Print Header Logo

Institutional Research

Survey Tips

Sample size

  • Keep sample small
  • Can tell respondents that they are part of a select group (appeals to scarcity)
  • If using a sample of the population, you can administer multiple surveys, using different people, so that folks don’t get tired of being surveyed.

How many times do you contact folks?

  • Send at least 3 reminders
  • Change the day of the week and time of day that you send reminders, because students and professors tend to have times that they are more likely to respond.
  • Be sure to remove those who have already responded from your reminder list.

Use a mixed-method of contact

  • Use of both paper and email contacts has been found to be effective. For example, send a letter from the provost before the survey.
  • Follow-up with a postcard
  • A study has shown that sending a post-card before the survey has increased response rate by 9% (Public Opinion Quarterly)

Structure of email

  • Subject line
    • Don’t use the word “survey”
    • Mention EMU to make your email stand out from SPAM
    • On the final reminder, one study found that having the subject line blank got the most responses.
  • Body of email
    • Keep it short and use text only (not html because some SPAM filters look for html)
    • Put the hyperlink near the top of the email.
    • Include contact information out of courtesy.
  • Techniques
    • Tell respondents that they are part of a select group.
    • Give a deadline; limited amount of time to complete.
    • Ask for their help (i.e., appeal to their helping tendencies).
    • Tell them to let you know if they don’t want to be contacted again.
    • Give a rough estimate of completion time.
    • Promise confidentiality, if you can. But, don’t overemphasize it. Overstating confidentiality can decrease response rates and leads to suspiciousness.
    • Make sure that you can defend the content of your email to the Provost and President.
  • Things that typically cause emails to go into ‘Junk Mail’ as SPAM
    • Capital letters
    • “from” name contains underlines or numbers (e.g., sally_52)
    • A large number of blank lines or spaces in the body of the text
    • “Click here” or “Click below”; it is better to use the word “submit”

Response Rate

  • Rule of thumb – at least 50%
  • If your response rate is low, check to see how the demographics of your sample compare to the population.
  • Send follow-up emails not more than 1 week apart.
  • After the first 2-3 days, the response rate typically tends to drop off.
  • Once the response rate drops off, send another follow-up email.

Web Design

  • Appearance
    • Have a short introduction at the top
    • Avoid drop-down boxes (they are easy to miss); use radio buttons or check boxes instead.
    • Make sure that the layout is vertical and that respondents don’t have to scroll horizontally.
    • Use plenty of white space and headings.
    • Number your questions so that people don’t get lost.
    • Use yes/no instead of having them check all that apply (will be easier to use the data and forces respondents to read all of the options).
        Yes   No  
    Library?   Library   Library  
    Athletic Facilities?   Athletic Facilities   Athletic Facilities  
    Dining Hall?   Dining Hall   Dining Hall  
  • Length
    • Don’t ask demographic information if you already have it.
    • If it is long use multiple pages (downloads quicker).
    • Don’t force responses.
    • Keep the survey under 20-30 minutes to complete.
  • Salience of the survey topic
    • This is one of the best predictors of survey response.
    • Put interesting questions near the top and boring questions near the end.
  • Open-ended questions
    • Use of “other” often indicates that you have not formulated your responses well.
    • Keep short, for your own sake.
  • End of survey
    • Make sure the “submit” button is visible.
    • Have a “thank you” page after the submit button.
      • Lets them know the survey is over.
      • Can also use it to direct them to another website, or collect another survey.
  • Timeline
    • Start early, at least a month before you want to the survey to go out.
    • Developing a survey is a long process and there are often many refinements along the way.
    • Allow time for pilot testing.
    • Take feedback seriously.


Modified from Steven Porter’s presentation for an Academic Impressions webinar on February 15, 2005.