目的对汉化版Rivermead 行为记忆测验第3 版(RBMT-Ⅲ)进行信度和效度检验。方法对RBMT-Ⅲ进行汉化。对131名正常人应用汉化版RBMT-Ⅲ进行测验，间隔2 周后从中抽取30 名用不同内容进行重测。结果汉化版RBMT-Ⅲ 14 项分测验的内部一致性Cronbach α=0.799；各分测验分及记忆指数的复本相关系数均>0.65。因子分析抽取到6 个主成分，累计贡献率64.47%，依次为新技能学习、空间记忆、听觉记忆、视觉记忆、前瞻性记忆、日期和定向。结论汉化版RBMT-Ⅲ具有良好的信度、效度。

Objective To analyze the reliability and validity of the Chinese version of the Rivermead Behavioural Memory Test-Third Edition (RBMT-III). Methods The items of RBMT-III were translated in Chinese and adjusted. A total of 131 healthy subjects were tested with Chinese version of RBMT-III, 30 of them were tested again with other materials two weeks later. Results The Cronbach's α coefficient was 0.799. The correlation coefficient of parallel-forms in subtest scale and meory index were all more than 0.65. There were six main factors extracted, which explained 64.47% of the accumulated variance, those were novel task learning, spatial memory, verbal memory, visual memory, prospective memory and orientation/date. Conclusion RBMT-III is a reliable and validated instrument for assessment of memory.