Zar (1999) describes a setting in which it\'s desired to compare two different d
ID: 3231815 • Letter: Z
Question
Zar (1999) describes a setting in which it's desired to compare two different drugs to see which drug makes blood clot faster in samples taken from 12 people with hemophilia (one sample of blood per person) who agreed to be part of the study. Here are three possible experimental designs for this situation. 6 of the people are assigned to drug D and 6 to drug G at random. The number of years the person's body has been unable to produce a normal amount of blood coagulation (a measure of severity of illness) is identified as a potential confounding factor (PCF), and the 12 people are grouped into 6 pairs in such a way that within each pair both people have the same value of the PCF; the assignment to drug B or G within pairs is done at random. The time to clotting is measured for all 12 people on four separate occasions; 6 of the people (chosen at random) are observed on the schedule (1) measure clotting time. (2) administer drug D for an appropriate amount of time, (3) measure clotting time, (4) wait an appropriate time until all effects of the drug have disappeared, (5) measure clotting time, (6) administer drug G for an appropriate amount of time, (7) measure clotting time and the other 6 are treated similarly except that drug G is given first followed by drug B. Identify which, if any, of these designs can be described by the following terms: completely randomized, randomized-blocks, matched-pairs, repeated measures. Then briefly discuss which, if any, of these designs are valid for arriving at correct causal conclusions about the comparative effects of the two drugs, and (if any of them are valid) which are likely to be more accurate at estimating those effects.Explanation / Answer
Answers
Design 1 is completely randomized since it focuses on only the treatment (two different drugs) effects making it a single-factor experiment. The allotment of subjects is randomized.
Design 2 is matched pairs because by choosing each pair having the same value of the PCF, the effect of PCF, if any, is the same for both members of the pair and hence can be considered as identical entities. Consequently, the difference in the response variable for each pair becomes the statistic of concern.
Design 3 does not fit into any of the listed designs.
Clearly Design 2 is the best of the lot in terms of experimental error minimization. It recognizes the existence of PCF and also takes due care of that. The effect of PCF, if any, is nullified by pairing. Random assignment minimizes bias.
Related Questions
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.