The classic Likert item uses a 5-point response scale, but you can use 7, 9, or other points, too.
When numbers are associated with each response option, the Likert item can be seen as a special case of the linear numeric scale. The classic Likert scale has participants agree or disagree (or approve/disapprove) to multiple statements. (Labeling, neutral points, and number of response options are the topics of other articles.) Linear numeric scales usually have at least the endpoints labeled. The Single Ease Question (SEQ) and likelihood to recommend item used in the NPS are examples of linear numeric scales. This can include things like satisfaction, ease, brand favorability, feature importance, or likelihood to recommend. In a linear numeric scale, participants provide some numeric response to a question or statement. Here are 15 scales, in roughly the order of most to least commonly used. I’ve adapted a classification scheme based on our experience at MeasuringU and from the classic text on survey research by Alreck and Settle. For example, our MUIQ platform offers over 30 question types but I’ve identified 15 distinct ones. There are different ways of classifying rating scales and slight variations can result in different looking rating scales, even though they’re variations on the same scale. The item selection can have a big impact on both the responses and interpretation. Rating scales however usually involve asking participants to rate abstract concepts, such as satisfaction, ease, or likelihood to recommend. When properly written, they’re also straightforward to the respondents as they usually involve concrete selections.
age, education level, or electronic devices owned) are usually more straightforward for researchers with little survey experience. Closed-ended questions themselves can be classified into multiple choice questions or rating scales. In surveys, questions can be broadly classified as open-ended (free responses) or closed-ended. Having many ways to ask questions provides better options to the researcher to assess the opinion of the respondent.īut the wrong type of question can fail to capture what’s intended, confuse respondents, or even lead to incorrect decisions. Variety in question types can be both a blessing and a curse. Use question logic so that only those who are unhappy skip to a question asking for improvement suggestions.It seems like there are endless ways to ask questions of participants in surveys. For example, let’s say you want to ask how much your patron enjoyed your restaurant, but you only want more details if they were unhappy with something.
Keep it inclusive. Scales should span the entire range of responses.This can be tricky when using word labels instead of numbers, so make sure you know what your words mean. Keep it continuous. Response options in a scale should be equally spaced from each other.Our methodologists recommend five scale points for a unipolar scale, and seven scale points if you need to use a bipolar scale. If you provide more than seven response choices, people are likely to start picking an answer randomly, which can make your data meaningless. How many options should you give people? Respondents have difficulty defining their point of view on a scale greater than seven. Keep it odd. Scales with an odd number of values will have a midpoint.Keep it labeled. Numbered scales that only use numbers instead of words as response options may give survey respondents trouble, since they might not know which end of the range is positive or negative.