0
$\begingroup$

I have a continuous dependent variable and one categorical independent variable (with three levels). For the three levels, I have 42, 33, and 45 observations respectively. So, in short, I have unequal sample sizes.

I want to run a linear model. But I am unsure how to account for the three levels' different sample sizes.

Is using weights in a linear model the right approach? If so, how to use assign them properly?

Improve this question
$\endgroup$
1
  • 3
    $\begingroup$ Welcome to Cross Validated! Is there a reason you believe you have to do something to account for the different sample sizes? There are advantages when it comes to experimental design in balancing the categories, and if what’s the budget to get more observations, sure, it could be reasonable to allocate that to observing members of the smallest class instead of the largest, but equal sample size is not an inherent assumption of regression on categories. $\endgroup$
    – Dave
    Mar 27 at 12:06

1 Answer 1

Reset to default
1
$\begingroup$

There are advantages when it comes to experimental design in balancing the categories, and if there is a budget to get more observations, sure, it could be reasonable to allocate that to observing members of the smallest class instead of the largest, but equal sample size is not an inherent assumption of regression on categories.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.