Some Statistical Issues in the Measurement of the Top Quark Charge