Survey's without Response Rate - Surveys distributed by discussion forums

P

podge79

Hi,

Quick question in relation to Survey's if they are distributed via discussion forums you then will not know what the response rate. Can this be classed as a valid why to conduct a survey as really you can'teasure survey bias.

Would appreciate anyone view here.

Thanks in Advance.
 

Marc

Fully vaccinated are you?
Leader
Re: Survey's without Response Rate

Polls in forums are wildly inaccurate. Rounding errors, ballot stuffers, dynamic IPs, firewalls. Obviously you can not use these poll numbers to do anything important. Same with surveys *distributed* by way of discussion forums.
 
P

podge79

Re: Survey's without Response Rate

Hi Marc,

Thanks for the response, the reason I'm asking is as part of a masters thesis I wanted to get a handle on basic industry preceptions in realtion to an area I am studying and wanted to get access to a wider audience, as using email surveys tends to deliever a very low level of responses and thus response rate.

Do you think results from distributing it via a discussion would not be valid in this context then?
 

Marc

Fully vaccinated are you?
Leader
Re: Survey's without Response Rate

To be honest I'm not that well versed in analysis, of or validity of, such surveys to answer. What I do believe is that you can not get any data with regard to response rate.

Maybe there's someone here who can give you a real answer.
 
P

podge79

Re: Survey's without Response Rate

To be honest I'm not that well versed in analysis, of or validity of, such surveys to answer. What I do believe is that you can not get any data with regard to response rate.

Maybe there's someone here who can give you a real answer.

No Problem Marc, I appreciate the input you have given :thanx:
 
M

mguilbert

To help with response rates I have seen one company add a drawing for a prize ($50 gift card) for submited responses. I understand as a student that may be difficult. But by using a prize you will also know more about your respondents.
 

BradM

Leader
Admin
Well... in short, it depends. :D

I wanted to get a handle on basic industry preceptions in realtion to an area I am studying and wanted to get access to a wider audience, as using email surveys tends to deliever a very low level of responses and thus response rate.

Who are you going to be performing a survey with? When you state non-response bias, what are you assessing that against?

Response rates from surveys are typically pretty low. One thing you can do is compare the companies that did respond to the companies that did not respond. So if the "typical" subject that did not respond is similar to the subjects that responded, you might be OK.

If you find that a majority of the companies that responded are large mfg., and very few small service providers responded, you will have a difficult time generalizing your results over a broad organization structure.

Also, I hope you have done a field survey, focus groups, validation and reliability studies, etc. on your survey. If you are going to do it, it needs to be good.

That said, calculate what your non-response rate is. Compare to equivalent studies to yours. If your rate is the same/ lower, quote their study to give your results some strength. Compare the non-responders to the responders and determine if there are significant differences. And as always, increase how many people conduct your survey. Increased sample size helps.
 

BradM

Leader
Admin
Re: Survey's without Response Rate

What I do believe is that you can not get any data with regard to response rate.

An astute point :agree1:. It does depend on how the survey is conducted. If you send a survey request to a particular mail database, SIC code, or whatever, then you know how many people received it. That's your total. Then you calculate based on how many respond.

If you put a survey out and don't have a particular targeted group, then you're correct. It's impossible to know who did not respond; only those who responded. :agree1:
 
P

podge79

Re: Survey's without Response Rate

An astute point :agree1:. It does depend on how the survey is conducted. If you send a survey request to a particular mail database, SIC code, or whatever, then you know how many people received it. That's your total. Then you calculate based on how many respond.

If you put a survey out and don't have a particular targeted group, then you're correct. It's impossible to know who did not respond; only those who responded. :agree1:

Hi Brad,
Just wondering if what I propose to do in distributing could fall into what is termed Snowball sampling?

Rather than distributing the survey over directly over a forum maybe ask for people to respond and may suggest others that might respond.... forum with private messaging would be better for this approach??
 
Top Bottom