The structure of a closed-ended question

In order to structure a scaled question, you will need to look at these 3 elements: the wording of the question, the scaling, and the labels of the answers. A closed-ended question is used to validate and frame a need.

 

1- The formulation of the question

The question may be expressed in either affirmative or interrogative form. It is recommended to use simple, direct, and precise questions. 

Let's take a closer look at the different elements to consider in order to get good results for a better statistical analysis!

 

Assess one topic at a time

For example, 

❌ instead of having a single question as follows:

"I have a good relationship with my team and my manager."

✅ Opt for two separate questions: 

"I have a good relationship with my team"

"I have a good relationship with my manager."

 

Formulate the questions in such a way as not to bias the answers.

Identify any wording that is not neutral, and try not to bias the answers.

For example, 

❌ in the following questions you make an assumption about the participants' tastes:

"What type of beer do you prefer?"

"Do you prefer beer or a non-alcoholic drink?"

✅ Instead, opt for a broader question: 

"What is your favorite drink?

 

Avoid absolutes

Participants would find themselves in the obligation of "all or nothing", and cannot qualify their answer and enrich it.

❌"Would you like to generalise telework? Yes / No"

✅"How many days per week would you like to telework?"

How to find these questions? They often call for Yes/No answers and contain words like "always", "all" and "every".

 

Turning the question in a positive direction

 

At Supermood we use affirmative statements, which then allow us to compare scores between questions.

For example,

❌ if you formulate your question in the negative as follows: 

"I have to make an effort to hide my emotions at work."

A score of 1 is a good score and 5 is a very bad score. Your reading will therefore be reversed and you will not be able to compare the question with those proposed by Supermood.

 

✅ Whereas if you formulate your question in a positive way : 

"I am free to express my emotions at work".

A score of 1 will correspond to a very bad score and 5 to a good score. 

 

Similarly, it is recommended to avoid marking a positive evolution; this would also risk reversing the reading of the scale. 

 

For example,

A high score on the question below implies that the sense of freedom was previously low:

❌"I feel freer to express my emotions at work"

 

Avoid notions of time

It is easier for the respondents to answer about their feelings presently rather than since a certain date.

❌ "Since our last meeting, the relationship in my team has improved."

✅ "I have a good relationship with my team"

If you ask a question on a regular basis, you will be able to observe the evolution of the employees' feelings on the same question. 

 

For more information, see our article on Managing survey frequency.

 

Keep it simple and straight to the point

It is always best to use simple, clear, and concise language in your surveys. In general, avoid acronyms, jargon, and technical terms. Also try to avoid subjective notions such as 'I think', 'I believe', ...'.

In addition, the Supermood questions have been formulated in such a way that a score of 5 (on a scale of 1 to 5) is rated as excellent. If your question is worded with too high a standard, you risk destabilizing respondents and getting low scores.

 

For example: 

❌ "I have the best relationship with my manager"

✅"I have a good relationship with my manager"

 

❌ "I think my company is making the changes necessary to be competitive in the future"

✅ "My company is making the necessary changes to be competitive in the future"

 

2- Graduation

This is the number of possible answers given to the respondent (from 1 to 5 or from 0 to 10, for example).

 

It is best to use a question in the form of a statement and remember to provide context; this will allow the respondent to clearly identify what each scoring point corresponds to. If you do not explain how to interpret the numbers on your scale, you risk distorting the results. 

 

Having a sufficient number of response points allows the respondent to express his or her opinion more accurately. However, it is advisable to keep the number of points moderate; too many points can make the exercise difficult. We recommend using a five-point scale from 1 to 5.

 

To make your analysis easier, make sure to use only one scale for your entire survey! Results are only comparable across multiple questions if the scale is the same. Also, once a respondent has adjusted to one scale, it will be difficult for them to adjust to another scale. 

 

For more information, see our article: How to choose a rating scale?

 

3- Labels 

Labels are texts that clearly explain to respondents what each answer point means.

Supermood offers 2 models of labels : 

• For questions from 0 to 10 (or 1 to 10) : 

The labels "Not at all likely" and "Very likely" at both ends of your scale (to avoid a profusion of labels).

• For the other questions : 

The degree of agreement or disagreement with the question asked; ranging from Strongly Disagree to Strongly Agree. 

 

Do not be intimidated by the design phase. You can launch a survey in a few minutes using one of the Supermood models designed by specialists. And if you would like advice on how to create your survey, don't hesitate to contact your account manager!