Attitudinal Questions and Likert Scales

An attitudinal question is a special type of closed-ended question that specifically assesses a respondent’s disposition or emotional response to a topic. Like open-ended questions, these are commonly used in qualitative surveys and give the researcher insight that a standard multiple-choice survey may not. The main benefit of the attitudinal question is that it provides the researcher with a way to measure the respondents’ emotional responses to a construct. A drawback of the attitudinal question, however, is that it requires extensive inquiries to obtain a complete picture with quality data. For example, Siegle (2015) recommends writing attitudinal questions in batches of 8-10 that each measure a different element in the construct of interest to capture a complete picture of the respondents’ attitudes.

Likert Scales

Attitudinal questions are generally provided in a format of a scale ranging from one extreme to the other. This linear range of options is called a Likert Scale and can be used to gauge qualitative responses, such as general frequency as well as attitude towards a topic, and convert them into quantitative data for analysis using methods that are scientifically accepted and tested. Historically, Likert scales have been recommended to contain between 5 and 7 response options (Lietz, 2010). Recent studies show, however, that the reliability and validity of the data collected from these question types increases when using specifically a 7-category scale due to the increased spread of options to best represent respondents’ viewpoints (Taherdoost 2022 and Joshi 2015).

The examples in this guide include numbers to illustrate the connection between response options and data analysis. However, it is recommended that the number values associated with each option be hidden from the respondents to include periodically flipped series questions as a test for internal reliability and validity (e.g., if students check the leftmost option the whole way down, the flipped question will stand out).

Table 1. An example Likert Scale Showing a Range of Attitudinal Response Options

1	2	3	4	5	6	7
Strongly Disagree	Disagree	Slightly Disagree	Neither Agree Nor Disagree	Slightly Agree	Agree	Strongly Agree

Pragmatics of Likert Questions

Though traditional Likert scales frame the prompts as statements: e.g. ‘I am happy with my level of understanding,’ Likert-type questions can also be posed to the respondent in the form of a question by varying the wording of the scale options (Tables 2 and 3). Whether you use statements or questions, the options for the respondent should align as a natural language type response. Consider the examples below; while both address the happiness of understanding, the response options vary due to the structure of the prompt.

Table 2. A statement-based Likert-type scale question with modified responses to accommodate a question format.

***"I am happy with my level of understanding."***
1	2	3	4	5	6	7
Strongly Disagree	Disagree	Slightly Disagree	Neither Agree Nor Disagree	Slightly Agree	Agree	Strongly Agree

As long as the options are still a linearly increasing scale, the validity of the Likert instrument remains intact. Both formatting options are valid and commonly used. Bunce and Cole recommend that all survey prompts should be uniform throughout a survey, which would require the statement format. Conversely, Qualtrics recommends asking questions over making statements to avoid acquiescence bias (tendency of respondents to agree with established statements). The frequent re-wording of the scale options will, however, increase the mental load associated with your survey, contributing to survey fatigue and making your respondents less likely to complete the survey (more about this in Survey Delivery and Participation). It is once again up to the researcher to consider these factors and decide which format better suits the research goal.

Ordering the Options

Table 3. An example of a Likert scale that does NOT conform to the recommended numbering format.

7	6	5	4	3	2	1
Strongly Disagree	Disagree	Slightly Disagree	Neither Agree Nor Disagree	Slightly Agree	Agree	Strongly Agree

Studies have shown that when the numbers are shown to respondents, the directionality of the scale does not influence the responses received (Rammstedt and Krebs, 2007); however, the number values should increase with increasing positivity so that the most desirable answer is associated with the highest number on the scale. Assigning higher numbers to the negative/disagree choices discourages respondents from answering in the negative which are already more difficult to select. Therefore, to obtain the most accurate data and not bias your survey results toward a more favorable outlook, the lowest number should correspond to the negative/disagree statements (Bunce and Cole, 2008).

