When is a good time to poll?

If you poll continuously this question is irrelevant, because you’d be running fieldwork all the time.

For polls that are not run continuously, timing is an important consideration. Below are some things I consider when advising clients on fieldwork timing. Some of these considerations apply more to random probability surveys, designed to drive as high a response rate as possible. Quota surveys are a bit different, because they are designed to ensure the final sample includes the correct proportion of people who meet key predetermined criteria (they are not designed to target a high response rate).*

  • Firstly, people are at home at different times and on different days, and it’s possible that these home/away patterns differ by party support. A poll should ideally cover a range of days and times, over a number of fieldwork days. That way, if the eligible respondent isn’t home when you first call, you can call back or arrange a time to call when they are home. (My view is that it’s really important to try and get in touch the selected respondent in the household – that could mean making an appointment to call on a weekday morning, or calling on an alternative number or cell number if someone has given you one.)
  • Secondly, consistency is fairly important. Because people are at home at different times and on different days, and those patterns might differ by party support, it’s ideal if your polls cover the same days and times. That way, if there is a bias introduced by your standard fieldwork day and shift patterns, at least that bias will be consistent across all your polls and the results will be comparable. (Another view I have is that consistency is preferable to volatility.)
  • Thirdly, it’s good to avoid times when you know people are less likely to be home. This means avoiding peak holiday season, public holidays, and long weekends. Achieving a good response rate at these times can be very difficult, and if you want your results over time to be comparable then it’s good to be targeting a similar response rate each time.
  • Finally, it can be useful to plan your poll around potential key events, such as state visits, important speeches, the release of the budget, a party leadership vote, a party conference or, ahem, an election. You can poll before the event to get a prior read on public sentiment, or you can poll after the event to try and gauge it’s impact.

Something else I should mention is that not everything is within a pollster’s control. For example, there may be a very good reason to change fieldwork days between one poll and another, and we can never control the darn weather!

*Both of these approaches are trying to collect a representative sample, but they go about it in different ways.


Make a comment...

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s