So my coursebook says beta is normally distributed in large samples due to CLT. But why is it not chi squared distributed? I'm dealing with variances which are second power of normally distributed random variable so should I not end up with chi squared distribution?