If you are choosing to use the format that does not correlate number values to options, place the options students will be less eager to choose (the negative connotation statements) on the left (which will be seen first as respondents read left-to-right). Respondents have a limited attention span, so placing the easier to select options later (to the right) helps to avoid your participants selecting the first choice they see.

Non-Opinion Options

Another element to consider when generating your response options is whether or not you include an odd or even number of options. So far in this guide, we have used the seven-option example format. The odd number of answers serves to leave a truly central option that shows no preference to either the positive or the negative response (Table 5). If this option is removed, respondents are forced to give at least a marginal opinion on each of the options (Table 6). By including a non-opinion option, respondents can avoid making a decision. If your survey is trying to gauge attitudes, this may prove counterproductive. It is important to consider two notable types of respondents when choosing which scale to employ: Floaters and Fence-Sitters.

A floater is a respondent who chooses to respond with an opinion even if they don’t necessarily have one. Fence-sitters, on the other hand, are the respondents who choose the neutral non-response option even when they have an opinion. This is particularly common in respondents whose opinions are not socially desirable (Mauldin 2020). (See Guide on Survey Delivery and Participation for more on Social Desirability Bias).

By using an even-number choice scale, you eliminate the ability for the fence sitters to avoid giving an opinion, however, you are in turn eliminating the ability for those who truly do not have an opinion to answer accurately, essentially forcing them to become floaters.
By using an even-number choice scale, you eliminate the ability for the fence sitters to avoid giving an opinion, however, you are in turn eliminating the ability for those who truly do not have an opinion to answer accurately, essentially forcing them to become floaters.

You may also consider including an N/A or other non-engaged option. Each of the options has its own pros and cons, so consider what best suits the study you are conducting.

Table 4. A Likert scale with an odd number of response choices, giving a non-response option.

1	2	3	4	5	6	7
Strongly Disagree	Disagree	Slightly Disagree	Neither Agree Nor Disagree	Slightly Agree	Agree	Strongly Agree

Table 5. A Likert scale with an even number of response choices, eliminating the non-response option.

1	2	3	4	5	6
Strongly Disagree	Disagree	Slightly Disagree	Slightly Agree	Agree	Strongly Agree

References

Bunce, D. M., & Cole, R. S. (Eds.). (2008). Nuts and Bolts of Chemical Education Research (Vol. 976). American Chemical Society. https://doi.org/10.1021/bk-2008-0976

Joshi, A., Kale, S., Chandel, S., & Pal, D. K. (2015). Likert Scale: Explored and Explained. Current Journal of Applied Science and Technology, 396–403. https://doi.org/10.9734/BJAST/2015/14975

Lietz, P. (2010). Research into Questionnaire Design: A Summary of the Literature. International Journal of Market Research, 52(2), 249–272. https://doi.org/10.2501/S147078530920120X

Mauldin, R. L. (2020). 7.4 Designing effective questions and questionnaires. https://uta.pressbooks.pub/foundationsofsocialworkresearch/chapter/7-4-designing-effective-questions-and-questionnaires/

Rammstedt, B., & Krebs, D. (2007). Does Response Scale Format Affect the Answering of Personality Scales?: Assessing the Big Five Dimensions of Personality with Different Response Scales in a Dependent Sample. European Journal of Psychological Assessment - EUR J PSYCHOL ASSESS, 23, 32–38. https://doi.org/10.1027/1015-5759.23.1.32

Siegle, D. (2015, June 30). Developing an Attitude Instrument | Educational Research Basics by Del Siegle. https://researchbasics.education.uconn.edu/developing_an_attitude_instrument/

Taherdoost, H. (2022). What Is the Best Response Scale for Survey and Questionnaire Design; Review of Different Lengths of Rating Scale / Attitude Scale / Likert Scale. International Journal of Academic Research in Management, 8(1), 1–10.

What is a likert scale? (n.d.). Qualtrics. Retrieved April 29, 2023, from https://www.qualtrics.com/experience-management/research/likert-scale/

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.