I'm having trouble unifying these two ideas. When I read examples using, say, qiskit, it's typical to measure a qubit, so that to find the probability of measuring a state $|\psi \rangle$ in the state $|x\rangle$ is $p(|x \rangle) = (\langle x | \psi \rangle)^2$.
But when I read quantum computing textbooks, they usually refer to some observable that is getting measured. So what is the operator that gets measured when we "measure a qubit"? Is it the identity operator or something?