Survey Best Practices
Thanks to everyone who attended our recent webinar, “Top Ten Survey Best Practices.” Here is the follow-up info we promised.
What is a good way to create a subset data for the ones who participated the survey? And what is a good way to create a pool of people who agreed to participate in an interview?
If you are looking to separate participators in the email invitation process, our email invite tool helps you segment your list based on whether they answered the survey or not.
If you export your survey data you’ll have only the folks who participated; you then use this as the basis of a follow survey, email invite or survey data pre-population. As for Interview participation, you could also ask a question in the survey about their willingness and then use that field to sort or filter the results.
Can you collect emails for reward incentives SEPARATELY from main responses so that anonymity is preserved?
Yes, you could redirect the Thank You page of the first survey to another survey to collect the personal information.
This would make them separate data stores. The transfer could be seamless to the user.
Does SurveyGizmo allow you to separate those who have taken the survey with those who haven’t taken it? That way those who have completed the survey will not get the reminder.
If you use our Email Invitation tool, you can easily create email follow-ups based on status, i.e. send reminders to non-takers or share results only with survey takers in a separate email.
Can pictures be presented as a part of the survey question? (Think: Which is better at communicating the idea of safety? for example.)
Yes, you can use the SurveyGizmo Image Chooser question type.
Is it a good idea to send an email asking if the person is interested in being included in surveys or is it more accurate to send the survey out to the membership as a whole? Will SurveyGizmo create a random sampling list of my contact list?
Send to everyone or sample at random. Of course you can never make someone respond so in some respects samples are self-selected.
Is it possible to convert answers from a 5 point scale to a 10 point scale and maintain validity in the results? For questions that use a numeric scale for their answers, do you have a recommendation for what that scale should be (0-4, 1-10, etc…)?
You can easily make a linear transformation to go from a 5 point scale to a 10 point scale; by multiplying by two.
As answered during the webinar, I recommend a 5-pt scale or in some cases a 7-pt scale.
Is there a specific time of week or day in the US that is proven to be more successful to send surveys? Where can we get information as to how to formulate questions as to not guide the answers we want.
Yes, for email invitations it is best to avoid the beginning and end of the week, i.e. Tues through Thursday is best.