Comments on: Podcast survey: FIR listeners are influential, educated, mobile and global http://www.nevillehobson.com/2006/05/02/podcast-survey-fir-listeners-are-influential-educated-mobile-and-global/ Business, Communication, Technology Sat, 22 Nov 2008 08:57:48 +0000 http://wordpress.org/?v=2.6.3 By: neville http://www.nevillehobson.com/2006/05/02/podcast-survey-fir-listeners-are-influential-educated-mobile-and-global/#comment-1512 neville Wed, 03 May 2006 20:37:24 +0000 http://www.nevillehobson.com/2006/05/02/podcast-survey-fir-listeners-are-influential-educated-mobile-and-global/#comment-1512 Thanks for pointing that out, Derek. Our survey results are the opinions of 126 people who say they listen to the podcast and chose to tell us what they think of it. Those opinions are good enough for Shel and I to consider in our planning for how we develop the show. But I take your point re using the word "representative" which clearly has specific meaning in surveys. Thanks for pointing that out, Derek.

Our survey results are the opinions of 126 people who say they listen to the podcast and chose to tell us what they think of it. Those opinions are good enough for Shel and I to consider in our planning for how we develop the show.

But I take your point re using the word “representative” which clearly has specific meaning in surveys.

]]> By: Derek Hodge http://www.nevillehobson.com/2006/05/02/podcast-survey-fir-listeners-are-influential-educated-mobile-and-global/#comment-1490 Derek Hodge Wed, 03 May 2006 09:38:53 +0000 http://www.nevillehobson.com/2006/05/02/podcast-survey-fir-listeners-are-influential-educated-mobile-and-global/#comment-1490 I’d be very careful with any claims you make for your sample being representative. You have no way of establishing whether it is representative or not and using a self-selecting sample, as you did, is not a good start. And even a truly random sample of that size would give you margins of error of more than or - 8% for results around the 50% mark. Interesting results and they provide would seem to provide valuable feedback C5rossposted from For Immediate Release podcast blog. I’d be very careful with any claims you make for your sample being representative. You have no way of establishing whether it is representative or not and using a self-selecting sample, as you did, is not a good start.

And even a truly random sample of that size would give you margins of error of more than or - 8% for results around the 50% mark.

Interesting results and they provide would seem to provide valuable feedback

C5rossposted from For Immediate Release podcast blog.

]]>