Survey or Poll? What’s the Difference and When to Use Which

Now that QuestionPro has added PollBob to their product family, you might find yourself logging into QuestionPro to do a survey and then wonder if you should do a poll instead.  So which will it be?  Survey or Poll and how will you know when to do which?

The quick and dirty difference

  • Polls are short, surveys are long
  • Surveys are for making decisions and getting insights, polls are for taking a snapshot
  • Polls give you results right away — with surveys, you have to wait for the analysis

How quickly do you need the results?

  • NOW!  Use a poll.  Online polls are the ideal choice when you have a single question with straightforward choices for answers using a poll software.  For example:  Will you be driving more than 300 miles in your car over the Thanksgiving holiday? (yes, no, don’t know)
  • LATER.  I can wait.  If you can wait for the results – then use a survey.

How many questions do you need answered?

  • One?  Use a poll  – Polls are a quick check on the pulse of an audience.  Unless you’re the Gallup Organization, there isn’t a whole lot of academic theory or design that has to go into the questions.  Ask a question – get an answer – see what everyone else said.
  • Many? Use a survey.  Surveys are more structured and advanced forms of questions.  You give a survey when you’re trying to make a decision or a series of decisions about what to do.

Do you have a BIG decision to make?

  • No.  Use a poll to get a snapshot or pulse of how people view a topic
  • Yes. Use a survey to gather quantitative or statistical data that predicts how people will act.

Are there finite, specific choices?

  • Yes.  Use a poll.  If the list of all the potential options is less than 7 — use a poll.
  • No.  Use a survey.  Surveys are ideal if there is some fuzziness around the answers and you need to gather some open ended responses.

Is this serious or fun?

  • Serious stuff.  Use a survey.  Granted polls can be about serious topics, but the structure doesn’t lend itself for developing a more nuanced context of where the answers come from — that is hard to achieve with one question.
  • Fun stuff.  Polls give you the opportunity to insert a little fun into your research.

Do you need to control the respondents?

  • Yes.  Use a survey.  With multiple questions and demographics, surveys allow you to control the respondents.  If you need a specific number of respondents from a specific demographic – then surveys are the mechanism to achieve that.  You can ask qualifying questions and only allow those respondents who fit the criteria to finish the survey.
  • No.   Polls are not as good as surveys at parsing out specific respondent attributes.  You can’t really qualify your respondents.  Your responses will generally be an overall measure.

The Exceptions

QuestionPro’s new polling feature allows you to do a few things that traditional polls don’t; namely the ability to link polls together and to provide some open ended responses.  This can be good — but not if you use a poll to do a survey’s work.

My recommendation is to use a poll when you have a single question and you want to build engagement around a topic or generate a conversation around a topic.

Use a survey to gather data so that you can make decisions.

What to you say?

If you’re still on the fence, you’ll want to get more details from a FREE Webinar that we’re holding on our new polling feature Thursday October 11, 2012.