The possibility that classical gamma ray bursts (GRB) occasionally repeat from the same locations on the sky provides a critical test of GRB models. There is currently some controversy about whether there is evidence for burst repetition in the BATSE data. We introduce a gamma ray burst ``pair matching" statistic that can be used to search for a repeater signal in the BATSE data. The pair matching statistic is built upon the reported position errors for each burst and is more sensitive than previously used statistics at detecting faint repeating bursts or multiple burst recurrences. It is also less likely to produce (false) evidence of burst repetition due to correlations in the positions that are inconsistent with repeating bursters. We find that the excesses in ``matched" and ``antipodal" pairs seen with other statistics are caused by an excess of pairs with a separation smaller than their error bars would indicate. When we consider all separations consistent with the error bars, no significant signal remains. We conclude that the publicly available BATSE 1B and 2B data sets contain no evidence for repeating gamma ray bursters.