Homogeneity tests for scale parameters of 2-parameter exponential populations under time censoring

Comparing two or more scale parameters of exponential distributions is often important in life-testing problems. This paper considers tests for the equality of several exponential scale parameters in presence of unspecified location parameters based on time-censored data. Using a conditional approach, various test statistics are derived. The Monte Carlo studies were carried out for comparing the powers of these statistics with the power of the usual likelihood ratio statistic. The power of the usual likelihood ratio test was better than those of the other tests for various combinations of mean number of failures