Refer to Exercise 18.30. Consider as your response variable the proportional change in the mean percentage leakage at time 3 hours and at time 0. That is,
where P0 and P3 are the percentage leakage values at times 0 and 3 hours, respectively. Run an analysis of variance on y and test for significant interaction and/or main effects due to CCl4 and CHCl3. Do you reach similar conclusions to those obtained in Exercise 18.31?
Refer to Exercise 18.30.
a. Run a repeated measures analysis of variance and determine if there are significant interaction and/or main effects due to CCl4 and CHCl3. Is there a significant time effect?
b. Do the conditions necessary for using a split-plot analysis of repeated-measures data appear to be valid?
The following data are from Gennings, Chinchilli, and Carter, Journal of the American Statistical Association 84 (1989): 805 – 809. An in vitro toxicity study of isolated hepatocyte suspensions was conducted to study the impact of combining carbon tetrachloride (CCl4) and chloroform (CHCl3) on the toxicity of cells. Cell toxicity was measured by the amount of lactic dehydrogenase (LDH) enzyme leakage. The study involved randomly assigning four flasks to each of the 16 treatments obtained by combining four levels of CCl4: 0, 1.0, 2.5, 5.0 mM with four levels of CHCl3: 0, 5, 10, 25 mM. The percent LDH leakage from the cells in each of the 64 flasks was measured just prior to applying the treatment to the flasks and at .01, .25, .5, 1, 2, and 3 hours after applying the treatment. The percent LDH leakage is given in the following table.
a. Plot the mean percentage LDH leakage by time for the 16 treatments. Does there appear to be an effect due to increasing the levels of CCl4 or CHCl3?
b. From the plot, does there appear to be an increase in the mean percentage leakage as time after treatment increases?
c. Plot a profile plot of the mean percentage LDH leakage separately for each time period. Does there appear to be a difference in the profile plots?