Experiments, Good and Bad An experiment deliberately imposes some treatment on individuals in order to observe their responses. The purpose of an experiment is to study whether the treatment causes a change in response. Definitions A RESPONSE VARIABLE is a variable that measures an outcome or result of a study. response variables are sometimes called dependent variables An EXPLANANTORY VARIABLE is a variable that we think explains or causes changes in the response variable. explanatory variables are sometimes called independent variables A TREATMENT is any specific experimental condition applied to the subjects. Treatments are different levels (or combination of different levels) of the explanatory variable(s). The individuals in an experiment are often called SUBJECTS. Example 1 You are planning an experiment to study the effect of gasoline brand and engine size on the gas mileage (miles per gallon) of sport utility vehicles. In this study, gas mileage is A. The response variable. B. The explanatory variable. C. A lurking variable. In this study, the explanatory variable(s) is(are) A. Miles per gallon B. Sport utility vehicles C. Gasoline brand D. Engine size E. (C) and (D) Suppose the experiment consists of 3 gasoline brands and 2 engine sizes. How many treatments are there? A. 2 B. 3 C. 5 D. 6 What are the subjects in this experiment? A. Gasoline Brand B. Engine Size C. Sport Utility Vehicles D. Drivers of Sport Utility VehiclesChapter 5 Page 2 A LURKING VARIABLE is a variable that has an important effect on the relationship among the variables in a study, but is not one of the explanatory variables studied. Two variables are CONFOUNDED when their effects on a response variable cannot be distinguished from each other. Confounded variables may be either explanatory variables or lurking variables. What is a possible lurking variable for the study of gasoline brand and engine size on gas mileage? A. Price of the gasoline B. Size of the SUV C. Prior engine maintenace D. (B) and (C) E. All of the above A PLACEBO is a dummy treatment. If the treatment is delivered in the form of a pill, it has no active ingredients The PLACEBO EFFECT is the response to a placebo (a dummy treatment). Some subjects improve while taking a placebo! The placebo effect can be confounded with the effect of a treatment. Confounding in experiments with only one treatment (“one track experiments”) masks the effect of the explanatory variable on the response variable. A RANDOMIZED COMPARATIVE EXPERIMENT (RCE) is an experiment in which • subjects are randomly assigned to treatments – Label subjects, use random digits – Randomization produces groups of subjects that should be similar in all respects before we apply the treatments. • Two or more treatments are compared – Usually to a control; comparing to a control manages the effects of lurking variables – Comparative design ensures that influences other than the experimental treatments operate equally on all groups. Therefore, differences in the response variable must be due to the effects of the treatments.Chapter 5 Page 3 Diagram of a Randomized Comparative Experiment with 3 treatments Principles of Experimental Design Control the effects of lurking variables on the response, most simply by comparing two or more treatments. Randomize – use impersonal chance to assign subjects to treatments. Use enough subjects in each group to reduce chance variation in the results. When doing a randomized comparative experiment, we compare the results of the two treatments. Do you expect the different treatments to give you the exact same results in all the experimental treatment groups? NO! How different do the results have to be to decide if one treatment is better than another? An observed difference “large enough” that it would rarely occur by chance is called a STATISTICALLY SIGNIFICANT difference. We’ll talk about how we qualify differences as “large enough” later…..Chapter 5 Page 4 Fruits and vegetables are rich in antioxidants such as vitamins A, C, and E. A clinical trial is conducted to determine whether or not antioxidants prevent colon cancer. The researcher recruited 864 people who were at risk for colon cancer. They were either given daily beta-carotene, daily vitamins C and E, all three vitamins every day, or a daily placebo. After four years, the researchers were surprised to find no significant difference in colon cancer among the groups. What is the response? What is the explanatory variable? What are the subjects? What are the treatments? What does “no significant difference” mean in the outcome of the study? The researchers did not find significant evidence that taking antioxidants daily prevents colon cancer. What does this mean? The difference in colon cancer rates between the two groups was not large enough to conclude anything other than chance variation. (So, even if antioxidants really didn’t prevent colon cancer we could still see these sorts of results.) Living with Observational Studies Good studies are comparative even when they’re not experiments. We can use matching to control for lurking variables. We can measure and adjust for confounding variables (by using statistical
View Full Document
Unlocking...