How I made my histograms

 histocum.p min=0 max=100 bins=20 out=H10 P10.txt
 histocum.p min=0 max=100 bins=20 out=H65 P65.txt
!histocum.p min=0 max=100 bins=100 out=H10h P10.txt
!histocum.p min=0 max=100 bins=100 out=H65h P65.txt

min = 0.000000, max = 100.000000, bins = 20.    gnuplot
plot 'H10' u 1:2 w boxes, 'H10' u 1:3 w l
plot 'H10' u 1:6 w l , 'H10' u 1:7 w l
count = 25, mu = 20.380000, SumSquareDevns = 6005.640000, sd about zero = 25.604101, sd = 15.818818, range = 1.000000,70.000000
 histocum.p min=0 max=100 bins=20 out=H65 P65.txt
min = 0.000000, max = 100.000000, bins = 20.    gnuplot
plot 'H65' u 1:2 w boxes, 'H65' u 1:3 w l
plot 'H65' u 1:6 w l , 'H65' u 1:7 w l
count = 23, mu = 38.173913, SumSquareDevns = 7961.304348, sd about zero = 42.466355, sd = 19.023077, range = 16.000000,85.000000

See also the file gnu for how I plotted these.

! histocum.p min=-1 max=26 bins=27 out=Mother c=1 children.txt
! histocum.p min=-1 max=26 bins=27 out=Aunt c=2 children.txt
plot 'Mother' u 1:2 w boxes, 'Mother' u 1:3 w l
plot 'Mother' u 1:6 w l , 'Mother' u 1:7 w l
count = 49, mu = 5.510204, SumSquareDevns = 192.244898, sd about zero = 5.855400, sd = 2.001275, range = 2.000000,10.000000
!
gnuplot> ! histocum.p min=0 max=25 bins=25 out=Aunt c=2 children.txt
min = 0.000000, max = 25.000000, bins = 25.    gnuplot
plot 'Aunt' u 1:2 w boxes, 'Aunt' u 1:3 w l
plot 'Aunt' u 1:6 w l , 'Aunt' u 1:7 w l

===========================

 ResponsesG.gnumeric 

is where I have made graphs using gnumeric

raw data are in Responses.gnumeric

===========================

Sampling theory for the Aunts versus Mothers
--------------------------------------------
To analyse the 'significance' of the "16 above, 29 below" result
exactly, we can compute the 'p-value' 
like this.  The null hypothesis is H0: p=1/2.
The alternative says 'no, mothers have larger numbers of children'.

load oct
(oct.m)

octave
> a=[0:45];
> y=binomial_pdf(a,45,0.5);
> a(17)
ans = 16
> sum(y(1:17))
ans = 0.036227

So the 'p-value' is 0.036.

This is normally viewed as 'significant' evidence.