There’s always a risk that if you increase sample size without addressing selection bias, you’ll just end up with more confidence in an incorrect conclusion. (“Dewey defeats Truman!”)