I don't know Mellie but I have a theory lol
If the youngest 2 children get 6 each then there is 1 peice left that can got to any one of the other 3 children.
so there is 3 ways this can happen 3 ways
If the youngest 2 children get 5 peices each then the other 3 peices can be allocated as
(1,1,1 ) [1way] (1,2,0 ) [3!=6ways] or (3,0,0) [3ways] = 10 ways
If the youngest 2 children get 4 peices each then this is how the other 5 peices can be allocated
(5,0,0) [3!/2!=3 ways] (4,1,0) [3!=6 ways] (3,2,0) [6ways] (3,1,1) [3ways] (2,2,1) [3ways] = 21 ways
If the youngest 2 children get 3 peices each then this is how the other 7 peices can be allocated
(7,0,0) 3ways (6,1,0) 6 ways(5,2,0) 6 ways (5,1,1) 3 ways (4,3,0)6 ways (4,2,1) 6 ways (3,3,1) 3 ways (3,2,2) 3 ways = 3+6+6+3+6+6+3+3 = 36 ways
Now I am going out on a ledge and say that I can see a pattern forming here
3 = 3C3 10= 5C3 21= 7C3 36= 9C3 .............
So maybe if the 2 youngest get 2 peices each then there will be 11C3 = 165 ways to distribute the other 9 peies
and maybe if the 2 youngest get 1 peice each there will be 13C3 = 286 ways to distribute the other 11 peices between the other 3 children.
and if the yougest 2 get none each then there will be 15C3 = 455 ways to distribute the other 13 peices between the other 3 children.
Total = $${\mathtt{455}}{\mathtt{\,\small\textbf+\,}}{\mathtt{286}}{\mathtt{\,\small\textbf+\,}}{\mathtt{165}}{\mathtt{\,\small\textbf+\,}}{\mathtt{36}}{\mathtt{\,\small\textbf+\,}}{\mathtt{21}}{\mathtt{\,\small\textbf+\,}}{\mathtt{10}}{\mathtt{\,\small\textbf+\,}}{\mathtt{3}} = {\mathtt{976}}$$ ways altogether. 
--------------------------------
Mmm
If this is correct, I have found that n identical peices of candy can be distribued
between 3 children $$^{(n+2)}C_3$$ It would be good if someone could walk through the logic of why this is so :/
I shall think upon it 