Conducting Surveys in Dental Education Research: Guidelines and Reminders

Abstract

The use of surveys is popular in dental education research. However, designing and conducting a survey can have many pitfalls. This article aims to prepare a new researcher or one with little experience to undertake survey research. It covers points such as survey design (including question construction), pilot testing for validity and reliability, sampling strategy, methods to increase response rates, logistical considerations, and items to include when writing the manuscript. Careful consideration of a survey from beginning to end can help one design and conduct a successful study that meets its research aims and adds valuable evidence to the literature.

Evaluating a dental sleep apnea mini‐residency program using the Kirkpatrick model

“…Some of the questionnaire items were “double‐barreled,” that is, they addressed more than one aspect of satisfaction within a single question. Double‐barreled questions should be avoided because they require the participant to provide one answer, even though the participant might have different levels of agreement regarding the different aspects of the question 52 . For instance, Q1 of the satisfaction surveys (“Course objectives were defined and met in the presentations”) asked about both the objectives being defined and the objectives being met.…”

“…See Magnuson et al. for guidelines on surveys in dental education research 52 . Additionally, as mentioned previously, the participants were not tracked longitudinally.…”

“…Doublebarreled questions should be avoided because they require the participant to provide one answer, even though the participant might have different levels of agreement regarding the different aspects of the question. 52 For instance, Q1 of the satisfaction surveys ("Course objectives were defined and met in the presentations") asked about both the objectives being defined and the objectives being met. Therefore, if a participant had one level of agreement about the objectives being defined, and a different level of agreement about the objectives being met, a single answer would not capture the nuances of their satisfaction level.…”

“…See Magnuson et al for guidelines on surveys in dental education research. 52 Additionally, as mentioned previously, the participants were not tracked longitudinally. All levels of the Kirkpatrick model, except for Level 4, were evaluated concurrent with the program and might not reflect long-term outcomes.…”

Correlates of burnout in predoctoral dental students in the United States

“…Additionally, although the sample size of 631 was large compared with many previous studies, and the exact response rate could not be calculated due to uncertainty about the total number of predoctoral dental students in the US during the data collection period, the response rate was certainly low when considering that all 66 US dental schools were invited to take the survey and an estimate of 25,000 total students was obtained based on data from ADEA 37 . While these types of limitations frequently occur in survey research within the field of dental education, 42 they can result in selection bias 43 . Another limitation was that burnout was measured via self‐report rather than a clinical interview; however, some researchers have expressed that self‐report measures have greater validity than is commonly believed 44 .…”