i agree that it's a design issue rather than a stats issue. maybe for mags below a certain sample size, it's hidden by default, & you have to click to expand an infobox with the disclaimer that the sample size is small, or something like the duotrope disclaimer about response bias?