Estimating Measures of Pass-Fail Reliability From Parallel Half-Tests