一个理想的调查技术是可重复的,即研究对象再次被检查时可得到同样结果,而且也是有效的,即它能达到所要测量的目的。重复性差意味着有效性差,因为只可能有一个答案是正确的。但是一个一致的答案也可以是错误的,例如一项试验可能持续产生假阳性结果,或者采用了一种不敏感的测量方法。为了评定流行病学研究结果的份量,需要对其重复性和有效性进行数值估计。 An ideal survey technique is repeatable, that is, when the subject is examined again, the same result can be obtained, and it is also valid, that is, it can achieve the purpose of measurement. Poor repeatability means poor effectiveness because only one answer may be correct. But a consistent answer can also be wrong, for example, an experiment may continue to produce false positive results, or adopt an insensitive measurement method. In order to assess the weight of epidemiological studies, numerical estimates of their reproducibility and validity are needed.