04.Data collection04.Email surveys - sporedata/researchdesigneR Wiki
1. Use cases: in which situations should I use this method?
When questions are to be asked to a specific patient or stakeholder population by email, and response rates are expected to be adequate (usually at least 40%). You can ask questions with alternative responses or with open responses.
To investigate practice patterns - see Physician Confidence in Neck Ultrasonography for Surveillance of Differentiated Thyroid Cancer Recurrence
Respondent-driven sampling is a subtype of email surveys used whenever the population of interest is rare. For example, studies involving transgender individuals, because their percentage in the community is relatively low, as well as rare diseases.
To analyze patient's perception - see [Oncology patients’ perceptions of and experiences with COVID-19] (https://pubmed.ncbi.nlm.nih.gov/32809060/)
2. Input: what kind of data does the method require?
- A list of email addresses from individuals in the target population.
3. Algorithm: how does the method work?
One of the key elements in survey design is to create incentive mechanisms, so that response rates are as high as possible (30 to 40% is extremely high these days). These mechanisms can include rewarding respondents with personalized information of interest to them. The information that will be provided back to participants as an incentive for their response should make use of truthful answers to the survey. In that way, participants will connect their responses to the benefit that they will get by participating in the project.
It is an excellent practice to only to write survey questions based on the results of a qualitative study. In that way, the insights coming out of emergent themes resulting from the qualitative studies can be explored quantitatively through the surveys in the end for qualitative study.
Surveys with a prominent sponsor achieve high response rates with a balanced sample composition and high quality of responses, just because participants deem the sponsor as trustworthy. Hence, they consider their participation as useful 1(#1). For general perspectives on incentives in surveys, see [Incentives in surveys] (https://forscenter.ch/fors-guides/fg-2019-00008/).
Email surveys of the general population can be subject to significant biases resulting from under coverage and nonresponse. Not everyone has an email account or access to the internet, and there are significant demographic differences between those who do have access and those who do not. People with lower incomes, less education, living in rural areas, or age 65 and older are underrepresented among internet users and those with high-speed internet access. See [Internet Surveys] (https://www.pewresearch.org/politics/methodology/collecting-survey-data/internet-surveys/).
The Checklist for Reporting Results of Internet E-Surveys (CHERRIES) gives peer reviewers and readers a better understanding of the sample selection and its possible differences from a “representative” sample. Although focusing on Web-based surveys, many CHERRIES items are also valid for surveys administered via e-mail. In total, the checklist has 30 items referring to design, IRB approval, consent process, development, pre-testing, recruitment process, description, administration, response rates, and analysis. An explanation of each item is available at [Improving the Quality of Web Surveys:The Checklist for Reporting Results of Internet E-Surveys (CHERRIES)] (https://pubmed.ncbi.nlm.nih.gov/15471760/).
More Comprehensive Reporting of Methods in Studies Using Respondent Driven Sampling Is Required: A Systematic Review of the Uptake of the STROBE-RDS Guidelines 2(#2).
Improving the quality of Web surveys: the Checklist for Reporting Results of Internet E-Surveys (CHERRIES) 3(#3).
Suggested companion methods
- Use of traditional scales or IRT/CAT methods if the survey intends to collect information on latent variables such as self-report concepts scales of quality of life, mental health, physical activities, among others.
- Use of randomized trials if the goal is to test an intervention, such as in behavioral studies.
- Use of Mechanical Turk as a mechanism for data collection if the condition of population of interest is widely prevalent among the population.
- Use conjoint analysis if the goal is to measure preferences.
- Use sampling mechanisms if the intention is to have the survey representing a specific target population.
- Use of chatbots to collect information and provide respondents with learning material that might be of interest to them.
- Use in mixed methods as a component of qualitative studies.
- Internet, phone, mail, and mixed-mode surveys: the tailored design method 4(#4).
4. Output: how do I interpret this method's results?
[Association methods] (https://github.com/sporedata/researchdesigneR/wiki/%2301.Association)
[Exploratory Analysis] (https://github.com/sporedata/researchdesigneR/wiki/01.Association01.Exploratory-analysis)
 Lipps O, Herzing JME, Pekari N, Ernst Stähli M, Pollien A, Riedo G, Reveilhac M. [Incentives in surveys] (https://forscenter.ch/fors-guides/fg-2019-00008/). FORS Guide No. 08, Version 1.0. Lausanne: Swiss Centre of Expertise in the Social Sciences FORS. 2019. doi:10.24449/FG-2019-00008
 Avery L, Rotondi M. More comprehensive reporting of methods in studies using respondent driven sampling is required: a systematic review of the uptake of the STROBE-RDS guidelines. Journal of clinical epidemiology. 2020 Jan 1;117:68-77.
 Eysenbach G. [Improving the quality of Web surveys: the Checklist for Reporting Results of Internet E-Surveys (CHERRIES)] (https://pubmed.ncbi.nlm.nih.gov/15471760/). Journal of Medical Internet Research. 2004 Sep 29;6(3):e34. doi: 10.2196/jmir.6.3.e34. Erratum in: doi:10.2196/jmir.2042. PMID: 15471760; PMCID: PMC1550605.
 Dillman, Don A., Jolene D. Smyth, and Leah Melani Christian. Internet, phone, mail, and mixed-mode surveys: the tailored design method. 4th ed. John Wiley & Sons, 2014.
 Brasel K, Haider A, Haukoos J. Practical Guide to Survey Research. JAMA surgery. 2020 Apr 1;155(4):351-2.