You’ve planned your research design and identified the type(s) of information you want to gather. Now you’re ready for the next step in developing your research survey: deciding which measurement and scaling techniques you want to use to collect your data.

**Characteristics of Market Research Scales: Description, Order, Distance, Origin**

There are four basic characteristics of market research scales: description, order, distance, and origin. All scales have **description**, or the unique labels that are used to define each value of the scale.

**Order** refers to the relative sizes or positions of the descriptors, signified by labels like “greater than”, “less than”, or “equal to”. Not all scales have order. Gender scales have no order, but age scales do (age 18 comes before/ is “less than” age 25).For a scale to have **distance**, there must be an absolute difference between the scale descriptors that can be expressed in units. For example: “The number of children under the age of 18 living in your household” is a scale that has distance because a household with three children is one more than a household with two children. If a scale has distance, it also has order. But, the reverse is not necessarily true.

**Origin** is the fourth and highest-level characteristic. It means that the scale must have a fixed beginning or true zero point, such as in a scale asking for personal income (where it is possible for someone to report $0). Many scales used in market research do not possess a true zero. If a scale has origin, it also has distance, order, and description.

## Primary Scales of Measurement: Nominal, Ordinal, Interval, Ratio

The primary scales of measurement are very similar to their characteristics. A **nominal scale**, like the description characteristic, is the most basic. The numbers (or letters or symbols) in a nominal scale only serve to identify objects, like gender classification. It does not matter if females are assigned a 1 or a 2 (or any other number, letter, or symbol), and order is not a factor.

Conversely, the numbers assigned in an **ordinal scale** do have meaning: they indicate the relative position between objects (but not the differences between them). One of the most commonly used ordinal scales in market research is preference rankings.

**Interval scales** possess all the information in an ordinal and nominal scale, while also enabling you to compare the differences between objects. There is a constant equal interval between scale values, such as in a Fahrenheit temperature scale.

Lastly, **ratio scales** feature all the characteristics of nominal, ordinal, and interval scales, plus an absolute zero point (origin). You can do the most advanced statistical techniques during analysis with a ratio scale.

**Putting Scaling Techniques to Work in Your Survey**

Now that you understand the characteristics and primary types of scales available to you, how do you know which ones to choose? This will depend on the type of information you want to collect and how you want to analyze it. It may be helpful to think backwards and ask yourself: “How will I use or measure this data?”

All scales are technically nominal scales, but they may also be ordinal; or ordinal and interval; or ordinal, interval, and ratio if they possess the right characteristics. As you move from a nominal scale to an ordinal scale and so on, more advanced analytical techniques are available to you, should you need them.

Nominal-only scales are the most limited: you can only perform frequency counts and determine things like percentages and mode. You can count the 50 females and 75 males who took your survey and easily create percentages using these numbers – in this example, 40% of survey takers were female.

Use an ordinal scale, such as a reorder-ranking question type, when you want to calculate percentiles or medians, for example. This is commonly used for ranking objects, like favorite fast-food restaurants, logos, or package designs, and is also used in conjoint analysis.

If you also want to understand the extent of differences between ratings – in other words, how much more or less a respondent prefers concept A to concept B – a distribution question type (a type of interval scale) allows you to compute range, mean, and standard deviation.

Data related to age, income, costs, sales, and market shares would be best gathered using a ratio scale where any type of statistical analysis can be performed.

**The Takeaway**

Choosing the right question types for your survey is wholly dependent on the scaling and measurement techniques you wish to employ. Having a strong foundation in the advantages and disadvantages of each type of scale will enable you to build a better survey and obtain more useful data from respondents.