# Framing Bias: Why Your Questionnaires Already Contain the Answers Learn how framing bias and leading questions distort questionnaire results. Discover concrete examples and better formulations to collect more honest, useful answers. - Canonical URL: https://www.harmate.com/en/blog/framing-bias-why-your-questionnaires-already-contain-the-answers - Author: Harmate Team - Published: 2026-05-26 - Updated: 2026-05-26T06:14:32.670993+00:00 - Language: en ## Content # Framing Bias: Why Your Questionnaires Already Contain the Answers Imagine a blue house. What do you see? Probably a blue house. Not a red house, not a green house, not a stone house. A blue house. Now, if I mention a house again in a few seconds, there is a good chance that your brain will keep that color in the background. Even if I do not say the word "blue" again. This is what many questionnaires do. They install an image, a frame, an expectation, then ask the respondent to answer. The problem is that this image does not come from the respondent. It comes from the question. Strictly speaking, the blue house example is first a form of priming: an idea is introduced and influences what comes next. In questionnaires, this priming becomes a framing problem when the wording of the question shapes how the respondent understands the topic and formulates the answer. A poorly worded question does not just collect an opinion. It can produce one. ## A question is never neutral We like to believe that a questionnaire works in a simple way: you ask a question, the person answers, and you get their opinion. In reality, a question does more than ask. It frames the topic, selects what seems important, sometimes suggests a direction, and makes some answers easier than others. Take these three formulations: **"What did you think of this training?"** **"What did you appreciate in this training?"** **"What were the strengths of this training?"** They refer to the same object, but they do not produce the same type of answer. The first one opens a fairly broad space. The second already pushes toward the positive. The third assumes that strengths exist and mainly asks the respondent to identify them. The difference seems small. It is not. In the first case, you ask for an opinion. In the second, you ask for appreciation. In the third, you almost ask for validation. The house is already blue. ## What framing bias means in a questionnaire Framing bias appears when the way a question is presented influences the answer obtained. This is not a matter of style. It is a data quality problem. A biased questionnaire can give you the impression that you have collected sincere opinions, when in reality you have mostly measured the respondents' reaction to the frame imposed by the question. Example: **"Is our new tool easy to use?"** This question is not neutral. It already introduces the idea that the tool should be easy. The respondent evaluates the experience through that expectation. A cleaner formulation would be: **"How would you describe your experience using the tool?"** Better still: **"At what point did the tool feel clear, difficult, or confusing?"** In this case, you do not force an interpretation. You let the person describe what they actually experienced. ## Bias often starts in your intention Framing bias rarely appears by accident. It often appears when you already want the answer to go in a certain direction. You created a training program, launched a tool, organized an event, designed an onboarding process, or implemented a new organization. You want to know what people think about it. But without realizing it, you may ask a question that protects your intention. You ask **"What did you like?"** instead of **"What stood out to you?"**. You ask **"How can we improve this experience?"** instead of **"What helped, what was missing, or what slowed you down in this experience?"**. You ask **"Did you understand the value of this initiative?"** instead of **"How did you understand this initiative?"**. In each case, your expectation slips into the question. You are no longer simply asking what the person thinks. You are already showing them the direction. ## Words bias answers Words are not neutral. They carry an interpretation. Calling something a **problem** does not produce the same response as calling it a **point of attention**. Calling something an **innovation** does not produce the same response as calling it a **change**. Calling something **resistance** does not produce the same response as calling it **disagreement**. Calling something an **improvement** does not produce the same response as calling it a **correction**. Take this question: **"Why are some employees resisting the change?"** It already contains a hypothesis: the change is legitimate, the issue comes from employees, and their reaction should be interpreted as resistance. A more useful formulation would be: **"What makes this change difficult to adopt?"** This time, the issue may come from the pace, the lack of explanation, the tools, management, field constraints, or the deployment method. Same topic. Different question. Different answers. ## Leading questions often produce clean answers The most dangerous thing about a leading question is that it does not necessarily produce absurd answers. Quite the opposite. It often produces clean, coherent, easy-to-classify answers. That is exactly the trap. If you ask: **"What are the benefits of this new organization?"