Abstract For the linear combinator of the slope vector in the multiple measurement error model, existing confidence interval is so seriously affected by Gleser–Hwang effect that it is subject to have poor empirical coverage and unacceptable length. Moreover, it is frequent absence when the data is unbelievable. This article therefore construct a new confidence interval which is always available and slightly affected by Gleser–Hwang effect. Besides, we test the equality of the components of the slope vector. Simulation results demonstrate that the generalized fiducial method often outperforms the existing method. A real example is also provided to illustrate our approach.

中文翻译：

---

具有测量误差的多元回归模型中的广义基准推断

摘要 对于多重测量误差模型中斜率向量的线性组合器，现有置信区间受Gleser-Hwang效应影响严重，经验覆盖率差，长度不可接受。此外，当数据令人难以置信时，经常缺席。因此，本文构建了一个新的置信区间，该区间始终可用且受 Gleser-Hwang 效应的影响较小。此外，我们测试斜率向量的分量的相等性。仿真结果表明，广义基准方法通常优于现有方法。还提供了一个真实的例子来说明我们的方法。