After months of trying to get responses to an 89-item self-administered survey returned from 230 people, only 13 have been returned. Since I cannot directly approach the people, I must work though department chairs (after approval from district personnel and school principals), the low response rate must be addressed.

Do you think the survey should be changed to an online administration rather than pen and paper survey instrument? Is there support in the literature to suggest online is getting better response rates?

Since a Web survey also is self-selective as is one via paper then don't think this will have any real effect.

Have you sent out follow-up reminders? Is there a way the intermediators can help through encouragement, etc?

I was thinking a mixed-mode survey approach might catch some of the non-responders by making it easier to complete the survey at home.

Two phases to the research:

Phase I: Mailings / reminders / etc
Phase II: Send invitations for responding via Internet.



Absolutely. Emails, phone calls, postcard reminders -- distanced to give them time between. .

Sounds like an audience issue in that they simply don't want to participate. Adding a Web option might add in some new responses so if do the only call those non-respondents, as sure you know, to avoid duplications

Check the work of Dillman (e.g., see link below), he has written extensively on increasing responses to surveys. Many of you have posted good suggestions in line with Dillman's recommendations like mailing reminders and so forth.

There is a chapter in his new book on internet surveys. It is not stellar, but it would be useful for someone new to the area.

http://www.psypress.com/shopping_cart/p ... 857528&pc=

We've (my research group) have had success with mailouts (Canada Post mailouts by postal code), telephone interviews, and most recently we have internet data on 500 postmenopausal women.

Response rates have been falling over the years, however your response is quite low. However, this is difficult to judge since response rates will depend on who you are getting to respond (An undergrad class vs a random mailout, for example). So, if you are randomly sending out 230 surveys, 13 is low but not shocking.

Anyway, let me know if you want any specifics. Feel free to email me.

When I posted this I was looking for the latest copy of Dr. Dillman's textbook. I now own it and am in the middle of reading/re-reading sections.

I now have 25 surveys returned. The audience is not receptive because many of these new teachers received RIF notices (reduction in force) notices. Yet a bigger challenge is the number of teachers reported to the state and actually calling the schools and finding out how many new teachers are actually working. This is quite a challenge. Numbers are rather 'cloudy' with the reported numbers ...

Now, I have run and re-run the sample size numbers and must get 129 -- 104 to go

An electronic version of the survey is now available and my latest letter to the teachers states they can complete the survey online now. Hopefully this will help.