One of the most common market research strategies used on surveys is the Likert solution; it is a sort of question that calculates the respondent’s degree of agreement or disagreement with a particular declaration. Every item is comprised of a stem and a scale
For example, look at the following item…
“The federal minimum wage should be higher.”
1. Strongly disagree
2. Disagree
3. Undecided
4. Agree
5. Strongly agree
The stem of a Likert item is the declaration (i.e. “The federal minimum wage ought to be greater.”). The item’s scale is the collection of possible replies. While this is a relatively simple technique for soliciting the ideas of your market, it often leads to skewed data due to poor design.
In this article, we will have a look at some of the issues natural with using Likert items in your online study. The purpose is not to deter you from utilizing them. On the contrary, by revealing the potential traps, you’ll be able to stay away from them when developing your surveys.
Possible Issues With Utilizing Likert Products
A critical piece of creating an efficient survey is knowing the prospective biases of your market and reducing their effect on your information; a particularly typical phenomenon with Likert items is acquiescence bias
This happens when answerers communicate a natural inclination to concur with the item’s stem. From our previous example, participants will be inclined to choose options 4 or 5 (“agree” and “strongly agree,” respectively). In case they disagree with the stem, they might choose selection 3 (“undecided”).
This can happen for a variety of motives. First, the respondent may have an inborn penchant for agreeing with other folks – in this case, the questionnaire’s designer – motivated by a desire to be courteous or amenable.
Second, the prejudice will happen because of an appeal to presumed authority. That is, the participant may think the surveyor is an expert on the topic, and thus very likely to be much better educated than he or she. From this point of view, agreement with the stem would seem correct.
Third, acquiescence tendency may occur due to the respondent’s wish to complete the questionnaire. Disagreement requires more justification than agreement; that requires more effort and time to cautiously consider the possible replies.
An additional potential problem with utilizing Likert items involves a tendency known as social desirability bias; this occurs when the respondent’s selection is influenced by his or her need to be regarded in a favorable light by others
Based on the stem’s topic, this individual might be motivated to agree or disagree based on how they comprehend an additional person’s expectations.
Of the two types of bias, acquiescence bias is simpler to minimize (although hard to eradicate completely); every Likert item’s scale should be dispersed evenly across positive and negative responses to prevent stimulating the participant’s prejudice.
Studying Your Outcomes: Interval Vs. Ordinal Information
Initially, studying the information from a Likert item appears simple; with a 5-level scale – for example the one utilized in our previous example – compiling and reporting the outcomes is easy. Nevertheless, it is worth taking a closer look.
A frequent mistake is to assume the scale replies reflect interval as opposed to ordinal data. Interval information assumes the range between replies is equal. That is, option 1 is equidistant from option 2 as 2 is to option 3
While an argument may be made that this is indeed the case for a distinct Likert scale, most scale responses mirror ordinal data. This is info that does not indicate distance – equal or otherwise – between responses. Any presumption to the contrary is most likely to lend to a misunderstanding of your questionnaire’s outcomes.
Likert items are beneficial and ought to be incorporated within your online surveys. The tip is to generate them in a way that reduces the impact of acquiescence bias, and to avoid presuming the info indicates something it does not.
Leave a Reply