I plotted the census results given in Bamidbar, and it struck me that the tribes descended from Rachel are very small. In contrast, those descended from Leah have population counts between average and very large. (Levites are an exception, being a very small tribe descended from Leah, but they are explicitly set apart as not to be counted with the other tribes.)
Initially I thought it was simply a matter of timing: Ephraim and Manasseh are from a generation later, and Benjamin was born late. So their tribes have had less time to grow. But this doesn't work: if it were true, Leah's younger sons Issachar and Zevulun would have smaller tribes than Reuven, but that's not the case.
Then I thought it might have to do with Leah's merit. But then I'd expect the same pattern for the handmaids. Instead, it's reversed: tribes descended from Bilhah are larger than those descended from Zilpah.
Plotting the average tribe count for each matriarch shows that the gap between the smaller and larger groups is more than the standard error. If the tribe of Levi is included, the standard deviation for Leah increases and the two groups barely overlap.
Total counts of descendants are mostly driven by the number of children: 48% of everyone descends from Leah. (50% if the tribe of Levi is included.) The other groups have similar populations, which makes sense since each is descended from two sons. (Joseph has many descendants split between the two tribes, but Benjamin has few.)
Overall, these plots suggest there's a specific pattern and not simply random variance. Is there a reason for this?