This is the study guide for the first exam in Biological Data Analysis, spring 2017. The exam will be on Tuesday, March 7. You may not use your notes or books during the exam; if English is your second language, you may use a dictionary.
You will not be doing any statistical tests or other calculations for any exam in this class. Therefore, you will not need to use a calculator.
Study your lecture notes and the web pages (available at the class syllabus page) on all the topics covered before the exam.
There will be three kinds of question on the first exam:
1. Identify the variables. I will describe an experiment and ask you to identify the variables and say what kind they are (nominal, measurement, or ranked). Do not include variables that aren't mentioned in the question; for example, if I say "I had 20 people eat spaghetti and 20 people eat steak for dinner, then timed them running 5 kilometers the next morning," the only variables are dinner type (nominal) and running time (measurement). Age, sex, height, weight, running experience, etc. would be important variables if you were really doing this experiment, but they are not mentioned in the question, so don't include them in your answer.
2. Questions about probability, hypothesis testing, and power analysis. You should know what a P-value means and how to interpret the results of statistical tests. You should understand why we do power analysis, what is needed to do one, and how to interpret the results.
3. Choose a statistical test. I will describe an experiment, and you should say which test would be most appropriate. At this point, your choices are exact test of goodness-of-fit, chi-square test of goodness-of-fit, G-test of goodness-of-fit, Fisher's exact test, chi-square test of independence, G-test of independence, and Cochran-Mantel-Haenszel test.
In class, we mostly talked about the exact binomial test of goodness-of-fit (only two values of the nominal variable), but the exams may include exact tests of goodness-of-fit with more than two values. We mostly talked about Fisher's exact test for 2x2 tables, but the exams may include Fisher's exact tests for larger tables. Remember the rule of thumb for this class: if the total N is less than 1000 for a test of goodness-of-fit or independence, you must use an exact test.
You must give only one answer, even if more than one would be correct; don't say "chi-square test of independence or G-test of independence," pick one or the other. You must specify test of goodness-of-fit or independence; don't just say "chi-square test."
If a question uses biological words or concepts that are unfamiliar to you (for example, you don't know what a "promoter element" is), please ask me for help. You are being tested on your knowledge of methods used to analyze biological experiments, not on your knowledge of biology.
The exam will consist of about 15 to 20 short-answer questions. Here are some examples. To see the correct answer to a question, click on the "Answer" button. Don't bother typing your answer into the box, it won't do anything.
Return to the Biological Data Analysis syllabus