In my estimates I used a biased urn model. Given the number of remaining journalists is very small when compared to total population (hundreds vs. millions), the final proportions should still be similar, unless the risk disparity is huge enough to overcome the effect of four order of magnitude difference in absolute number between groups.