Generating confidence intervals for both height and weight with the same sample?

[downthesun01]

Given that height and weight tend to be correlated, would it be okay for me to use a single simple random sample of people to create confidence intervals for both their height and their weight?

 

[hlsmith]

Are you talking about just calculating CI based on SE from descriptive statistics? If so, I see no issues in doing this.

 

[downthesun01]

Are you talking about just calculating CI based on SE from descriptive statistics? If so, I see no issues in doing this.

Thanks. I was thinking that because of the correlation I'd need to draw wto separate samples.

 

[hlsmith]

Side note, if you were running a regression model you would not have to draw two samples but the SE associated with a prediction term would likely be wider if collinearity was present.