PDA

View Full Version : Survey Sez…



Chuck Bao
3/2/2007, 06:47 PM
Anybody have any idea how to analyze an online survey?

I just got the results of an online poll and I’m at a loss what to do with the information.

Two weeks ago, we sent out 40,000 emails to only our active client list to participate in our online survey and provided the link. Nobody else was given the URL to the online poll. In total, we got 842 to participate in my 12-question survey. All matching IP addresses were eliminated.

Would those who responded be representative? I’m seriously thinking about doing this every three months and developing an index of the confidence of retail stock market investors.

That is if my survey results can be trusted. Maybe only the really disgruntled people with recent stock market losses felt more inclined to respond and maybe it’s not a leading indicator at all but a seriously lagging one.

The response to the first question could be indicative.

1) How confident are you of getting a satisfactory return on your stock market investments this year?

a) Very confident – 4.99%
b) Slightly confident – 38.12%
c) Not confident at all – 56.89%

So, do you think my survey is representative and how confident are you in making a good return on your online stock trades?

TheUnnamedSooner
3/2/2007, 06:56 PM
842 out of 40,000 is not a good representation. If my company sent me an email about a survey (and i get them quite often from different places) I usually only fill out the survey if I'm unhappy with them for whatever reason. If I'm perfectly happy, I typically don't spend the time with the survey b/c I don't want them to change anything. I know how retarded that sounds, but that is the way that your average person will respond (i.e. a no response is a good one)

proud gonzo
3/2/2007, 07:07 PM
yeah, the problem is that the 56.89% who said they're not confident of their returns only represent 1.19% of the 40,000...

Chuck Bao
3/2/2007, 07:33 PM
yeah, the problem is that the 56.89% who said they're not confident of their returns only represent 1.19% of the 40,000...

But, that's what polls do, isn't it? Sort of extrapolate.

Okay, I see. It isn't a random polling.

Thanks PG and Unamed Sooner for the responses.

TheUnnamedSooner
3/2/2007, 08:03 PM
np, if you had like 5,000 responses, then that would be a much better representation