This is a nice Word Problems / Algebra setup. The concept being tested is building a tax equation with two different rates and then equating it to a single blended rate.
Let x = Leonor's annual income (and x > $42,000, since she's paying the higher bracket rate too).
1. Tax formula from the brackets: p% of $42,000 + (p+3)% of (x - 42,000)
2. Tax formula from the blended rate: (p + 0.75)% of x
Set them equal and divide everything by 0.01 to drop the percentage notation:
42000p + (p+3)(x - 42000) = (p + 0.75)x
3. Expand the left side:
42000p + px - 42000p + 3x - 126000 = px + 0.75x
The 42000p terms cancel:
px + 3x - 126000 = px + 0.75x
4. The px terms cancel too:
3x - 126000 = 0.75x
2.25x = 126000
x = 56,000
Answer is D.
The trap most people fall into here is getting intimidated by the variable p and assuming you can't solve without knowing its value. But when you expand and simplify, p completely cancels out. That's the whole trick. If you spent time trying to find p or trying to substitute answer choices into the p term, you burned time for nothing.
I'd actually flag this type as a "two-rate blended average" problem. They show up on GMAT Focus in both PS and Data Sufficiency, and the algebra almost always collapses neatly if you're patient with the expansion.