Home > Statistics Every Writer Should Know > The Stats Board > Discusssion
point biserial & sample size
Message posted by Lilian (via 212.137.183.217) on June 25, 2001 at 5:06 PM (ET)
Does it matter what my sample size is if I'm doing a point biserial correlation?
READERS RESPOND:
(In chronological order. Most recent at the bottom.)
Re: point biserial & sample size
Message posted by JG (via 128.8.23.12) on June 25, 2001 at 5:55 PM (ET)
What is biserial correlation ? What type of problem are we talking about ? Usually, sample size is not important if the sample is either large enough or way too small.
Re: point biserial & sample size
Message posted by Peter (via 131.193.141.96) on June 26, 2001 at 5:00 PM (ET)
I would say that you would need a minimum of 30 subjects, with at least 15 subjects for each category of your predictor variable. why this number? there is a consensus that a minimum of 30 subjects is needed for any general linear model to ensure robustness, that is, that the sample size has enough power that the result you obtain is a trustworthy result and is representative of the population at large. i agree with the previous responder, it would help to know the background of your question to determine an approprate sample size.
Re: point biserial & sample size
Message posted by Lilian (via 212.137.186.88) on June 28, 2001 at 11:32 AM (ET)
Thanks for the replies, that's exactly what i wanted to know. I have a sample of 30, but not all of the predictor variables have 15 subjects in each category.
Re: point biserial & sample size
Message posted by alexandra (via 198.81.17.182) on October 31, 2001 at 2:41 AM (ET)
i am working on my dissertation proposal. i have a dichotomous independent variable and 3 dependent variables. i was advised by a professor that i need 300 subjects. this seems too large. i would kindly appreciate your input.
Your $5 contribution helps cover part the $500 annual cost of keeping this site online.