Q1 CPU Prices
In a 2010 survey the prices of server rentals per hour were reviewed. The file cpup.txt contains two variables the Price per hour in US dollars and a variable Bmark which measures with a benchmark score the computational power of each server (a greater score indicates a more powerful server). We will try to test if an increased price indicates a more powerful server.
(a) Draw a scaterplot of the computational power v.s. the Price. Explain how you would choose a “value for money” server using the previous scatterplot?
(b) Using an appropriate test investigate whether there is a significant linear association between Price and Bmark.
(c) Find a 95% confidence interval for the slope and give an informative interpretation.
(d) Calculate a 95% prediction interval when x = $1 and comment on this interval. Include Rcmdr’s
input and output.
(e) Using the appropriate plots (include them) would you recommend the statistical model?

Q2 Fertilisers
A field was treated with three fertilisers A,B or C. The following Table shows the yield in bushes per acre of a wheat variety grown into that field. We are interested in testing whether the differences in mean yields between the different treatments (Fertilisers) can be attributed to chance considering a significance level of 5% and 1%.

Fertiliser Observations
A 3, 4, 5, 4
B 2, 4, 3, 3
C 4, 6, 5, 5
Table 1: Observations of wheat yield from plots treated with fertilisers.

(a) Formulate the relevant hypothesis using the appropriate notation and proceed with the test. Include the relevant Rcmdr input and output.
(b) Using the previous results state your conclusion considering both significance levels.
(c) Using the results from (a), calculate the 95% confidence intervals of the mean yield for the group
B by hand and comment on your findings. Show your calculations.