Data Collection

Chapter 1 Data Collection

Smartphones Leaking Information

Share this YouTube video with your students, which deals with how smartphones "leak" information.

The Placebo Effect

A popular method for defining a control group is through the use of a placebo. A placebo is an innocuous medication, such as a sugar tablet, that looks, tastes, and smells like the experimental medication. A placebo might also be a procedure that follows the same steps as the experimental procedure, but leaves out a key intervention. For example, a procedure called vertebroplasty where medical cement is pumped into a spine fracture was tested through a designed experiment. All subjects went through a surgery to repair the spine, but only half received the medical cement.

An interesting outcome results from the vertebroplasty experiment. A subject in the placebo group found that the procedure resulted in complete abatement of the back pain even though she did not receive the medical cement! This type of phenomena in an experiment is referred to as the placebo effect. A book entitled Cure by Jo Marchant explores the placebo effect.   In the book, she suggests that placebo treatments can have measurable effects. For example, in patients with Parkinson’s disease placebos caused an increase of the neurotransmitter dopamine. In fact, in a study of 459 migraine sufferers, it was found that the placebo effect accounted for about 60% of the benefit of the drug Maxalt. Of course, the placebo effect will not account for improvements in someone with a tumor or replace insulin with someone with diabetes. However, the Maxalt study suggests that remedies for pain, nausea, or depression rely extensively on the placebo effect.

Random Sample versus Random Assignment

In random sampling, individuals are randomly selected from some population. In random assignment, it is often individuals who happen to be available for a study who are then randomly assigned to one of n treatment groups.

When there is random selection of individuals from a population, inference allows conclusions that apply to the population to be made.   When individuals are randomly assigned to a treatment group, causal inferences may be drawn.

Completely Randomized Design Example

Benefits of Exercise. Researchers enrolled 247 women between the ages of 70 and 93 in a study. They women were randomly assigned into three groups. The individuals in Group 1 attended a computer class three times a week for 90 minutes each session; the individuals in Group 2 enrolled in an exercise program three times a week for 90 minutes each session; the individuals in Group 3 were told to stick to their normal routines. After six months, the women were issued a survey that results in an age-satisfaction score.

  • Who are the experimental units in this study?
  • Is this a designed experiment or observational study? State the type of experimental design or observational study.
  • What is the response variable?
  • What variables were controlled and fixed?
  • What variable was manipulated? How many levels of this variable are there?