Research Methods Guide: Survey Research

"There's never an option that reflects what I exactly want to say."

"Sorry, sir, but we don't have a category for that."

"I'm an honest person but when I take an online survey, I'm a big liar."

"93% of the people who took the survey thought it was 100% waste of time."

"Contrary to popular opinion, most people's opinions aren't so popular."

"According to the latest poll, 64% of the public don't pay any attention to polls..."

"Do you really care about my opinion or am I just a random sample?"

"Having conducted a survey of 32.4% of the 56.6% of tax experts about 43% of the budget we found an 87.6% probability that we haven't got the foggiest ideas of what it means."

You can find more survey jokes and cartoons here.

Surveys are designed to:

• Collect information from a small number of people to be representative of a larger number of people to be studied, for instance, the information about their...
  • Attributes
  • Behavior
  • Preferences
  • Attitudes
  • Opinions
• Systematically draw information from a certain population in order to...
  • Describe demographic information (e.g., age, gender, school year, affiliation)
  • Draw patterns from the population studied
  • Explain trends out of phenomenon

There are several types of surveys, including:

• Phone
• Face-to-Face
• Paper and Pencil
• Online (Web-based): e.g., SurveyMonkey, Qualtrics, Question Pro

How can I develop effective survey questions?

Clarity, simplicity, length, and acceptability are keys to create effective survey questions.  In crafting your survey, you should avoid complicated, long, and ambiguous questions.  Try not to address too many issues in one question - thus avoid double-barreled questions.  Try not to ask too many questions in one survey.  Try not to ask too difficult questions.

What question formats are there?

There are two basic formats: 1) Forced (closed-ended), and 2) Open-ended.  You can use a combination of both in a survey depending on your goal.

Can you give me some examples of closed- and open-ended questions?

Let's say you are interested in studying people's environment concerns through shopping habits:

• Forced question examples:
  • "Do you use a reusable shopping bag when you go grocery shopping?"
    • Yes"  or "No" (Note: Survey respondents are forced to choose one of the two.)
  • "How frequently do you use a reusable shopping bag when you do grocery shopping?"
    • "Very infrequently", "Somewhat infrequently", "Occasionally", "Somewhat frequently", "Very frequently" (Note: Survey respondents are forced to choose one out of multiple choices.)
• Open-ended question examples:
  • "In the last 30 days, how many times have you used a reusable shopping bag when doing grocery shopping?" (Note: Survey respondents are allowed to use either numeric values or text entries.) 
  • "Please describe why (or why not) you use a reusable shopping bag for grocery shopping." (Note: Survey respondents can freely write their answers if they want to).

Does ordering of questions influence survey results?

Yes, it does. The natural and logical flow of survey is important to collect good survey results.  Start and end the survey with easy questions.  Start the survey with most familiar questions.  Keep in mind that a high response rate does not guarantee a high survey completion rate - in many online surveys, people do not always complete a survey. 

Do you have any other suggestions for conducting a good survey research?

• Invitation: Creating a good invitation to participate your survey is important.  In your survey invitation (email) letter, try to include 1-2 sentences describing a purpose or goal of your survey.
• Disclosure:
  • Mention length of survey
  • Ensure that responses are confidential
  • Ensure that participation is voluntary
  • Provide contact information in the cases where participants have questions about survey
• Pretesting: Before administering a survey, make sure you test your survey in advance.  Survey pretesting will help you determine the effectiveness of your survey.