A researcher has conducted the 99% confidence interval that is 10 points wide, based upon the sample of 36 scores. Supposing that variability among scores remains same, how large the sample would be needed to produce a 99% confidence interval that is only 9 points wide?