AP Statistics / Mr. Hansen 2/11/1999
Group Project #3: Critique of a Scientific Article
Objective Your group will carefully analyze the statistical methodology in an article of your choice from a scientific journal. Your writeup should address the areas of control, randomization, and replication in depth and should include other relevant concerns, such as ethical concerns.
Deadlines By no later than the end of class on Thursday, Feb. 11, your group leader should tell me what article you are planning to work on. Your writeup is due on Thursday, Feb. 18.
Length I anticipate that a good paper will need to be at least 2 pages long, but a precise estimate is difficult since everyone will be analyzing different articles. Anything longer than 5 pages is probably too long. Avoid wordiness; make every sentence say something important. Points will be deducted for a paper that is longer than it needs to be.
You may prefer to make a slide presentation instead of a paper in paragraph form. If you choose this option, be aware that good slide shows tend to be much longer in page count, since you can make only a few points per slide. I estimate that 20-25 slides would be required. To save paper, please make your hard copy using "6-up" format (6 slides per printed page).
Contents At a minimum, you must address four areas:
control
randomization
replication
ethical concerns
Some sample questions to ask yourself are shown below. Note: There is no need to answer all of these questions. The list below is intended only to start your thinking process. You will probably come up with different and (in many cases) better questions to answer.
control
Did the researchers use blocking or other methods to account for lurking variables? Did they include a control group? Was the control group as similar as possible to the experimental group? Are there additional lurking variables, or possible lurking variables? Did the researchers discuss any cost limitations or practical considerations that may have reduced the effectiveness of the control process? Was the experiment blind or double-blind?
randomization
Did the researchers mention how subjects were assigned to experimental or control groups? Is there any reason to suspect systematic bias? If the process was not entirely random, do you still have reason to believe the results found? (In certain cases, you might.)
replication
Is the methodology described sufficiently clearly that another researcher with similar lab conditions or similar sample populations should be able to duplicate the results? Discuss the sample sizes: Are they large enough to support the confidence levels or P values claimed? (You need not repeat the computations, and in most cases you won’t be able to since you don’t have access to the raw data, but are the results reasonable? A quick estimate of the computations the researchers must have performed to obtain a confidence interval or P value would be most impressive. See me if you need help with c 2 distributions; you should already be able to handle Z and t calculations on your own.)
ethical concerns
Did the researchers act properly toward their human subjects, if any? If they used animal subjects, what can you say about the use of animals to obtain these results? Summarize the results in one sentence and answer this question: Did the ends justify the means? Were the results obtained (or not obtained) of sufficient value to justify the animals used, or the humans inconvenienced? There are no right or wrong answers to this section. Your opinions and your ability to discuss them are what will be considered.
Grading This project is shorter and simpler than the previous two and is therefore worth a maximum of 225 points, not 300. Points will be awarded as follows:
writing quality 60 points
consistency 30 points (should read as if one author wrote it)
statistical analysis 100 points
format/neatness 25 points
group leader’s justification of point split 10 points (note: an explanation is needed; merely
saying "we all did equal work" is not sufficient)