What is the primary reason that some districted electoral systems using proportional representation formulas exhibit bias toward specific parties? While malapportionment is widely regarded as the leading cause of partisan bias—where equal parties are not treated equally—we argue that electoral magnitude variance has an equally significant, if not greater, role in producing such bias. Magnitude variance, akin to gerrymandering in uniform single-member districts, influences vote efficiency. A positive correlation between magnitude and electoral support leads to wasted votes, while a negative correlation enables more efficient vote allocation. To test these propositions, we evaluate the effects of both malapportionment and magnitude variance using simulated and real data. Using Spain as a case study, we demonstrate that, contrary to common belief, variance in electoral magnitude is the primary driver of conservative party bias in the system.