Step 1: Recall the formulas for a binomial distribution.
For a binomial distribution,
\[
\mu=np,
\]
and
\[
\sigma^2=npq,
\]
where
\[
q=1-p.
\]
Hence,
\[
p=1-\frac{\sigma^2}{\mu}.
\]
Also,
\[
n=\frac{\mu}{p}.
\]
For a valid binomial distribution,
\[
n
\]
must be a positive integer.
Step 2: Check each option.
\[
\boxed{(A)
\]
\[
p=1-\frac26=\frac23,\qquad
n=\frac6{2/3}=9.
\]
Valid.
\[
\boxed{(B)
\]
\[
p=1-\frac{12}{16}=\frac14,\qquad
n=\frac{16}{1/4}=64.
\]
Valid.
\[
\boxed{(C)
\]
\[
p=1-\frac{16}{20}=\frac15,\qquad
n=\frac{20}{1/5}=100.
\]
Valid.
\[
\boxed{(D)
\]
\[
p=1-\frac5{15}=\frac23,\qquad
n=\frac{15}{2/3}
=\frac{45}{2}
=22.5.
\]
Since
\[
n
\]
is not an integer, this is not possible.
Step 3: Write the final answer.
Hence,
\[
\boxed{\text{Mean }15,\ \text{variance }5}
\]
is not the correct data of a binomial distribution.
Thus,
\[
\boxed{(D)}
\]
is the correct answer.