This paper proposes a method that tests for the existence of low-value spatial clustering while accounting for the influence of high-value clustering. Although the method was developed in reference to the Tango test, it can be extended to other testing methods. The simulation results showed that the proposed method is able to effectively detect low-value clustering with substantially lower rates of type I errors than those of the Tango test, while maintaining comparable statistical power. Applying the method in a case study of leukemia in Minnesota demonstrated an overall tendency toward low-value clustering of leukemia mortality for males but provided inconclusive results for females.