For the structure of the Survey:

• Layout:​
  • Create an all-inclusive survey.
  • Use consistent scales for Likert question types.
  • Question order matters:
    • Easy before complex
    • Behavior before attitude
    • General before specific
    • Awareness before usage
    • Sensitive questions should go near the end (e.g., demographics)
  • Make randomized questions within a section if there’s no logical sequence.
• Typefaces and styles:
  • Use text styles (italics, bold, underline, etc.) but not all at ONCE.
  • Highlight important instructions with colors but sparingly and consistently. Think also about perceptions.
  • Choose fonts so that they provide a clear hierarchal order to the survey.
  • When in doubt, use online guides for which fonts to mix.
• Graphics:
  • Include an image of the sponsoring organization’s logo.
  • Use visuals to support or clarify questions.
  • Be wary of the different ways graphics can be used to interpret a question.
  • May impact page loading.
• Skip Logic:
  • Base the question each respondent receives on answers from previous questions.
  • Preview several times and try all possible paths to ensure the logic is correct.
  • Advantages:
    • Personalizes the survey.
    • Typically less confusion (especially in internet-based surveys).
    • Can lead to stronger results.
  • Disadvantages:
    • May confuse respondents in paper-based surveys due to lack of continuity.
    • Increase possibility of error when creating survey.
• Appeal to the respondent:​
  • Create a good first impression: include a clear introduction, contact information and the name of any sponsoring organizations.
  • Produce a sense of altruism.
  • Time the release of the survey with outside events.
  • Consider incentivizing with an appropriate gift or drawing.
    • Should be group-specific.
    • Not so great in value that it might be lead to biased answers.
    • May lead to replicates/multiple submissions by one individuals.
  • Always remember to thank for their time.
• Confidentiality:​
  • Be honest about if/how confidentiality will be maintained.
  • Inform the respondents on how data collected will be used.
    • Should be reasonably accurate but not necessarily too specific.
  • Determine which questions are “required” and which are not.
    • Assure respondents they can skip questions if they are sensitive.