I will take a different tack, which might be called the "Vi$\grave{e}$te approach". I will note that for resolving the choices in this problem, the numbers 3, 5, and 7 appearing in the coefficients have no special significance, other than that they do not change the signs of the coefficients from what is given by $ \ p \ , \ q \ , \ \text{and} \ r \ . $ (They do have the effect of making the discriminant of the quadratic polynomial look more intimidating to analyze.)
Since $ \ p \ , \ q \ , \ \text{and} \ r \ $ are real, the coefficients of our polynomial are real; as the door has been "left open" to consider imaginary zeroes, this means that such roots must form a "conjugate pair". If we call the zeroes $ \ \alpha \ \text{and} \ \beta \ , $ then the factored form of this polynomial is $ \ 3p \ (x - \alpha) \ (x- \beta) \ , $ which "multiplies out" as
$$ 3p x^2 \ - \ 3p \ [\alpha + \beta] x \ + \ 3p \alpha \beta \ = \ 3px^2 + 5qx + 7r \ \ . $$
We can eliminate choice (D) immediately. If the roots were pure imaginary, the real coefficients require that $ \ \alpha = b i \ \text{and} \ \beta = - b i \ . $ This would force their sum to be zero and, consequently $ \ q = 0 \ ; $ further, their product would be positive, requiring that $ \ p \ \text{and} \ r \ $ have the same sign. But under the condition $ \ p + q + r = 0 \ , \ q \ $ being zero forces $ \ p = -r \ . $ So we have an outright contradiction in this case.
If both $ \ \alpha \ \text{and} \ \beta \ $ are positive, then their sum and product are positive, requiring that $ \ p \ \text{and} \ r \ $ have the same sign and $ \ p \ \text{and} \ q \ $ have opposite signs. Under the specified condition, we can write $ \ p + r = -q \ , $ but this by itself can be satisfied by values of $ \ p \ \text{and} \ r \ $ of opposite sign as well. This is to say that the specified condition, $ \ p + q + r = 0 \ , $ does not impose a strong enough constraint to require that the two roots both be positive. (That is to say, they could be, but don't have to be.)
A stronger argument follows for having both $ \ \alpha \ \text{and} \ \beta \ $ negative: here the sum of the roots is negative and their product is positive, causing $ \ p \ , \ q \ , \ \text{and} \ r \ $ all to have the same sign, which is impossible under the imposed condition.
Finally, we show that choice (C) is correct, also by contradiction. Were the two roots equal $ ( \ \alpha = \beta \ ) , $ our polynomial would be
$$ \ 3p \ (x - \alpha)^2 \ = 3p x^2 \ - \ 6p \alpha x \ + \ 3p \alpha^2 \ . $$
We would thus have $ \ 5q = -6p \alpha \ \Rightarrow \ q = -\frac{6 \alpha}{5} p \ \ $ and $ \ 7r = 3p \alpha^2 \ \Rightarrow \ r = \frac{3 \alpha^2}{7} p \ . $ The specified condition implies that $ \ p \ (1 - \frac{6}{5} \alpha + \frac{3}{7} \alpha^2 ) \ = \ 0 . \ $ It is impermissible to have $ \ p = 0 \ . $ But the discriminant of the quadratic equation for $ \ \alpha \ $ is $ \ (-\frac{6}{5})^2 - 4 \cdot \frac{3}{7} \cdot 1 \ = \ \frac{36}{25} - \frac{12}{7} \ < \ 0 . $ So the roots cannot be equal and real. [This is the only place in the entire discussion where the numbers 3 , 5 , and 7 have much importance.]