Advertisement
If you have a new account but are having problems posting or verifying your account, please email us on hello@boards.ie for help. Thanks :)
Hello all! Please ensure that you are posting a new thread or question in the appropriate forum. The Feedback forum is overwhelmed with questions that are having to be moved elsewhere. If you need help to verify your account contact hello@boards.ie

Survey Response Rate

  • 30-08-2010 10:16am
    #1
    Closed Accounts Posts: 384 ✭✭ David Some Numerology


    Me again :D,

    Is there a rule of thumb for survey response rates? I.e. if it is at least x it is deemed acceptable.

    I've read around on the internet that its 30+ respondents. But also read it can be 100+.

    Obviously this depends on what type of sample your using, but for a general population survey is there a rule that is well known, justified and quotable by some researcher. I'm trying to explain my desired minimal response rate and want to put a reference in to justify a number i write down.

    I know response rates go by percentage of people responding to survey, but I sent my survey out on facebook so gods knows how many people had a chance to do it.


Comments

  • Registered Users Posts: 3,483 ✭✭✭Ostrom


    Me again :D,

    Is there a rule of thumb for survey response rates? I.e. if it is at least x it is deemed acceptable.

    I've read around on the internet that its 30+ respondents. But also read it can be 100+.

    Obviously this depends on what type of sample your using, but for a general population survey is there a rule that is well known, justified and quotable by some researcher. I'm trying to explain my desired minimal response rate and want to put a reference in to justify a number i write down.

    I know response rates go by percentage of people responding to survey, but I sent my survey out on facebook so gods knows how many people had a chance to do it.

    If you used a stratified or cluster sample, you might be concerned if the response rate left you with an imbalance on the stratifying variables. For simple random or nth samples, it doesn't matter as much. Sample size becomes important when conducting certain tests, as some will require a particular sample size - or if your cell counts fell below the limit for chi-square for example.

    Also, if you had specified a certain desired confidence interval for a particular variable at the design stage, your response rate would be more important, but I'm assuming you didnt do this either. Mark Leary's Introduction to Behavioural Research Methods has a table somewhere in chapter 2 on sample sizes that might be worth a look.


  • Closed Accounts Posts: 384 ✭✭ David Some Numerology


    cheers for the help efla (again!), i'm able to get my hands on that book so i'll take a look (that rhymes!):p


  • Registered Users Posts: 3,483 ✭✭✭Ostrom


    cheers for the help efla (again!), i'm able to get my hands on that book so i'll take a look (that rhymes!):p

    No problem!

    Also, if using regression, there are a few rules of thumb on suitable sample sizes - I cant remember them offhand but they should be in that book. It involves multiplying the number of predictors by a certain constant to give a recommended sample size. There are also effect size tables you can look up to determine how big a sample you will need to detect at certain levels. Dont worry too much about the sample size in this respect, as there are a few different ways of reckoning suitability - one of them is bound to suit :)

    Come back if you need these later and i'll dig out the references


Advertisement