levels require different approaches to statistical testing.
Higher levels of measurement (interval/ratio) generally allow for more sophisticated statistical analyses, whilst lower levels (nominal) restrict the types of tests that can meaningfully be applied to the data.
Remember!
Key Points to Remember:
- NOIR mnemonic helps you remember the three levels: Nominal, Ordinal, Interval/Ratio
- Nominal data counts frequencies in categories - least informative but still useful for basic comparisons
- Ordinal data ranks information in order but gaps between ranks can vary significantly
- Interval/ratio data uses equal, standardised measurement units and provides the most precise information
- Your measurement level determines which statistical tests you can use to analyse your results