Aaaah! A-HA! I get it.
In this case, we need to utilise a
jacobhinds said:
Depends on what proportions you're looking at.
The diagram tells us that
T ≥
F ≥
Fu ≥
S ≥
Su, which is obvious. But what's the question?
If we're looking at the "success to all" ratio for first and second marriages, then
Fu /
T ≥
Su /
T, so second marriages are, by definition from
F ≥
S, more successful (or in the worse case scenario, just as successful as the first ones).
I'm pretty sure that's the wrong ratio though, and for the real one, we really need to sample real data.
The real ratio is probably
Fu /
F <=>?
Su /
S. The proportion of unsuccessful first marriages to all first marriages, compared to the ratio of unsuccessful second marriages to all second marriages. "Tried" to make the diagram display the possibility of
Fu /
F <
Su /
S - if 30% of all first marriages are unsuccessful, but 60% of all second marriages are unsuccessful, then second marriages are proportionally less successful than first ones, no? (Numbers pulled right out of my arse.)
Also I can't write "equals with question mark above it" and that's infuriating, because I love that symbol.