A friend of yours tells you he has just performed an experiment to find out what effect exercise has on people’s visual acuity. In his experiment he got a group of 20 friends to come to his house at 8 a.m., Saturday morning. He checked their visual acuity by having them identify 15 different symbols on an eye chart from 10 feet away. The average number of errors for these 20 people was 5. After the first visual acuity test, he had all of them run two miles. After they had all finished running, he tested each of their visual acuity using the same method. The average number of errors for the group this time was only 2. He tells you that his experiment shows that exercise improves people’s visual acuity. You realize, of course, that the “experiment” he performed was in fact not a true experiment, but rather a type of pre-experimental design. Write a letter to your friend (approximately 400-600 words) in which you identify the type of pre-experimental design he used and explain to him the problems with drawing a conclusion based on such a design. In particular you want to discuss issues of internal invalidity (which sources of internal invalidity are present in his “experiment” – explain why they may invalidate his conclusion). Then explain to him a proper experimental design that would allow for an appropriate test of his hypothesis (i.e., exercise improves visual acuity). Explain how a true experimental design improves on his pre-experimental design. Explain the role of each component of the design (e.g., experimental and control groups, pretesting and post testing, randomization/matching etc.) and how these relate to issues of internal validity.
Sample Solution