The present study uses an acceptability judgment experiment designed to explore the nature of Korean why-stripping with reference to voice. First, it investigates whether the acceptability of Korean why-stripping is affected by the voice mismatch between antecedent and ellipsis clauses. Second, it examines whether there is a difference in acceptability between passive antecedents with active ellipsis clauses (Passive-Active) and active antecedents with passive ellipsis clauses (Active-Passive). One finding indicates that Korean why-stripping prefers voice matches over voice mismatches: voice matches make it easier to link the remnant to its correlate in the antecedent clause. The other finding indicates that Passive-Active why-stripping is more acceptable than Active-Passive why-stripping. The study also evaluates whether the second finding supports Arregui et al.’s (2006) Recycling Hypothesis according to which passive clauses are more likely to be remembered as active than the other way around, creating an illusion of grammaticality or Poppels and Kehler’s (2019) hypothesis according to which there exists a penalty against passive ellipsis clauses, which applies to both matched and mismatched cases of English VP ellipsis. The results confirm the hypothesis of Poppels and Kehler in that Active-Active why-stripping is more acceptable than Passive-Passive why-stripping and that Passive-Active why-stripping is more acceptable than Active-Passive why-stripping.