论文部分内容阅读
上文介绍了三种简单的准实验设计:单组仅施后期测试;非对等的比较组设计;单组先期一后期测试。这三种简单设计,依库克和坎培尔的意见,是“一般不能说明问题”的,下文则将介绍一些被他们二人称为“一般能够说明问题”的准实验设计。不过,即使这些“能说明问题”的设计,它们的可靠性也只是相对地稍高一些而已。换言之,从以下这些设计中所得出的自变量—因变量间因果关系的结论,正确的机率要高一些。但是和实验相比,它们又具有较大的含混性。
Three simple quasi-experimental designs are described above: single-group post-test only; non-equivalent comparative group design; and single-group pre-post-test. The three simple designs, according to Cook and Campbell, are “generally unable to explain the problem,” and the following is a description of some of the quasi-experimental designs that they and their two describe as “generally capable of explaining the problem.” However, even with these “problem-solving” designs, their reliability is only marginally higher. In other words, the correct probability is higher from the following conclusions drawn from these designs: the causal relationship between the dependent variables. But compared with the experiment, they have a greater ambiguity.