Constant Sum Question Type: Definition, Examples, and How to Use It

A constant sum question type asks respondents to split a fixed number of points, dollars, or percentage units across several options so the total always adds up to the same number.

Researchers use it when a standard rating scale lets everyone call every option “important” and hides what actually matters most.

This guide breaks down what a constant sum question is, how it differs from rank order and Likert scale questions, and when it fits your survey better than a rating scale. You will also find real examples, common setup mistakes, and the steps to build one in a survey tool.

What is a constant sum question type?

A constant sum question is a survey question that asks a respondent to divide a set total, usually 100 points, dollars, or percent, across a list of items. Every allocation must add up to that fixed number before the respondent can move on.

A constant sum question type collects ratio data, meaning the numbers show both order and the exact distance between items, not just which one wins.

That distance is what sets it apart from most other scales. A respondent who gives 40 points to price and 10 points to design has told you price matters four times as much, not just that it ranks higher. That level of detail is hard to get from a standard rating scale, and it is the whole reason researchers reach for this question type.

How does a constant sum question work?

A constant sum survey question follows the same basic structure whether you are asking about budgets, time, or importance. Here is the sequence a respondent goes through:

1. They see a list of options, usually four to seven items.
2. They see the total they must reach, most often 100 points or 100 percent.
3. They type a number into each option’s field, or drag a slider.
4. A running total updates as they go.
5. The survey blocks submission until the numbers add up to the required total.

This built-in math check is what makes the data usable. If a tool let respondents submit 85 or 130 points, the resulting averages would not mean anything at the analysis stage. That validation step is why most survey platforms display a live running total next to the question as respondents fill it in.

Constant sum vs. Rank order and Likert scale questions

Constant sum, rank order, and Likert scale questions are three of the most common survey question types, but they capture very different amounts of detail. Picking the wrong one is a frequent reason researchers end up with data they cannot use.

Rank order questions show you the sequence of preference, nothing more. A Likert scale question tells you how strongly someone feels about one item at a time, using ordinal data, meaning the values have order but not a fixed distance between them. Constant sum adds a layer neither format offers: both the order and the size of the gap between items.

Choose constant sum when you need to know not just that price beats design, but by how much.

When should you use a constant sum question?

A constant sum scale earns its place in a survey when the research question is really about proportion, not a single winner. It works best in a handful of recurring situations.

• Budget allocation.
  Asking how a respondent would split a fixed budget across categories or features in a budget allocation survey.
  
• Time distribution.
  Measuring how someone divides hours across tasks, channels, or activities in a typical day.
  
• Feature trade-offs.
  Getting importance weighting across product attributes when respondents cannot rate everything as equally important.
  
• Share of wallet.
  Finding out what percentage of category spending goes to each brand a customer already uses.

Avoid this format when you only need a single top choice, or when your list runs past seven or eight items. Past that point, the arithmetic becomes the hardest part of the question, not the actual answer.

Real-world examples of constant sum questions

Seeing the format in context makes it easier to write your own. These three examples cover common research scenarios.

• Household budget example.
  “You have $100 to spend on groceries this month. Divide it across produce, meat, dairy, and packaged goods based on how you actually shop.” This version works well for average order value research, since it mirrors how customers really allocate spending.
  
• Feature importance example.
  “Distribute 100 points across these five software features based on how important each one is to your purchase decision: ease of use, price, integrations, customer support, and reporting.”
  
• Time allocation example.
  “Divide 24 hours across a typical workday: focused work, meetings, email, breaks, and administrative tasks.”

Each example keeps the list under seven items and states the total up front. Those two details are what keep completion rates high.

How to create a constant sum question in QuestionPro

Building a constant sum question takes only a few steps inside a survey builder. Here is the process in QuestionPro’s survey software:

1. Open your survey and click Add Question.
2. Select Constant Sum from the question type library.
3. Enter the question text and add each option as a row.
4. Set the required total, such as 100 points or 100 percent.
5. Choose whether respondents enter numbers directly or use a slider.
6. Turn on the running total display so respondents can track their progress.

Keeping the option list short and the total round, such as 100 or 10, makes the question faster to finish on both desktop and mobile.

Common mistakes to avoid with constant sum questions

A constant sum question produces clean data only when it is built carefully. A few recurring mistakes account for most of the unusable data researchers see from this format.

• Listing too many options: Long lists raise the cognitive load on respondents. Pew Research has found that higher cognitive burden drives up item nonresponse in surveys, and constant sum questions with eight or more rows show the same pattern.
  
• Skipping randomization: Showing options in the same order for every respondent introduces position bias into the results.
  
• Using an odd total: Totals like 47 or 83 make the mental math harder than round numbers like 100.
  
• Ignoring mobile respondents: A long list of number fields is harder to complete accurately on a small screen than on a desktop.
  
• Treating the output as ordinal data: The whole value of this format is the ratio data it produces. Analyzing it like a ranking question throws that advantage away.

How to measure and analyze constant sum data

Once responses come in, the goal is turning individual allocations into a clear priority picture across the full sample. This table shows the metrics researchers rely on most.

Mean allocation is the number most teams report, but standard deviation matters just as much. A high mean with high variance means the group is split, not aligned, and that split is often more useful to a product or customer experience team than the average by itself.

Getting more than a ranked list from your survey

A constant sum question will not replace every rating scale in your toolkit, and it should not try to. It earns its place when the real question is about trade-offs, not top picks. Used with a short, well-randomized list of options, it turns a simple survey question into a weighted priority map you can act on directly.

If you want to test one in your next study, QuestionPro’s survey templates include ready-made examples you can adapt in minutes.

Frequently Asked Questions (FAQs)