This post is inaccurate, see the following post for details.
So Chris wanted to see the distribution as to when the errors in the 1back were occurring. We know that position is a huge factor, more errors for rotation than stay, but when do the errors actually occur?
Again, no graphs.. screw the manual copying (damn you SPSS, why don’t you have a viable cut & paste option from your crappy output tables?)
I split the nback responses into temporal thirds. RMANOVAs aren’t really needed to see how this plays out: virtually all the errors occur with the first third (four responses). Error rates by thirds: 3.5%, 0.34%, 0.05% (p<0.000). Error rates by position: 0.3%(stay), 2%(rotate). The interaction is significant as well: Stay x When : 1%, 0%, 0%; Rotate x When : 6%, 0.7%, 0.09% (p<0.002). Again, no difference between spatial or verbal.
I’ll ponder this stuff later.
