About the Survey Response Scale
All closed-ended questions in the WES use a 5-point Likert-type bipolar response scale ranging from strongly disagree on one end to strongly agree on the other. This type of scale was originally developed by an educator and organizational psychologist Rensis Likert in 1932 to improve measurement in social research by standardizing response categories in questionnaires. The Likert scale is widely used for surveys of all types, and has become a standard for employee surveying across public sectors in Canada.
For every question measured by this scale, employees can answer the question by selecting one of seven possible responses. Five of the responses measure the intensity of an opinion (level of agreement) using the Likert scale. Two responses are for those who wish to opt out of answering because they either do not have an opinion (‘don’t know’) or believe they can’t answer the question because it does not apply to them (‘not applicable’).
Responses labelled ‘strongly disagree’ are assigned 1-point, while responses labelled ‘strongly agree’ are assigned 5-points. The middle points ‘2’, ‘3’, and ‘4’ are not labelled, but permit employees to select their level of agreement as somewhere in the middle in relation to strongly disagree and strongly agree.
The Three Categories
When reporting percentages, we combine response options ‘1’ and ‘2’ on the survey into the ‘Disagree’ category in the reports and combine ‘4’ and ‘5’ into the ‘Agree’ category.
We use the three reporting categories for two reasons:
Reason 1: To reduce the volume of numbers when reporting percentages.
Reporting the results using the full 5-point scale makes for overwhelmingly large tables. Often, the extra detail is not helpful during interpretation and we find that most people simply want to know how many people disagreed and how many agreed.
Reason 2: To protect confidentiality.
In small work units of 20 to 30 employees, there may be only a few people who chose ‘1’ (strongly disagree) or ‘5’ (strongly agree). While there is no way to identify a specific person if the percentages are reported using the full 5-point scale, some people may feel uncomfortable with that level of detail, particularly for the more sensitive questions.
Then why don't we use a 3-point survey response scale? The extra detail available with the 5-point response scale is very important when developing the engagement model. There is a considerable difference between simply agreeing with a question and strongly agreeing with a question. Similarly, strongly disagreeing with a question is more emphatic than simply disagreeing.
The model building process is sensitive to the intensity of these responses and requires the full range of responses to properly identify drivers and identify connections.
The Neutral Response Option
The middle category can be selected by people who have a neutral opinion or are undecided about the topic/issue (they neither agree nor disagree). A neutral or undecided opinion is valid and forcing people to pick a side (either agree or disagree) using a 4-point scale may not accurately reflect how they feel about the topic. Also, forcing a choice may cause people to bias their answer towards agree; this behaviour is called ‘acquiescence bias’.
Having a neutral opinion in the middle of the scale reduces the likelihood of people agreeing for the sake of agreeing, or disagreeing for the sake of disagreeing. People also have the option of choosing the ‘don't know’ or ‘not applicable’ response options, listed at the end of the survey scale. These two response categories separate people who truly have no opinion on a topic from those who have a neutral or undecided opinion.
Dr. Don Dillman, a sociologist who has made major contributions to survey methods, recommends having a middle category within the scale and a "no opinion" (for example, ‘don't know’ and ‘not applicable’) category at the end to increase the accuracy of responses collected during a survey.