Hi vinay1983,
For the real GMAT, questions go through a rigorous testing process before they are considered "fair" and appear as "questions that count" on a live GMAT. During the testing phrase, questions appear as "experimental" questions (which do not count) on live GMAT exams. In that way, the questions can be statistically measured and compared against expectations.
For example, if an experimental question is expected to be answered correctly by 40% of Test Takers AND 40% of Test Takers DO answer it correctly, then it is considered "fair" and then put into the pool of active questions. The process is fairly standardized and remarkably accurate, although sometimes there are anomalies and questions are removed. There have also been cases in which a question was put into the active pool and then later determined to be "unfair" (because the active numbers didn't match the expected numbers).
All of this means that It's possible that this type of SC question could show up on Test Day (either as an experimental question or an active one), but if the results don't match the expectations, then the question would be removed. Questions with an "interpretational bias" don't get very far on the real GMAT though, so I wouldn't worry about it. Also, if this is the only SC that you got wrong on the Official GMAT, then there'd be a chance that it would be experimental (and thus wouldn't count). Even if it DID count, getting one SC wrong wouldn't hurt your score.
GMAT assassins aren't born, they're made,
Rich
