1. Before you conduct a test, first draw a dot plot to find out potential outliers. 2. To check normality, you can either use some graphical tool like q-q plot or conduct a formal test (search "check normality" in wikipedia). 3. Compared to the normality assumption, the independence assumption is more important because t distribution is close to normal distribution as long as sample size is large enough (when the data are i.i.d.). On the other hand, if each subject is randomly assigned to one of the two group, then using t distribution won't be a bad idea, even when the data are dependent and the sample size is small. 4. If the assumption of equal variance between the two groups is a concern, then a F-test can be used. ※ 引述《clara830 (...)》之铭言： : 我现在有2组data是要作为之後心理实验要用的stimulus items : 我现在想要比较2组data在某些控制的variables上是否有显着差异 : (检定出来是希望不要有差异.... : 也就是说希望这2组data在某些控制变项上是matched的 : 因为才能避免2组data在之後作实验的时候是受到其他variables的影响) : 我知道现在是应该要跑t-test去作检定 : 可是再仔细查一下资料後 : 我有几个问题不太确定怎麽作.. : 1. 好像跑t-test前 必须要先确定常态分配跟同质性才能做 : 不过我现在不太知道怎麽用SPSS来作常态分配检定和同质性假设检定?? : 2. 如果检定出来不是常态分配的话...那要用什麽检定去比较呢?? : 3. 如果T检定出来2组data有显着差异 那要怎麽作筛选的动作?? : 找出可能影响的outliers并剔除... 才能使我的2组data是matched的呢? : 谢谢解答!! --

※ 发信站: 批踢踢实业坊(ptt.cc) ◆ From: 76.124.84.13