Sorry my stats are rusty--question would seem to be what's the 95% confidence interval for a 50% chance event when drawn 100 times? I'm guessing that 40-60 times would be about normal and that in this particular case you just ended up toward the end of the normal range. Next time could be 43/57, 48/52 etc.

To avoid this, you could randonly pre-assign your subjects to conditions and then use an autostart file. See autostart in the users guide or online here:

www.empirisoft.com/medialab/help/preferences.htm#prefs_auto