In: Statistics and Probability
In the “Probable Error of a Mean” by William Gossett the following example was used. Yields of corn in pounds per acre were calculated for kiln-dried seed and seed that was not dried. (It was believed in the day that kiln-dried seed produces a higher yield). The yield values for dried seed were:
2009, 1915, 2011, 2463, 2180, 1925, 2122, 1482, 1542, 1443, 1535
The yields achieved with regular seed were:
1903, 1935, 1910, 2496, 2108, 1961, 2060, 1444, 1612, 1316, 1511
Construct a 95% confidence interval estimate of the mean yield in these two cases. Do they look similar?
Let X represent Kiln dried
seeds
Let Y represent undried
seeds
n = sample size = 11
Using Excel functions Average and stdev.s we find the mean and
standard deviations of X and
Y
Mean X = Mx = 1875.1818
Standard Deviation X = Sx =
332.8501
Mean Y = My = 1841.4545
Standard Deviation Y = Sy =
342.7373
Confidence interval for mean is given
by
Confidence Interval for
Kiln dried seeds
For 95% confidence interval, α = 0.05, α/2 =
0.025
From the z-tables, or Excel function
NORM.S.INV(α/2)
z = NORM.S.INV(0.025) = 1.96 (We take the
positive value for calculations)
Confidence interval is
= (1678.48, 2071.88)
95% Confidence Interval for mean yield of Kiln dried seeds =
(1678.48,
2071.88)
Confidence Interval for
non dried
seeds
For 95% confidence interval, α = 0.05, α/2 =
0.025
From the z-tables, or Excel function
NORM.S.INV(α/2)
z = NORM.S.INV(0.025) = 1.96 (We take the
positive value for calculations)
Confidence interval is
= (1638.91, 2043.9993)
95% Confidence Interval for mean yield of non dried seeds is
(1638.91,
2043.9993)
Although there is not a very big difference, it can be seen
that since the lower limit of yield of kiln dried seeds
is higher than the lower limit of the yields of non-dried
seeds
the kiln dried seeds produced a little higher yield than the
undried seeds.