The effect is typical of hypothesis tests. Consider the situation where you toss a coin 10 times and get 60% heads. Would you worry that the coin was biased?
Now toss a coin 10,000 times. If you get 60% heads, you might worry (I sure would!) that the coin may be biased. At different sample sizes you get the same effect size but different test statistics - as n gets smaller, the average size of (say) a z-statistic gets smaller for a given effect size. Glen . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
