Survey Scales: Definition, Types, Examples
A survey scale is a tool for converting subjective experience into scale points.
Key direct measurement methods:
- association of the answer with a scale point;
- ranking of objects based on a given criterion;
- comparison of paired objects based on a certain principal;
- distribution of objects along a scale of “equal” intervals;
- scoring of objects with a limited number of points.
Depending on what mathematical operations are applicable to data processing, scales fall into the following categories: nominal, ordinal, interval, and ratio scales.
The nominal scale identifies whether an object has a certain attribute. On this scale, different attributes cannot be ranked or put in order, as they cannot be compared in terms of “bigger–smaller” or “better–worse”.
Nominal data can be used to calculate the percentage of objects with a certain attribute, and the mode (the most common value of the attribute). Two rows of nominal data can be used to create a contingency table and test a hypothesis on the link between attributes.
Sometimes, to facilitate the analysis, the nominal scale is transformed into binary, where 1 means that the object has a given attribute, and 0 means it does NOT have the attribute.
An example of the nominal scale:
Which of the following car brands are you aware of?
The ordinal scale helps to identify whether an object has one of ordered attributes.
Ordinal data can be used to calculate the percentage of objects with a certain attribute, the mode, and the median.
Types of ordinal scales:
- A ranking scale: respondents are asked to rank objects in an ascending or descending order of its attribute.
- A point scale: respondents are asked to score objects with points from a given range.
An example of the ordinal scale:
Please rank the drinks from the list in the order of preference, where 1 is the most preferred, and 5 is the least preferred:
- Coke 3
- Dr Pepper 1
- Pepsi 4
- Seven Up 5
- Sprite 2
On the interval scale, attribute values are ordered with equal intervals between them, and have an assigned measurement unit.
The feature of interval scales is that they do not have a true zero. The zero value is arbitrary and does not mean that the object does not have a measured attribute. Examples include calendars with different starting points, or temperature scales with different zero values (Celsius, Fahrenheit, Kelvin).
Since the attribute is measured in certain units, its values can be added or subtracted. However, multiplication and division operations do not make sense due to the lack of a true zero that represents the absence of the attribute. For instance, we can say that it is 7 degrees hotter today than yesterday, but not X times hotter.
Interval data can be used to calculate the percentage of objects with a certain attribute, the mode, the median, the mean, the variance, and the standard deviation.
Moreover, the interval scale can be compressed and expanded indefinitely. For instance, a 0 to 100 scale can be compressed to a 0 to 1 range or shifted 50 points to the left to get a -50 to +50 range.
An example of the interval scale:
On a scale of 1 to 10, please rate your satisfaction with the service:
Not at all satisfied 1 2 3 4 5 6 7 8 9 10 Extremely satisfied
The ratio scale has all the same properties as the other scales, plus a true zero that represents complete absence of an attribute. Thanks to this feature, ratio scales can be used to compare different objects. For example, a respondent can be 1.5 times older, go to malls 2 times less often, or make several times more expensive purchases than other respondents.
Intervals between values do not have to be equal, but they need to have distinct boundaries. For instance, if one interval on an age scale ends on 30 years, the next one has to start with 31 years.
An example of the ratio scale:
How old are you?
After creating a scale for measuring an attribute, the researcher needs to ensure that it meets the key requirements to scales:
- The scale values accurately capture the essence of what is being researched.
- The scale includes all relevant answer options, including “Other” and “Do not know / No answer”.
- The scale is precise enough to measure attributes.
- The scale produces similar results with minimal errors for each measurement.
- Measurement results are consistent with real characteristics of an object of interest.
Next, let’s take a look at the scales most commonly used in marketing research: Likert scale, semantic differential, Stapel scale, NPS and CSI, Thurstone scale, and mnemonic scales.
With the Likert scale, we ask respondents to rate how much they agree (or disagree) with a statement on a scale of 1 to 5 (or 7).
