Some considerations for creating a research questionnaire

Incorrectly designed questionnaires further impact data extraction when the definition of the objectives of a study are not clear. Here, I will only talk about the questionnaire itself without going into details about the study design.

Some errors can lead to unwanted results at the end of the process. These errors can be derived from problems with data capture or improper processing. The following are five important issues to be aware of when developing a questionnaire:

1. Do not ask the study participant to identify himself/herself: the participant should be identified under very special situations such as in longitudinal surveys for which individuals will be followed over time;

2. Do not directly ask the participant about income: in general, it is common for people not to tell the truth about income, in particular, when it is very low or high. Increasing the value to avoid constraining for those with very low financial standards or decreasing it to avoid curiosity about a high financial standard of life are not rare. Rather than directly asking about this variable, one can use either broader income ranges or identify the socioeconomic strata by using indirect measures. Anyhow, it is necessary to think about whether the income is an important variable for the study or it is there to just comply with a standard question protocol;

3. Avoid open-ended questions: open-ended questions are those in which the respondent freely writes his answer, without having to meet a standard of choice between options. These data are unlikely to be organized and used in a quantitative statistical analysis. As a descriptive form it works but if it is an important variable for the key objectives of the study, use multiple choice questions;

4. Always make the option “I don’t want to answer” available for any question: The participant has the right not to answer questions that he/she does not feel comfortable with. In many online questionnaires, the survey organizer places the instruction that all responses are mandatory but does not leave the person’s option not to respond. This can generate two problems: a) the interviewee gives up answering and b) the interviewee lies to move on;

5. The Likert scale. The Likert scale (^{Likert, 1932}) is an attitude measurement scale that measures the meaning and intensity of agreement or disagreement about an affirmative. Usually it is a scale of 5, 7 or even 10 points. This scale must also allow the participant not to respond, at the risk of the interviewee assigning a low value due to his discontentment in being forced to answer the question. Again, a bias that will be included in the results and may seriously and negatively impact the conclusions.

It is worth mentioning that there is no perfect survey and research, but it is always possible to track down some common problems, and avoid them

REFERENCE

• ¹
  Likert R. A technique for the measurement of attitudes. New York : The Science Press, 1932. Series: Archives of psychology, no. 140.

Publication Dates

• Publication in this collection
  18 Dec 2020
• Date of issue
  2020