> So, using a battery of test questions divided into five subjects, INCLUDING TIME AND EVENTS, there was no significant difference between the two groups
It says right in the part you quoted that the Kindle readers scored significantly lower on the event time questions.
They also did poorly on the separate plot ordering test.
I fail to see any problem with this, since I would expect the to be good correlation between how well people do on questions about event timing in the story and how well they can order a timeline of the plot.
I think parent is pointing out that there could be a problem of "mutliple testing" in the statistics used. If you do enough tests for statistical significance, and do not correct for multiple testing, then some fraction (your cutoff, typically .05) will be expected to test positive even when nothing is happening (null hypothesis holds).
It says right in the part you quoted that the Kindle readers scored significantly lower on the event time questions.
They also did poorly on the separate plot ordering test.
I fail to see any problem with this, since I would expect the to be good correlation between how well people do on questions about event timing in the story and how well they can order a timeline of the plot.