To create a Likert scale in the PeakPoll survey builder, the Scale or Matrix question types should be used. The Scale type of questions allows you to assess one statement, and Matrix — a group of statements.
Pro tip: For Matrix, we recommend that you enable random question order in the question settings, as respondents tend to focus more on the first couple of options on a list.
Please evaluate our brand in terms of the following statements:
I would recommend this brand to others.
I have had a positive experience with this brand’s customer service.
I would purchase from this brand again.
- Strongly Disagree
- Strongly Agree
With the semantic differential method, we ask respondents to assess an object using bipolar scales where the two opposite end points are represented by antonyms. Such as, “classical–contemporary”, “adult–teenage”, “reserved–expressive”. The results give you a descriptive image of how the target audience sees the object.
To apply this method, the Scale question type should be used. In the question settings, you should put the antonym pairs into the Ranges field, and the scale points — into Options. The most commonly used scale has 7 points — from -3 to +3, including 0. But you can put smaller or bigger fractions.
Pro tip: To help respondents to understand your question, you can simplify the scale by replacing the numbers with generic symbols (for example, a dot or stars).
Please rate the app based on the following criteria:
- Unattractive 0 1 2 3 4 5 Attractive
- Irrelevant 0 1 2 3 4 5 Relevant
- Slow 0 1 2 3 4 5 Fast
- Confusing 0 1 2 3 4 5 Stimulating
- Difficult to use 0 1 2 3 4 5 Easy to use
Recommendations for working with semantic differential
Create a list of associations the object evokes — adjectives, collocations or phrases. For example, a watch can be digital, luxury, or water-proof. Then, think of antonyms for each association to set extreme ends of the range.
Do not put all negative adjectives on one side of the scale. That way you can avoid situations where respondents with a strong opinion on the object automatically choose outermost points on one side of the scale without reading hints.
Calculate the average score on each scale to see the list of associations for the perceived object.
[picture with zigzags on the scales]
To assess the attitude to an object using the Stapel scale, respondents are asked to rate how well each given statement describes the object of interest. In essence, this scale is a simplified version of the semantic differential method.
Features of the Stapel scale:
- associations are presented on a unipolar scale (without antonyms);
- scale points are assigned numerical values;
- the scale has a total of 10 points: from -5 to +5.
Pros of the Stapel scale:
- you do not have to come up with antonyms,
- the scale points have more fractions.
Cons of the Stapel scale:
- the wording (positive, negative, or neutral) affects the respondent’s ability to give coherent answers.
Please rate how well these statements describe our company:
- Polite employees -5 -4 -3 -2 -1 +1 +2 +3 +4 +5
- Good location -5 -4 -3 -2 -1 +1 +2 +3 +4 +5
- Convenient business hours -5 -4 -3 -2 -1 +1 +2 +3 +4 +5
- High interest rates -5 -4 -3 -2 -1 +1 +2 +3 +4 +5
NPS is one of the most popular customer experience metrics. This index reflects the audience’s willingness to recommend certain goods or services to others. To calculate NPS, respondents are asked to rate the likelihood that they will recommend the product to their friends and acquaintances on a scale of 0 to 10. Based on the results, consumers are divided into three groups: promoters (a score of 9–10), passives (7–8), and detractors (0–6). The final NPS is calculated by the formula: NPS = (number of promoters – number of detractors) / number of respondents * 100. The range varies from -100 to +100.
To measure the index, use the Scale question type in our survey builder. Fill in the Question text field and put 0 to 10 in answer options. If necessary, you can disable the text labels for the lower and upper range limits (“not at all likely” and “extremely likely”) in the question settings.
Pro tip: You can replace the numerical scale with a color gradient, stars, or emojis.
How likely are you to recommend our services to your friends and acquaintances?
