Effect of Ethosuximide on IQ Data
IQ after
Placebo |
IQ after
Ethosuximide |
|
97
|
113
|
|
106
|
113
|
|
106
|
101
|
|
95
|
119
|
|
102
|
111
|
|
111
|
122
|
|
115
|
121
|
|
104
|
106
|
|
90
|
110
|
|
96
|
126
|
This data set resulted from an experiment to determine if ethosuximide, a drug useful for treating a mild form of epilepsy,
could be useful as a "learning facilitator" for children with learning and behavioral problems. In the study, each of 10 children
were given a placebo for a three week period, and then given some parts of a standard IQ test. Also, each of the same 10 children
were given ethosuximide for three weeks, followed by an IQ test. Because a child might be expected to do better on an IQ
test the second time he or she took it, the order in which the placebo and the ethosuximide were administered was randomized.
This way if there is a learning/familiarization effect due to taking an IQ test for the second time, and the drug had no effect,
the randomization would result in the apparent "IQ boost" benefitting the placebo score with probability 1/2 and benefitting
the ethosuximide score with probability 1/2. (So here a good experimental design should result in the test statistics having
the proper null sampling distribution if there is no difference in the way the two treatments affect the IQ scores.)