Wednesday 15 June 2016

EXPERIMENTAL GROUPS AND CONTROL GROUPS

EXPERIMENTAL GROUPS AND CONTROL GROUPS:

Experimental research requires, then, that the responses of at least two groups be compared. One group will receive some special  Treatment —the manipulation implemented by the experimenter—and another group will receive either no treatment or a different treatment. Any group that receives a treatment is called an Experimental group ; A group that receives no treatment is called a Control group. (In some experiments there are multiple experimental and control groups, each of which is compared with another group.)  By employing both experimental and control groups in an experiment, researchers are able to rule out the possibility that something other than the experimental manipulation produced the results observed in the experiment. Without a control group, we couldn’t be sure that some other variable, such as the temperature at the time we were running the experiment, the color of the experimenter’s hair, or even the mere passage of time, wasn’t causing the changes observed. For example, consider a medical researcher who thinks he has invented a medicine that cures the common cold. To test his claim, he gives the medicine one day to a group of 20 people who have colds and finds that 10 days later all of them are cured.  Eureka? Not so fast. An observer viewing this fl awed study might reasonably argue that the people would have gotten better even without the medicine. What the researcher obviously needed was a control group consisting of people with colds who  don’t get the medicine and whose health is also checked 10 days later. Only if there is a significant difference between experimental and control groups can the effectiveness of the medicine be assessed. Through the use of control groups, then, researchers can isolate specific causes for their findings—and draw cause-and-effect inferences. Returning to Latané and Darley’s experiment, we see that the researchers needed to translate their hypothesis into something testable. To do this, they decided to create a false emergency situation that would appear to require the aid of a bystander. As their experimental manipulation, they decided to vary the number of bystanders present. They could have had just one experimental group with, say, two people present, and a control group for comparison purposes with just one person present. Instead, they settled on a more complex procedure involving the creation of groups of three sizes—consisting of two, three, and six people—that could be compared with one another.