Not at all likely 0 1 2 3 4 5 6 7 8 9 10 Extremely likely
CSI is not as popular as NPS, but you can still often find it in surveys. The index measures how satisfied a customer is with a particular product feature, taking into account its importance to the customer. This method consists of two questions: first, respondents are asked to rate the importance of listed parameters, and then — their satisfaction with each of them. The audience is split into four segments for each parameter: “important and satisfied”, “important and not satisfied”, “not important and satisfied”, and “not important and not satisfied”.
Quite often, survey creators ask about satisfaction with product features, but omit the question about their importance to the client. This skews the results: you get a more generalized picture of “satisfied–not satisfied” without subtleties.
To measure CSI, use the Scale questions (for assessing one feature) or Matrix questions (for several features) in our survey builder. Fill in the Question text field, put 0 to 10 in answer options and select their appearance (numerical scale, color gradient, stars, or emojis).
Pro tip: We recommend that you put questions about importance and satisfaction on different survey pages to reduce the influence of the first answer.
Please rate how much the quality of the product matters to you.
Doesn’t matter at all 0 1 2 3 4 5 6 7 8 9 10 Matters a lot
Please rate how satisfied you are with the product features below:
NOT satisfied at all 0 1 2 3 4 5 6 7 8 9 10 Extremely satisfied
- Product mix
This method was developed to assess the attitude towards objects in psychological and social research. It was first described in 1929 in a study of public attitudes towards the church.
The scale is created in several steps:
First, you should generate a large pool of statements about the subject of research. You can source them from publications, opinions of people around you, group discussions, etc. Then, make a short list, leaving only the statements that fit the following criteria:
- not too wordy,
- without professional jargon,
- relevant to the present, not the past,
- without generalizing words like “all”, “always”, “no one”, “never”, etc.
At the second stage, the statements are reviewed by representatives of the target audience (experts). Their task is to split the statements into 11 categories removed from each other at subjectively equal intervals, based on positive or negative stance towards the research subject and its intensity.
Then, each statement is assigned a score equal to the median of the points it received from the experts. The statements with divided expert opinions (check for those using the interquartile range) are discarded.
In the end, you get a handy scale for measuring what you are interested in. The survey respondents are given 11 statements and asked to choose the ones they agree with. Based on scores of selected statements, you can calculate the mean value which will indicate the respondent’s attitude towards the object.
An example of the Thurstone scale:
|The interests of the top management and the staff do not coincide in the majority of cases||Agree||Disagree|
|A manager should not discuss his/her personal problems with subordinates as it undermines his/her authority||Agree||Disagree|
|When a manager keeps subordinates at arm’s length, they perform their tasks more diligently||Agree||Disagree|
|In order to be an effective manager, you should be aware of your subordinates’ problems||Agree||Disagree|
|The company must fulfill its obligations towards the staff despite financial difficulties||Agree||Disagree|
|An employee’s salary should be based on the financial situation of his/her family||Agree||Disagree|
|One company employee should not receive a salary 10x bigger than that of his/her colleagues||Agree||Disagree|
|Promotions are dictated primarily by the management’s feelings towards you, rather than your professionalism||Agree||Disagree|
|Diligent performance of your job duties is a necessary and sufficient condition for your promotion||Agree||Disagree|
|It is okay for a manager to accommodate a request from a good employee in breach of the general rules||Agree||Disagree|
|A close-knit team is nice to work in, but hard to manage||Agree||Disagree|
Mnemonic, or visual, scales take care of two things: they make it easier for the researcher to create hints and for the respondent to select answers. Unlike text descriptions, images serve as unambiguous representations of an array of possible responses to a question.
In the PeakPoll survey builder, for Scales and Matrix question types, you can change the settings to replace numerical values with a color gradient, stars, or emojis. What’s more, instead of default scale point images, you can use custom ones. To do that, you need the images themselves and CSS coding skills. Learn more about how to add a CSS code to your survey here. Keep in mind the number of points in the scale to make sure you replace all of them. The default image size is 48*48 px.