Statistics 323 Lab # 9: Hypothesis Testing - Answers/Solutions
Lab Exercise 1:
No Water Meters
(c) The water usage, for both populations consisting of households (1) with water meters and (2) without
water meters appears to be normally distributed, as the medians are roughly in the middle. In addition, the
whiskers coming out of each box appear to be of equal length...this is something one would expect when the
sample is taken from a population of values that is (or near) normally distributed.
(d) The variation in the water usage appears not to be the same. It seems as if there is more variation in
monthly water usage for households without water meters when compared to households with meters. The
size if the interquartile range - the difference between the third-quartile (top of the box) and the first-quartile
(bottom of the box) does not appear to be the close. The size of the interquartile range seems to be close to
twice that in the ‘nometerusage’ compared to the ‘meterusage’.
. From the boxplots, one can infer that the variation in the monthly
water usage of households that have water meters is less than the variation in the monthly water usage of
households that do not have water meters.
As a result, one will employ the ‘Non-Pooled T’ test, or the
Welch-Satterthwaite T. This can be done by manual calculations, producing a test statistic that takes on
a value of
The degrees of freedom would be computed to be
Being on the