This study used the posterior time limit data to explore whether the more stringent time limit leads the test to measure speed and the reliability changing. The three kinds of data mad from posterior time limit approach are (1) time data, (2) time-accuracy data, and (3) Incomplete accuracy data. The result showed the speed and power measurements are different concept because the correlation between power data and stringent time limiting data is low and the angle of their component vectors are almost orthogonal. The difficulty and reliability change weakly from power to speed. The more stringent time limit result the time-accuracy data trend to speed measurement because the Cronbach and Warrington’s index of speediness are increasing.
1) Those index and linear method is very interesting for me to explain the relationship between data and human behavior.
2) The posterior time limit is useful for us to model the time and accuracy data. However, the real speedness situation doesn't correspond that.
3) The time data fit log-normal distribution more referred in previous studies. Therefore, the binary time data should model by log-normal Ogive model rather than logit model as response accuracy.