The constant sum scale asks respondents to allocate a fixed number of points among a set of options. It helps you understand how respondents prioritize different aspects of a product, service, or concept.
Unlike other preference scales such as Likert scales, constant sum scales capture relative preferences by making respondents compromise.
Let’s take an in-depth look at what constant scale is, its applications, and best practices.
Understanding the Constant Sum Scale
A CMS (constant sum scale) question is a survey that asks respondents to attribute a single number to a set of options. For example, asking respondents to attribute 100 points to four different product features according to how important each element is to them.
The constant sum scale is based on the psychological principle of trade-off. Unlike other rating scales, such as Likert scales or semantic differential scales, the constant sum scale forces respondents to make trade-offs among the options, which reveals their relative preferences.
The constant sum scale also follows the psychological principle of relative importance. People don’t place the same value on everything; some things mean more to us than others. With the constant sum scale, you can find out what matters most to your respondent.
Applications of the Constant Sum Scale
A constant sum scale allows you to see how users rank different features or benefits and how they trade them off between them. This allows you to design, position, and price products or services to be a perfect fit for user needs and expectations.
- Decision-making
You can also use the constant sum scale for assessing different options and factors in your decision-making process. It helps you to make smart, consistent judgments that provide you with the most value.
For example, you can use a constant sum scale to see how students rank their college choices using factors such as academic quality, cost of attendance, location, school culture, and more.
- Policy Evaluation
As a policy-maker, you can use the constant sum scale to assess the impact of different policies on different outcomes or stakeholders. It allows you to compare and rank policies or interventions based on their effectiveness or efficiency.
For example, you can use the constant sum scale to measure how people distribute their support among different environmental goals, such as reducing greenhouse gas emissions, preserving biodiversity, promoting renewable energy sources, etc.
Designing and Administering Constant Sum Scale Surveys
- Define the Question’s Purpose
Clearly state what you want to learn from respondents. What are the key factors or items that you want them to rank or prioritize?
For example, instead of asking “How important are these factors to you when buying an air condition system?”, ask “How would you allocate 100 points to each of these factors based on how important they are in your choice of AC system?”
Also, add clear instructions to ensure respondents understand how to use the constant sum scale. For example, “the more points you assign to a factor, the more important it is”.
- Use Clear and Mutually Exclusive Response Options
The options should not overlap or contradict one another and should cover different aspects of the concept or product. Also, avoid using options that are too vague or overly specific.
For example, if you want to know how important specific features of a headphone are, you can use options such as “battery life”, “noise cancellation”, “speaker”, “microphone”, and “price”. The options are distinct and tackle the most important components of a headphone.
- Choose a Consistent and Appropriate Scale for the Response Options
Ensure the scale is user-intuitive and configure it to fit the total number of points or units respondents have to assign. You should also avoid using scales that are too large or too small, such as 0 to 1000 or 0 to 10.
User-friendly scales allow respondents to easily assign points to each option without having to do complex calculations or conversions. For example, if you want respondents to assign 100 points to four different options, you could set up a scale that goes from 0 to 100, in steps of 10.
- Consider the Level of Precision You Want to Achieve
The total and response options should reflect the degree of differentiation that you want to measure among the options.
For example, if you want to rank 5 product features with 100 total points, a scale from zero to 100 with 1-point increments is much more precise than 10-point increments.
- Take Note of the Cognitive Load the Response Options Imposes
The total and answer options shouldn’t be too complex for respondents to figure out. It is recommended to use a scale of 100 points because it works with percentages and is easy to figure out.
So, if you are asking respondents to allocate 100 points among five options, don’t use a scale that ranges from 0 to 1 in increments of 0.01. This would require respondents to do complex calculations to assign points to each option.
Related – Questionnaire Design: 10 Questioning Mistakes to Avoid
Analyzing and Interpreting Constant Sum Scale Responses
Methods and Techniques for Constant Sum Scale Data
- Mean Scores
This is the simplest way to analyze the constant-sum scale data. All you have to do is divide the total number of points of an option by the number of respondents, and you have the mean score.
- Proportion of Total Points/Percentages
This method provides a more nuanced understanding of how respondents rank the items. For example, if an item receives 20% of the total points, this means that it is twice as important as an item that receives 10% of the total points.
- Factor Analysis
This method groups response options into factors based on their correlations. It enables you to simplify the data and identify the underlying factors that drive respondents’ preferences.
- Regression Analysis
You can also model the interaction between the points assigned to one option (dependent variables) and the points assigned to other options (independent variables). This allows you to figure out how much each option affects another option and forecast future outcomes.
How to Interpret and Report Constant Sum Scale Results
- Focus Item Relative Importance
Instead of focusing on the absolute score of each item, look at the relative importance of each option. Then compare the average score, percentage, or factor load for each option.
- Visual presentation
Use graphs and diagrams to display your data to make it easier to understand. You can bar charts, pie charts, line graphs, scatter plots, or radar charts, depending on your data and message.
- Highlight Data Limitation
Constant sum scale data only captures the relative order of importance, not the absolute importance. So, you have to be cautious when inferring absolute values for response options.
Common Challenges and Considerations
- Higher Cognitive Load
Constant sum scale is cognitively demanding and time-consuming for respondents, especially if there are many options to rate. This can lead to respondent fatigue, frustration, or dropoff.
Respondents could allocate points based on the order or presentation of the options or attributes, rather than their true preferences or priorities. They could also give equal or similar points to all options to avoid making difficult trade-offs or expressing negative opinions.
- Range Sensitivity
Using very small or large points affects the level of detail respondents can provide when assigning points. It can also make it difficult to compare or group results when different points are used for different questions or groups of respondents.
However, the most effective way to mitigate constant sum scale challenges is to validate and pre-test the constant sum scale before using it in your research project. Validation ensures that the constant sum scale measures what is expected to be measured and is free of bias and measurement errors.
Alternatives to the Constant Sum Scale
1. Paired Comparison
This method asks respondents to compare two items one at a time and select the one they like the most. It’s a great way to get an idea of how important different items are, but it is very time-consuming when there are a lot of items to compare.
2. Rank Order
With rank order, respondents rank items in order of preference. It helps you understand respondents’ overall order of preference.
For example, you can use rank order to ask respondents to rank their favorite lunch combination. However, it is not very effective when you want to understand the relative importance of items that are close together in rank.
3. Discrete Choice model
It’s a statistical model for estimating the relative significance of various items, particularly large numbers of items. It’s a more complex and holistic model than the others, but it’s also more challenging to understand and explain.
Conclusion
Constant sum scales offer better and more accurate data about respondents’ preferences compared to other scale rating systems. However, it is important to understand its limitations and best practices to make informed decisions and product or service improvement.