Mistakes in survey questions distort results in ways that cannot be fixed in analysis. The most common are asking about two things in one question, leading the answer, using negative wording and asking for opinions on something the employee cannot know. In this article, we cover the 10 most common mistakes in employee surveys, each with a wrong and a right example.
- The most common survey question mistakes are asking about two things in one question, leading the answer and asking for opinions on something the employee cannot know.
- A badly written question produces a wrong result, and that error cannot be fixed in analysis. The quality of the questions determines the quality of the survey.
- Testing the survey with 8–10 people from different units and levels before launch is the most effective way to catch faulty questions.
1. Asking about two things in one question
This is the most common mistake. If an employee agrees with one part and disagrees with the other, they cannot give an accurate answer.
- Wrong: My manager listens to me and supports me.
- Right: My manager listens to my ideas. / My manager supports me when I need it in my work.
If a question contains "and", "with" or "or", check whether it is asking about two separate things.
2. Leading the answer
If the wording pushes employees toward a particular answer, the result reflects how the question was written, not what people really think.
- Wrong: How satisfied are you with the comprehensive training opportunities our company offers?
- Right: I am offered enough training opportunities to develop in my job.
Adjectives such as "comprehensive", "excellent" or "innovative" pull the question in a positive direction.
3. Using negative wording
With negative statements, answering "disagree" creates a double negative, and employees easily misread the question.
- Wrong: My manager does not give me enough feedback.
- Right: My manager gives me feedback regularly.
4. Using vague wording
Words such as "often", "enough" or "usually" mean different things to different people.
- Wrong: I am often recognized.
- Right: In the last month, I was recognized for something I did well.
Specifying a time frame or a concrete situation makes answers comparable.
5. Asking about something the employee cannot know
Employees can only assess their own experience. Questions about the whole organization or about what others think produce answers based on guesses.
- Wrong: Employees at our company are satisfied with their pay.
- Right: I feel I am paid fairly for the work I do.
6. Corporate jargon and complex language
For someone working on the frontline, "stakeholder value", "strategic alignment" or "agile transformation" may mean nothing. Faced with a question they do not understand, employees either pick the middle option or answer at random.
- Wrong: I feel aligned with the organization's strategic objectives.
- Right: I know how my work contributes to the company's goals.
Plain language matters even more in surveys that include blue-collar employees. For details, see our article on how to survey blue-collar employees.
7. Unbalanced answer scales
Scales with more positive than negative options inflate the results artificially.
- Wrong: Poor / Fair / Good / Very good / Excellent
- Right: Strongly disagree / Disagree / Neutral / Agree / Strongly agree
Both ends of the scale should carry equal weight, and the same scale should be used throughout the survey. We cover scale choice in our article on what a Likert scale is.
8. Demographic questions that undermine anonymity
When unit, job title, age, gender and tenure are all asked together, employees in small groups become easy to identify. That lowers both participation and the honesty of answers.
The right approach: Ask only for the breakdowns you will actually use in reporting, use broad ranges and do not report results for groups of fewer than 10 people. For details, see our article on anonymity in employee surveys.
9. Asking questions that cannot lead to action
Asking for opinions on something the organization will not change under any circumstances creates expectations and damages trust when nothing happens. Every question should have an answer to "what could we do based on this result?" We explain how to link questions to goals in our article on setting goals before an employee survey.
10. Changing the questions every year
Rewording statements every year makes it impossible to compare results with the previous period. Even a small change in wording can affect the score.
The right approach: Keep the statements in the engagement index and core categories fixed, and update only the questions for focus topics.
Catch faulty questions before launch
- Read each question on its own and ask whether it could be understood in two different ways.
- Test the survey with 8–10 people from different units and levels.
- Ask them which questions they found hard to understand and which they could not answer.
- After the first run, review questions with many "neutral" or blank answers.
For ready-made sample questions, see our article on employee engagement survey questions, and for writing open-ended questions well, our article on open-ended employee survey questions.
Frequently asked questions
What is a double-barreled question?
A question that asks about two different things at once. If an employee disagrees with "My manager listens to me and supports me", you cannot tell whether they feel unheard, unsupported or both.
Should negatively worded questions be used?
Generally not. With statements such as "My manager does not give me feedback", answering "disagree" creates a double negative and employees may misread the question. Positive, clear statements give more reliable results.
Should the scale have a middle option?
In employee surveys, usually yes. Forcing a genuinely undecided employee toward a positive or negative side reduces the accuracy of the data. That is why the 5-point Likert scale is the most common choice.
How can we spot faulty questions?
Test the survey with a small group before launch and ask which questions they found hard to understand and which they could not answer. After the first run, review questions that received many "neutral" or blank answers.
