We frequently receive requests on our smartphones to respond to surveys on opinions,
attitudes, and practices related to our profession. Survey studies account for an
increasing proportion of published research.[1] With the widespread use of the Internet and smartphones, the past two decades have
witnessed a massive rise in these studies, particularly since the COVID-19 pandemic.
It is relatively easy to get quick data, cost-effective, and much easier to conduct
and analyze with user-friendly web tools. They are ideal for measuring elements that
are not directly observable, such as attitudes, opinions, and beliefs, including satisfaction
and feedback. Well-constructed and validated survey studies can provide vital and
reliable insights, sometimes the only tool for some research questions.
Online surveys are the simplest form of research, gaining popularity among trainees
and novice researchers. Unfortunately, many such surveys are poorly designed, resulting
in incorrect inferences and biased reporting, casting questionable reputation on survey
studies. Indeed, the enthusiasm to quickly conduct a survey and get published in their
names is appreciated and understandable. It is, however, imperative for authors to
dive deep into the principles of these studies and understand and apply them for good
outcomes. It is beyond the scope of this editorial to provide detailed steps to carry
out such studies but encourage and assist authors in following all essential components
of the research and include them in the manuscript preparation. By adhering to fundamental
principles, the critical questions raised by the authors and the responses should
accurately reflect reality. Studies with ineffective research questions and incomplete
or flawed designs are considered inadequate reporting and most likely to be rejected.
Essential Steps
There are a few key principles in survey research that should be followed.[2] They include the following:
-
Clear research question: Like in any other research, the foremost step in designing the survey is to define
what you are looking for, the primary research question. The survey should focus on
eliciting specific essential information than a blanket approach to get as much information
as possible. Keep the survey as small as possible, focused on the intended information,
without redundant questions.
-
Designing the study and questionnaire: Decisions regarding the format, structure number, and order of survey questions
can be quite complex, and developing them from scratch is a huge task. Most novice
researchers may not understand the principles of valid questionnaire design. A good
literature review, referring to previous similar study formats and adopting them with
suitable modification, without copyright infringement, is quite useful.
-
Open versus closed question formats: Closed-ended questions provide a set of options to choose from, while open-ended
questions allow responders to come up with their own answers. The former is easier
to evaluate and the data are more valid statistically, while the latter is more useful
when seeking expert opinions or feedback from a small group. However, the closed-ended
questions need initial efforts to frame appropriate responses, while in the open-ended
format, although it easier to set up initially, it is much more difficult to analyze
the data. The questions should be clear, concise, and without bias, intendant to address
the research question.
-
Pretesting the survey instrument with pilot test: Although not mandatory, it is useful to test the completed questionnaire among a
small group of target subjects so that few unconsidered problems, such as ambiguity,
language errors, etc., are resolved.
-
Validate the survey instrument: Establishing the validity of the survey instrument is the responsibility of the
investigator. It simply means how accurate and reliable the data are to measure what
is intended. Adopting already validated and published tools from peer-reviewed studies
is rather easier. If the survey is new, establishing validity is essential based on
different types. These include the following: face validity (does it make sense simply by looking at it?); content validity (a panel of experts reviews and revises it for relevant content); criterion validity (how does it compare with an established standard in the field?); predictive validity (how does it compare with the actual outcomes of interest?); and construct validity (does the internal construct/structure conform to scientific survey study principles?).
Limitations and Biases
The major limitations of survey studies are the participants' willingness to respond,
and the response rate is crucial to evaluate the validity of the study.[3] There are no defined criteria for an adequate response, but a response rate of anything
between 50 and 70% is considered adequate. The method of survey too plays a crucial
role, with the highest response to telephonic interviews (>80%) against e-mail surveys
(24%).[3] Poor response rate leads to bias in the study, such as “nonresponders bias,” calculated
from the variance between responders and nonresponders. Studies with less than 20%
response rate are not considered for publication.
The Manuscript Preparation
Based on the intent and construct of the study, the manuscript could be submitted
as an original research paper, brief communication, or a letter with limited information.
Thorough and complete details of essential steps are mandatory for peer reviewing
and considerations for acceptance. These are the following[2]
[4]:
-
The introduction: It should describe the rationale for using the survey, how the research question
will be addressed, and the target population.
-
Methods: The methodology is crucial and describes whether the survey questionnaire was developed
by the authors or adopted from previous studies and the details about the framing
of the questions, the experts involved, and the relevance and rationale behind it.
The description of whether the survey was pretested or a pilot study was undertaken
before conducting the survey. The execution of the survey, including mode of contact
(web-based, interview, telephone, or postal), the timing of the study, whether anonymity
was maintained, how long was the response time, reminders, incentives, etc., should
be mentioned in detail.
-
The results: Describe who the respondents were, the response rate, and how it was calculated,
and how nonresponse bias was assessed and corrected. Describe how score reliability
and validity were assessed and the statistical methods used.
-
Discussion: Authors construct their arguments on why their study findings should be believed
and their survey is relevant to the current literature. Discuss the new insights and
whether they adequately addressed the primary research question. It is essential to
address all limitations and biases and refrain from overinterpretation of results.
-
Data submission: Besides providing concise tabulated data, it is desirable to submit the entire survey
questionnaire as an appendix or supplementary file for reviewing the manuscript and
future research.
Final Thoughts
Surveys are a critical tool in educating health professionals. Compliance with the
fundamental steps and guidelines of the survey studies is reemphasized here. A detailed
manuscript with transparent data, validated tools, addressing biases will deter advocating
an invalid hypothesis. An in-depth approach is expected from authors than merely treating
surveys as desktop research.
