Originally Posted by
astrohog
At what point, statistically, do they extrapolate sufficient information?
Last I checked e-mail was free, so the only added expense would be to read and assimilate the data collected. The AMA was formed by its members and for its members, seems like listening to ALL its members (or at least giving them all a chance to voice their opinions) would be a prudent way to spend its members' dollars!
Regards,
Astro
You won't like the answer, but..it depends. the size of the survey population plays a large part, and past participation in similar surveys can play a part too. Segmentation is a factor as well, are they looking for the total demographic here, from coast to coast, or looking for key areas. Dunno. Perhaps it was based on membership density in geographic areas. The degree of variance in the responses from the surveyed population is also a factor. If a lot of responses are similar in a clustered area, chances are the more surveys they send out, the more similar the responses will be. Typically the more than respond the better, more data equals more stats to analyze, but there is a point of diminishing returns as well. I've seen survey response rates of 10% that were statistically relevant (tested against a control survey for level setting), to needing to get into the 25% response rate. It's not a perfect science.
As for e-mail, no, actually it's not free. The e-mail fairy doesn't exist anymore, he/she has been replaced with programmers, servers, bandwidth, and infrastructure. The AMA was formed for the membership, who has elected officials to act on their behalf. While in a perfect world ALL would get to voice their opinion, and as it happens they can. They have direct lines to their local district leadership, regional AVP, and they can even right directly to the senior leadership. I'm not aware of anything stopping them.
In the case of the survey, it's not realistic or cost effective to send out 150,000 surveys. In the very rare case even 50% of people responded, consider the time it would take to review the results. If they were formatted as answer 1-10, that would be easier to do. One or two people could slice and dice those results easier. Tack on the free form answers, that ads a level of complexity that requires much more labor and man-hours. That cost money, and god knows the AMA has already been criticized for they way the spend our dues. I could hear the complaints now, right?
As is, I'd ballpark the cost of this in the $5,000k range, at least. Keep in mind the AMA has a researcher on staff too.