Improvements to ratings
Improvements to ratings
I believe that the averages for performance are skewed by the first few attempts an individual has at any given test. My response time has gotten better since I first began with the C.Fun activities, but the average score is still elevated by a few miliseconds because of how terribly I did on the first few trials when I began training.
I posit that it would be useful for accuracy purposes to set it so that new accounts can not start recording stats until they have done a minimum number of trials, so that stats won't be skewed by the initial basal lack of skill.
I posit that it would be useful for accuracy purposes to set it so that new accounts can not start recording stats until they have done a minimum number of trials, so that stats won't be skewed by the initial basal lack of skill.
reddisert | 3 years ago
Reply
? | 3 years ago
Reply
reddisert | 3 years ago
Reply
I mean, you could pose that question to cognitivefun, too. Average is too general for a more specifiable idea. Hence the "mean, median, mode" bit.
? | 3 years ago
Reply
The averages are means. reddisert is right about the initial skew but I haven't gotten around to fixing it. The "fix" will probably end up as the mean within a 95% confidence interval, or a 1.96 standard deviation.
cognitivefun | 3 years ago
Reply
Surely another problem is that, even if you had been able to approach your current efforts on your first three attempts, the progressive set-up of many of the tasks would have prevented you from doing so.
For instance, you can't prove you can do 5-back until you've succeeded on the lower levels, by which time your 3 goes would have been used up.
For instance, you can't prove you can do 5-back until you've succeeded on the lower levels, by which time your 3 goes would have been used up.
cevapcici | 3 years ago
Reply
