&textDisplay0=Here's our two by five table with the row percentages in each cell, importantly, percentages that add up to one hundred within each of the categories of the causally prior variable, music genre. If you want to check this, you can sum them along each of the rows and they will total one hundred.

&textDisplay1=The best way to compare percentages is to start to look down the columns. So first of all, to take an example, we'll look at column one with rates of accidental death and you can see that approximately a third of R%26B musicians died accidentally whereas only about a quarter of the rock musicians did. So taking this by itself we can see there may be a slight difference here - it's not a big difference - but maybe the R%26B musicians are more accident prone, or perhaps they live more risky lives than rock stars.

&textDisplay2=One can analyse the table just by doing that separately for each of the columns, going through comparing, eyeballing the percentages one at a time to see what the differences are. However, we make things much easier for ourselves if we present a graph of these percentages and then those differences are going to stand out much more clearly for that. So we can see dropping down those percentages into bars forms our graph. If we look at that graph we can see some patterns have emerged very clearly now. We can see we do get that small difference in accidental deaths but the much more staggering difference is in that second column for alcohol and drug related deaths where we've got a very high proportion for the rock musicians but exactly 0% for the R%26B musicians. Suicides seem to be at about the same rate for both groups. But again for murder we can see quite a strong difference emerging for the rock musicians: four percent of their premature deaths were by murder but for the R%26B musicians the rate was four times as high at16 percent. That is a very big difference and certainly worthy of further investigation. Finally we can look at that large category of deaths from medical causes, and again we can see not many differences between the two.