Unit III - Gathering Data

Ch. 11: Understanding Randomness

1. Conduct a simulation to simulate random outcomes for a real-world situation
  1. Be able to recognize random outcomes in a real-world situation
  2. Be able to recognize when a simulation might usefully model random behavior in the real world
  3. Know how to perform a simulation either by generating random numbers on a computer or calculator, or by using some other source of random values such as dice, a spinner, or a table of random numbers
  4. Be able to discuss the results of a simulation study and draw conclusions about the question being investigated
  5. Be able to describe a simulation so that others could repeat it

Ch. 12: Sample Surveys

1. Know the basic concepts and terminology of sampling
  1. Understand the value of randomization as a defense against bias
2. Understand the value of sampling to estimate population parameters from statistics calculated on representative samples drawn from the population.
  1. Recognize population parameters in descriptions of populations and samples
  2. Understand the size of the sample (not the fraction of the population) determines the precision of the estimates
3. Know how to draw a simple random sample from a master list of a population, using a computer or a table of random numbers
  1. Know what to report about a sample as part of your account of a statistical analysis
4. Report possible sources of bias in sampling methods. Recognize voluntary response and nonresponse as sources of bias in a sample survey

Ch.13: Experiments and Observational Studies

1. Design and conduct a retrospective and prospective observational study
  1. Identify the subjects, how the data were gathered, and any potential biases or flaws you may be aware of. Identify the factors known and those that might have been revealed by the study
  2. Recognize when an observational study would be appropriate
  3. Be able to identify observational studies as retrospective or prospective, and understand the strengths and weaknesses of each method
  4. Know how to report the results of an observational study.
2. Be able to design a completely randomized experiment to test the effect of a single factor
  1. Know the 4 basic principles of sound experiment design:control, randomize, replicate, and block, and be able to explain each
  2. Be able to recognize the factors, treatments, and the response variable in a description of an experiment
  3. Understand the essential importance of randomization in assigning treatments to experimental units
  4. Understand the importance of replication to move from anecdotes to general conclusions
  5. Understand the importance of a control group and the need for a placebo treatment in some studies
  6. Understand the importance of blinding and double-blinding in studies on human subjects, and be able to identify blinding and the need for blinding in experiments
  7. Understand the value of a placebo in experiments with human participants
  8. Understand that your description of an experiment should be sufficient for another researcher to replicate the study with the same methods
3. Be able to design an experiment in which blocking is used to reduce variation
  1. Understand the value of blocking so that variability due to differences in attributes of subjects can be removed
  2. See all other principles of experimental design <Skill 13.2>
4. Interpret the results and make conclusions about an experiment
  1. Know how to use graphical displays to compare responses for different treatment groups
  2. Understand that you should never proceed with any other analysis of a designed experiment without first looking at box plots or other graphical displays
  3. Know how to compare the responses in different treatment groups to assess whether the differences are larger than could be reasonably expected from ordinary sampling variability
  4. Tell who the subjects are and how their assignment to treatments was determined. Report on how the response variable was measured and in what measurement units
  5. Be able to report on the statistical significance of the result in terms of whether the observed group-to-group differences are larger than could be expected from ordinary sampling variation