[From Samuel Saunders (950312:2340 EST)]
I got Luis, my 5 year old son, to do 10 runs of the SD3 program. I
analysised the last 3 using the original Bruce Abbott SD3an program. Runs
8 and 10 are listed here; on run 9 he missed one of the stimulus changes
and so error is much higher. The data are for the best delay. I
analysised in steps of 10, then backeted to a final step of 2 on the delay:
run delay k Correlation RMS error
8 57 0.0146 0.897 50.741
10 60 0.0205 0.873 59.742
Luis is color blind, so I asked him after the runs if he could clearly
distinguish the color of the cursor, and he said he could. He said he
would like to try some more runs, so I should have more data later to see
if his delay drops with practice. I thought the runs would be a little
long for him, but he asked "why are the runs so short; I wish they were
longer". He also wished he could "earn" more points.
Looking over his shoulder, I noticed that he often moved the cursor about
half-way between the two targets, then moved on to the new target on a
switch. I also noticed that he seemed slowest at detecting the change in
color when he was tracking closest to the target at the time of color
switch. From run 2 to about run 4, he called out the color each time it
switched. I suppose some of these observations may suggest hypotheses for
determining why his delay is so long, if that finding holds up. It may
just be that he need more practice.
I didn't get a chance to test the 2 year old yet. She and I both were
suffering with flu this weekend. I will try to get here to do a few runs
and get more data from Luis in the next few days.
···
-------------------------------------------------
On the overshoot question- Is everyone using some sort of mouse
acceleration? I am using the mildest settings available in OS/2, which I
understand are milder than most, but I still suspect some of the oveshoot
I have may be due to the difficulty in estimating the effect of the
accelerator.
//----------------------------------------------------------------------------
//Samuel Spence Saunders,Ph.D.