Test for Effectiveness of Appetite-Reducing Drug - Essay Example

 A good experiment possesses four characteristics. It must be internally valid, meaning that the experiment assures that changes made to the independent variable are the sole cause of variations in the dependent variable. The experiment must be reliable, such that the outcome is not caused by chance factors. The experiment must be sensitive, capable of detecting even small effects of the independent variable. Finally, the experiment must be externally valid if its findings hold for individuals and conditions beyond the experiment’s scope. In the experiment described, External validity is compromised because while the experiment was supposed to be valid for the average college student at KU, the 200 students that comprised the sample were all part-time students, and all students of the professor conducting the experiment. The sample does not include full-time students who comprise the majority of the student body, or students taking other courses, so the sample does not represent the average student. Also, the data gathering is done by the students themselves, which leaves a wide margin of possible variation in the recording of the number of hours, therefore the internal validity and reliability are both compromised as chance and other factors (errors, and lapses, difficulty of subjects) tend to affect the dependent variable. Finally, sensitivity may not be adequate because of estimations and approximations that subjects may have resorted to in recording their hours. Twelve hours versus eight hours is significant enough to conclude that the experimental group studies more than the control group, but due to the lack in internal and external validity, sensitivity and reliability, then the results are inconclusive. The sample must be correctly stratified (full-time vs part-time, course variations) proportional to the actual distribution to be representative of the average student. There must also be a way that the accuracy of the readings (number of hours) could be ascertained; if not, then this is at best only a quasi-experiment that may not yield conclusive results without the necessary qualification. Describe the four different threats to internal validity [intact groups, extraneous variables, subject loss, and expectancies]. How do these threats apply to the following study? There are four threats to internal validity in experiments. The assignment of intact groups, rather than individuals, compromises the randomness of the sample and therefore skews the results. The presence of extraneous variables confounds the findings and renders them uninterpretable. Subject loss due to any one particular condition more than others renders some subjects unable to continue. Finally expectations held by either the subjects or experimenters, or both, tend to affect their conduct during the experiment which may affect the results. Internal validity is compromised in this experiment. The biggest threat to the experiment’s internal validity is the presence of expectations of the experimenter which may create a bias in the results obtained. In the given situation, the experimenter is not blind as to who among the students are given the placebo, the tested drug, and the new experimental drug. While it is mentioned that the experimenter is determined to remain unbiased, this is not enough as psychological bias may be unintended but nevertheless present in the results. External validity does not seem to be an issue here, because the sample chosen was randomly selected and therefore may be assumed to represent the relevant population of KU students. Also, since all groups are given the same examination that covers the contents of the full ten week course, and no other variable had been varied except the test variable (the experimental drug), then chance variables did not materially affect the results. On the basis of the given information, there is no reason to presume that external validity is compromised. PROJECT: Test for Effectiveness of Appetite-Reducing Drug The objective of the experiment to be designed should determine whether or not the experimental drug is effective in reducing the appetite of the individual ingesting it. The procedure would involve administering the drug to the test group and monitoring the progress in the food consumption (proxied by the relative change in weight of the subjects) of the test group against a control group. In order for the control group to be blind to which group they belonged, they will be given placebos which are known to have no effects on the appetite. The experimental method to be used would be the independent group design, the alternative to the repeated measures design group. Repeated measures design is not applicable to this assignment because for repeated measures, the same subjects repeatedly undergo all the procedures at intervals, and reactions to the test condition are observed when the condition is applied and when it is withdrawn. This will not do because the successful administration of the drug would presumably include a weight loss, and repeated treatment (to reduce appetite) may reasonably cause a drop in weight. The loss in weight would be a permanent change to the subject, whether as a result of the treatment or not, and withdrawal of the treatment does not necessarily cause a reversal to pre-treatment conditions. It therefore defies the reason for the alternate application and removal of treatment. Furthermore, it is ethically questionable to withdraw a treatment for the purpose of getting the individual to recede from a state of better health to one of poorer health (Shaughnessy, et al. 2008). For these reasons, the repeated measures design would be inappropriate for the experiment to derive the best possible information. The target population are comprised of the intended beneficiaries of the drug, who are clinically overweight adults constrained to reduce their food intake for health purposes. It is important to specify “adults,” as medication for that age range would essentially be expected to have a different formulation that that for children The sampling method shall be stratified random sampling. The reason for this is that obesity (or the overweight condition) is rather widespread among the populace that a broad random sample would be necessary to rule out other accidental variables that may affect the results. By randomizing the groups, the only difference that would account for the results would be the difference in the independent variable. In this case, the independent variable is isolated such that it provides the only explanation of the effects on the random variable. In this study, the most important ethical issue is related to the welfare of the subjects, as to whether the experiment may result in harm to the individual due to the medical side-effects that may result in the subject (Newman, et al., 2008). It is for this reason that full disclosure about the experimental drug, information on its active ingredient, and possible side effects should be told the subjects, so that they may verify from their own doctors (if they so wish) the safety of the drug for them. The expected results are findings backed by statistical confirmation significant differences existed between the mean weight changes in the control group and the experimental group. There is little threat to the external validity of the study as long as random sampling is assured. This would allow the extension of the findings to situations beyond the study. There is a slight threat to internal validity since it is difficult to measure the appetite of a person from his exact caloric consumption (i.e. due to the variability of different foods and their preparation). Weight is being used as proxy, but even this is not foolproof as there are many reasons why weight should change that is not related to food consumption, such as amount of physical exercise. For future research on the topic, a clinically controlled experiment with closely monitored results would be best before releasing the product into the market. References Newman, Barbara M. & Newman, Philip R. (2008) Development Through Life: A Psychoshocial Approach, Tenth edition.. Wadsworth, Cengage Learning. Zechmeister, Jeanne, Eugene Zechmeister & John Shaughnessy (2001) Research Methods in Psychology. McGraw-Hill. Read More