We are given the following information:
\[
n(A \cup B) = 50, \quad n(A \cap B) = 12, \quad n(A - B) = 15.
\]
We need to find \( n(B - A) \).
Using the principle of set theory, the number of elements in the union of two sets can be expressed as:
\[
n(A \cup B) = n(A) + n(B) - n(A \cap B).
\]
Also, we know that:
\[
n(A) = n(A - B) + n(A \cap B), \quad n(B) = n(B - A) + n(A \cap B).
\]
Substituting the known values into the equations:
\[
50 = n(A) + n(B) - 12,
\]
where
\[
n(A) = 15 + 12 = 27 \quad {(since } n(A - B) = 15 { and } n(A \cap B) = 12 {)}.
\]
Now substitute \( n(A) = 27 \) into the first equation:
\[
50 = 27 + n(B) - 12 \quad \Rightarrow \quad n(B) = 35.
\]
Next, we calculate \( n(B - A) \):
\[
n(B) = n(B - A) + n(A \cap B) \quad \Rightarrow \quad 35 = n(B - A) + 12 \quad \Rightarrow \quad n(B - A) = 23.
\]
Thus, the number of elements in \( B - A \) is \( \boxed{23} \).