** respondents will look for benefits. Even if they mostly experienced confusion. Even if they have concerns. Even if they would have spontaneously talked about overload, unsuitable tools, or loss of reference points. You will get a neat list of benefits. Then you may conclude: "Employees clearly identified several benefits." But that is not necessarily true. You only know that they answered a question that pushed them to look for benefits. That is not the same thing. ## A questionnaire should not be there to reassure you A good questionnaire is not there to confirm what you hope is true. It is there to help you learn what you do not know yet. That difference matters. A useful questionnaire must allow an answer that disturbs you. Otherwise, it is too controlled. Instead of asking: **"How was this training useful to you?"** ask: **"What did this training change, or not change, in the way you work?"** Instead of: **"Which parts of the event did you enjoy?"** ask: **"What do you remember from this event?"** Instead of: **"How did our solution help you save time?"** ask: **"At what points did the solution make you save or lose time?"** The best question leaves room for a positive, negative, neutral, or unexpected answer. It does not decide in advance what the respondent should see. ## How to recognize a leading question A leading question often contains a judgment, even a subtle one. For example: **"Did our intuitive tool help you?"**. The word "intuitive" evaluates the tool before the respondent does. A better formulation would be: **"How did you experience using the tool?"** It can also assume an outcome: **"How did this training help you progress?"**. Here, progress is already treated as a fact. A better question would be: **"What changed, or did not change, after this training?"** It can lock the answer into one direction: **"What did you like in this session?"**. The expected response is positive. A more useful formulation would be: **"What helped you, what was missing, or what slowed you down during this session?"** Finally, it can turn a hypothesis into a fact: **"Why are teams resisting this new process?"**. Resistance is already assumed to be real. It is better to ask: **"How do teams perceive this new process?"** ## Examples of biased questions and better formulations **Biased question:** "What were the strengths of this training?" **More useful question:** "What helped you, what was missing, or what slowed you down during this training?" **Biased question:** "Is our new tool easy to use?" **More useful question:** "At what point did the tool feel clear, difficult, or unnecessary?" **Biased question:** "What did you appreciate in your onboarding?" **More useful question:** "What helped you or was missing during your first weeks?" **Biased question:** "Why is this new organization more efficient?" **More useful question:** "What did this new organization make simpler, harder, or unchanged?" **Biased question:** "Are you satisfied with this experience?" **More useful question:** "Which moment of this experience most influenced your opinion?" ## How to write a less biased question There is no such thing as a perfectly neutral question. But there are questions that are much less leading. **Rule 1: remove unnecessary adjectives.** A phrase like **"our simple and fast new tool"** already contains two judgments. Prefer **"How would you describe your experience with this new tool?"** **Rule 2: open several possible directions.** **"What did you appreciate in this training?"** pushes toward the positive. **"What helped, what was missing, or what slowed you down?"** allows several realities to emerge. **Rule 3: ask for facts before opinions.** **"Did you find the process efficient?"** calls for a general impression. **"At what point did this process make you save or lose time?"** produces more usable information. **Rule 4: test the predictability of the answer.** Before sending a question, ask yourself: **"What answer does this formulation make more likely?"** If the answer is obvious, the question is probably too leading. ## A precise question is not necessarily biased Avoiding framing bias does not mean asking vague questions. A question like **"What do you think?"** is open, but often too broad to produce a useful answer. A good question can be precise without being leading. For example: **"At what exact moment did you need help during the exercise?"** This question is targeted, but it does not force a positive or negative answer. Another example: **"Which part of the process required more effort than expected?"** It is precise, but it lets the respondent identify the friction point. The quality of a question does not come from maximum openness. It comes from its ability to produce useful information without having written the answer in advance. ## The blue house is everywhere The blue house is not just an image. It is what you do every time you write **"What did you like?"** instead of **"What stood out to you?"**. It is what you do when you write **"Why is this change useful?"** instead of **"What does this change concretely modify for you?"**. It is what you do when you write **"Did our solution help you save time?"** instead of **"At what points did you save or lose time?"**. You think you are asking a question. But you have already chosen a color. ## Conclusion: do not paint the answer before asking for it Imagine a blue house. Now remove the color. That is the work of a good questionnaire: do not paint the answer before asking for it. A poorly worded question turns a survey into a mirror. You do not see other people's opinions. You see your own expectations. A good question does the opposite. It removes your hypothesis, reduces your ego, lets the respondent formulate their experience in their own words, and accepts answers that may disturb you. It is less comfortable. But that is when your questionnaires start producing real answers.