Partial differentiation and the chain rule: the setup mistake that costs the whole question
Partial differentiation questions in an engineering mathematics paper rarely fail because a student cannot differentiate. They fail because the student differentiated the wrong thing — held the wrong variable constant, or applied the chain rule for a dependency that was not there. Four checks catch most of it.
- Write down what depends on what, before anything else.
If z depends on x and y, and x and y depend on t, the chain rule for dz⁄dt needs both ∂z⁄∂x and ∂z⁄∂y, multiplied by dx⁄dt and dy⁄dt, then added. A single line at the start — “z = z(x, y), x = x(t), y = y(t)” — prevents the most common error: a chain rule with a missing term.
- Treat “hold this variable constant” as a decision, not a habit.
In ∂z⁄∂x, y is held constant. In ∂z⁄∂y, x is held constant. Students go wrong under pressure by treating both the same way instead of consciously switching which variable is frozen.
- Do both second-order derivatives independently.
For most functions, ∂²z⁄∂x∂y equals ∂²z⁄∂y∂x, and examiners often ask you to verify exactly this. Arriving at that equality by two different, differently error-prone routes is the point of the question — do not do one and assume the other matches.
- Watch for an extra hidden variable in implicit questions.
A question giving F(x, y, z) = 0 and asking for ∂z⁄∂x wants you to treat z as an implicit function of x and y, then rearrange. Forgetting this and treating z as a constant produces an answer with the wrong number of variables in it.
If you can differentiate correctly but keep losing marks on this kind of question, the problem is almost never your calculus. Bring the specific past paper questions catching you out to Sector 12 and we will work through the setup together.
