Social desirability bias: A confounding factor to consider in survey by self-administered questionnaire.

Sir,

We read a short communication titled “Perception of pharmacists regarding over-the-counter medication: A survey” by Ravichandran A and Basavareddy A published in your journal.[1] The article evaluated the knowledge, attitude, and practice of the pharmacists toward dispensing the over-the-counter (OTC) drugs. Overall, the article was able to express the study protocol. However, we felt that more clarity about sample population, details of the instruments (i.e., survey questionnaire), and some additional discussions would make the article superior. Hence, we intended to share this correspondence.

The study was aimed at testing the knowledge, attitude, and practice of the “pharmacists” as evident from the Abstract and Introduction of the article. However, the authors had distributed the survey questionnaire to “pharmacists and pharmacy attendants.” Furthermore, according to the result of the study, only the responses of “pharmacists” were taken. The reason for exclusion of responses obtained from pharmacy attendants was not explained. Hence, “pharmacy attendants” might be a typing error!

According to the Materials and Methods section of the abstract, the survey response rate was 80% (80 out of 100). Surprisingly, in the Materials and Methods section in the article body, it was described as 74.6% (112 out of 150). The authors could be more cautious about this discordant sample size. While writing article in IMRaD format,[2] it is better to write the sections of the article as shown in Figure 1.[34] A thorough revision by all authors might help detecting this error.

Figure 1

Flowchart of writing style of sections of an article in IMRaD format

For the study, a pretested questionnaire was used as the survey instrument with 16 questions. More details about the questionnaire, especially the response options (e.g., yes, no, and don't know tick box; Likert-type scale) for the questions, would give us a better idea about the instrument. The authors had taken responses of 10 subjects for calculating the Cronbach's alpha. However, those who responded during pretesting of the scale for internal consistency – “pharmacists only” or “pharmacists and pharmacy attendants” – remain unanswered. The internal consistency (alpha) obtained for a questionnaire from a particular sample may be invalid if administered to a different sample.

The description of the survey questionnaire provides an obscure view about its structure. According to the authors' description, the first part of the questionnaire was used to record demographic details, qualification, specialization, and “knowledge” of OTC medication. Investigators forgot to mention the number of questions used to test the “knowledge.” According to the authors, the second part of the questionnaire was intended to gather information about the “acceptance and popularity” of OTC drugs among the “Kolar population.” In this statement, there were two confusing terms. First one was the “acceptance and popularity.” If this was the second part of the questionnaire, where is the described 16-item questionnaire to test the “attitude and practice?” Another erroneous broad term “Kolar population” was the second one. The study was concerned about pharmacists, not the population.

It is an alarming issue that many of the pharmacy shops run with a rented certificate of a pharmacist, while the certificate holder is working in another setting.[5] Hence, in the sample, some respondents might not have Bachelor of Pharmacy degree or Diploma of Pharmacy. In addition, the worker in the pharmacy might be reluctant to admit that illegal activity. In the study, the researchers used a self-administered questionnaire. They collected demographic data in the questionnaire. This increases the chance of reluctance, even though the aim of the study is described to the participants. Hence, “social desirability bias” may be a confounding factor in the study.[6] An additional discussion about this might enrich the article.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.